|Home | About | Journals | Submit | Contact Us | Français|
Newly recognized Fic family virulence proteins may be important in many bacterial pathogens. To relate cellular mechanisms to pathogenesis and immune protection, we studied the cytotoxicity of the Histophilus somni immunoglobulin-binding protein A (IbpA) direct repeat 2 Fic domain (DR2/Fic) for natural host target cells. Live virulent IbpA-producing H. somni strain 2336, a cell-free culture supernatant (CCS) of this strain, or recombinant DR2/Fic (rDR2/Fic) caused dramatic retraction and rounding of bovine alveolar type 2 (BAT2) epithelial cells. IbpA-deficient H. somni strain 129Pt and a Fic motif His298Ala mutant rDR2/Fic protein were not cytotoxic. The cellular mechanism of DR2/Fic cytotoxicity was demonstrated by incubation of BAT2 cell lysates with strain 2336 CCS or rDR2/Fic in the presence of [α-32P]ATP, which resulted in adenylylation of Rho GTPases and cytoskeletal disruption. Since IbpA is not secreted by type III or type IV secretion systems, we determined whether DR2/Fic entered the host cytoplasm to access its Rho GTPase targets. Although H. somni did not invade BAT2 cells, DR2/Fic was internalized by cells treated with H. somni, CCS, or the rDR2/Fic protein, as shown by confocal immunomicroscopy. Transwell bacterial migration assays showed that large numbers of strain 2336 bacteria migrated between retracted BAT2 cells, but IbpA-deficient strain 129Pt did not cross a monolayer unless the monolayer was pretreated with strain 2336 CCS or rDR2/Fic protein. Antibody to rDR2/Fic or passively protective convalescent-phase serum blocked IbpA-mediated cytotoxicity and inhibited H. somni transmigration across BAT2 monolayers, confirming the role of DR2/Fic in pathogenesis and corresponding to the results for in vivo protection in previous animal studies.
New mechanisms of virulence due to Fic family proteins may be significant since many bacterial pathogens have Fic gene sequences in their genomes (27). Mechanisms of action were reported for a few of these pathogens for the first time in 2009 based on data obtained using cell lines (14, 18, 27, 28), but their importance in relevant models of pathogenesis and immune protection remains to be demonstrated. We studied the immunoglobulin-binding protein A (IbpA) DR2/Fic cytotoxin of Histophilus somni (formerly Haemophilus somnus ) because we previously reproduced pneumonia and septicemia in animals with this pathogen (7, 9) and demonstrated that immunizing animals with IbpA DR2/Fic provided protection (7). H. somni is an economically important pathogen of cattle and other ruminants that causes respiratory disease, septicemia, thrombotic meningoencephalitis, myocarditis, arthritis, and abortion (5, 11, 16, 21, 26). This organism also can be a member of the normal flora of the lower reproductive tract and, to a lesser extent, the upper respiratory tract (5, 12). The pathogenesis of H. somni pneumonia, the most commonly reported syndrome in H. somni infections, and the mechanisms by which the bacteria spread into the systemic circulation from the respiratory tract are not clearly defined.
One of the virulence factors of H. somni is immunoglobulin-binding protein A (IbpA), a secreted and surface-associated fibrillar protein consisting of 4,095 amino acid residues. This protein is transported to the bacterial surface by a two-partner secretion pathway (13, 23). All isolates of H. somni tested produce IbpA, except for four carrier strains, including strain 129Pt, which lacks the entire ibpA gene locus (4, 25). IbpA-producing strain 2336 has been shown to be virulent in a bovine pneumonia model (9, 10). Convalescent-phase bovine serum which recognizes IbpA (6, 19, 29) passively protects calves against pneumonia (8). The N-terminal region of IbpA has several putative adhesin domains with homology to the domains of filamentous hemagglutinin (FHA) of Bordetella pertussis (23). The C terminus of IbpA contains several repeat sequences, including two large (400-residue) direct repeats (DR1 and DR2) (6). Each direct repeat contains a conserved Fic (filamentation induced by cyclic AMP [cAMP]) motif (27). This motif was originally described in Escherichia coli as a stress response protein associated with filamentous bacterial growth in the presence of excess cAMP (15). The Fic family proteins all contain a conserved Fic motif, HXFX(D/E)(A/G)N(K/G)R, which is involved in the virulence of several pathogens (14, 18, 27, 28).
