Host-Pathogen Interactions of Actinobacillus pleuropneumoniae with Porcine Lung and Tracheal Epithelial Cells

Abstract

Host-pathogen interactions are of great importance in understanding the pathogenesis of infectious microorganisms. We developed in vitro models to study the host-pathogen interactions of porcine respiratory tract pathogens using two immortalized epithelial cell lines, namely, the newborn pig trachea (NPTr) and St. Jude porcine lung (SJPL) cell lines. We first studied the interactions of Actinobacillus pleuropneumoniae, an important swine pathogen, using these models. Under conditions where cytotoxicity was absent or low, we showed that A. pleuropneumoniae adheres to both cell lines, stimulating the induction of NF-κB. The NPTr cells consequently secrete interleukin 8, while the SJPL cells do not, since they are deprived of the NF-κB p65 subunit. Cell death ultimately occurs by necrosis, not apoptosis. The transcriptomic profile of A. pleuropneumoniae was determined after contact with the porcine lung epithelial cells by using DNA microarrays. Genes such as tadB and rcpA, members of a putative adhesin locus, and a gene whose product has high homology to the Hsf autotransporter adhesin of Haemophilus influenzae were upregulated, as were the genes pgaBC, involved in biofilm biosynthesis, while capsular polysaccharide-associated genes were downregulated. The in vitro models also proved to be efficient with other swine pathogens, such as Actinobacillus suis, Haemophilus parasuis, and Pasteurella multocida. Our results demonstrate that interactions of A. pleuropneumoniae with host epithelial cells seem to involve complex cross talk which results in regulation of various bacterial genes, including some coding for putative adhesins. Furthermore, our data demonstrate the potential of these in vitro models in studying the host-pathogen interactions of other porcine respiratory tract pathogens.

Porcine respiratory diseases have heavily impacted the economy of the pig rearing industry worldwide. Actinobacillus pleuropneumoniae, an exemplar of these porcine respiratory pathogens, causes porcine pleuropneumonia, a very contagious and often fatal disease characterized by necrotic and hemorrhagic lung lesions, coughing, and severe respiratory distress. Fifteen serotypes of this gram-negative facultative anaerobic bacteria are presently known based on surface polysaccharides (56). The virulence of this pathogen is accomplished with the help of many factors, including exotoxins, endotoxin, capsule polysaccharides, adhesins, and outer membrane proteins, such as iron acquisition systems.

The four pore-forming exotoxins of A. pleuropneumoniae, called Apx, are cytolytic and/or hemolytic (20, 51). In fact, the virulence of the different serotypes coincides greatly with the presence of the Apx toxins, particularly ApxI and ApxII. Serotypes 1, 5, 9, and 11 are known to be especially virulent, and all express both ApxI and ApxII (19).

As demonstrated by Jacques et al., lipopolysaccharides (LPS) are the major molecules responsible for adhesion, principally the core oligosaccharide region (45, 48). However, other putative adhesins have also been described, such as type IV fimbriae expressed under specific conditions in different serotypes (63), a 60-kDa collagen-binding protein (14), a 55-kDa outer membrane protein (61), and an autotransporter protein (4).

A close relative of A. pleuropneumoniae, Aggregatibacter (Actinobacillus) actinomycetemcomitans, was found to be invasive of the human KB cell line and primary gingival cells (17). The invasiveness of this strain has been shown to be related to the colonial morphology, since a switch from a rough to a smooth morphology leads to the loss of its invasive capacity. Haemophilus parasuis has also been shown to be invasive of porcine brain microvascular endothelial cells (43, 60). Moreover, A. actinomycetemcomitans has been demonstrated to induce cell death by apoptosis of numerous cell types, while the cytolethal distending toxin of Haemophilus ducreyi has been shown to induce apoptosis of Jurkat T cells (23, 37). Adherence, invasion, toxin secretion, and other mechanisms involved in the pathogenesis of Pasteurellaceae lead to changes in cellular processes, including the induction of nuclear factors and cytokine production. In fact, A. pleuropneumoniae stimulates the production of proinflammatory cytokines such as interleukin 1β (IL-1β), IL-8, and tumor necrosis factor alpha (TNF-α), which are detected in alveolar fluid and tissue lesions (56). Likewise, a study by our group demonstrated that the production of IL-6, TNF-α, IL-1β, MCP-1, and IL-8 by porcine alveolar macrophages is induced by purified serotype 1 A. pleuropneumoniae LPS and by heat-killed bacteria (48). IL-8, a neutrophil chemoattractant, is of particular interest, since neutrophil accumulation at the infection site is characteristic of porcine pleuropneumonia (56).

Changes in bacterial gene expression also occur during infection. Studies have been conducted to investigate the gene expression of A. pleuropneumoniae under conditions mimicking that of the host. A study by our group used microarray technology to detect changes in gene expression of A. pleuropneumoniae serotype 1 grown under iron-restricted conditions (13). In this study, many genes involved in iron acquisition were shown to be upregulated, while genes involved mainly in energy metabolism were downregulated. In vivo studies based on selective capture of transcribed sequences, in vivo expression technology, or signature-tagged mutagenesis (3, 4, 22, 31, 53) have allowed the detection of adhesin and toxin genes and of genes involved in metabolism, stress, regulation, and transport.

Epithelial cells play an important role as the interface between the host mucosal surfaces and the surrounding environment and are the initial site of colonization for most bacterial pathogens. Two porcine respiratory tract epithelial cell lines have been established and described in the literature, namely, the newborn pig trachea (NPTr) (15) and St. Jude porcine lung (SJPL) (52) cell lines. The NPTr cell line was established from a 2-day-old piglet from a pathogen-free herd, while the SJPL cell line was spontaneously established from the lung of a normal 4-week-old female Yorkshire pig (15, 52).

The use of these cell lines has the possibility of generating a great amount of information on the infection mechanism of A. pleuropneumoniae, as well as that of other swine bacterial or viral respiratory tract pathogens. Consequently, we developed in vitro models using these cell lines and investigated host-pathogen interactions, including adherence, invasion, and bacterial transcriptomic profile, as well as cell death, nuclear factor expression, and cytokine production by the epithelial cells. This is the first report of models in which immortalized cell lines are used to study interactions of A. pleuropneumoniae with respiratory tract epithelial cells of porcine origin.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

All strains used in this study are listed in Table 1. All A. pleuropneumoniae strains and the Pasteurella multocida capsular type A and D strains were grown in brain heart infusion (BHI) broth and/or agar (Gibco, Burlington, VT) supplemented with 15 μg/ml NAD at 37°C in 5% CO₂. The Actinobacillus suis strain was grown under the same conditions, with the addition of 25 μg/ml nalidixic acid and 5 μg/ml chloramphenicol. Both H. parasuis strains were grown on pleuropneumonia-like organism medium broth and on chocolate agar at 37°C without CO₂.

TABLE 1.

Bacterial strains used in the present study

Strain	Serotype	Source or reference
A. pleuropneumoniae S4074	1	K. R. Mittala
A. pleuropneumoniae L20	5b	K. R. Mittal
A. pleuropneumoniae WF83	7	K. R. Mittal
A. pleuropneumoniae FMV91-6514	1 (rough field strain)	32
A. pleuropneumoniae 05-4817	5a (field strain)	K. R. Mittal
A. pleuropneumoniae 05-6501	5b (field strain)	K. R. Mittal
A. pleuropneumoniae 05-3695	7 (field strain)	K. R. Mittal
H. parasuis Nagasaki	5	M. Gottschalka
H. parasuis 29755	5	E. Thackerb
A. suis H91-0380	O2/K2	J. MacInnesc
P. multocida 88-761	A	K. R. Mittal
P. multocida 1703	D	K. R. Mittal

^aFaculté de Médecine Vétérinaire, Université de Montréal, St.-Hyacinthe, Canada.

^bFaculty of Veterinary Medicine, Iowa State University.

^cDepartment of Pathobiology, Ontario Veterinary College, University of Guelph, Guelph, Canada.

Cell culture.

The NPTr cell line (Instituto Zooprofilattico Sperimental, Brescia, Italy) (15) was grown at 37°C in 5% CO₂ in minimum essential medium (Gibco) supplemented with 10% fetal bovine serum (Gibco) and 1% sodium pyruvate (Gibco). The SJPL cell line (St. Jude Children's Hospital, Memphis, TN) (52) was grown at 37°C in 5% CO₂ in Dulbecco's modified Eagle's medium (DMEM) (Gibco), supplemented with 10% fetal bovine serum, 1% sodium pyruvate, and 1.5% minimal essential medium nonessential amino acids solution (Gibco). Both cell lines were tested by PCR using porcine-specific primers, and amplicons were sequenced to ensure their origin (41).

Cytotoxicity detection assay.

