ABSTRACT
Actinobacillus pleuropneumoniae is the etiological agent of porcine pleuropneumonia, an economically important disease of pigs. The hfq gene in A. pleuropneumoniae, encoding the RNA chaperone and posttranscriptional regulator Hfq, is upregulated during infection of porcine lungs. To investigate the role of this in vivo-induced gene in A. pleuropneumoniae, an hfq mutant strain was constructed. The hfq mutant was defective in biofilm formation on abiotic surfaces. The level of pgaC transcript, encoding the biosynthesis of poly-β-1,6-N-acetylglucosamine (PNAG), a major biofilm matrix component, was lower and PNAG content was 10-fold lower in the hfq mutant than in the wild-type strain. When outer membrane proteins were examined, cysteine synthase, implicated in resistance to oxidative stress and tellurite, was not found at detectable levels in the absence of Hfq. The hfq mutant displayed enhanced sensitivity to superoxide generated by methyl viologen and tellurite. These phenotypes were readily reversed by complementation with the hfq gene expressed from its native promoter. The role of Hfq in the fitness of A. pleuropneumoniae was assessed in a natural host infection model. The hfq mutant failed to colonize porcine lungs and was outcompeted by the wild-type strain (median competitive index of 2 × 10−5). Our data demonstrate that the in vivo-induced gene hfq is involved in the regulation of PNAG-dependent biofilm formation, resistance to superoxide stress, and the fitness and virulence of A. pleuropneumoniae in pigs and begin to elucidate the role of an in vivo-induced gene in the pathogenesis of pleuropneumonia.
INTRODUCTION
Actinobacillus pleuropneumoniae is an encapsulated Gram-negative pleiomorphic coccobacillus in the family Pasteurellaceae and is the etiological agent of porcine pleuropneumonia (1). This disease is characterized by a fulminating fibrino-hemorrhagic bronchopneumonia, which is often fatal. Although the incidence of outbreaks has decreased in the developed world, porcine pleuropneumonia remains a major global cause of economic loss in intensive swine production (2). A. pleuropneumoniae produces several well-defined virulence factors, including the Apx toxins, capsular polysaccharides, and lipopolysaccharides, that enhance evasion of clearance by phagocytes and induce tissue damage, resulting in edema, hemorrhage, and necrosis within the lung (1).
To identify additional virulence factors, an in vivo expression technology study was performed previously in our laboratory and A. pleuropneumoniae genes that are specifically upregulated during infection of the porcine lungs were identified (3). A total of 32 genes, including the hfq gene that encodes host factor Q-β (Hfq), were identified in this screen (3, 4). A total of 25% (8/32) of the in vivo-induced genes, including hfq, were upregulated under limitation of the branched-chain amino acids (BCAAs) isoleucine, leucine, and valine in defined laboratory medium (5). Limitation of BCAAs is one of the host signals encountered by A. pleuropneumoniae in porcine lungs, and the ability to synthesize BCAAs is essential for the survival and virulence of A. pleuropneumoniae during experimental infection (6).
Hfq was originally identified as a factor required for the replication of RNA bacteriophage Q-β in Escherichia coli (7). Hfq is a pleiotropic posttranscriptional regulator which modulates translation and transcript stability by acting as an RNA chaperone in bacteria (8). Homohexamers of Hfq bind to the A/U-rich regions in the 5′ untranslated regions (UTR) of transcripts and small regulatory RNAs (sRNAs) to facilitate formation of mRNA-sRNA duplexes by incomplete base pairing (8). This interaction either enhances or blocks the access of ribosomes to the translation initiation region, and the mRNA-sRNA duplex may be targeted to degradation, although inhibition of translation alone is sufficient for silencing gene expression (9). Small RNAs play a number of regulatory roles in the physiology as well as the virulence of bacterial pathogens by acting as switches in adaptation to ever-changing environmental conditions (10). However, Hfq can also act as a regulator, independent of sRNAs. For instance, in E. coli, Hfq binds to the 5′ UTR of its own transcript and regulates translation (11).
Bacteria are frequently found as sessile communities embedded in a self-produced extracellular polymeric matrix described as a biofilm. Bacterial pathogens growing as a biofilm have enhanced resistance to host defense mechanisms and antimicrobial agents (12). Many field isolates of A. pleuropneumoniae form robust biofilm on abiotic surfaces (13). Poly-β-1,6-N-acetylglucosamine (PNAG) is the major biofilm matrix component that facilitates autoaggregation and biofilm formation in A. pleuropneumoniae (14). The pgaABCD operon encodes the proteins involved in the biosynthesis and export of PNAG (14). A. pleuropneumoniae also produces dispersin B, a hexosaminidase which specifically degrades PNAG (15). Hfq is implicated in biofilm formation by uropathogenic E. coli (16), Vibrio cholerae (10), and Yersinia pestis (17).
Hfq is also involved in resistance to oxidative stress and virulence in a number of bacterial pathogens (18). However, the effects of Hfq on the fitness and virulence of bacterial pathogens during experimentally induced pneumonia have not been reported; here, the role of Hfq in the competitive fitness and virulence of A. pleuropneumoniae during porcine pleuropneumonia is described. In this report, we provide evidence for the regulation of PNAG-based biofilm formation by Hfq. Studies to identify additional Hfq-regulated factors involved in biofilm formation led us to the finding that cysteine synthase, CysK, was not expressed at detectable levels in outer membranes of the hfq mutant strain. Since CysK is involved in resistance to oxidative stress, we tested the ability of the hfq mutants to survive under oxidative stress and found that Hfq is required for resistance to superoxide stress in A. pleuropneumoniae. Finally, we tested the role of Hfq in the fitness and virulence of A. pleuropneumoniae in a porcine pleuropneumonia infection model. Competitive index analysis revealed that the hfq mutant is defective in survival during infection of porcine lungs compared to the wild type. To our knowledge, this is the first report of the role of Hfq in the virulence of a respiratory tract pathogen in the lungs.