We recently showed that expression of the DR2 Fic motif (DR2/Fic) in HeLa cells resulted in disruption of the cellular cytoskeleton due to adenylylation and subsequent inactivation of the Rho GTPases (27). The Fic motif in DR2 was critical since replacement of the conserved His residue in this motif by Ala eliminated cytotoxicity (27). In that study, we transfected human HeLa cells with DR2/Fic. This did not reflect a physiologically relevant interaction of H. somni with its natural host cells. Therefore, we developed systems for assessing the IbpA DR2/Fic function that are relevant to natural disease and protective immunity. Based on the previously reported attachment of H. somni to bovine turbinate (BT) cells (24) and the location of the organism in the alveolus during pneumonia (3, 9), we compared the cytotoxic effects of recombinant DR2/Fic (rDR2/Fic) in BT cells, primary bovine alveolar type 2 (BAT2) cells, and the human HeLa cell line used previously (27). We report here that treatment of these cells with virulent H. somni strain 2336, IbpA-enriched cell-free culture supernatant (CCS), or recombinant DR2/Fic (rDR2/Fic) caused cell rounding and retraction. BAT2 cells were the most susceptible cells, and HeLa cells were the least susceptible cells. Retraction was shown to be due to adenylylation of Rho GTPases that resulted in cytoskeletal disruption in BAT2 cells. Paracellular migration of virulent H. somni across a BAT2 monolayer was observed, which suggested a route for invasion of the bloodstream. A carrier strain of H. somni which lacks IbpA, 129Pt, did not cause cytotoxicity or transmigrate across an alveolar epithelial monolayer. Convalescent-phase serum or antibody to DR2/Fic neutralized toxicity and prevented migration across BAT2 monolayers, indicating that the IbpA DR2/Fic domain is relevant to in vivo pathogenesis and immune protection against disease in the natural host.
H. somni strain 2336, a virulent pneumonic isolate previously used to induce experimental pneumonia in calves (8, 9), and strain 129Pt, an asymptomatic carrier strain isolated from the prepuce of a normal bull (4, 25), were grown on brain heart infusion (BHI) (BD Diagnostics, Sparks, MD) agar containing 5% bovine blood in Alsevers solution (Colorado Serum Co., Denver, CO) at 37°C in a candle jar. Culture supernatant was prepared from H. somni cells scraped from an 18-h BHI blood agar plate, inoculated into BHI broth supplemented with 0.1% Tris base and 0.01% thiamine monophosphate, and grown for 6 h at 37°C with shaking at 200 rpm. The inocula used for the cultures were standardized spectrophotometrically and confirmed by plate counting to contain approximately 5 × 107 CFU. Six-hour cultures of H. somni shed minimal detectable lipooligosaccharide into the culture supernatant in preliminary studies (unpublished data). Each culture was centrifuged at 5,000 × g for 15 min, and the supernatant was filtered through a 0.22-μm-pore-size filter and then concentrated 40-fold in a centrifugal filter concentrating device (Amicon Ultra) with a 10-kDa cutoff (Millipore, Billerica, MA) by about 2 h of centrifugation at 3,000 × g. The retentate was washed twice in phosphate-buffered saline (PBS) to produce a cell-free culture supernatant (CCS). Each preparation was monitored using Western blotting against rabbit antibodies to rDR2/Fic and bovine convalescent-phase serum for the presence of IbpA.
