Listeria monocytogenes Uses Listeria Adhesion Protein (LAP) To Promote Bacterial Transepithelial Translocation and Induces Expression of LAP Receptor Hsp60

Bacterial strains and growth conditions. L. monocytogenes F4244 (wild type [WT]; serovar 4b), the isogenic lap-deficient insertion mutant KB208 (lap strain), and lap-complemented CKB208 (lap⁺) were used as described previously (6, 32) (Table 1). An inlA deletion mutant (ΔinlA; AKB301) and its complement (inlA⁺; AKB302) were generated for this study. All L. monocytogenes strains were grown in brain heart infusion (BHI) broth (Becton Dickinson) at 37°C, unless indicated otherwise. The lap strain was grown with erythromycin (5 μg/ml) at 42°C, the lap⁺ strain was grown with erythromycin (5 μg/ml) and chloramphenicol (5 μg/ml) at 37°C, and the inlA⁺ strain was grown with chloramphenicol (5 μg/ml) at 37°C.

Bacterial mutagenesis and complementation.A modification of splicing by overlap extension (SOE) was used to generate a 5′-3′ in-frame deletion of the inlA open reading frame (ORF) (ΔinlA) (10). Oligonucleotide primers InlAUSF and InlAUSR (Table 1) generated a 403-bp product from the 5′ end of the inlA locus. A 400-bp 3′ product was amplified using primers InlADSF and InlADSR (Table 1). The 5′ and 3′ products were combined in equimolar concentrations in a ligation reaction mixture to yield an 803-bp 5′-3′ SOE fragment. This product was digested with BamHI and SacI, cloned into pGEM-T Easy (Promega, Madison, WI), and subcloned into the temperature-sensitive pAUL-A shuttle vector (11). Electrocompetent L. monocytogenes F4244 WT cells were transformed with the construct and subjected to temperature-dependent allelic exchange (10, 11). In-frame deletion of inlA was confirmed by PCR, using primers InlAUSF and InlADSR. Immunoblotting confirmed the absence of InlA but the presence of LAP in the L. monocytogenes ΔinlA strain.

To complement the ΔinlA mutant with the inlA gene (inlA⁺), oligonucleotide primers InlABamF and InlASphR (Table 1) were used to amplify the entire inlA ORF from L. monocytogenes F4244 genomic DNA. Amplified inlA was cloned via BamHI and SphI sites into pGEM-T Easy and subcloned into pMGS101 (22). Electrocompetent L. monocytogenes ΔinlA cells were transformed with pMGS101-inlA, and positive transformants were selected by plating in the presence of chloramphenicol.

Cell culture.Secondary human enterocyte-like Caco-2 cells (HTB37; American Type Culture Collection) were grown in Dulbecco's modified Eagle medium (DMEM; Invitrogen) containing 10% fetal bovine serum (D10F) (Atlanta Biologicals) at 37°C under 7% CO₂ in a humidified incubator. Cells (passages 25 to 35) were seeded at approximately 5 × 10⁴ to 10 × 10⁴ cells/well into 12- or 24-well plates (Corning), and cell monolayers were used between 10 and 14 days later. Confluence was typically achieved at 4 to 5 days, and by 10 days, Caco-2 monolayers were polarized (24).

Antibodies.An anti-Hsp60 antibody was generated via immunization of a New Zealand White rabbit with purified human Hsp60 (Assay Design, Ann Arbor, MI). Antisera were collected at the Purdue University Small Animal Care Facilities, and anti-Hsp60 antibody was purified using a protein G affinity column. An anti-InlA antibody was generated via immunization of a New Zealand White rabbit with purified truncated InlA (amino acid residues 36 to 496) (54). Antisera were collected, and anti-InlA antibody was purified by use of a protein A column. Other antibodies included an anti-β-actin (43 kDa) monoclonal antibody (MAb) (Abcam), an anti-ZO1 polyclonal antibody (PAb) (Invitrogen), anti-LAP MAb H7, and IgG control MAb C11E9, from our lab.