MATERIALS AND METHODS
Bacterial strains and growth conditions.The bacterial strains and plasmids used for this work are listed in Table 1. A. pleuropneumoniae strains were grown in brain heart infusion (Becton, Dickinson and Company, Sparks, MD) supplemented with 10 μg/ml of NAD (BHIV). The hfq mutant and complemented mutant strains were grown in BHIV containing 1.5 μg/ml of chloramphenicol and 50 μg/ml of ampicillin, respectively. Agar plates were incubated at 35°C under 5% CO2; broth cultures were incubated at 35°C in a water bath shaken at 160 rpm. Dispersin B (Kane Biotech Inc., Winnipeg, Canada) was added to A. pleuropneumoniae cultures (250 ng/ml) to prevent autoaggregation during preparation of competent cells for natural transformation and for growth curves. Chemically defined medium (CDM) was used for biofilm experiments and for RNA extraction (5). For pig infection experiments, A. pleuropneumoniae was grown in heart infusion broth (Becton, Dickinson) supplemented with NAD, CaCl2 (5 mM), and dispersin B to an optical density at 520 nm (OD520) of 0.8. Cells were washed and resuspended in phosphate-buffered saline (PBS) (127 mM NaCl, 7 mM Na2HPO4, and 4 mM NaH2PO4·H2O; pH 7) to the required number of CFU/ml.
Strains and plasmids used in this study
E. coli XL1-Blue was used for cloning and grown in Luria-Bertani medium at 37°C. For antibiotic selection, 30 μg/ml of chloramphenicol or 100 μg/ml of ampicillin was used.
DNA manipulation.Oligonucleotide primers used in this study are listed in Table 2. The nucleotide sequence of the hfq gene (APL_1961) from the A. pleuropneumoniae serotype 5b strain L20 genome sequence (GenBank accession no. CP000569.1) was used to design oligonucleotide primers. A knockout construct (pSS103) containing the 3′ end of miaA [encoding tRNA delta(2)-isopentenyl pyrophosphate transferase], the 5′ end of hfq, the promoter region of nptI (neomycin phosphotransferase from pUC4K), the chloramphenicol acetyltransferase (cat) cassette, the 3′ end of hfq, and the 5′ end of hflX (GTP-binding protein HflX), in tandem with an internal deletion of 153 bp from the 279-bp hfq gene, was constructed in the pUC18 vector (Fig. 1). The 5′ end of hfq and the upstream flanking region (644 bp) and the 3′ end of hfq and the downstream flanking region (617 bp) were amplified with the primer pairs MM 721-MM 722 and MM 724-MM 725, respectively, from A. pleuropneumoniae ATCC 27088 genomic DNA and cloned into pUC18 to generate the plasmid pSS102. This construct has a unique NsiI recognition site between the 5′ and 3′ ends of hfq. A cat cassette expressed from the nptI promoter, derived from pRL100 (6), was cloned into the NsiI site of pSS102 to engineer pSS103. A deletion disruption mutation at the hfq locus in AP 93-9 (a biofilm-positive serotype 1 field isolate) was introduced by allelic exchange using natural transformation (19). To validate this mutant construct, genomic DNA isolated from chloramphenicol-resistant transformants was screened by PCR.
Oligonucleotide primers used in this study
Construction of an A. pleuropneumoniae hfq mutant strain. (A) Alignment of amino acid sequences of Hfq from A. pleuropneumoniae (A. p) and E. coli (E. c) reveals the conservation of Sm1 and Sm2 motifs (underlined). (B and C) Schematic representation of the hfq locus in the wild-type strain and in the hfq mutant strain, respectively. The flanking genes miaA, encoding tRNA delta(2)-isopentenyl pyrophosphate transferase, and hflX, encoding the GTP-binding protein HflX, are depicted here. The promoter region (P) of the nptI gene encoding neomycin phosphotransferase from pUC4K is upstream of the chloramphenicol acetyltransferase (cat) cassette. (D) Verification of the hfq mutant using PCR. Primers complementary to the hfq start and stop codon regions (marked with arrows in panels A and B) were used with the following templates: wild type (lane 1), Δhfq (lane 2), knockout plasmid (lane 3), complementation plasmid (lane 4), no-template control (lane 5), and molecular size marker (M). (E) Effect of mutations of hfq on growth. Results presented here are means and standard errors of the means (SEM) (error bars) from four independent experiments with the following strains: WT, wild type; Δhfq, hfq mutant; Δhfq + hfq, complemented mutant. The SEM are small, and therefore, the error bars are not readily apparent.
The entire hfq gene and 595 bp of the upstream region to include the native promoter were amplified with primers MM 721 and MM 729 and cloned into the A. pleuropneumoniae-E. coli shuttle vector pGZRS19 for trans complementation (20). The complementation construct, pSS104, was verified by PCR, restriction mapping, and sequencing.
Biofilm assays.For initial studies on the time course of biofilm formation, wild-type biofilm-positive A. pleuropneumoniae (AP 93-9) was inoculated into either BHIV or CDM to an OD520 of ∼0.05 in 6-well polystyrene plates (Corning, Corning, NY) containing 3 ml of culture. Biofilm biomass was quantified at 0, 3, 6, 9, 12, and 24 h using a crystal violet binding assay as described previously (13, 21). Results reported are the mean and standard error of triplicate samples from four independent experiments.
RNA extraction.Wild-type (AP 93-9) and hfq mutant (AP 371) strains were grown in CDM in 6-well plates for 6 h. Ice-cold methanol was used to prevent changes in transcript levels after sample collection. From wells containing the wild-type strain, planktonic and biofilm cells were collected separately. Since the hfq mutant did not form biofilm, only the planktonic cells were available for harvesting. A total of 500 μl of planktonic cells, from wells with the wild type or hfq mutant, was mixed with 500 μl of 100% ice-cold methanol and centrifuged at 4°C and 16,000 × g for 2.5 min. To collect the biofilm, the supernatant was completely removed and 3 ml of 50% ice-cold methanol was added to each well. A cell scraper was used to detach the sessile cells. Aliquots of 1 ml were centrifuged at 4°C and 16,000 × g for 2.5 min. The cell pellets were stored at −80°C until RNA extraction.