Expression and purification of recombinant DR2/Fic protein and mutant DR2/Fic have been described previously (27). Briefly, the ibpA DR2/Fic-encoding region was PCR amplified using primers AAATCATCTCCGCAAGAAGGA and TTTTGCCAACTCTTTTAAAAAC (ibpA GenBank accession no. CP000947, locus HSM_1489) and cloned into glutathione S-transferase (GST)-tagged plasmid vector pET41a (Novagen, Madison, WI). Recombinant DR2/Fic expressed in E. coli BL21 cells was purified by glutathione affinity chromatography (Sigma, St. Louis, MO). A mutant with a site-directed point mutation that replaced the histidine residue at position 298 in DR2/Fic (residue 3717 in the IbpA sequence) with alanine (rDR2/Fic H/A) was constructed as previously described (27) and was purified as described above.
Polyclonal antibodies against the rDR2/Fic protein were produced by immunizing two rabbits with purified recombinant GST-DR2/Fic protein emulsified with Freund's adjuvant (Cocalico Biologicals, Reamstown, PA). Blood was collected before and after immunization for preparation of pre- and postimmunization sera. Convalescent-phase bovine serum was obtained from two calves (calves E5 and E7) with experimental H. somni pneumonia induced by intrabronchial inoculation of strain 2336 (9). Serum samples were collected before infection (preimmune serum) and 6 weeks after induction of pneumonia, during convalescence. The specificities of these sera were evaluated by Western blotting.
Primary BAT2 cells were isolated from a newborn calf lung collected in cold Hanks' balanced salt solution (HBSS) containing gentamicin (40 μg/ml), fungizone (5 μg/ml), and cefoperazone (35 μg/ml). Finely minced tissue was stirred to release macrophages and then digested with 0.3% Dispase II (Boehringer, Mannheim, IN) in HBSS and centrifuged at 500 × g for 10 min, and the supernatant was decanted. The pellet containing BAT2 cells was suspended in the last drop of Dispase solution and gently mixed with PBS containing 10% horse serum. After centrifugation at 500 × g for 5 min, the pellet was resuspended in 10 ml Dulbecco modified Eagle medium (DMEM)-keratinocyte medium (1:1) (Invitrogen, Carlsbad, CA), tissue clumps were sedimented, and the top layer was transferred to a tissue culture flask precoated with 0.1% gelatin containing 20% (vol/vol) fetal bovine serum (FBS) (gelatin/FBS). Cells were incubated at 37°C in the presence of 5% CO2 and fed fresh DMEM-keratinocyte medium supplemented with 2% FBS, 5 mM l-glutamine, 0.02% lactalbumin hydrolysate, penicillin (100 U/ml), and streptomycin (100 μg/ml) (Invitrogen) every 2 days. BAT2 cells were identified by their star-shaped appearance before they were flattened and by their Nile Red-stained cytoplasmic vacuoles (lamellar bodies) (20), and they were used at a maximum of 13 passages. BT cells (kindly provided by L. J. Gershwin, University of California Davis) were grown in DMEM-keratinocyte medium, and HeLa cells were grown in DMEM supplemented with 10% FBS, penicillin (100 U/ml), and streptomycin (100 μg/ml) at 37°C in a humidified atmosphere containing 5% CO2. Cells were grown in 75-cm2 culture flasks precoated with gelatin/FBS and were harvested by brief digestion with a 0. 05% trypsin-10 mM EDTA solution (Invitrogen). For assays, cells were seeded either into 24-well tissue culture plates (5 × 104 cells/well), into 12-well culture plates containing coverslips (5 × 105 cells/well), or into 8-well chambered glass slides (Nunc, Naperville, IL) (5 × 103 cells/well). Glass slides and glass coverslips were precoated with10 μg/ml bovine plasma fibronectin (Invitrogen) overnight at 4°C.