RNA interference and overexpression of Hsp60.Short hairpin RNA (shRNA) constructs targeting human hsp60 mRNA (SureSilencing shRNA-HSPD1) and a nontargeting control shRNA vector (SABiosciences, Frederick, MD) were used to generate a Caco-2 cell line with stable suppression of Hsp60 expression (Table 1). To create a Caco-2 cell line exhibiting overexpression of Hsp60, full-length human hsp60 cDNA was subcloned from pOTB7-hsp60 (Open Biosystems) into the expression vector pcDNA3.1 (Invitrogen) (Table 1). The integrity of hsp60 within the expression construct (pCDNA3.1-hsp60) was verified by sequencing. A pCDNA3.1 plasmid without an insert was used as a vector control. Caco-2 cell transfections were performed using Lipofectamine LTX Plus reagent according to the manufacturer's protocol (Invitrogen). Stable transformants were selected by growth in D10F containing 800 μg/ml Geneticin sulfate (G418) (Sigma). Hsp60 expression levels were monitored by reverse transcription-PCR (RT-PCR) and Western blotting.

Analysis of Hsp60 expression in Caco-2 cell fractions by SDS-PAGE and immunoblotting.Caco-2 cell monolayers were washed twice with cold phosphate-buffered saline (PBS) (0.1 M; pH 7.0), and membrane and intracellular proteins were isolated using a Mem-Per eukaryotic protein extraction kit (Pierce) (64). Protein preparations were desalted (Zeba desalting spin columns; Pierce), precipitated using acetone, and resuspended in sample solvent (4.6% SDS, 0.5% β-mercaptoethanol, PBS, pH 7.0). The protein concentration was determined by a reducing agent-compatible bicinchoninic acid (BCA) assay (Pierce). A lactate dehydrogenase (LDH) assay was performed with acetone-precipitated protein fractions dissolved in DMEM to rule out contamination of membrane fractions with intracellular proteins (data not shown). Equal protein concentrations (15 μg/lane) from Caco-2 membrane and cytosolic fractions were separated by SDS-PAGE (10% acrylamide). Proteins were transferred to Immobilon-P membranes (Millipore) and immunoprobed with anti-Hsp60 PAb (1.4 mg/ml; 1:1,000 dilution) and β-actin-specific MAb (1.8 mg/ml; 1:1,000 dilution). Bands were detected using horseradish peroxidase-coupled anti-mouse or anti-rabbit antibodies (Jackson Immuno Research, West Grove, PA) with enhanced chemiluminescence substrate (Pierce) and were developed on X-ray films. To compare reaction intensities, average band densities were determined with Quantity One software (Bio-Rad).

Bacterial adhesion and invasion assays.Fresh bacterial cultures were washed and resuspended in D10F and then added to Caco-2 monolayers at a multiplicity of infection (MOI) of ∼10. To measure bacterial adhesion, monolayers were washed after 1 h of infection, and adherent bacteria were enumerated by being plated on BHI agar as previously described (32). For bacterial invasion, monolayers were washed after 1 h of infection and incubated with D10F containing gentamicin (50 μg/ml) for 1 h (47). Caco-2 cells were lysed with 0.1% Triton X, and internalized bacteria were enumerated by plate counting as described before (47).

Transepithelial bacterial translocation assay.Caco-2 monolayers were grown to confluence on Transwell filter inserts (4-μm pore size; Corning) and then placed in 12-well tissue culture plates. Transepithelial electrical resistance (TEER) of polarized monolayers was measured (Voltometer; Millipore), and those with a minimum TEER of about 200 Ω/cm² (range, 190 to 209 Ω/cm²) were used for translocation experiments (15, 27). Furthermore, to ensure monolayer integrity on Transwell inserts, Caco-2 cells from representative experiments were immunostained using antibodies specific to the tight junction proteins ZO-1, occludin, and claudin and Cy5- and fluorescein isothiocyanate (FITC)-conjugated secondary antibodies and were examined by confocal microscopy as described in “Fluorescence microscopy.” Bacteria (MOI, ∼10) were added to the apical well of the Transwell system, and after 2 h at 37°C in 7% CO₂, liquid was collected from the basal well and translocated bacteria were enumerated by plate counting (15). Initial experiments were conducted to determine the appropriate length of time for monitoring bacterial translocation through Caco-2 monolayers, and 2 h was chosen because at this time point bacteria in the basal well most closely represented the number of translocated microorganisms, with minimal posttranslocation replication (data not shown).