RNA extraction and qPCR.Pellets were resuspended in 400 μl of TE (10 mM Tris HCl and 1 mM EDTA; pH 8). A total of 100 μl of 50-mg/ml lysozyme solution was added, and the samples were incubated for 10 min at room temperature. RNA was extracted using Qiagen RNeasy mini columns (Qiagen, Valencia, CA) with an on-column DNase treatment by following the manufacturer's instructions. The concentration, purity, and integrity of RNA samples were determined using a NanoDrop spectrophotometer (NanoDrop Products, Wilmington, DE) and agarose gel electrophoresis. Additional contaminating DNA was removed by DNase treatment with the Turbo DNA-free kit according to the manufacturer's instructions (Ambion, Austin, TX). RNA was reverse transcribed using a Superscript II reverse transcription kit (Invitrogen, Carlsbad, CA). Three 10-fold dilutions of cDNA were used in quantitative PCR (qPCR) with SYBR green PCR core reagents in a StepOnePlus real-time PCR system (Applied Biosystems, Foster City, CA). Data were normalized using the 16S rRNA gene as the endogenous control, and the results were analyzed by a threshold cycle (ΔΔCT) method described by Pfaffl (22) to calculate fold change in gene expression. Each sample was measured in triplicate, and the experiment in its entirety was repeated two times.
Determination of PNAG levels.The Congo red binding property of PNAG was utilized to customize a quantitative assay for PNAG. A. pleuropneumoniae strains were grown for 6 h in 3 ml of BHIV supplemented with ampicillin, with or without dispersin B (1.5 μg/ml), in polystyrene 6-well plates. Crude polysaccharide extracts were prepared as described previously (14) with the following changes. For the wild-type and complemented mutant strains, planktonic and biofilm cells were harvested and processed separately. Only planktonic cells were available from wells treated with dispersin B. Biofilms were collected from the culture wells with a cell scraper and resuspended in 500 μl of acetate buffer (50 mM sodium acetate, pH 5.8, with 100 mM sodium chloride). Cells were harvested by centrifugation in 1.5-ml tubes at 16,000 × g for 3 min. Pellets were resuspended in acetate buffer (4 μl buffer/mg of pellet), and an equal volume of Tris buffer (10 mM Tris-HCl, pH 8, 5 mM EDTA, and 0.5% SDS) was added. Tubes were vortexed briefly and placed in a boiling water bath for 10 min before centrifugation at 16,000 × g for 1 min. Pellets contain the crude polysaccharides and were resuspended in acetate buffer (2 μl buffer/mg of pellet). A total of 10 μl of crude polysaccharide suspension was applied to a well in a polystyrene 96-well tissue culture-treated plate in duplicate and dried at 55°C for 1 h. Wells were stained with 200 μl of 1% Congo red in water (Sigma, St. Louis, MO) for 5 min, rinsed three times with 200 μl of water, and dried. PNAG-bound Congo red was solubilized in 200 μl of dimethyl sulfoxide (DMSO), and absorbance (OD500 nm) was measured (23).
Preparation of outer membrane fractions.A. pleuropneumoniae strains were grown in BHIV for 9 h and harvested by centrifugation at 4°C. Cell pellets were processed according to the method described by Cruz et al. to obtain membrane fractions (24). Briefly, pellets were treated with lysozyme to generate spheroplasts, which were subsequently ruptured by sonication in a Branson sonifier (Branson, Danbury, CT). A total membrane fraction was isolated by ultracentrifugation of whole-cell sonicate at 150,000 × g for 1 h. Cytoplasmic membrane and outer membrane (OM) were separated from the total membrane fraction by sucrose density gradient centrifugation. Aliquots were stored at −20°C for further analysis.
SDS-PAGE.The protein concentration in the OM fractions was determined by Bradford's method (Bio-Rad, Hercules, CA). A total of 2.5 μg of OM samples from the wild type and hfq mutant was separated on 12% SDS-PAGE gels using the discontinuous buffer system of Laemmli (25). Combined Coomassie blue-silver staining was used to compare the membrane protein profile of the hfq mutant to that of the parent strain as described by Cruz et al. (24). The experiment was repeated three times. A differentially expressed protein, found only in the wild type, was excised from the gel and submitted for matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) mass spectrometry analysis at the Michigan State University Research Technology Support Facility.
Oxidative stress protection assays.Cumene hydroperoxide, hydrogen peroxide, methyl viologen (paraquat), and potassium tellurite (Sigma) were used to induce oxidative stress. Growth inhibition assays on top agar plates were performed as described previously (26). Briefly, A. pleuropneumoniae strains grown in BHIV for 16 h were diluted 30-fold in 0.7% BHIV agar and cast on top of BHIV agar in plates. After solidification, filter paper discs (Whatman Paper Ltd., Maidstone, England) previously soaked in 10 μl of 200 mM cumene hydroperoxide, 880 mM hydrogen peroxide, 74 mM methyl viologen, or 390 mM potassium tellurite were placed on the plates. Plates were incubated for 22 h, and the diameter of zones of growth inhibition around the discs was recorded. Three independent experiments with at least three technical replicates were performed. To determine the effect of Hfq on oxidative stress in the absence of PNAG, experiments were done as described above except for the addition of dispersin B (1.5 μg/ml) to medium used for overnight growth. The entire experiment was repeated four times independently.
Experimental porcine pleuropneumonia model.The protocol and procedures used in pig infection experiments were reviewed and approved by the Michigan State University Institutional Animal Care and Use Committee. Eight-week-old clinically healthy pigs free of A. pleuropneumoniae, Mycoplasma, and porcine reproductive and respiratory syndrome (PRRS) virus were purchased from the Michigan State University Swine Research Facility. Infection experiments were conducted in two phases, and all animals were inoculated by percutaneous intratracheal injection with 10 ml of inoculum or PBS. This infection model was previously established in our laboratory (27). Prior to inoculation and euthanasia, pigs were sedated by intramuscular injection with a mixture of telazol (tiletamine plus zolazepam), ketamine, and xylazine (50 mg of each compound/ml of the mixture) at a dosage of 1 ml of mixture per 34 kg of body weight.
In the first phase, one pig was infected with 5 × 106 CFU of the wild-type strain (AP 93-9). Two pigs were infected with the hfq mutant strain (AP 371): one pig received 3 × 106 CFU and another pig received 6 × 107 CFU. Uninfected controls (2 pigs) received PBS. Pigs were monitored for the development of clinical signs of porcine pleuropneumonia at 2-h intervals as described previously (27). The respiratory rate was measured as the number of breaths in a 15-s period. The degree of respiratory distress, dyspnea, was scored on a scale of 0 to 3, with 0 being normal and 3 being extreme dyspnea. Physical inactivity (depression) was scored on a scale of 0 to 3, with 0 being normal and actively maintaining flight distance and 3 being moribund. Rectal temperature and appetite, indicated by feeding readily upon provision of feed, were also recorded.