The cellular cytotoxicity was measured in the different assays using the lactate dehydrogenase (LDH)-measuring CytoTox 96 nonradioactive cytotoxicity assay (Promega, Madison, WI) as prescribed by the manufacturer. Noninfected cells were used as a negative control, while total lysis of cells by a treatment with 2% Triton X-100 represented the 100%-cytotoxicity positive control. Optical densities at 490 nm were measured with a Power Wave X340 (Biotek Instruments Inc, Winooski, VT) microplate reader and used to calculate the percentage of cytotoxicity.

Apoptosis detection assays.

Apoptosis assays were performed using the cell death detection enzyme-linked immunosorbent assay (ELISA) (Roche, Laval, Québec, Canada) and the caspase-3 Western detection kit (Cell Signaling Technology Inc., Beverly, MA) as prescribed by the manufacturers. A bacterial suspension at an optical density at 600 nm (OD₆₀₀) of 0.6 was added to a confluent monolayer of cells at a multiplicity of infection (MOI) of 10:1 and incubated for 3 h at 37°C in 5% CO₂. Uninfected cells were used as negative controls, and cells treated with 20 μg/ml camptothecin (Sigma) for 4 h at 37°C in 5% CO₂ were used as positive controls. Following the infection, the supernatant and adherent cells were recovered. Plates for the cell death detection ELISA were read at 405 nm in a Power Wave X340 (Biotek Instruments Inc.) microplate reader. For the caspase-3 Western blots, the samples were loaded on a 12% (wt/vol) polyacrylamide gel, migrated at 100 V, and transferred to a nitrocellulose membrane (Bio-Rad, Hercules, CA). The membrane was blocked for 1 h at room temperature in 2% skim milk and then incubated overnight (O/N) at 4°C with polyclonal rabbit antibodies against cleaved caspase-3 and caspase-3. Membranes were washed three times in Tris-buffered saline and incubated with mouse anti-immunoglobulin G antibodies conjugated with horseradish peroxidase for 1 h at ambient temperature and revealed with 3,3′,5,5′-tetramethylbenzidine (Sigma).

Microscopy.

Cells were seeded to semiconfluence in four-well LabTekII chamber microscopy slides (Nunc, Naperville, IL) and incubated O/N. One ml of a 2.5 × 10⁶-CFU/ml suspension of A. pleuropneumoniae S4074 was added, and slides were then incubated for 2 h at 37°C in 5% CO₂. Following four washes with Dubelcco's phosphate-buffered saline (DPBS) (Gibco), cells were fixed for 10 min in methanol and stained for 30 min with Giemsa stain (Sigma). Four washes with DPBS were performed, and the slides were left to dry. Noninfected cells were also stained as controls. Observation was done at a ×1,000 magnification on a Leica DMR microscope.

Adherence assay.

To quantify the adherence of the different strains on both cell lines, 2.5 × 10⁵ epithelial cells were seeded into wells of 24-well tissue culture plates (Sarstedt, Numbrecht, Germany) and incubated O/N. Bacteria from an O/N culture grown at an OD₆₀₀ of 0.6 were resuspended in the adequate cell culture medium to a concentration of 2.5 × 10⁶ CFU/ml. One ml of this suspension was added to each well at an MOI of 10:1, and the plates were incubated from 1 to 3 h. Nonadherent bacteria were removed by washing four times with DPBS. Cell with adherent bacteria were released from the wells by adding 100 μl of 1× trypsin-EDTA (Gibco) and resuspended in 900 μl DPBS buffer. Serial dilutions were performed, with plating on agar plates to determine the number of bacteria that adhered to the epithelial cells.

Statistical analysis.

Data were log transformed prior to analysis. A two-way analysis of variance, with the cell line and the bacterial strain as factors, was used at each time point separately. The level of statistical significance was set at 0.05 throughout. Analyses were carried out using the SAS software program, version 9.1 (SAS, Cary, N C). Contrasts were used to examine differences between pairs of means.

Invasion assay.

Epithelial cells (2.5 × 10⁵) were seeded into wells of 24-well tissue culture plates and incubated O/N. One ml of a 2.5 × 10⁷-CFU/ml bacterial suspension was added to the wells (MOI of 100:1). Plates were incubated for 1 to 3 h. Nonadherent bacteria were removed by washing four times with DPBS buffer. One ml of DMEM containing 100 μg/ml of gentamicin was added to each well, followed by a 1-h incubation period at 37°C in 5% CO₂. Killed bacteria were removed by washing twice with DPBS buffer. Cells were then lysed with 100 μl of sterile distilled H₂O, and samples were plated on agar plates and incubated O/N.

Protein profiling of SJPL and NPTr cells in contact with A. pleuropneumoniae.

Protein profiling of the cells was performed to detect differentially expressed proteins. Two T175 flasks were seeded with a confluent monolayer of cells, and 25 ml of DMEM culture medium with or without 1 × 10⁷ CFU/ml of bacteria grown at an OD₆₀₀ of 0.6 was added to the flasks. Flasks were incubated for 3 h at 37°C in 5% CO₂ and then washed three times with DPBS, and 500 μl of a lysis solution (20 mM morpholinepropanesulfonic acid, 0.5% triton X-100, and protease inhibitors) was added. Cells were removed from the flasks and transferred to microcentrifuge tubes on ice. Sonication at ∼180 J was performed using an ultrasonic processor (Cole-Parmer, Vernon Hills, IL) in order to lyse the cells. The samples were then ultracentrifugated at 50,000 rpm for 30 min in a Sorvall RC M100 ultracentrifuge. The supernatant was preserved and analyzed for protein concentration using the Bradford assay (Bio-Rad). Samples were diluted to 2 mg/ml and frozen at −80°C. The samples were then analyzed using the Kinexus antibody microarray, which tracks changes in protein expression of 608 different cell signaling proteins in duplicate, including phosphorylation sites and kinases (Kinex Bioinformatics Inc., Vancouver, British Columbia, Canada). Fifty μg of proteins from both the untreated (control) and treated cells were labeled with the same proprietary fluorescent dye. Each sample was separately applied to opposite sides of the antibody microarray, which contains a dam to prevent mixing of the samples. Following incubation of samples with the Kinex chip, the unbound proteins were washed away and the chips were scanned with a Perkin-Elmer ScanArray Express reader. Image analysis of the TIF files that were produced was performed using the ImaGene 7.0 software program from BioDiscovery (El Segundo, CA). Qualitative and semiquantitative analyses of the expression and phosphorylation states of the cell signaling proteins were performed.

EMSA for detection of NF-κB and AP-1.

Cells were infected at an MOI of 1:10 for 30 min, 1 h, or 3 h with A. pleuropneumoniae grown at an OD₆₀₀ of 0.6. Uninfected cells were used as a control. Cell stimulation was terminated by the addition of cold phosphate-buffered saline. Six μg of nuclear proteins, extracted as described by Blanchette et al. (9), were incubated for 20 min at room temperature in 1 μl binding buffer (100 mM HEPES [pH 7.9], 40% [vol/vol] glycerol, 10% [wt/vol] Ficoll, 250 mM KCl, 10 mM dithiothreitol, 5 mM EDTA, 250 mM NaCl, and 10 mg/ml bovine serum albumin), and 200 ng/μl poly(dI-dC)-0.02% bromophenol blue with 1 μl of the labeled oligonucleotide containing a consensus sequence of NF-κB/c-Rel homodimeric and heterodimeric complexes (5′AGTTGAGGGGACTTTCCCAGGC-3′; Santa Cruz Biotechnology, Santa Cruz, CA) or of AP-1 complexes (5′CGCTTGATGACTCAGCCGGAA-3′; Santa Cruz Biotechnology) which were previously labeled using T4 polynucleotide kinase and [γ-³²P]dATP (GE Healthcare, Piscataway, NJ). After incubation, DNA-protein complexes were resolved by electrophoresis in a 5% (wt/vol) nondenaturing polyacrylamide gel. Subsequently gels were dried and autoradiographed. The nonspecific probes (SP-1 [5′ATTCGATCGGGGCGGGGCGAG-3′] used to confirm the specificity of the DNA/nuclear protein reactions were synthesized in our laboratory. Cold competitor assays were conducted by adding a 100-fold molar excess of homologous unlabeled oligonucleotide for NF-κB or AP-1 and noncompetitor SP-1. For supershift assays, 2 μg of nuclear proteins were incubated with binding buffer, poly(dI-dC), 0.02% bromophenol blue, labeled ologonucleotide, and 4 μg specific antibody (anti-p50 and anti-p65N, both from Santa Cruz Biotechnology) at room temperature for 1 h and complexes resolved on a standard nondenaturing 5% (wt/vol) polyacrylamide gel. For the IRAK inhibition assays, the cells were preincubated for 1 h with 50 μM IRAK 1/4 inhibitor (Calbiochem, Darmstadt, Germany) before addition of the bacteria for an incubation of 3 h. A nuclear protein electrophoretic mobility shift assay (EMSA) was then performed as mentioned above.