Effect of Hsp60 expression on Listeria interaction with Caco-2 cells.To evaluate the role of Hsp60 expression in LAP-mediated infection, Caco-2 cells were subjected to heat stress (41°C) or tumor necrosis factor alpha (TNF-α) (Sigma) exposure at 37°C (100 U/ml D10F) (20) for 1 h, followed by a 3-h recovery period at 37°C with 7% CO₂. Flow cytometric analyses were performed to determine that 3 h was the optimum recovery period required for surface Hsp60 expression in Caco-2 cells following heat shock (data not shown). Hsp60 expression in Caco-2 cells was monitored by immunoblotting. Bacterial adhesion and translocation were measured as described above.

Influence of direct or indirect exposure of Caco-2 cells to Listeria on Hsp60 expression.Caco-2 cells were exposed to various doses (10⁴, 10⁶, and 10⁸ CFU/ml, corresponding to MOIs of 0.1, 10, and 100, respectively) of the L. monocytogenes WT or lap strain or of Listeria innocua for 1 h, followed by a 3-h recovery in D10F containing gentamicin (50 μg/ml). Initial flow cytometric analyses determined that 3 h was the optimal recovery period for surface Hsp60 expression following L. monocytogenes infection (data not shown). To monitor Hsp60 expression following direct bacterial exposure, bacteria were added to the apical side of Caco-2 monolayers grown in a 12-well plate. For indirect exposure, bacteria were separated from monolayers via a Transwell filter insert (0.4-μm pore size; Corning). Hsp60 expression in Caco-2 cells was monitored by immunoblotting as described above.

Influence of initial (primary) Listeria infection on susceptibility to subsequent (secondary) LAP-mediated adhesion and transepithelial translocation.A series of experiments was conducted to determine whether Listeria-induced Hsp60 expression affects host cell susceptibility to subsequent (secondary) LAP-mediated infection.

The first experiment was conducted to determine whether initial exposure to WT L. monocytogenes or L. innocua would affect susceptibility to subsequent L. monocytogenes adhesion. L. monocytogenes and L. innocua were chosen for the initial infection because they differed in the ability to induce Caco-2 cell Hsp60 expression. Untransfected Caco-2 cells were exposed to 10⁶ CFU/ml (MOI, ∼10) of the L. monocytogenes WT or lap strain or of L. innocua for 1 h (primary infection) or were left untreated. Noninternalized bacteria were killed by gentamicin (50 μg/ml) treatment, and after a 3-h recovery period in gentamicin-containing D10F, cells were thoroughly washed and exposed again to WT L. monocytogenes (10⁶ CFU/ml) as a secondary infection. Adherent bacteria from the secondary infection were enumerated at 1 h postinfection as described above. To account for the number of bacteria that were internalized during the primary infection and had undergone intracellular replication, control wells were included which had been subjected only to the 1-h primary infection and then maintained in D10F with gentamicin instead of being exposed to the secondary infection. Intracellular bacteria were enumerated from these control wells, and the average value was subtracted from the number of adherent bacteria after the secondary infection.

A second set of experiments was conducted to determine the roles of Hsp60 and LAP in mediating susceptibility to secondary infection. In these experiments, control shRNA (control-sh; mock-transfected)- or shRNA-Hsp60-transfected Caco-2 cells were exposed to L. monocytogenes in a primary infection or were left uninfected. Caco-2 cells were exposed to L. monocytogenes in a primary infection or were left uninfected. Cells were recovered for 3 h as in the first experiment, as described above, and then were exposed to the WT, lap, lap⁺, ΔinlA, or inlA⁺ strain in a secondary infection. Secondary bacterial adhesion and translocation were measured in separate experiments. As in the first experiment (above), control wells were included to account for intracellular bacteria resulting from the primary L. monocytogenes infection.

Fluorescence microscopy.To evaluate surface localization of Hsp60 by immunofluorescence microscopy, Caco-2 cells were grown to confluence in D10F on LabTek chamber slides (Nunc, ThermoFisher Scientific). Cells were washed with Hanks balanced salt solution (HBSS) (Cellgro) and then incubated with anti-Hsp60 PAb (diluted 1:250 in HBSS) for 1 h at 37°C. After being washed with HBSS containing 1% bovine serum albumin (BSA), cells were incubated with FITC-conjugated monovalent secondary Fab fragment (diluted 1:250 in HBSS) for 1 h at 37°C. Propidium iodide (1 μg/ml in HBSS) was added for 10 min to stain dead cells, and then monolayers were washed with HBSS and fixed with 4% paraformaldehyde diluted 1:1 in HBSS at room temperature for 10 min. Cells were mounted with ProLong Gold antifade reagent (Molecular Probes) and examined under an epifluorescence microscope (Leica, Wetzlar, Germany) equipped with Spot software (Sterling Heights, MI).