Pigs were euthanized at 17 h postinfection (hpi), and parts of the lung exhibiting pleuropneumonia-like lesions were collected for culture. Samples were plated on plain or chloramphenicol-containing BHIV plates. The wild-type and hfq mutant strains isolated from the pigs were confirmed as A. pleuropneumoniae by a Gram stain, a requirement for V factor and PCR with primers that are specific for the A. pleuropneumoniae omlA gene (28). Coagglutination was used to verify that the wild type and hfq mutant reisolated from pigs belonged to serotype 1 (29). PCR was performed using MM 728 and MM 729 to confirm that the strains reisolated from pig lungs retained the wild-type or the mutated version of the hfq gene.
In the second phase, wild-type (AP 93-9-1) and hfq mutant (AP 371-1) strains reisolated from pigs in the first phase were used for individual and coinfection experiments. Ten pigs were divided into three groups with 3 to 4 pigs per group using a randomly stratified sampling method, balancing each group for body weight and sex. Littermates were allotted to different groups. One group (3 pigs) was infected with 4 × 106 CFU of the wild-type strain. The second group (3 pigs) received 4 × 106 CFU of the hfq mutant strain. The third group (4 pigs) received a 1:1 mixture of the wild type and the hfq mutant containing a total of 4 × 106 CFU.
Clinical evaluation was performed as described above at 2-h intervals for the first 12 h and at 4- to 6-h intervals afterwards until the end of the experiment. Pentobarbital was administered intravenously for euthanasia. Pigs were euthanized either at specific times (16, 40, and 64 hpi) or when a dyspnea score of 3, a physical inactivity score of 3, or a score of 2 in both dyspnea and physical inactivity was reached. Tracheotomy was performed to collect bronchoalveolar lavage fluid (BALF) by infusing 200 ml of PBS into the lungs via the trachea. Lungs were scored for gross lesions. The percentages of the lungs exhibiting gross pathological changes suggestive of pleuritis and pneumonia were estimated using a formula that accounts for the contribution of each lung lobe to the total volume of the lung (27). Representative samples from all six lung lobes from each pig were collected for culture and histopathological examination.
Determination of bacterial load in the lungs and BALF.BALF and lung samples were used for quantitative cultures. The samples were labeled with the animal identifier and, therefore, were not blinded to the observers. Lung samples were homogenized in heart infusion (HI) broth using a Stomacher 80 (Seward Laboratory Systems, Bohemia, NY). The total number of viable bacteria was determined by counting colonies on BHIV plates. The number of hfq mutants was enumerated by counting colonies on BHIV with chloramphenicol. The number of wild-type bacteria was deduced by subtracting the number of hfq mutants from the total number of viable bacteria. Select colonies were verified as A. pleuropneumoniae by assessing the NAD requirement and by PCR with primers for the omlA gene and the hfq gene. The competitive index for each sample was calculated as the ratio of the number of hfq mutant bacteria to the number of wild-type bacteria at any time divided by the ratio of the number of hfq mutant bacteria to the number of wild-type bacteria in the inoculum. Competitive indices were calculated for each of the six lung lobes and BALF.
In vitro competitive indices were determined by growing the wild type and hfq mutant in a mixed culture in HI broth. Cultures were serially diluted and plated to enumerate the number of wild-type and hfq mutant bacteria. Competitive indices were calculated essentially as described above.
Statistical analysis.Effects of Hfq on biofilm formation, PNAG production, and resistance to oxidative stress were analyzed by one-way analysis of variance (ANOVA) with Bonferroni's multiple-comparison test. CFU counts from individual infection experiments were analyzed using the Mann-Whitney test. In all cases, a P value of less than 0.05 was considered a statistically significant difference.
RESULTS
Construction of an hfq mutant strain.The predicted A. pleuropneumoniae Hfq protein exhibits a high degree of identity to Hfq proteins from several bacteria and contains both Sm1 and Sm2 motifs (Fig. 1A), which are conserved among bacterial Hfqs (30). A BLASTP search revealed that the A. pleuropneumoniae Hfq protein has 88% identity to Hfq from E. coli. A deletion disruption mutation at the hfq locus in A. pleuropneumoniae was introduced by allelic exchange using natural transformation (Fig. 1B and C). Chloramphenicol-resistant transformants were selected and verified as A. pleuropneumoniae by PCR using oligonucleotide primers specific for the omlA gene (data not shown) (28). The transformants were further analyzed by PCR to verify the presence, location, and size of the insert in the chromosome (Fig. 1D). PCR with primers to amplify the full-length hfq gene revealed 279-bp amplicons from the wild-type and the complementation plasmid templates, while 1,300-bp amplicons were observed with the hfq mutant and the knockout construct as templates (Fig. 1D), indicating that the cat cassette and the nptI promoter were inserted by a double-crossover event at the chromosomal hfq locus. The full-length hfq gene, including the native promoter region, was cloned to generate a complemented mutant strain.
Effect of Hfq on growth.Growth patterns of the wild-type, hfq mutant, and complemented mutant strains were determined in BHIV, a rich medium used for cultivating A. pleuropneumoniae (Fig. 1E). The hfq mutant exhibited a slightly reduced growth rate in the exponential phase compared to the wild type and complemented mutant, but all the strains used in this experiment reached similar optical density values at stationary phase. Provision of hfq on a low-copy-number plasmid reversed the growth defect, demonstrating that the reduced growth rate observed in the hfq mutant strain is due to the loss of Hfq and not caused by potential polar effects of the mutation (Fig. 1E).
Hfq is required for biofilm formation.Changes in A. pleuropneumoniae biofilm biomass over time under static conditions were measured to understand the kinetics of biofilm formation and determine a suitable time point to conduct further studies on biofilm formation. Biofilms were indirectly quantified using a crystal violet binding assay (21) at 0, 3, 6, 9, 12, and 24 h of incubation (Fig. 2A). Biofilm biomass peaked after 6 h of incubation and remained high until 12 h in BHIV (Fig. 2A). In CDM, maximum biofilm biomass was reached at 3 h and was retained up to 12 h. After 24 h, the amount of biofilm was lower than the amount of biofilm at 12 h but remained higher than the baseline value in both BHIV and CDM. Generally, higher biofilm biomass was observed in BHIV than in CDM (Fig. 2A).