Stimulation of cytokine production.

Induction assays were performed with both cell lines as described by Ramjeet et al. (48). Briefly, 1 ml culture medium containing 1 × 10⁹ CFU A. pleuropneumoniae S4074, heat killed at 60°C for 45 min, was added to wells of 24-well tissue culture plates containing a monolayer of epithelial cells. The plates were incubated from 30 min to 48 h at 37°C in 5% CO₂. The supernatant was then collected and analyzed by ELISA to detect the amounts of IL-1β, TNF-α, IL-6, and IL-8 produced by the stimulated epithelial cells as described by Ramjeet et al. (48). The stimulation tests were also performed using 35 to 3,500 endotoxin units/ml of A. pleuropneumoniae serotype 1 S4074 LPS. These LPS concentrations were shown to induce a response in porcine alveolar macrophages (48). As a control, NF-κB inhibition assays were performed where cells were preincubated for 1 h at 37°C in 5% CO₂ with 25 μg/ml caffeic acid phenethyl ester (Sigma) before addition of the bacteria for an additional incubation of 12 h. The supernatant was collected and tested by ELISA for the IL-8 concentration.

RNA extractions for microarray experiments.

Monolayers of SJPL cells in T175 flasks were infected for 3 h with 100 μl of an A. pleuropneumoniae culture with an OD₆₀₀ of 0.6 (MOI of 10:1). Planktonic bacteria were harvested from the culture supernatant, while adherent bacteria were harvested with the epithelial cells following two washes in phosphate-buffered saline buffer. Ice-cold RNA degradation stop solution (95% ethanol, 5% buffer-saturated phenol) was added to all samples at a 1:10 (vol/vol) ratio, and samples were then frozen at −80°C after a 5-min centrifugation at 4,000 × g. The isolation of bacterial RNA was carried out using the Qiagen RNeasy minikit with an in-column DNase treatment, as prescribed by the manufacturer. RNA was further treated with Turbo DNase (Ambion, Tx) as prescribed by the manufacturer.

Transcriptomic microarray experiments.

The AppChip1 design was part of the A. pleuropneumoniae 5b L20 genome sequencing project led by the team of John Nash (NRC-IBS, Ottawa, Canada). The microarrays used in this study contain PCR amplicons representing all the open reading frames that were identified in the A. pleuropneumoniae 5b L20 genome (13, 18). RNA was reverse transcribed to cDNA using Invitrogen Superscript II reverse transcriptase. cDNA was indirectly labeled with monofunctional Cy3 or Cy5 N-hydroxysuccinimide ester (Amersham Biosciences, Piscataway, NJ). Samples from planktonic growth or adhesion versus growth in DMEM medium were combined and cohybridized on the microarray. Four hybridizations were conducted for each condition, including a pair of microarrays for which the Cy3 and Cy5 dyes were swapped. Microarray analysis was carried out using the TM4 software suite (The Institute for Genomic Research) and the significant-analysis-of-microarrays algorithm, using a false-discovery-rate value of 0% (50). For the planktonic and adhesion experiments, the Cy5 signal was compared to the Cy3 signal in order to obtain a list of significantly differentially expressed genes. In order to obtain the list of differentially expressed genes between planktonic growth and adhesion, log₂ ratios were compared in TM4, also using the significant-analysis-of-microarrays algorithm. Functional classification was performed using The Institute for Genomic Research's Comprehensive Microbial Resource (47).

Microarray data accession number.

Data files have been submitted to the Gene Expression Omnibus (GEO accession number GSE12009).

RESULTS

Effect of bacterial infection on viability of epithelial cells.

A. pleuropneumoniae expresses potent exotoxins. To ensure cell viability in our experiments, cell death assays were performed at different MOIs (10:1, 100:1, and 1,000:1) and with different incubation periods (0.5 h, 1 h, 2 h, 3 h, and 4 h) with A. pleuropneumoniae strain S4074 representing serotype 1. LDH cytotoxicity assays to detect necrosis were first performed. Important cytotoxicity was observed after an incubaction of 4 h (up to 80% at an MOI of 10:1) or with an MOI of 100:1 (up to 40% at 3 h) (data not shown). An MOI of 10:1 and incubation times not surpassing 3 h were chosen for subsequent tests as a result of low cytotoxicity under these conditions (Fig. 1). Since cell death by apoptosis cannot be detected by the LDH test, apoptosis assays were also preformed. An ELISA assay detecting DNA degradation and a Western blot assay detecting caspase-3 activation were carried out and demonstrated that neither cell line undergoes apoptosis after 3 h of bacterial infection at an MOI of 10:1 (data not shown).

FIG. 1.

SJPL (filled bars) and NPTr (empty bars) cells were assessed for cytotoxicity following an infection with A. pleuropneumoniae strain S4074 at an MOI of 10:1.

Adherence of A. pleuropneumoniae.

Standardization of adherence models was performed using both cell lines and the A. pleuropneumoniae reference strain S4074. Microscopy assays visually demonstrated the adhesion of the bacteria to the cells (Fig. 2). The increase of adherence over time was well demonstrated in the adherence assay, with an increase of about 1 log every hour (Fig. 3).

FIG. 2.

NPTr (a and c) or SJPL (b and d) cells stained with Giemsa stain in the presence (c and d) or absence (a and b) of A. pleuropneumoniae S4074, seen through a Leica DMR microscope at an original magnification of ×1,000.

FIG. 3.

Adherence of A. pleuropneumoniae S4074 to SJPL (filled bars) or NPTr (empty bars) cells from 1 to 3 h.

Protein profiling of SJPL and NPTr cells incubated with A. pleuropneumoniae.

Protein profiling of SJPL and NPTr cells incubated with A. pleuropneumoniae was performed as a rapid screening method using the Kinexus antibody microarray. Six hundred eight cell signaling proteins, including 250 phospho sites, 240 protein kinases, and 110 cell signaling proteins that regulate cell proliferation, stress,and apoptosis, were represented on the microarray. Only proteins with an n-fold change of ±1 and higher on a log₂ scale were deemed differentially expressed. Twenty proteins were upregulated for the SJPL cells, in contrast to 21 for the NPTr cells, while 25 proteins were downregulated for both the SJPL and NPTr cells (see Tables S1 and S2 in the supplemental material). Among the upregulated proteins, most were implicated in stress response. Mostly proteins implicated in cell growth and proliferation were observed to be downregulated. Among the proteins differentially expressed, IκB kinase α (IKKα), IKKβ, IRAK4, and three different MEKK proteins were detected and directed our focus toward the NF-κB and AP-1 pathways and ultimately to an examination of the production of cytokines by the epithelial cells.

NF-κB induction and cytokine production.

EMSAs detecting the induction of NF-κB and AP-1 were performed on both cell lines following incubation with A. pleuropneumoniae S4074, since proteins involved in these pathways were observed to be differentially expressed in our previous experiment. In comparison to the basal level of uninfected cells, a clear induction of NF-κB was noticed for the NPTr cells as soon as at 1 h postinfection which is represented by the appearance of bands of higher density in the upper part of the gel, corresponding to a band shift (Fig. 4A). Only a slight increase in density was observed for the SJPL cells at 3 h postinfection. In order to assess the specificity of the NF-κB induction, a supershift assay using specific antibodies was carried out for the detection of two subunits of NF-κB, p50 and p65. The p50 subunit was found to be induced for the SJPL cells but not the NPTr cells after 3 h of incubation with A. pleuropneumoniae S4074, and inversely, the p65 subunit was induced in the NPTr cells only (Fig. 4B). No induction of the AP-1 transcription factor was detected for either cell line in comparison to results for the uninfected cells (data not shown).

FIG. 4.

EMSA (A) or supershift assay (B) performed on nuclear proteins of SJPL and NPTr cells following an incubation (30 to 180 min) with A. pleuropneumoniae S4074 or no treatment for the control (Nil). For controls, proteins were incubated with specific oligonucleotides (S) and nonspecific oligonucleotides (NS). The probe alone was also loaded on the gel (P). For the supershift assay (B), proteins were incubated with p50 antibodies (p50), p65 antibodies (p65), or no antibodies (-). Arrows demonstrate the subunit band shifts (B).