To evaluate intracellular Hsp60 levels, confluent Caco-2 cells were fixed with paraformaldehyde as described above and then permeabilized with 0.01% Triton X in HBSS. Monolayers were immunoprobed at room temperature with anti-Hsp60 PAb (1:250) and FITC-conjugated monovalent secondary Fab fragment (1:250), as well as with anti-ZO1 PAb (1:250) (Molecular Probes) and Cy5-conjugated monovalent Fab fragment (1:250), as described above. In some cases, cells were treated with Hoechst dye (0.5 μg/ml in HBSS) for nuclear (blue) staining. Images were visualized at the Purdue Life Science Fluorescence Imaging Facility, using a Zeiss LSM 710 confocal fluorescence microscope (Carl Zeiss, Jena, Germany) with a 63×/1.4 water immersion objective. Images were acquired and processed (colocalization) using Zeiss LSM Image Browser software (Carl Zeiss, Jena, Germany).

Statistical analysis.The SAS program (Cary, NC) was employed for statistical analyses. Differences between pairs were assessed by Wilcoxon rank sum analysis, and the Proc GLM test was used for comparison of data from more than three groups (SAS, Cary, NC). P values of <0.05 were considered significant.

LAP-Hsp60 association promotes L. monocytogenes adhesion to Caco-2 cells.We previously demonstrated that a lap-deficient mutant exhibited reduced adhesion to Caco-2 cells (32, 58). In this study, to confirm the role of the Hsp60-LAP interaction in promoting adhesion, we evaluated LAP-mediated adhesion when levels of cellular Hsp60 were altered. Induction of Caco-2 Hsp60 expression via heat stress (41°C) or exposure to TNF-α (100 U/ml) (20) led to increased levels of Hsp60 in both intracellular (44% increase for heat shock and 50% increase for TNF-α exposure) and plasma membrane (69% increase for heat shock and 80% increase for TNF-α exposure) protein fractions (Fig. 1A). Adhesion of the WT, but not the lap strain, to heat-stressed and TNF-α-treated Caco-2 cells increased 2.2-fold (P = 0.01) and 1.9-fold (P = 0.03), respectively, compared to that with unstressed Caco-2 cells (Fig. 1B). The contribution of Hsp60 to this interaction was further confirmed by pretreating stressed cells with an Hsp60-specific antibody prior to infection, which reduced adhesion of the WT to levels similar to those with unstressed Caco-2 cells (P = 0.01) (Fig. 1B).

• Open in new tab
• Download powerpoint

FIG. 1.

Influence of Caco-2 Hsp60 expression on LAP-mediated adhesion of L. monocytogenes. (A) Immunoblot showing the level of Hsp60 expression in Caco-2 plasma membrane (Mem-Hsp60) and intracellular (Int-Hsp60) fractions after heat stress (41°C, 1 h) and TNF-α exposure (100 U/ml, 1 h). β-Actin (43 kDa) was used as an internal control. Proteins were loaded at 15 μg/well. (B) Adhesion of WT, lap (mutant), and lap⁺ (lap-complemented) L. monocytogenes strains to Caco-2 monolayers following Caco-2 heat stress (41°C) and TNF-α (100 U/ml) exposure and after treatment with anti-Hsp60 PAb (1 μg/ml). Adhesion data are averages for at least three independent experiments performed in quadruplicate (n ≥ 12). Error bars represent standard errors of the means (SEM). Lowercase letters (a and b) represent significant differences (P < 0.05) in adhesion of bacterial strains.