Effect of Hfq on biofilm formation and production of PNAG. (A) Crystal violent binding was used to determine biomass of wild-type biofilm in BHIV (brain heart infusion) and CDM (chemically defined medium). (B) Loss of Hfq results in loss of biofilm production. (C) Quantification of PNAG in the crude polysaccharide extracts with a Congo red binding assay. WT, wild type; Δhfq, hfq mutant; Δhfq + hfq, complemented mutant; DspB, dispersin B. (A, B, and C) The data presented here are the means and standard errors (error bars) from four independent experiments. *, P < 0.0001. One-way ANOVA with Bonferroni's multiple-comparison test.
Biofilm formation by the wild-type, hfq mutant, and complemented mutant strains was assessed in BHIV after 6 h by using the crystal violet binding assay. The wild-type and complemented mutant strains produced 54-fold and 49-fold more biofilm, respectively, than the hfq mutant strain (Fig. 2B), clearly indicating that Hfq is required for biofilm formation in A. pleuropneumoniae. Similar results were observed when strains were cultured in CDM (data not shown).
Hfq regulates the levels of pgaC and dspB mRNA.Since the hfq mutant is defective in biofilm formation, we tested whether loss of Hfq affects the abundance of transcripts of genes known to be involved in biofilm formation in A. pleuropneumoniae. PgaC is an inner membrane-associated glycosyl transferase that synthesizes the major biofilm matrix polysaccharide PNAG (14). DspB is a PNAG-specific hexosaminidase that promotes the dispersal of cells from a biofilm (15). The levels of pgaC and dspB transcripts in the wild-type biofilm, wild-type supernatant, and hfq mutant were measured to test whether there are Hfq-dependent changes in the levels of these transcripts. Real-time PCR analyses revealed a 14-fold reduction in the level of pgaC mRNA and a 6-fold reduction in the level of dspB mRNA in the hfq mutant compared to the wild-type biofilm (Table 3). A 7-fold reduction in the level of pgaC mRNA and a 2-fold reduction in the level of dspB mRNA in the hfq mutant compared to those of the wild-type planktonic cells were also observed (Table 3). The level of pgaC and dspB transcripts in the wild-type planktonic cells was 2-fold lower than that in the wild-type biofilm. Collectively, these results demonstrate that the levels of both pgaC and dspB transcripts were reduced in the hfq mutant compared to the wild type and that these differences are specifically due to a loss of Hfq and cannot be attributed to the differences between biofilm and planktonic phases of growth (Table 3).
Real-time PCR analysis of pgaC and dspB transcripts
Hfq is required for PNAG production in A. pleuropneumoniae.To test whether the Hfq-mediated difference in the pgaC transcript level was reflected in the amount of cell-associated PNAG content, crude polysaccharide preparations from the wild-type, hfq mutant, and complemented mutant cells were used in a Congo red binding assay. The Congo red binding assay was validated by comparing crude polysaccharide extracts from cells grown in the absence or presence of dispersin B (23). Crude polysaccharide extracts from dispersin B-treated cells revealed very low levels of Congo red binding compared to untreated cells, indicating that Congo red bound primarily to PNAG under our assay conditions (Fig. 2C). The wild-type and complemented mutant strains produced ∼10-fold more PNAG than the hfq mutant (Fig. 2C). PNAG contents of biofilm and planktonic cells from the wild-type and complemented mutant strains were similar (Fig. 2C). Our results clearly indicate that Hfq is required for the production of PNAG in A. pleuropneumoniae.
Hfq is required for the production of cysteine synthase.To identify Hfq-regulated surface proteins that might affect biofilm formation, membrane fractions were isolated from the wild-type and hfq mutant strains grown to stationary phase. Outer membrane fractions were separated by SDS-PAGE and stained with a silver-based stain. A 34-kDa protein was missing in the hfq mutant compared to the protein profile of the wild-type strain. The differentially expressed protein was identified with a peptide mass fingerprint obtained from MALDI-TOF mass spectrometry as CysK, a cysteine synthase (Fig. 3A). Cysteine synthases are known to be involved in protection against oxidative stress and tellurite in bacteria. Although we did not identify any biofilm-related proteins from this experiment, the observation that the hfq mutant exhibited reduced CysK levels led us to probe the significance of Hfq in resistance to oxidative stress in A. pleuropneumoniae.
Hfq is involved in resistance against superoxide stress. (A) SDS-PAGE analyses of outer membrane proteins revealed that CysK (34 kDa) (arrowhead) was found only in the wild-type strain. Locations of molecular mass markers are indicated. The experiment was repeated twice, and an image of a representative gel is shown here. (B) Growth inhibition assays were performed with potassium tellurite and methyl viologen (superoxide generators). The experiment was repeated four times independently (*, P < 0.001; one-way ANOVA with Bonferroni's multiple-comparison test). WT, wild type; Δhfq, hfq mutant; Δhfq+hfq, complemented mutant.
Hfq is involved in protection against superoxide stress and tellurite.A. pleuropneumoniae is exposed to several forms of oxidative stress induced by organic hydroperoxides, hydrogen peroxide, and superoxide during infection of the pig lungs (26). The following agents recapitulate various types of oxidative stress encountered by this pathogen in the porcine lungs: cumene hydroperoxide, an organic hydroperoxide; hydrogen peroxide; methyl viologen; and potassium tellurite. Top agar growth inhibition assays were used to test the role of Hfq in resistance to oxidative stress. Both methyl viologen and potassium tellurite are known to generate superoxide radicals within bacterial cells (26, 31). The hfq mutant displayed increased sensitivity to both methyl viologen and potassium tellurite compared to the wild-type strain. This phenotype was able to be readily reversed by complementation with hfq (Fig. 3B). The wild-type and hfq mutant strains displayed similar sensitivity to cumene hydroperoxide and hydrogen peroxide (data not shown). These results indicate that Hfq is involved in protection against superoxide stress but not against cumene hydroperoxide or hydrogen peroxide stress in A. pleuropneumoniae.