Cells grown to about 90% confluence in chambered slides or on glass coverslips in 12-well culture plates were treated with live bacteria, CCS, purified rDR2/Fic protein, or the rDR2/Fic H/A mutant protein. Bacteria were harvested from an 18-h culture on BHI blood agar, resuspended in tissue culture medium without antibiotics, and incubated for 2 h at 37°C before they were added to the cells. Bacterial counts were estimated with a spectrophotometer and were confirmed by counting CFU. Cells were treated with bacteria at a multiplicity of infection (MOI) of 10:1 or 100:1 (ratio of bacteria to cells), with 20× CCS (final concentration) in culture medium, or with a recombinant protein at a concentration of 20 μg/ml of culture medium without FBS. After treatment for 4 h, the cells were washed twice with PBS and fixed with 4% fresh paraformaldehyde for 20 min at 4°C before they were permeabilized with 0.1% Triton X-100 for 5 min and stained with rhodamine-phalloidin (Invitrogen) for 30 min at room temperature. Nuclei were counterstained with 4′,6′-diamidino-2-phenylindole (DAPI), and slides were mounted in Prolong antifade reagent (Invitrogen). The number of rounded or retracted cells and the total number of cells in a field (as determined by DAPI nuclear staining) were determined for 10 separate microscope fields by using fluorescence microscopy, and means were calculated. Mitotic figures were used to distinguish normal retracted or rounded cells due to mitosis from the rounded or retracted phenotype due to cytotoxicity. Toxicity refers to cell rounding or retraction, not necessarily cell death. Experiments were repeated at least twice.
Mammalian cell extracts were prepared by lysing cells in lysis buffer (50 mM Tris [pH 7.5], 500 mM NaCl, 1% Triton X-100, 0.1% SDS, 10 mM MgCl2, 1 mM Pefabloc, 1 mM benzimidine hydrochloride, 1 μM leupeptide, 1 μM E64) and collecting the supernatant after centrifugation at 13,000 rpm in a microcentrifuge for 10 min. Approximately 1 μg of GST-DR2 was incubated with 30 μg of HEK293T or BAT2 cell extract in 40-μl adenylylation reaction mixtures containing 25 mM Tris-HCl (pH 7.5), 3.0 mM MgCl2, 1 mM dithiothreitol (DTT), 0.5 mM EDTA, and 1 μCi [α-32P]ATP for 1 h at 30°C. Reactions were stopped by adding Nupage loading buffer (Invitrogen). Extract samples containing 30 μg protein were used for Western analyses. Reaction products were separated by SDS-polyacrylamide gel electrophoresis and visualized by Coomassie blue staining (loading controls), autoradiography, or Western blot analysis with antibody to RhoA (Cell Signaling Technology, Danvers, MA), Cdc42 (BD Transduction Laboratories, Lexington, KY), or Rac1 (Abcam Inc., Cambridge MA).
To determine whether antibody neutralizes the cytotoxicity of live viable bacteria or CCS for BAT2 cells, heat-decomplemented filtered-sterilized rabbit anti-DR2/Fic serum, rabbit preimmune serum, bovine convalescent-phase serum, or bovine preimmune serum was added to a suspension of H. somni cells (106 CFU/ml) or 20× CCS to obtain a final serum concentration of 1:100. The mixtures containing bacteria or CCS and serum were incubated with shaking for 45 min at room temperature for the preparation containing bacteria or at 4°C for the preparation containing CCS. The mixtures were then added to the BAT2 cells. After 4 h of incubation cells were washed and fixed, and the cytotoxicity was quantified by microscopy as described above.
Confluent BAT2 and BT cell monolayers in 24-well culture plates were washed three times in DMEM and infected with H. somni in tissue culture medium without antibiotics using an MOI of 50. Plates were centrifuged at 500 × g for 5 min and incubated for 1.5 h at 37°C in a humidified 5% CO2 incubator. Cells were then washed gently five times with PBS to remove nonadhering bacteria. The numbers of viable cell-associated bacteria were determined by lysing cells with 0.5% Triton X-100 for 5 min at room temperature, diluting the preparations in PBS with vigorous pipetting, plating the preparations, and counting the CFU. Preliminary experiments showed that treatment of H. somni with 0.5% Triton X-100 for 5 min did not decrease the viable cell counts (data not shown). Invasion was assessed by quantifying the intracellular bacteria using a gentamicin protection assay. After infection for 1.5 h and five washes to remove nonadherent bacteria as described above, fresh medium containing gentamicin (100 μg/ml) was added. The cells were then incubated for 1 h, washed three times with PBS, and lysed, and dilutions were plated to count the internalized bacterial CFU.