To further examine the effect of Hsp60 on Listeria adhesion, we reduced the levels of endogenous Hsp60 by transfecting Hsp60-specific shRNA into Caco-2 cells, which resulted in reductions of Hsp60 protein expression of approximately 70% in intracellular fractions and 90% in plasma membrane fractions (Fig. 2 B). Hsp60 suppression in Caco-2 cells was also confirmed by RT-PCR (Fig. 2A). Partial suppression of Hsp60 by RNA interference was reported by others (9, 13); complete Hsp60 suppression may not be possible because it has important antiapoptotic functions in cultured cell lines (25). Adhesion of the WT decreased >4.5-fold in Hsp60-suppressed cells, whereas adhesion of the lap strain was unchanged (Fig. 2C). Overexpression of Hsp60 via pcDNA3.1-hsp60 resulted in a 60% Hsp60 increase in intracellular protein fractions and a 75% increase in plasma membrane protein fractions (Fig. 2B). Adhesion of the WT increased 2.8-fold (P = 0.006) for cells overexpressing Hsp60, while adhesion of the lap strain was unaffected (Fig. 2C).

• Open in new tab
• Download powerpoint

FIG. 2.

Analysis of L. monocytogenes adhesion to Caco-2 cells following Hsp60 knockdown by shRNA (shRNA Hsp60−) or Hsp60 overexpression (Hsp60+) or adhesion to control-sh-transfected cells (Caco-2 cells transfected with noncoding shRNA). (A) RT-PCR analysis of hsp60 expression (23 cycles) in Caco-2 cells following shRNA-Hsp60 knockdown. 18S RNA was used as an internal control. (B) Immunoblot showing Hsp60 suppression (shRNA Hsp60−) and overexpression (Hsp60+) in Caco-2 plasma membrane (Mem-Hsp60) and intracellular (Int-Hsp60) fractions. β-Actin (43 kDa) was used as an internal control. Proteins were loaded at 15 μg/well. (C) Adhesion of L. monocytogenes WT, lap (mutant), and lap⁺ (lap-complemented) strains to control-sh, shRNA-Hsp60, and Hsp60⁺ Caco-2 monolayers. Adhesion data are averages for at least three independent experiments performed in quadruplicate (n ≥ 12) and are presented with SEM. Lowercase letters (a, b, and c) indicate significant differences (P < 0.05) in adhesion of individual bacterial strains to Caco-2 monolayers expressing varied levels of Hsp60.

LAP-Hsp60 interaction promotes Listeria monocytogenes translocation through epithelial monolayers.Although adhesion is a crucial step in L. monocytogenes pathogenesis, binding of bacteria to epithelial cells does not always directly correlate with dissemination to target organs (30). We therefore investigated whether LAP was implicated in Listeria invasion of intestinal epithelial cells, which could contribute to the differences in systemic dissemination observed between LAP-deficient and WT L. monocytogenes strains (6, 31). In gentamicin protection invasion assays, internalization of the WT was greater than that of the lap strain (P = 0.003), but both strains were significantly more invasive than the ΔinlA strain (Fig. 3). However, neither Hsp60 suppression nor overexpression (Fig. 3) influenced invasion of the WT, lap, or ΔinlA strain (Fig. 3), indicating that the LAP-Hsp60 interaction may not be directly responsible for invasion of L. monocytogenes into epithelial cells.

• Open in new tab
• Download powerpoint

FIG. 3.

Analysis of L. monocytogenes WT, lap, lap⁺, ΔinlA, and inlA⁺ strain invasion of Caco-2 cells following Hsp60 knockdown by shRNA (shRNA Hsp60−) or Hsp60 overexpression (Hsp60+) or of control-sh-transfected cells (Caco-2 cells with scrambled shRNA). Invasion results are averages for at least three independent experiments performed in quadruplicate (n ≥ 12) and are presented with SEM. Lowercase letters (a to c) indicate significant differences (P < 0.05) in invasion among bacterial strains.