To further investigate the bacterium-based induction of the NF-κB transcription factor, we evaluated cytokine production by both cell lines under stimulation conditions. Incubations for up to 48 h of the SJPL and NPTr cells with heat-killed A. pleuropneumoniae were then performed to quantify the production of IL-1β, IL-6, IL-8 and TNF-α, proinflammatory cytokines involved in innate immunity, by the epithelial cells. ELISAs performed on the supernatant samples showed that under these conditions, the NPTr cells but not the SJPL cells secrete IL-8. Production of IL-8 by the NPTr cells increased over time to reach 2,500 pg/ml at 48 h (Fig. 5), in comparison to 800 pg/ml with purified LPS. However, no IL-1β, IL-6, or TNF-α was detected in the samples from both cell lines (data not shown). Following inhibition of NF-κB by caffeic acid phenethyl ester in the NPTr cells, IL-8 concentrations decreased to basal levels (data not shown). This demonstrates that the production of IL-8 observed for the NPTr cells is indeed due to the induction of NF-κΒ.

FIG. 5.

Production of IL-8 by NPTr cells following an induction with heat-killed A. pleuropneumoniae S4074 (▪) or when not stimulated (▴).

To further investigate the mechanism of NF-κB induction in both cell lines, we performed an EMSA on cells pretreated with an IRAK1/4 inhibitor. IRAK1/4 is recruited by the MyD88 protein immediately after Toll receptor activation at the beginning of the NF-κB pathway (27). The level of NF-κΒ induction in comparison to that of noninhibited cells consequently demonstrated indirectly if the activation of Toll-like receptors by the bacterium is responsible for this induction. Our EMSA results indicate that for the SJPL cells, NF-κΒ induction occurs through the Toll receptor pathway (data not shown), but for the NPTr cells, NF-κΒ induction occurs through a different pathway.

A. pleuropneumoniae transcriptomic profiling.

To assess the transcriptional response of A. pleuropneumoniae to both planktonic life over and adherence to SJPL cells, transcript profiling experiments using DNA microarrays were performed after an incubation time of 3 h. Overall, 170 genes were significantly differentially expressed during planktonic growth (Tables 2 and 3), with this number dropping to 131 during adhesion to SJPL cells (Tables 4 and 5). While some genes showed similar patterns of expression during the two conditions, 150 were differentially expressed(see Table S3 in the supplemental material).

TABLE 2.

A. pleuropneumoniae genes which are upregulated during planktonic life over SJPL cells^a

Locus tag	Gene	Description	Fold change
Hypothetical/unclassified/unknown
APL_1839	udp	COG2820: uridine phosphorylase, probable outer membrane protein, possible efflux protein	3.937
APL_1833		COG2717: predicted membrane protein, conserved hypothetical protein	3.388
APL_1435		Unassigned protein	3.214
APL_1285		Hypothetical protein	2.894
APL_0145		COG1611: predicted Rossmann fold nucleotide-binding protein	2.726
APL_1976	yedF	COG0425: predicted redox protein, regulator of disulfide bond formation	2.669
APL_0116		Hypothetical protein	2.523
APL_0093		DUF1260 domain-containing protein	2.020
APL_1374		Unassigned protein	1.946
APL_0465		DUF74 domain-containing protein	1.943
APL_1919		Predicted enzyme related to aldose 1 epimerase	1.939
APL_1138		DUF526 domain-containing protein	1.933
APL_0443		COG5295: autotransporter adhesin	1.918
APL_1656		COG0561: predicted hydrolases of the HAD superfamily	1.842
APL_1203		DUF479 domain-containing protein, hypothetical protein	1.811
APL_1609		DUF533 domain-containing protein	1.724
APL_1881		Putative carbamoylphosphate synthase large subunit	1.703
APL_1244		Hypothetical protein	1.555
Biosynthesis of cofactors
APL_1622	cbiM	Predicted ABC cobalt transport permease protein CbiM	2.419
APL_0931	iscS	Cysteine desulfurase, iron-sulfur cluster assembly	2.374
APL_0930	nifU	NifU-like protein, involved in Fe-S cluster formation	2.011
APL_1555	hemL	Glutamate-1-semialdehyde 2,1-aminomutase	1.619
Energy metabolism
APL_1832		COG2041: sulfite oxidase and related enzymes	13.073
APL_0892	fdxG	Formate dehydrogenase, nitrate inducible, major subunit	12.558
APL_0895	fdnI	Formate dehydrogenase, cytochrome b556 subunit	10.946
APL_0894	fdxH	Formate dehydrogenase, iron-sulfur subunit	6.0988
APL_0687	dld	d-Lactate dehydrogenase	5.442
APL_0100	nrfA	Ammonia-forming cytochrome c552 nitrite reductase	5.039
APL_0106	putA	Bifunctional protein PutA	4.634
APL_0101	nrfB	Nitrate reductase, cytochrome-c-type protein	4.295
APL_1091	aspA	Aspartate ammonia lyase	4.254
APL_1528	frdC	Fumarate reductase subunit C	3.713
APL_1137	pgi	Glucose-6-phosphate isomerase	3.136
APL_1959	adhI	Alcohol dehydrogenases 1	2.992
APL_1414	mqo	Putative malate:quinone oxidoreductase	2.802
APL_1379	ccp	Cytochrome c peroxidase	2.796
APL_1450	fbp	Fructose-1,6-bisphosphatase	2.528
APL_1529	frdA	Fumarate reductase, flavoprotein subunit	2.420
APL_0187	pykA	Pyruvate kinase	2.322
APL_1197		3-Hydroxyacid dehydrogenase	2.270
APL_0102	nrfC	Nitrate reductase	2.250
APL_1526	frdD	Fumarate reductase subunit D	2.178
APL_0486	maeA	Malic enzyme (NADP dependent)	2.074
APL_1674	dmsA	Anaerobic dimethylsulfoxide reductase chain A precursor	1.934
APL_0483		Predicted nitroreductase	1.912
APL_0084	trxB	Thioredoxin reductase	1.486
Transport and binding proteins: cations and iron
APL_1265	copA	Copper-transporting P-type ATPase	1.518
Transport and binding proteins: others
APL_0107	putP	Na+/proline symporter	6.218
APL_1173	pnuC	Nicotinamide mononucleotide transporter	2.448
APL_0377	glpT	Glycerol-3-phosphate transporter	2.331
APL_1254		COG0471: di- and tricarboxylate transporters	2.237
APL_1319	ptsB	PTS system sucrose-specific EIIBC component	2.206
APL_0447	lctP	l-Lactate permease	2.204
APL_0262	modA	Molybdate-binding periplasmic protein precursor	2.049
APL_1620	cbiO	Predicted ABC-type cobalt transport system, ATPase component	1.902
APL_1627		Putative di- and tricarboxylate transporters	1.876
APL_1624	cbiK	Putative periplasmic binding protein CbiK	1.823
APL_1902	yrhG	COG2116: formate/nitrite family of transporters	1.718
APL_1371	ccmB	Heme exporter protein B, cytochrome-c-type biogenesis protein	1.596
Regulatory functions
APL_108	iclR	Putative HTH-type transcriptional regulator	2.832
APL_0395	rseA	Putative sigma E factor negative regulatory protein	2.260
APL_0823	glpR	Glycerol-3-phosphate regulon repressor	2.175
APL_0997	lacZ	Beta-galactosidase	1.654
Transcription
APL_1475	rpoD	RNA polymerase sigma 70 factor	2.135
APL_0560	rhlB	ATP-dependent RNA helicase RhlB	1.911
Purines, pyrimidines, nucleosides, and nucleotides
APL_0646	cpdB	2′,3′-cyclic-nucleotide 2′-phosphodiesterase precursor	2.087
Protein fate
APL_0742	degS	Protease DegS precursor	1.768
APL_0871	pepE	Peptidase E	1.686
Protein synthesis
APL_0484	rimK	Ribosomal protein S6 modification enzyme	3.093
APL_0146	dusA	tRNA-dihydrouridine synthase A	2.456
Cellular processes
APL_0004	sodC	Cu/Zn superoxide dismutase precursor	3.003
APL_0251	sodA	Manganese superoxide dismutase	2.356
APL_1241		Probable carbon starvation protein A, predicted membrane protein	2.098
APL_1405	oapA	Cell envelope opacity-associated protein A	1.793
Cell envelope
APL_1494	ftpA	DNA-binding ferritin-like protein (oxidative damage protectant), fine-tangled pilus major subunit (24-kDa surface protein)	6.378
Fatty acid and phospholipid metabolism
APL_0397	lcfA	Acyl-CoA synthetases (AMP forming)/AMP-acid ligases IIb	1.746
Mobile and extrachromosomal element functions
APL_1501		Transposase	2.078
APL_0990		Transposase	1.814
APL_0985		Transposase	1.680
DNA metabolism
APL_1143	recA	RecA recombinase	1.491
Central intermediary metabolism
APL_0109		Possible 5-formyltetrahydrofolate cycloligase	2.607
APL_0375	glpK	Glycerol kinase	2.316

^aEighty-two genes.

^bCoA, coenzyme A.