BAT2 cells were grown on glass coverslips in a 12-well culture plate to 80% to 90% confluence. Cells were treated with virulent strain 2336 or IbpA-negative strain 129Pt using an MOI of 10, with 20× CCS, or with rDR2/Fic protein (20 μg/ml) for 4 h as described above for the cytotoxicity assay. This step was included to determine attachment and uptake of IbpA protein. After treatment with the IbpA-containing preparations described above, cells were washed three times with PBS, fixed with 4% paraformaldehyde, and permeabilized or not permeabilized with 0.1% Triton X-100 for 5 min before they were treated with rabbit anti-DR2/Fic (1:100 in 0.5% BSA in PBS) for 2 h at room temperature. The Triton X-100 treatment was included to permeabilize the cells to the antibody in order to detect IbpA DR2/Fic which would have been taken up during the incubation step before fixation and Triton X-100 treatment. Cells were then washed, incubated for 2 h at room temperature with Alexa Fluor 488-conjugated goat anti-rabbit antibody at a 1:500 dilution, washed again, and stained with rhodamine-phalloidin for F-actin labeling and with TOPRO-3 for labeling of nuclei (Invitrogen). Slides were dried, mounted in Prolong antifade reagent (Invitrogen), and examined with a Leica TCS SP5 confocal microscope (Leica Microsystems, Bannockburn, IL). Alexa Fluor 488 was excited with a 488-nm laser, and emission was measured at wavelengths of 515 ± 30 nm; rhodamine was excited at 568 nm, and emission was measured at 600 ± 40 nm; and TOPRO-3 was excited at 635 nm, and emission was measured at wavelengths of >650 nm. A series of z-section images (0.25 μm) were collected and analyzed with ImageJ software (http://www.rsb.info.nih.gov).
BAT2 cells were grown on polycarbonate Transwell inserts (diameter, 6.5 mm) with 3 μm-pore-size filters in a 24-well plate (Corning, Cambridge, MA). The Transwell filters were first coated with gelatin/FBS, dried, and covered with DMEM for 1 h at 37°C before BAT2 cells were added at a concentration of 2 × 103 cells/well. Cells were grown for 5 days to obtain a complete monolayer. Then cells were washed, and 105 CFU of H. somni strain 2336 or 129Pt in 80 μl cell culture medium (without antibiotics) was added to each Transwell insert (MOI, approximately 10). The lower chamber was filled with 250 μl of cell culture medium. After 3 h of incubation, each Transwell insert was removed, and the contents of the lower chamber were diluted and plated on BHI blood agar plates to determine viable bacterial counts. In order to assess the effect of antiserum on H. somni 2336 transmigration across a BAT2 monolayer, bacteria were incubated with anti-DR2/Fic serum, convalescent-phase bovine serum, or the preimmune serum controls at a 1:100 dilution for 45 min at room temperature before they were transferred onto a cell monolayer. In a separate experiment to determine whether the presence of IbpA DR2/Fic resulted in migration of strain 129Pt, a BAT2 monolayer were pretreated for 4 h with either purified rDR2/Fic protein or the mutant protein (both at a concentration of 20 μg/ml), H. somni 2336 20× CCS, or cell culture medium alone before the bacteria were added.
Data were analyzed by one-way analysis of variance (ANOVA) using GraphPad Prism software (GraphPad Software, La Jolla, CA). Treatment groups were compared using Bonferroni's multiple-comparison tests.