Cells in a bacterial biofilm are exposed to conditions such as high cell density, nutrient deprivation, and accumulation of waste products. Therefore, cells in a biofilm respond differently to various stressors, including oxidative stress, than their planktonic counterparts (32). Since the hfq mutant was defective in biofilm formation, we tested whether the role of Hfq in superoxide stress resistance is independent of its role in PNAG-based biofilm formation. All strains were grown in the presence of dispersin B to eliminate PNAG-based biofilm, and top agar growth inhibition assays were performed. The hfq mutant remained hypersensitive to methyl viologen and potassium tellurite compared to the wild-type and complemented mutant strains under conditions nonpermissive for biofilm development (dispersin B treatment), indicating that the effect of Hfq on superoxide resistance in A. pleuropneumoniae is independent of its role in biofilm versus the planktonic lifestyle (data not shown).
Hfq is required for the virulence and fitness of A. pleuropneumoniae.To test the role of Hfq in the pathogenesis of porcine pleuropneumonia, pigs were experimentally infected with the wild type or hfq mutant or a mixed inoculum containing both the wild-type and hfq mutant strains in a 1:1 ratio. Pigs infected with the wild type and mixed inoculum (wild type and hfq mutant) manifested an increased respiratory rate, increased dyspnea, and increased signs of depression by 15 to 18 hpi (Table 4). One pig infected with the wild-type strain revealed blood-tinged nasal exudate at 16 hpi and was euthanized immediately. Pigs infected with either the wild type alone or mixed inoculum developed signs of moderate pleuropneumonia. Pigs infected with the hfq mutant alone, however, did not display the signs of pleuropneumonia (Table 4).
Clinical manifestation of pleuropneumonia in infected pigsa
Lungs from the pigs infected with either the wild type or a mixture of the wild type and hfq mutant revealed large areas of necrosis and hemorrhage concomitant with edema. Fibrin deposits were found on the pleural surface, indicating the onset of fibrinous pleuritis. Cut surfaces of the lungs were friable, and sequestration of the necrotic areas was evident. The lesions were distributed primarily on the dorsocaudal areas of the right caudal and left caudal lobes. There were no gross lesions in the lungs of pigs infected with the hfq mutant strain alone except for a small pneumonic area found in the left caudal lobe of one pig. Lungs from pigs infected with the wild-type strain revealed pneumonic changes in 28% of their lung surface area (Table 4). In the hfq mutant-infected group, pneumonic changes covered 2% of the lung area (Table 4). Pleuritis, a hallmark lesion of A. pleuropneumoniae infection, was found over 13% and 1% of the lung surface in the wild-type- and hfq mutant-infected groups, respectively (Table 4).
Histologic examination of lung sections from pigs infected with the wild-type strain revealed extensive areas of diffuse lobular pneumonia, with severe hemorrhage and necrosis sequestered by streaming neutrophils, as well as fibrinous pleuritis. Accumulation of bacteria, fibrin, blood, and necrotic debris, including lysed neutrophils, was evident in the affected areas. The alveolar septa were necrotic, along with a loss in nuclear detail of pneumocytes and a loss of delineation between alveoli. Vascular thrombosis was widespread within the affected areas, and a large number of polymorphonuclear cells were observed within blood vessels. Edema and hemorrhage were evident, resulting in fluid-filled alveoli and bronchi. In contrast, lung sections from the pigs infected with the hfq mutant were indistinguishable from those of PBS-inoculated controls. Lung sections revealed normal lung architecture and normal pleura, with well-defined alveolar septa and neutrophils noticed occasionally.
Bacterial burden in lung samples was determined for pigs infected with either the wild-type strain or the hfq mutant. The wild-type strain colonized lung tissue at high levels, whereas the hfq mutant exhibited poor colonization potential (Fig. 4A). This result indicates that Hfq acts as a virulence factor in A. pleuropneumoniae. We did not attempt to infect pigs with the complemented mutant strain because of poor plasmid maintenance in vivo. Determination of competitive indices (CI) is a sensitive technique to compare the fitnesses of two strains under a given condition (33). Since pigs used in our experiments are not inbred, it is important to compare the biological effects of a loss-of-function mutation within a single animal. Briefly, the CI was calculated as the ratio of the mutant to the wild type at any time divided by the ratio of the mutant to the wild type in the inoculum. If the hfq mutant exhibits a fitness defect, the CI value will be <1. The CI during coculture of the wild-type and hfq mutant strains in BHIV was close to 1, indicating that the hfq mutant does not exhibit a fitness defect in vitro. The median competitive index determined from data for BALF and all six lung lobes from 4 pigs was 2 × 10−5, indicating that Hfq is a critical contributor to fitness during infection (Fig. 4B). Notwithstanding the fact that the wild-type strain induces considerable tissue damage, the hfq mutant exhibits a significant fitness defect in vivo during cochallenge infections. Collectively, these data (Fig. 4) demonstrate that Hfq not only is critical for survival but also promotes virulence of A. pleuropneumoniae in an experimental model of porcine pleuropneumonia.
Virulence and fitness of an hfq mutant strain during experimentally induced porcine pleuropneumonia. (A) Colonization of pulmonary tissue by wild-type and Δhfq strains during independent infection. Data from all 6 lung lobes from 3 pigs infected with either strain are represented here (*, P < 0.001; Mann-Whitney U test). The limit of detection is indicated by the dotted line. (B) Hfq is required for in vivo fitness. Competitive indices (CI) were determined as described in Materials and Methods for pulmonary samples obtained from pigs cochallenged with a mixed inoculum. Each data point represents the CI for a sample derived from an individual animal. CI values of less than 0 indicate a fitness defect. BALF, bronchoalveolar lavage fluid; RCR, right cranial lobe; RM, right medial lobe; RCD, right caudal lobe; LCD, left caudal lobe; LCR, left cranial lobe; ACC, accessory lobe.