TABLE 3.

A. pleuropneumoniae genes which are downregulated during planktonic life over SJPL cells^a

Locus tag	Gene	Description	Fold change
Hypothetical/unclassified/unknown
APL_1387		Predicted metal-binding, possibly nucleic acid-binding protein	−2.431
APL_0053	typA	Predicted membrane GTPase involved in stress response	−2.386
APL_0173		Hypothetical cytosine deaminase and related metal-dependent hydrolases	−2.056
APL_1863		Predicted glycosyltransferase	−2.027
APL_1956		Hypothetical protein	−2.000
APL_0936		Putative integral membrane protein	−1.829
Energy metabolism
APL_1849	lldD	l-Lactate dehydrogenase (FMN dependent) and related alpha-hydroxy acid dehydrogenases	−3.284
APL_0592	guaA	GMP synthase	−2.629
APL_1219	fldA	Flavodoxin	−2.276
Transport and binding proteins: cations and iron
APL_1047	hgbA	Hemoglobin-binding protein A precursor	−9.346
APL_1571	tonB1	Periplasmic energy transducing protein TonB1	−5.319
APL_1952		Outer membrane receptor protein, mostly Fe transport	−3.517
APL_0077	exbD2	Energy transducing protein ExbD2	−3.499
APL_1953		Outer membrane receptor protein, mostly Fe transport	−2.874
APL_0670		Putative Fe2+/Pb2+ permease	−2.406
APL_1048	hugZ	Heme utilization protein	−2.387
APL_0272	yfeA	Iron (chelated) ABC transporter, periplasmic binding protein	−2.339
APL_0271	yfeB	putative chelated iron transport system ATP-binding protein	−2.251
APL_0714		Putative ABC-type enterochelin transport system, periplasmic component	−2.178
APL_1954		Outer membrane receptor proteins, mostly Fe transport	−2.163
APL_0715		COG4606: ABC-type enterochelin transport system, permease component	−1.993
APL_1955		Outer membrane receptor proteins, mostly Fe transport	−1.891
APL_0717		COG4604: ABC-type enterochelin transport system, ATPase component	−1.760
APL_0127	yfeD	Putative iron transport system membrane protein	−1.718
Transport and binding proteins: others
APL_0300	tolQ	Colicin transport protein	−2.584
APL_1392	ptnC	Mannose permease component IIC	−2.135
APL_1393	ptnD	Mannose permease component IID	−2.043
APL_1584	cpxB	Capsule polysaccharide export inner membrane protein	−1.863
APL_1585	cpxA	Capsule polysaccharide export ATP-binding protein	−1.572
APL_1583	cpxC	Capsule polysaccharide export inner membrane protein	−1.549
Transcription
APL_0638	nusA	Transcription elongation factor	−3.997
APL_0577	pnp	Polyribonucleotide nucleotidyltransferase	−2.375
APL_0757	rnb	Exoribonuclease II	−1.665
APL_0543	rnc	RNase III	−1.539
Purines, pyrimidines, nucleosides, and nucleotides
APL_0148	rnr1	Ribonucleotide reductase, alpha subunit	−2.556
APL_0593	guaB	Inosine-5′-monophosphate dehydrogenase	−2.474
APL_0147		Ribonucleotide reductase, beta subunit	−2.438
APL_0255	gpt	Xanthine phosphoribosyltransferase	−2.316
APL_0661	purK	Phosphoribosylaminoimidazole carboxylase ATPase subunit	−2.091
APL_1172	purD	Phosphoribosylamine-glycine ligase	−1.883
APL_2018	purC	Phosphoribosylaminoimidazole-succinocarboxamide synthase	−1.717
APL_0834	udk	Uridine kinase	−1.396
Protein fate
APL_1507	tig	FKBP-type peptidyl-prolyl cis-trans isomerase (trigger factor)	−3.849
APL_0364	Ssa1	Autotransporter serine protease	−2.279
Protein synthesis
APL_1718	rplK	50S ribosomal protein L11	−5.061
APL_0639	infB	Translation initiation factor 2 (IF-2)	−4.243
APL_1774	rplF	Ribosomal protein L6	−3.954
APL_1720	rplJ	Ribosomal protein L10	−3.833
APL_1558	rpsT	30S ribosomal protein S20	−3.720
APL_1765	rplV	50S ribosomal protein L22	−3.039
APL_0566	rpsB	30S ribosomal protein S2	−2.842
APL_1400	rpsG	30S ribosomal protein S7	−2.714
APL_0487	rplY	50S ribosomal protein L25 (general stress protein Ctc)	−2.364
APL_1762	rplW	50S ribosomal protein L23	−2.330
APL_1763	rplB	50S ribosomal protein L2	−2.265
APL_1761	rplD	50S ribosomal protein L4	−2.254
APL_0223	infC	Translation initiation factor 3 (IF-3)	−2.192
APL_1773	rpsH	30S ribosomal protein S8	−2.100
APL_1775	rplR	50S ribosomal protein L18	−2.020
APL_0399	ksgA	Dimethyladenosine transferase (rRNA methylation)	−1.950
APL_1401	rpsL	30S ribosomal protein S12	−1.914
APL_1972	rpmG	50S ribosomal protein L33	−1.902
APL_0040	yhbZ	Hypothetical GTP-binding protein	−1.721
APL_1228	infA	Translation initiation factor 1	−1.716
APL_0678	efp	Translation elongation factor P	−1.708
APL_1781	rpsM	30S ribosomal protein S13	−1.705
APL_1383		tRNA (guanine-N(7)-)-methyltransferase	−1.688
APL_0641	truB	tRNA pseudouridine synthase B	−1.660
APL_1476	tyrS	Tyrosyl-tRNA synthetase	−1.646
APL_1789	rplS	50S ribosomal protein L19	−1.521
Cellular processes
APL_0669		Predicted iron-dependent peroxidase	−2.630
APL_1443	apxIB	Toxin RTX-I translocation ATP-binding protein	−1.753
Cell envelope
APL_0933	ompP1	Putative long-chain-fatty-acid transport protein precursor	−8.199
APL_0651	galU	UTP-glucose-1-phosphate uridylyltransferase	−2.114
APL_1989		Predicted membrane protein	−1.951
APL_1071		Putative xanthine/uracil permease	−1.817
Fatty acid and phospholipid metabolism
APL_1864	accB	Biotin carboxyl carrier protein of acetyl-CoA carboxylaseb	−2.570
APL_1865	accC	Biotin carboxylase	−2.444
APL_1889	fabA	3-Hydroxydecanoyl-(acyl carrier protein) dehydratase	−2.426
APL_1385	plsX	Fatty acid/phospholipid biosynthesis enzyme PlsX	−1.922
Amino acid biosynthesis
APL_0194	aroK	Shikimate kinase	−2.401
APL_1499	thrC	Threonine synthase	−2.083
DNA metabolism
APL_0190	fis	Factor for inversion stimulation (Fis), transcriptional activator	−2.209
APL_0074	recR	Recombinational DNA repair protein (RecF pathway)	−1.634
Central intermediary metabolism
APL_1508		Putative rhodanese-related sulfurtransferase	−2.860
APL_0175	dksA	DnaK suppressor protein	−1.845
APL_0349	glgA	Glycogen synthase	−1.505

^aEighty-eight genes.

^bCoA, coenzyme A.

TABLE 4.