We previously showed that H. somni and its CCS are toxic for human HeLa cells (27), but it was clear that a model relevant for the host-specific disease was needed. Therefore, we developed an in vitro model using cells from the bovine upper respiratory tract (BT cells) and pulmonary alveolar epithelial cells (BAT2 cells). In stark contrast to the results obtained with HeLa cells, treatment with live virulent H. somni strain 2336, its CCS, or rDR2/Fic protein resulted in robust cell rounding and retraction of BAT2 cells (Fig. (Fig.1A).1A). Carrier strain 129Pt did not cause rounding or retraction (Fig. (Fig.1A),1A), nor did its CCS (data not shown). Strain 2336 CCS contained abundant high-molecular-weight proteins that reacted with anti-DR2/Fic or convalescent-phase serum in Western blots (Fig. (Fig.1B),1B), indicating that IbpA was present. CCS of strain 129Pt was negative for IbpA in the same blots. In cytotoxicity tests, the BAT2 cells were the most susceptible cells, followed by BT cells, and the human HeLa cells were the least susceptible cells (Fig. (Fig.1C,1C, D, and E). A mutant rDR2/Fic protein (rDR2/Fic H/A) with the critical His residue at position 298 in the Fic motif replaced by Ala did not cause cytotoxicity when it was used at concentrations equivalent to those of rDR2/Fic (Fig. (Fig.1E).1E). These results showed that an active Fic motif is required for IbpA-mediated cytotoxicity and that IbpA's effects were strongest against bovine targets relevant to H. somni disease. Since BAT2 cells were the most susceptible cells and were the cells most relevant to disease, these cells were used for the rest of this study.
Previously, we showed that the IbpA DR1 and DR2/Fic domains adenylylate Rho GTPases and disrupt their downstream signaling cascades in human HEK293T epithelial cell extracts (27). Therefore, here we determined whether Rho GTPases were targets for IbpA DR2/Fic-mediated adenylylation in BAT2 cells. Like control HEK293T cells, rDR2/Fic2 adenylylated endogenous proteins migrated at the size of Rho GTPases (Table (Table1)1) when they were incubated with cell lysates of BAT2 cells, as determined by autoradiography and Western blot analysis for RhoA (Fig. (Fig.2A)2A) or for Rac1 and Cdc42 (Fig. (Fig.2B).2B). Similar results were obtained when both cell lysates were treated with CCS from H. somni 2336 but not when they were treated with CCS from IbpA-deficient strain 129Pt (Fig. (Fig.2C).2C). Since bovine and human Rho GTPases are 100% identical at the amino acid level (Table (Table1)1) and, as previously shown (27), both rDR2/Fic and H. somni strain 2336 CCS adenylylate purified human GST-RhoA, -Rac, and -Cdc42 in vitro, data shown in Fig. Fig.22 confirm Fic-mediated adenylylation of bovine Rho GTPases. Interestingly, H. somni 2336 CCS and rDR2 displayed much higher intensities of radiolabeled Rho GTPases in BAT2 cell extracts than in HEK293T cell extracts, which correlates with the much higher concentrations of RhoA, Rac1, and Cdc42 in BAT2 cells than in HEK239T cells, as determined by Western blotting (Fig. (Fig.2A,2A, B, and C).
Convalescent-phase serum passively protects against bovine pneumonia, and immunization with rDR2 protein protects mice against septicemia, so we tested the abilities of bovine convalescent-phase serum and rabbit antibody to rDR2/Fic to neutralize cytotoxicity in vitro. Live H. somni strain 2336 and its CCS were treated with the antisera for 45 min prior to treatment of BAT2 cells. Both antisera reacted with high-molecular-weight IbpA from strain 2336 in Western blots (Fig. (Fig.1B).1B). The rabbit anti-DR2/Fic and convalescent-phase bovine sera significantly decreased the cytotoxicity of live bacteria or CCS for BAT2 cells compared with the corresponding preimmune sera (P < 0.05) (Fig. (Fig.3A3A and and3B3B).