DISCUSSION
Previous work on A. pleuropneumoniae in our laboratory led to the identification of hfq as an in vivo-induced gene that is upregulated during growth in CDM lacking BCAAs (3–5). Homologs of hfq are found in all members of the Pasteurellaceae family whose genomes have been sequenced, but the function of Hfq in this group of important pathogens has not been reported. To investigate the role of the predicted Hfq homolog in A. pleuropneumoniae, an hfq mutant strain was constructed (Fig. 1D). Loss of Hfq has variable effects on growth in other bacteria, ranging from extended doubling time to no effect on growth in laboratory medium (18). We observed a slightly reduced growth rate in the A. pleuropneumoniae hfq mutant strain (Fig. 1E). Nonetheless, the hfq mutant exhibited wild-type levels of fitness in vitro.
A. pleuropneumoniae strains are capable of forming biofilms on abiotic surfaces (14). We have previously reported that the biofilm formation phenotype is more common among field isolates of A. pleuropneumoniae than laboratory strains (13). PNAG is a major surface polysaccharide that acts as the extracellular matrix in A. pleuropneumoniae biofilms (14). PNAG is found in both Gram-positive and Gram-negative bacteria, such as Staphylococcus aureus and E. coli (14). However, only a few organisms, including A. pleuropneumoniae, encode a PNAG-specific hexosaminidase, DspB (15), which is involved in the dispersal of cells from a mature biofilm. Here, we demonstrate that Hfq affects the levels of pgaC transcript (Table 3), whose protein product is involved in PNAG biosynthesis, therefore affecting the PNAG content in the extracellular matrix (Fig. 2C). Potential regulatory proteins and cues that affect biofilm formation in A. pleuropneumoniae have been identified. Buettner et al. reported that an A. pleuropneumoniae serotype 7 arcA mutant strain was defective in autoaggregation and biofilm formation on glass surfaces, suggesting that ArcA positively regulates biofilm formation (34). ArcAB is a two-component global regulatory system involved in the adaptation to growth under anaerobic conditions and is required for full virulence of A. pleuropneumoniae and for persistence in chronic infection (34). A luxS mutant strain of A. pleuropneumoniae serotype 1, defective in the biosynthesis of quorum-sensing signal autoinducer 2, was shown to produce more biofilm than the parent strain (35). Zinc limitation has been reported to promote biofilm formation in A. pleuropneumoniae (36). While these studies have identified potential regulators and environmental conditions that affect biofilm formation, they did not identify molecular mechanisms that govern the biogenesis of biofilm in A. pleuropneumoniae.
In A. pleuropneumoniae serotype 1 strain 4074 (ATCC 27088), sigma E (σE; extracytoplasmic stress response sigma factor encoded by rpoE) and histone-like nucleoid structuring protein (H-NS) regulate biofilm formation at the transcriptional level by modulating the expression of the pgaABCD operon (37). Bosse et al. reported that a mutation in rseA (an anti-sigma factor that sequesters σE in the inner membrane) and hns genes induced biofilm formation in an otherwise non-biofilm-forming strain (37). Binding sites for H-NS (three) and σE (one) were identified in the promoter region of the pga operon, lending mechanistic insights into the transcriptional regulation of this operon in A. pleuropneumoniae (37). Therefore, σE and H-NS act as positive and negative transcriptional regulators of the pgaABCD operon expression, respectively.
Homologs of both hfq and the genes involved in PNAG biosynthesis are found in several bacteria, including Acinetobacter baumannii, A. pleuropneumoniae, Aggregatibacter actinomycetemcomitans, Bordetella pertussis, E. coli, Staphylococcus aureus, and Yersinia pestis (17, 18, 38). In Y. pestis, mutants that lack Hfq exhibit enhanced biofilm-forming capacity compared to that of parental strains and Hfq indeed acts as a repressor of PNAG production by regulating the intracellular levels of cyclic di-GMP (c-di-GMP) (17). By simultaneously increasing the concentration of HmsP, a phosphodiesterase, and reducing the level of HmsT, a diguanylate cyclase, Hfq reduces the intracellular levels of c-di-GMP in Y. pestis (17). In contrast, in A. pleuropneumoniae, Hfq appears to positively regulate PNAG production, although it is possible that Hfq may act indirectly by negatively regulating a negative regulator of PNAG biosynthesis. Bioinformatic searches of the genomes of the members of Pasteurellaceae, including A. pleuropneumoniae, do not reveal the presence of genes involved in the biosynthesis and/or degradation of c-di-GMP. Therefore, it is likely that c-di-GMP levels are not involved in the regulation of PNAG biosynthesis by Hfq in A. pleuropneumoniae. It will be interesting to probe the potential role of Hfq in regulating PNAG production in these diverse pathogens. Such studies will expand our knowledge of the regulatory proteins governing the development of medically important biofilms, and inhibitors of regulatory proteins, such as Hfq, may potentially be developed to curtail biofilm formation.
Bacterial pathogens, such as A. pleuropneumoniae, encounter oxidative stress in the form of oxygen radicals generated by professional phagocytes during infection. Hfq has been implicated in resistance to oxidative stress in bacteria (18). Brucella abortus hfq mutants exhibit poor expression of the sodC gene, which encodes a periplasmic Cu-Zn superoxide dismutase (39). Hfq mutants of Burkholderia cepacia (40), Moraxella catarrhalis (41), Neisseria meningitidis (42), Sinorhizobium meliloti (43), and uropathogenic E. coli (16) display enhanced sensitivity to methyl viologen, a redox-cycling agent which generates superoxide within bacterial cells. In contrast, hfq mutants of S. aureus (44) and Vibrio parahaemolyticus (45) exhibit higher resistance to oxidative stress induced by hydrogen peroxide and superoxide than does the parent strain. Here, we report that Hfq plays a role in protection against superoxide stress in A. pleuropneumoniae (Fig. 3B).
Our effort to identify Hfq-regulated membrane proteins that might be involved in biofilm formation led us to the finding that the cysteine synthase (CysK) protein is expressed at undetectable levels in the hfq mutant compared to the wild type (Fig. 3A). Copurification of CysK in the membrane fractions may represent a contamination in the membrane preparations or CysK may be another member of a growing family of moonlighting proteins. Nevertheless, CysK was consistently identified in membrane fractions of wild-type A. pleuropneumoniae. Cysteine synthases are known to be involved in resistance to oxidative stress and tellurite in bacteria (41, 46, 47). Cysteine is a precursor for glutathione, whose primary role is to maintain the redox status of the bacterial cytoplasm. Furthermore, cysteine residues are critical for the repair of an iron-sulfur cluster containing enzymes that are damaged by reactive oxygen species. A. pleuropneumoniae is exposed to several potent inducers of oxidative stress, including organic hydroperoxides, hydrogen peroxide, and superoxide anion, during infection of porcine lungs (26). Several genes involved in oxidative stress resistance, including ohr, encoding an organic hydroperoxide reductase, tehB, encoding a putative S-adenosyl methyltransferase involved in resistance to tellurite, and sodC, encoding a periplasmic copper-zinc superoxide dismutase, were identified as in vivo-induced genes in A. pleuropneumoniae (4, 26).