A. pleuropneumoniae genes which are upregulated during adherence to SJPL cells^a

Locus tag	Gene	Description	Fold change
Hypothetical/unclassified/unknown
APL_0568		Hypothetical membrane protein	3.083
APL_1459		Hypothetical protein	2.882
APL_1690		Predicted periplasmic/secreted protein	2.609
APL_1471		Putative sugar transferase	2.450
APL_1380		Uncharacterized conserved protein	2.310
APL_2002		Hypothetical protein	2.301
APL_0750		MscS family protein	2.181
APL_1575		Hypothetical protein	1.940
APL_1103		Predicted inner membrane protein	1.844
APL_1854	pqiA	Uncharacterized paraquat-inducible protein A	1.746
APL_0217		Hypothetical protein	1.640
Biosynthesis of cofactors
APL_0776	ispE	4-Diphosphocytidyl-2C-methyl-d-erythritol 2-phosphate synthase	2.913
Energy metabolism
APL_1849	lldD	l-Lactate dehydrogenase	6.756
APL_1820	rpe	d-Ribulose-phosphate-3 epimerase	6.687
APL_1191	namA	NADPH dehydrogenase	4.524
APL_1331	hyaA	Hydrogenase 2 small chain precursor	3.366
APL_1685	fucK	l-Fuculokinase	3.195
APL_1689	fucO	Probable alcohol dehydrogenase, class IV	2.887
APL_1684	fucI	l-Fucose isomerase	2.812
APL_1333	hybB	Putative Ni/Fe hydrogenase 2 b-type cytochrome component	2.796
APL_1019	kdgK	2-Dehydro-3-deoxygluconokinase	2.334
APL_0896	fdhE	Formate dehydrogenase accessory protein FdhE	2.007
Transport and binding proteins: cations and iron
APL_1793	fecE	Fe(III) dicitrate ABC transporter, ATP-binding protein	6.113
APL_1955		Outer membrane receptor proteins, mostly Fe transport	3.271
APL_1981	corA	Magnesium transport protein CorA	3.001
APL_0077	exbD2	Energy transducing protein ExbD2	2.346
Transport and binding proteins: others
APL_0066	dppC	Dipeptide transport system, permease components	12.073
APL_0870		Putative C4-dicarboxylate transporter	6.308
APL_1713		Putative oligopeptide transporter	4.639
APL_0191		Predicted Na+-dependent transporters of the SNF family	4.638
APL_0309	yheS	Putative ABC transporter ATP-binding protein YheS	2.331
APL_1848	cysA	Sulfate/thiosulfate import ATP-binding protein CysA	1.938
APL_1249	sapF	Peptide transport system ATP-binding protein SapF	1.858
APL_0260	modC	Molybdenum import ATP-binding protein ModC	1.799
APL_0582	sotB	Putative efflux transporter	1.773
Regulatory functions
APL_1099		Organic radical activating enzymes	7.175
APL_1962	hflX	GTP-binding protein HflX	2.005
Transcription
APL_0176	pcnB	Putative poly(A) polymerase	2.566
Purines, pyrimidines, nucleosides, and nucleotides
APL_0775	prsA	Ribose-phosphate pyrophosphokinase	8.463
APL_0425	purF	Amidophosphoribosyltransferase	1.812
Protein fate
APL_1742	srp54	Signal recognition particle protein (sigma 54 like)	6.861
APL_1905	dnaJ	Chaperone protein DnaJ	3.383
APL_1330	hypF	Carbamoyltransferase HypF	2.080
Protein synthesis
APL_0034	engD	GTP-dependent nucleic acid-binding protein EngD	17.924
APL_0575	deaD	Superfamily II DNA and RNA helicases	6.254
APL_0487	rplY	50S ribosomal protein L25 (general stress protein Ctc)	3.429
APL_1325		Putative 2-methylthioadenine synthetase	2.357
APL_1112	rumA	23S rRNA (uracil-5-)-methyltransferase RumB	2.145
APL_0853		Methionine synthase II (cobalamin independent)	2.078
Cellular processes
APL_1922	pgaB	Biofilm PGA synthesis lipoprotein PgaB precursor	7.257
APL_0011	ftsL	Cell division protein FtsL	3.349
APL_1923	pgaC	Biofilm PGA synthesis N-glycosyltransferase PgaC	2.454
Cell envelope
APL_0551	tadB	Tight adherence protein B	2.453
APL_1841	murI	Glutamate racemase	2.340
APL_0016	murD	UDP-N-acetylmuramoy-l-alanine-d-glutamate synthetase	2.113
APL_1598	mrdB	Rod-shape determining protein	2.096
APL_0419	lgtF	UDP-glucose-lipooligosaccharide glucosyltransferase	1.929
APL_1554	wecA	Putative undecaprenyl-phosphate α-N-acetylglucosaminyl 1-phosphate transferase	1.577
APL_0555	rcpA	Rough colony protein A	1.559
Fatty acids and phospholipids metabolism
APL_1864	accB	Biotin carboxyl carrier protein of acetyl-CoA carboxylaseb	8.665
APL_1865	accC	Biotin carboxylase	4.016
APL_0887	fadI	3-ketoacyl-CoA thiolase	3.675
Amino acids biosynthesis
APL_0320	metC	Cystathionine beta-lyase	9.278
APL_0099	ilvG	Acetolactate synthase isozyme II large subunit	6.897
APL_1452	serA	d-3-Phosphoglycerate dehydrogenase	3.896
APL_0469	trpB	Tryptophan synthase beta chain	3.274
APL_0139	leuC	3-Isopropylmalate dehydratase large subunit 2	3.253
APL_1951	proA	Gamma-glutamyl phosphate reductase	2.846
APL_1147	trpG	Putative anthranilate synthase component II	2.723
APL_0249	thrB	Homoserine kinase	2.459
APL_1125		Putative cysteine desulfurase	2.068
DNA metabolism
APL_1196		Type I site-specific restriction-modification system, R subunit	3.251
APL_1194		Type I restriction-modification system, methyltransferase subunit	2.816
APL_1146	rmuC	DNA recombination protein RmuC homolog	2.462
APL_0874	holA	DNA polymerase III, delta subunit	2.022
APL_0287	hsdM	Putative type I restriction-modification system, methyltransferase subunit	2.005
Central intermediary metabolism
APL_2045	sseA	Probable thiosulfate sulfurtransferase	1.942
APL_1843	cysJ	Sulfite reductase [NADPH] flavoprotein alpha component	1.780

^aSeventy-nine genes.

^bCoA, coenzyme A.

TABLE 5.

A. pleuropneumoniae genes which are downregulated during adherence to SJPL cells^a

Locus tag	Gene	Description	Fold change
Hypothetical/unclassified/unknown
APL_1887		Esterase domain-containing protein	−6.273
APL_1284		Putative DNA-binding protein	−3.076
APL_0049		Hypothetical protein	−2.982
APL_0970		Hypothetical protein	−2.980
APL_1100		Hypothetical protein	−2.928
APL_1437		Hypothetical protein	−2.751
APL_0116		Hypothetical protein	−2.566
APL_0389	ompP4	Lipoprotein E precursor, predicted secreted acid phosphatase	−2.509
APL_0704		Potential type III restriction enzyme	−2.293
APL_1365	Hly	Hypothetical protein	−2.174
APL_0110		Hypothetical protein	−2.144
APL_1396		Hypothetical protein	−2.125
APL_0756		Hypothetical protein	−2.098
APL_0889		Hypothetical protein	−1.757
Energy metabolism
APL_0434	gapA	Glyceraldehyde-3-phosphate dehydrogenase	−4.981
APL_0892	fdxG	Formate dehydrogenase, nitrate-inducible, major subunit	−4.853
APL_0771	lpdA	Dihydrolipoyl dehydrogenase	−4.765
APL_1379	ccp	Cytochrome c peroxidase	−4.447
APL_0983	tktA	Transketolase 2	−3.790
APL_0894	fdxH	Formate dehydrogenase, iron-sulfur subunit	−3.627
APL_1251	pgk	3-Phosphoglycerate kinase	−3.625
APL_1450	fbp	Fructose-1,6-bisphosphatase	−3.315
APL_0181	gloA	Lactoylglutathione lyase	−3.046
APL_1925	tpiA	Triosephosphate isomerase	−3.023
APL_1091	aspA	Aspartate ammonia-lyase	−2.722
APL_0688	torZ	Trimethylamine-N-oxide reductase precursor	−2.687
APL_1011	adh2	Aldehyde-alcohol dehydrogenase 2	−2.642
APL_0772	aceF	Dihydrolipoyllysine-residue acetyltransferase component of pyruvate dehydrogenase complex (E2)	−2.620
APL_1137	pgi	Glucose-6-phosphate isomerase	−2.595
APL_0486	maeA	NADP-dependent malic enzyme	−2.416
APL_1526	frdD	Fumarate reductase subunit D	−2.267
APL_0101	nrfB	Nitrate reductase, cytochrome-c-type protein	−2.231
APL_1250	fba	Fructose bisphosphate aldolase	−2.124
Transport and binding proteins: others
APL_1620	cbiO	Predicted ABC-type cobalt transport, ATPase component	−2.108
APL_0447	lctP	Putative l-lactate permease	−2.068
APL_0719		Putative phosphate permeases	−1.816
APL_0791		Transmembrane transport protein-permease	−1.702
Regulatory functions
APL_0656	hlyX	Regulatory protein HlyX	−2.727
APL_0629	cpxR	Transcriptional regulatory protein CpxR	−2.367
Purines, pyrimidines, nucleosides, and nucleotides
APL_0769	ushA	UshA precursor	−3.470
APL_0646	cpdB	2′,3′-Cyclic-nucleotide 2′-phosphodiesterase precursor	−2.778
APL_1014	deoD	Purine-nucleoside phosphorylase	−2.247
Protein fate
APL_1154		Putative zinc protease	−2.401
APL_1456	slyD	FkbP-type peptidyl-prolyl cis-trans isomerase	−2.313
Protein synthesis
APL_0740	rpsA	30s ribosomal protein S1	−1.608
Cellular processes
APL_1445	apxIC	RTX-I toxin-activating lysine-acyltransferase ApxIC	−4.117
APL_0956	apxIIA	RTX-II toxin determinant A	−4.052
APL_0004	sodC	Superoxide dismutase (Cu/Zn) precursor	−3.193
Cell envelope
APL_1494	ftpA	COG0783: DNA-binding ferritin-like protein (oxidative damage protectant)	−4.223
APL_0652	manB	Phosphomannomutase	−3.375
Central intermediary metabolism
APL_0645	ackA	Acetate kinase	−3.190
APL_1899	ppa	Inorganic pyrophosphatase	−2.207

^aFifty-two genes.