Since H. somni is found in alveoli during experimental and spontaneous pneumonia (3, 9), attaches to BT cells in vitro (24), and causes septicemia (5, 11), we investigated the possibility that H. somni attaches to BT and/or BAT2 cells and then invades the cells in order to cross the epithelial barrier. After 1.5 h of incubation with cells and washing, large numbers of H. somni bacteria were found to attach to both BT and BAT2 cells (Fig. (Fig.4).4). However, H. somni did not invade the cells in the gentamicin protection assay (Fig. (Fig.4).4). This raises two questions. First, does IbpA get into cells to access Rho GTPases? Second, how does H. somni cross the alveolar epithelial barrier to cause septicemia?
Confocal immunomicroscopy was used to determine whether the IbpA DR2/Fic protein binds to cell surfaces and whether it is internalized in BAT2 cells. Confocal images of BAT2 control cells, which were not treated with H. somni or CCS but were permeabilized and labeled with antibody to rDR2/Fic, showed no green immunofluorescence or retraction (Fig. (Fig.5A).5A). Cells infected with the IbpA-negative strain 129Pt also showed no immunofluorescence or retraction (Fig. (Fig.5B).5B). However, in cells infected with H. somni 2336 and immunolabeled for DR2/Fic antigen, DR2/Fic was detected on retracting cell surfaces in nonpermeabilized cells (Fig. (Fig.5C).5C). Much more DR2/Fic antigen was located in the cytoplasm in the z sections of retracting cells treated with H. somni, CCS, or rDR2/Fic protein and permeabilized with Triton X-100 before they were stained with the primary rabbit anti-rDR2/Fic antibody (Fig. (Fig.5D,5D, E, and F, respectively). Control cells treated with rGST protein alone were not stained, whether they were permeabilized or not permeabilized (data not shown). These studies showed that IbpA DR2/Fic attaches to the cell surface and is internalized in BAT2 cells. This model provides a means to investigate the mechanism of IbpA internalization in relevant host cells.
The data that we have obtained so far indicate that H. somni does not gain access to the bloodstream by invasion through BAT2 cells, so we determined whether H. somni crosses the alveolar epithelial barrier between retracted cells by paracellular migration to cause septicemia. Large numbers of virulent H. somni strain 2336 cells were recovered from the lower chambers of Transwell plates after 3 h of incubation (about 50% of the inoculum), whether cells were pretreated with medium or cells were pretreated with preimmune serum (Fig. (Fig.6A).6A). However, transmigration was significantly inhibited by the presence of antibody to rDR2/Fic or passively protective convalescent-phase serum (P < 0.01) (Fig. (Fig.6A).6A). Almost none of the IbpA-deficient strain 129Pt cells were recovered from the lower compartment unless the BAT2 monolayers were pretreated with the rDR2/Fic protein or the CCS of strain 2336 for 4 h to induce cell retraction before addition of the bacteria (Fig. (Fig.6B).6B). Pretreatment with the rDR2/Fic H/A mutant (with inactive Fic) did not significantly increase transmigration compared with the untreated control (P > 0.05). Both the antibody neutralization of strain 2336 transmigration and the facilitation of strain 129Pt transmigration by pretreatment with rDR2/Fic or strain 2336 CCS implicate DR2/Fic-induced BAT2 cell retraction in paracellular invasion through the alveolar epithelial barrier, as shown in a simplified model of H. somni migration between both alveolar epithelial cells and endothelial cells to cause septicemia (Fig. (Fig.77).