Here, we have provided evidence that Hfq is required for optimal resistance against oxidative stress induced by methyl viologen in A. pleuropneumoniae (Fig. 3B). In addition to affecting the levels of CysK, Hfq may also regulate the expression of superoxide dismutases. Sequenced A. pleuropneumoniae genomes encode two superoxide dismutases, sodA and sodC. Hfq may alter the expression of these key enzymes involved in detoxifying superoxide. In E. coli, Hfq is known to regulate the expression of sodB mRNA at the posttranscriptional level (48); however, A. pleuropneumoniae lacks a sodB homolog.
Genes conferring resistance to tellurite have been found in many bacterial pathogens, but tellurite is not found in significant amounts within mammalian hosts (31). Therefore, the true role of these tellurite resistance genes in the infection biology of bacterial pathogens remains to be clarified. The mechanisms involved in tellurite toxicity are also not fully understood; one known function of tellurite in bacteria is to trigger the production of superoxide (31). In Haemophilus influenzae, the tehB gene is essential for resistance against tellurite and oxidative stress and for virulence in a mouse model of invasive disease (49). TehB was also identified as an in vivo-induced gene in A. pleuropneumoniae (4). Together, these observations suggest that tellurite resistance genes might not only play a role in tellurite resistance but also affect oxidative stress resistance and virulence of bacterial pathogens. The role of Hfq in resistance to tellurite is a novel phenotype uncovered by this study. Cysteine synthases are known to be involved in bacterial resistance against tellurite, although the exact role of these enzymes in the protection against tellurite toxicity is not clear (46, 47). One possible mechanism may be the requirement of cysteine for production of molecules, including glutathione, that are involved in protection against oxidative stress. Expression of the cysK gene cloned from Bacillus stearothermophilus in E. coli confers ∼10-fold-higher resistance, indicated by changes in MIC, against tellurite (46). Furthermore, the cysK gene is upregulated when grown in medium containing tellurite, suggesting that cysteine biosynthesis is important for alleviating the toxic effects of tellurite. Since an A. pleuropneumoniae hfq mutant produced low levels of CysK, the effect of Hfq on tellurite resistance was tested and we found that Hfq is required for optimal resistance to tellurite in A. pleuropneumoniae (Fig. 3B).
Hfq is known to affect virulence in a diverse array of bacterial pathogens in various infection models (18). However, the role of Hfq in the fitness and virulence of a pulmonary pathogen during experimentally induced pneumonia has not been reported. Utilizing a natural host infection model, we have demonstrated that Hfq is essential for survival and virulence in porcine lungs (Fig. 4 and Table 4). The virulence and fitness defect observed in the A. pleuropneumoniae hfq mutant strain may be due to failure to produce PNAG and/or cysteine synthase and/or as-yet-unidentified factors.
PNAG contributes to the virulence of many bacterial pathogens. Polysaccharide intercellular adhesin, a polymer composed of PNAG, has been demonstrated as an important virulence factor in S. aureus and Staphylococcus epidermidis, in which it is required for the production of stable biofilm on catheters and implants (50). PNAG is also critical for the survival and virulence of uropathogenic E. coli in a septicemic mouse model of infection (51). Utilizing a rat model of Bordetella colonization, PNAG has been shown to promote colonization of B. pertussis in the upper respiratory tract (52). Further, in A. actinomycetemcomitans, a periodontal pathogen that is closely related to A. pleuropneumoniae, PNAG is required for survival within human and mouse macrophages (23). While it has been shown that A. pleuropneumoniae produces PNAG-based biofilms, the role of biofilm in the pathobiology of porcine pleuropneumonia remains to be elucidated. It is clear that production of biofilms is not essential for experimental infection models that utilize the intratracheal route, since A. pleuropneumoniae ATCC 27088, a biofilm-negative strain, is highly virulent (6, 27, 53). However, it is possible that the infectious dose for a biofilm-positive strain is lower than that for an isogenic biofilm-negative strain or that biofilm formation might be critical for colonization and infection via an intranasal route of infection.
Infection with live A. pleuropneumoniae leads to cross-protection against infection with heterologous serotypes, whereas vaccination with whole-cell bacterin preparations offers limited cross-protection (54). Given the morbidity and mortality caused by A. pleuropneumoniae, at least in the developing world, there is a clear need for better vaccines to protect against this economically important disease. Since the A. pleuropneumoniae hfq mutant is severely attenuated in pigs, the potential of this mutant as a live-attenuated vaccine candidate should be explored.
In summary, the function of an in vivo-induced gene, hfq, in the pathobiology of A. pleuropneumoniae was investigated, and Hfq was found to regulate PNAG-based biofilm formation and is required for resistance to oxidative stress. Because biofilm formation and oxidative stress resistance have been proposed to be important for the virulence of A. pleuropneumoniae, the effect of Hfq on virulence was tested and the hfq mutant was found to be defective in survival during infection of porcine lungs. Since Hfq is a pleiotropic regulator, expression of other virulence genes may be affected by Hfq. Comparative studies on the transcriptome and proteome in the hfq mutant and the wild type will shed further light on the scope of Hfq-mediated posttranscriptional regulation in this important bacterial pathogen.
ACKNOWLEDGMENTS
This work was supported by the Center for Microbial Pathogenesis and the Genetics Research Fund from the College of Veterinary Medicine at Michigan State University. S.S. is a recipient of the Rudolf Hugh Fellowship in the Department of Microbiology and Molecular Genetics at Michigan State University.
FOOTNOTES
- Received 27 March 2013.
- Returned for modification 25 April 2013.
- Accepted 28 May 2013.
- Accepted manuscript posted online 3 June 2013.
- Copyright © 2013, American Society for Microbiology. All Rights Reserved.