The genes that showed the highest level of upregulation during planktonic growth belonged to the “Energy metabolism” functional class, and this class was also the most affected, with 24 out of the 82 upregulated genes. Surprisingly, most of these genes are involved in anaerobic respiration. Various enzymes involved in anaerobic respiration using alternative electron acceptors were upregulated: genes encoding subunits of the formate dehydrogenase (fdxG, and fdnHI) and nitrate reductase (nrfABC), which are essential for anaerobic respiration on nitrate (35, 62), and genes encoding subunits of the fumarate reductase (frdACD), which allows fumarate to serve as a terminal electron acceptor (10), were all upregulated. Furthermore, the genes pgi, fbp, and pykA, which encode three enzymes involved at various steps of glycolysis, glucose-6-phosphate isomerase, fructose-1,6-bisphosphatase, and pyruvate kinase, respectively, showed an increase in transcription. Two dehydrogenases, alcohol dehydrogenase I (encoded by adhI) and malate:quinone oxidoreductase (encoded by mqo), are also involved in anaerobic respiration, the latter being controlled by the ArcA-ArcB two-component system (59). The genes aspA and dmsA were also upregulated. The “Transport and binding proteins” class was the second most affected, with 12 upregulated genes. Genes involved in the transport of l-lactate (lctP), formate and nitrite (yrhG), sucrose (ptsB), and glycerol (glpT) were all induced, as was the gene modA, which encodes a periplasmic protein involved in the ABC transport of molybdate (25). The gene APL_0443, coding for a putative autotransporter adhesin, showed a 1.9-fold induction.

Genes downregulated during planktonic growth mostly belonged to the “Protein synthesis” and “Transport and binding proteins: cations and iron” functional classes. The hgbA hemoglobin receptor gene and its hugZ heme utilization protein cotranscript were downregulated, and other well-established iron acquisition-related genes included tonB1 and exbD2. These genes are cotranscribed with other members of the TonB1 (exbB1, exbD1, tbpA, and tbpB) and TonB2 (exbB2 and exbD2) energy transduction system (7), but these were not identified in our study. At this time, the genes exbB1, exbD1, and tbpB are not present on AppChip1, and the gene tbpA was not downregulated. The genes exbB2 and tonB2, however, exhibit a twofold average downregulation, but variations between chips might have caused these to be ignored by our very stringent analysis parameter (false discovery rate = 0%). A high number of genes that were identified for the first time in A. pleuropneumoniae in our previous transcript profiling experiment under iron restriction (13) were also downregulated. The open reading frames APL_1952 to APL_1955, which code for a putative second hemoglobin receptor system and are likely transcriptionally linked, were repressed, as were the genes APL_0714-APL_0715-APL_0717, encoding a putative ABC-type siderophore transport system, and the genes yfeABD, likely responsible for the ABC-like periplasmic binding protein-dependent transport of chelated iron and possibly manganese. The genes cpxABC, coding for the capsule polysaccharide ABC-type export system, were also all downregulated, along with the gene ssa1, encoding a putative autotransporter serine protease.

Interestingly, some genes potentially involved in adhesion and biofilm biosynthesis were upregulated during adherence to SJPL cells. The genes rcpA and tadB, which belong to a large operon of 14 genes, were upregulated, as were the genes pgaBC, involved in poly-β-1,6-N-acetyl-d-glucosamine biofilm biosynthesis. A small number of genes involved in iron acquisition were also upregulated, the most notable being fecE and APL_1955. Once again, genes involved in anaerobic respiration were shown to be upregulated. Other enzyme genes coding for hydrogenases (hyaA and hybB) or dehydrogenases (lldD and fdhE) involved in energy metabolisms also showed upregulation. The gene fucO, essential for the anaerobic degradation of fucose, and the genes fucI and fucK, (11), involved in the general fucose degradation pathway, were also upregulated.

Only 52 genes were identified as downregulated, and most of them belong to the “Energy metabolism” functional class. The six enzymes which catalyze the first six steps of glycolysis (encoded by gapA, pgk, fbp, tpiA, pgi, and fba) were downregulated, as was the gene maeA, responsible for the first step of gluconeogenesis, and the gene tktA, which links glycolysis to the pentose-phosphate pathway. hlyX, coding for the A. pleuropneumoniae FNR anaerobic global regulator homolog, was repressed 2.72-fold. The toxin genes apxIC and apxIIA also showed downregulation during adhesion to SJPL cells.

Adherence and invasion of A. pleuropneumoniae and other members of the Pasteurellaceae.

Other serotypes of A. pleuropneumoniae and different swine-colonizing members of the Pasteurellaceae were tested using the adherence models. Differences in adherence were observed between strains for a given cell line and between the cell lines for a given strain (Fig. 6). We noticed that the field strains of A. pleuropneumoniae adhered significantly more to the cell lines than the reference strain of the same serotype. We also noticed that the level of adherence to a given cell line was strain dependent. Following the observation that all members of the Pasteurellaceae tested adhered to the cell lines, invasion tests were performed. A. pleuropneumoniae S4074 did not invade either cell line in our infection model, while the other members of the Pasteurellaceae tested showed invasion. H. parasuis showed the highest level of invasion, although at a reduced level compared to invasion seen with endothelial cells (Fig. 7) (60).

FIG. 6.

Adherence of 12 members of the Pasteurellaceae to the SJPL (filled bars) or NPTr (empty bars) cell line after 3 h of incubation. The strains include A. pleuropneumoniae serotype 1 S4074 and FMV91-6514, A. pleuropneumoniae serotype 5b L20 and 05-6501, A. pleuropneumoniae serotype 5a 05-4817, A. pleuropneumoniae serotype 7 WF83 and 05-3695, H. parasuis serotype 5 Nagasaki and 29755, A. suis serotype O2/K2 H91-0380, and P. multocida capsular type A 88-761 and capsular type D 1703. Asterisks represent statistical differences (P < 0.05) in adherence of the given strain between the two cell lines.

FIG. 7.

Invasion by A. pleuropneumoniae S4074 (▪), H. parasuis Nagasaki (▴), and H. parasuis 29755 (•) of SJPL (full line) or NPTr (dashed line) cells from 1 h to 3 h.

DISCUSSION

Using immortalized porcine lung and tracheal epithelial cells, we were able to study the host-pathogen interactions of A. pleuropneumoniae. In our models, A. pleuropneumoniae provoked cell death very rapidly through necrosis and not apoptosis. The presence of this bacterium causes many changes in the protein profiles of both epithelial cell lines. Indeed, using an antibody microarray as a rapid screening for differential protein expression, we were able to direct our efforts toward the NF-κB pathway, since numerous differentially expressed proteins were implicated in the NF-κB pathway, including IKKα and IKKβ. NF-κB consists of a homo- or heterodimer composed of the five mammalian Rel proteins, p65, c-Rel, p50, p52, and RelB (24), with the p50/p65 heterodimer being the most abundant and active of the NF-κB complexes (2). Out of the five subunits, only p65 (RelA), RelB, and c-Rel were found to contain the C-terminal transactivation domains essential for gene activation. In contrast, p50 and p52 do not possess the transactivation domains and therefore cannot act as transcriptional activators by themselves (39). Additionally, p50 and p52 are synthesized as precursor proteins that belong to the IκB family known as inhibitors of NF-κB, and homo- or heterodimers of p50 and p52 were also reported to repress κB site-dependent transcription in vivo (39). Interestingly, the p50 subunit was found to be induced in the SJPL cells but not in the NPTr cells after 3 h of incubation with A. pleuropneumoniae S4074, and inversely, the p65 subunit was induced in the NPTr cells only. It should be noted that the absence of detection of either p50 in the NPTr cells or p65 in the SJPL cells is not due to a weak bacterium-induced expression but is most probably due to the incapacity of the cell line to express the protein, since no basal level of expression was observed under unstimulated conditions for the subunits p50 in the NPTr cells and p65 in the SJPL cells. Those results suggest that in the absence of p65, inactive p50/p50 homodimers are more likely to form in the SJPL cells. The absence of IL-8 production by the SJPL cells might be explained by the weak NF-κB induction observed in the EMSAs but certainly correlates with the absence of the p65 subunit, necessary for binding to the IL-8 promoter. Previous studies have also shown that the binding affinity of p50 for the human IL-8 promoter is weak compared to the binding of the p65 subunit (38). Different pathways can activate NF-κB, the most frequent in gram-negative bacterial infection being the classical pathway through Toll-like receptor activation by LPS (44). We demonstrated that this is the case for the SJPL cells but not for the NPTr cells. A possibility is that an alternative pathway for NF-κB activation was used in the NPTr cells where IKKα homodimers are activated instead of the IKKβ in the classical pathway, leading to NF-κB2/p100 phosphorylation. This is a possibility, since this modification creates the production of p52 (44), a subunit which seems to be present in the stimulated NPTr cells, as seen in the EMSA, where a band slightly higher than p50 was detected, and since IKKα was upregulated in the NPTr antibody microarray. This pathway is generally triggered by TNF receptor family members, such as LTβR, BAFF-R, CD40, and CD30 (44). Additional experiments are necessary, however, to confirm this theory.