This study showed that IbpA-producing H. somni strain 2336, IbpA-enriched culture supernatant of this strain, and the rDR2/Fic protein are cytotoxic for relevant host BT cells and BAT2 cells but have only a minimal toxic effect on the human cervical carcinoma HeLa cell line, consistent with the strict ruminant host specificity of H. somni (5, 11). The greater sensitivity of BAT2 cells than of BT cells is consistent with the severe pulmonary disease but mild upper respiratory disease caused by bovine H. somni infection (5, 11). The previously reported detection of H. somni or the IbpA antigen primarily in the pulmonary alveoli of animals with pneumonia also is consistent with the high level of susceptibility of BAT2 cells (3, 9). Recently, we demonstrated that the IbpA DR2/Fic cytotoxicity in human cell lines is due to inactivation of the Rho GTPases by adenylylation, which results in disruption of the cytoskeletal network (27). Here we show that IbpA DR2/Fic adenylylates the Rho GTPases of BAT2 cells as well. The higher intensity of adenylylation in BAT2 cells than in HEK239T cells correlates with the higher concentration of Rho GTPases in BAT2 cells. Perhaps the severe retraction and rounding of BAT2 cells are related to the high concentration of Rho GTPases. Adenylylation of host Rho GTPases by bacterially secreted Fic family virulence proteins is a new frontier in microbial pathogenesis, as reported in the last year (14, 18, 27, 28). Two such Fic-containing virulence proteins, VopS of Vibrio parahaemolyticus and AnkX of Legionella pneumophila, have been shown to access the cellular cytoplasm by type III and type IV secretion systems, respectively (17, 28). However, IbpA is secreted to the bacterial surface by a two-partner secretion system (23), so it is not clear how IbpA gains access to the host cell cytoplasm to inhibit Rho GTPase signaling. Our experiments show that IbpA does not reach the cytoplasm by invasion of the bovine epithelial cells by H. somni. Rather, IbpA released from the bacterial surface is internalized by cells, as determined by detection of DR2/Fic antigen inside cells by confocal immunomicroscopy after treatment of cells with live bacteria, CCS, or rDR2/Fic protein.
The lack of invasion of bovine alveolar epithelial cells by H. somni also raised the question of how this organism reaches the bloodstream to cause septicemia. Experiments with BAT2 monolayers in Transwells indicated that H. somni passes between retracted BAT2 cells to cross the epithelial barrier by paracellular migration. A similar paracellular route may be used to cross the endothelial cells beneath the alveolar epithelial cells. Indeed, Behling-Kelly et al. (2) showed that H. somni increased the permeability of a bovine brain microvascular endothelial monolayer. It is likely that H. somni crosses the pulmonary microvascular endothelium in a similar fashion to cause septicemia, followed by myocarditis, where large numbers of the organism are found on the endothelial cells (16). Antibody to rDR2/Fic and convalescent-phase serum from calves with experimental H. somni pneumonia neutralized IbpA DR2/Fic cytotoxicity. This confirms that the Fic domain has a role in cell rounding and retraction. Antibody neutralization of in vitro cytotoxicity due to rDR2/Fic also is consistent with the previously demonstrated immunoprotection of animals against H. somni in active immunization studies with rDR2/Fic (7) and in passive immunization studies with the same convalescent-phase serum (8), which recognizes IbpA (29). The DR2/Fic antiserum and the bovine convalescent-phase serum also blocked bacterial transmigration across a BAT2 monolayer, suggesting that the protection against pneumonia and subsequent septicemia may be due to prevention of bacterial entry into the subepithelium at early stages of infection in the lungs. The neutralization of both cytotoxicity and paracellular migration by antibody in this relevant bovine respiratory alveolar barrier model along with the in vivo protection is convincing evidence that IbpA DR2/Fic is involved in both the virulence of H. somni and immunoprotection in the natural host.
In conclusion, IbpA DR2/Fic is cytotoxic for natural host epithelial target cells, and the relevant BAT2 cells are the most susceptible cells. IbpA adenylylates and inactivates BAT2 Rho GTPases, causing cell rounding and retraction due to collapse of the cellular cytoskeleton. The retraction of alveolar epithelial cells allows H. somni to cross the respiratory alveolar epithelial barrier by paracellular migration to reach the microvasculature. Neutralization by antisera recognizing IbpA DR2/Fic likely accounts for immunoprotection in vivo and confirms the role of IbpA DR2/Fic in pathogenesis.
This work was supported by USDA NRI grant 2005-35204-6257.
We thank J. E. Dixon for critical reading of the manuscript, L. J. Gershwin for kind provision of BT cells, and J. C. Mendez for technical assistance. We have no competing interests.
Editor: J. B. Bliska
Published ahead of print on 22 February 2010.