The presence of the epithelial cells stimulated differential expression of many A. pleuropneumoniae genes. Although it was shown previously, with the evidence of a putative involvement in virulence of genes dmsA and aspA, that genes involved in anaerobic respiration might in fact be essential for full virulence of A. pleuropneumoniae in the host (3, 5, 30), it is still unclear why genes involved in anaerobic respiration are upregulated under our experimental conditions. While such a metabolic switch might be important in vivo to adapt to the lack of oxygen in the deep lung tissues, it should not be necessary in our experimental setup unless this apparent aerobic/anaerobic shift is controled by a host cell-associated factor rather than by oxygen sensors. In fact, it is worth noting that this metabolic shift does not seem to be complete, since genes involved in aerobic respiration are not downregulated. The upregulation of the gene sodA, coding for a cytoplasmic Mn superoxide dismutase (21), also seems to indicate that aerobic respiration is not stopped, since these cytoplasmic superoxide dismutases are specifically useful in removing superoxide anions generated during the course of aerobic respiration (54).

A gene with possible involvement in virulence was also identified. Locus tag APL_0443 is described as an autotransporter adhesin. This protein shows a region of high homology with the A. actinomycetemcomitans extracellular matrix protein adhesin A (EmaA), an oligomeric autotransporter with a YadA domain (57), and the putative Mannheimia haemolytica Hsf protein (40). In Haemophilus influenzae serotype b, Hsf (Haemophilus surface fibrils) is considered the major nonpilus adhesin (55) and was found to be associated with adherence to human epithelial cells (6, 26). Whether this putative A. pleuropneumoniae Hsf protein has these properties remains to be seen, but the upregulation of this gene during planktonic life over SJPL cells might hint to a possible role in the initial steps of A. pleuropneumoniae adhesion during infection.

The fact that iron, in DMEM, is available only in the form of ferric nitrate (12) might explain why iron acquisition systems are more expressed in cell-free DMEM than during planktonic growth. Experiments conducted in our laboratory have shown that ferric nitrate cannot support growth of A. pleuropneumoniae in an ethylenediamine dihydroxyphenyl acetic acid iron-restricted medium (13). Lysis of SJPL cells, leading to the release of the intracellular content of those cells in the medium, supplies the planktonic bacteria with more readily available sources of iron, for example, iron-sulfur clusters in the catalytic core of enzymes and ferritin-bound ferric iron.

Some genes with possible involvement in virulence also came up as downregulated during planktonic growth. Downregulation of the cpxABC operon during planktonic growth over SJPL cells might indicate that when in contact with host cells, A. pleuropneumoniae might wear a thinner polysaccharide layer in order to unmask some adhesins. Surprisingly, this downregulation of the cpx operon was not seen during adhesion to SJPL cells. Upon verification of changes during adhesion, only cpxC showed a low level of downregulation (−1.24), although this was not statistically significant. Repression of the gene ssa1 was surprising, since this gene, also termed aasP, was shown to be expressed in vivo during the chronic stage of the disease (3). However, this gene was also shown to be iron responsive, as indicated by its upregulation during iron restriction (13). It might therefore simply follow the same trend as other iron-responsive genes which were downregulated during planktonic growth.

Our main focus, when looking at overexpressed genes during adherence to porcine lung epithelial cells, was to search for new potential adhesins. The genes tadB and rcpA are part of a large operon which, in A. actinomycetemcomitans, is composed of 14 genes (36) and mediates nonspecific adhesion to solid surfaces, whether they are biological surfaces or not (16). The genetic organization of the A. pleuropneumoniae tad locus is identical to that for A. actinomycetemcomitans (58). Although it is suspected that the tad genes might be transcribed as an operon, only two genes were identified as upregulated in our study. The 12 other genes are present on the microarray but are not significantly induced. Expression of the tad genes is responsible for the rough colony phenotype of A. actinomycetemcomitans, but smooth variants often arise after continued passage on rich medium (49) since mutations often appear in the promoter region of the gene flp-1 (58). We suspect that this might also be the case for A. pleuropneumoniae, since most field isolates exhibit this rough colony phenotype while the reference strains are often smooth colony variants. As is the case for A. actinomycetemcomitans, the Tad proteins might play an important role for the colonization of the respiratory tract by A. pleuropneumoniae, but this will have to be further investigated. Other genes possibly involved in adhesion were also upregulated during adhesion to SJPL cells. The genes pgaB and pgaC are both involved in poly-β-1,6-N-acetyl-d-glucosamine (PGA) biofilm formation. A pgaABCD cluster is present in the App5b L20 genome, and the gene pgaC has been shown to be present in 15 reference strains (29). These results are interesting since the only components that have been clearly shown to be involved in A. pleuropneumoniae adhesion to lung surfaces to date are LPS (1, 8, 34, 45, 46).

The gene hlyX was downregulated during adhesion to SJPL cells. This gene, which encodes the A. pleuropneumoniae FNR anaerobic global regulator homolog, was shown to be important for the colonization and persistence of A. pleuropneumoniae in the respiratory tracts of swine (5). The repression of hlyX probably explains the repression of aspA, which is presumably regulated by HlyX, as well as the downregulation of a few other genes linked with anaerobic respiration (fdxG, torZ, nrfB, and frdD). Genes putatively regulated by HlyX have been shown to be induced by bronchoalveolar lavage fluid from infected pigs (30), and it is possible that hlyX expression follows the same pattern. Also, putative HlyX-regulated genes were upregulated during planktonic growth over SJPL cells.

ApxI and ApxII have been shown to be major virulence factors in A. pleuropneumoniae. Not much is known about transcriptional regulation of those toxins in A. pleuropneumoniae. Studies have shown that levels of oxygen do not influence the levels of ApxI and ApxII (33) and that the iron response regulator Fur seems to have variable effects depending on the calcium concentration in the culture medium (28). Under high calcium concentrations, Fur seemed to act as an activator of the apxI operon, while it seemed to act as a repressor under low calcium concentrations. A previous microarray study conducted under iron restriction showed that Fur does have an effect on ApxI transcription (13). One would normally expect these toxins to be induced under conditions mimicking the in vivo environment, mostly after contact with epithelial cells. Downregulation of the genes apxIC and apxIIA was therefore intriguing. Perhaps smaller concentrations of RTX toxins are required when the bacteria are in close proximity to host cells, leading to the downregulation of the toxins following adherence.

Adherence is seen in both models for all A. pleuropneumoniae strains and serotypes tested. It is interesting to note that field strains adhere more to the cell lines than the reference strain of the same serotype. No invasion is noticed for A. pleuropneumoniae, even though close relatives, such as A. actinomycetemcomitans and H. parasuis, are known to be invasive (17, 42, 60).

Overall these results showed the efficacy of the models and allowed us to gain a great amount of knowledge of A. pleuropneumoniae host-pathogen interactions. Indeed, interaction of A. pleuropneumoniae with host epithelial cells seems to involve complex cross talk which results in the regulation of various bacterial genes. Many virulence genes were upregulated, including genes coding for the putative adhesins Hsf and poly-β-1,6-N-acetyl-d-glucosamine, while capsular polysaccharide-associated genes were downregulated, possibly exposing adhesins usually hidden by a thick capsule. Incubation with A. pleuropneumoniae then led, for both cell lines, to the induction of NF-κB. This is done through the activation of a Toll receptor for the SJPL cells but through an alternative pathway for the NPTr cells. The NPTr cells then secrete IL-8, which is known to attract neutrophils to the infection site, while the SJPL cells do not due to the absence of the p65 subunit of NF-κB. These models are a biologically relevant tool for studying porcine respiratory tract pathogens which could be further used, in the future, to evaluate the effect of a preinfection with agents such as mycoplasmas and viruses, often present with bacterial pathogens under field conditions.