We next sought to determine the role of LAP in translocation through Caco-2 monolayers. Caco-2 monolayer integrity was verified by confocal microscopy, with visible tight junction borders and no alteration in cellular distribution of tight junction proteins before and after treatment with L. monocytogenes (Fig. 4). Furthermore, TEER values for intact Caco-2 cell monolayers ranged from 190 ± 2 to 209 ± 6 (Table 2) and were consistent in all our experiments throughout this study, and these values are within the range reported by others (15, 27). As a positive control, treatment of monolayers with cytochalasin D induced a very low TEER (Table 2) (60). Following 2 h of bacterial exposure (MOI, ∼10), translocation of the lap strain through Caco-2 monolayers was 60% less than that of the WT (P = 0.02) (Fig. 5 B); however, there were no apparent differences in TEER values between WT- and lap mutant-treated Caco-2 cells (Table 2). Increasing Caco-2 Hsp60 expression via endogenous overexpression enhanced WT and lap⁺ strain translocation 2-fold (P = 0.001) and 1.6-fold (P = 0.01), respectively, but did not influence translocation of the lap strain (Fig. 5B). shRNA suppression of Hsp60 decreased translocation of the WT >3-fold (P = 0.004), but transit of the lap strain to the basal compartment (Fig. 5B) was not affected. Similarly, addition of exogenous Hsp60 to the surfaces of Caco-2 cells resulted in greater translocation of the WT (P = 0.001) but not of the lap strain (data not shown). These data indicate a clear role for the LAP-Hsp60 interaction in mediating transepithelial translocation of L. monocytogenes.

• Open in new tab
• Download powerpoint

FIG. 4.

Confocal microscopic analysis of tight junction integrity in uninfected (A) and L. monocytogenes-infected (2 h) (B) Caco-2 monolayers grown on Transwell filters. Monolayers were labeled with antibodies specific for the tight junction proteins ZO-1, occludin, and claudin (red) and for Hsp60 (green) and with a nuclear stain (blue). The images reveal visible tight junction borders in both uninfected and infected Caco-2 monolayers, with no alteration in cellular distribution of tight junction proteins or in cellular damage.

• Open in new tab
• Download powerpoint

FIG. 5.

Analysis of transepithelial translocation of L. monocytogenes through polarized Caco-2 cell monolayers grown on a Transwell filter insert (see the text for further details). (A) Immunoblot showing LAP and InlA expression in L. monocytogenes WT and ΔinlA strains. Each lane was loaded with 15 μg of total cell protein. (B) Transepithelial translocation of L. monocytogenes WT, lap, lap⁺, ΔinlA, and inlA⁺ cells through Caco-2 monolayers following Hsp60 knockdown by shRNA [shRNA Hsp60(−)] or Hsp60 overexpression (Hsp60+) or through control-sh-transfected cells (Caco-2 cells transfected with noncoding shRNA). (C) Translocation of WT, lap, lap⁺, and ΔinlA cells and of ΔinlA cells pretreated with anti-LAP MAb (1 μg/ml) or an IgG control (MAb C11E9) (1 μg/ml) through control-sh-transfected Caco-2 monolayers. Data are averages for at least three independent translocation experiments performed in quadruplicate (n ≥ 12) and are presented with SEM. Lowercase letters (a to c) indicate significant differences (P < 0.05) in translocation of individual bacterial strains through Caco-2 monolayers expressing various levels of Hsp60.

Surprisingly, translocation of the ΔinlA strain through Caco-2 monolayers was nearly 2-fold greater than that of the WT (P = 0.025) and was reversed to WT levels when the strain was complemented with inlA (Fig. 5B and C). Furthermore, ΔinlA strain translocation was influenced by Hsp60 expression, similar to that of the WT, as translocation decreased in shRNA-Hsp60-transfected Caco-2 monolayers (Fig. 5B). To test for LAP-mediated translocation in the ΔinlA strain, which expresses normal levels of LAP (Fig. 5A), the ΔinlA strain was coated with a LAP-specific antibody prior to Caco-2 cell infection. Antibody treatment resulted in a 60% reduction in translocation (P = 0.02) (Fig. 5C), indicating that LAP-mediated translocation occurs in the ΔinlA strain. The observance of increased translocation in the absence of epithelial cell invasion suggests that LAP-mediated translocation of the ΔinlA strain may occur via a noninvasive, or paracellular, mechanism. Such paracellular translocation might necessitate localization of Hsp60 near Caco-2 paracellular borders. Confocal microscopic analysis revealed low levels of Hsp60 colocalization with Caco-2 cell tight junctions (Fig. 6 C).

• Open in new tab
• Download powerpoint

FIG. 6.

Influence of L. monocytogenes infection on Hsp60 expression in Caco-2 cells. (A) Immunoblot analysis of the level of Hsp60 expression in Caco-2 plasma membrane (Mem-Hsp60) and intracellular (Int-Hsp60) fractions after exposure to L. monocytogenes (Lm) at 1 × 10⁴, 1 × 10⁶, or 1 × 10⁸ CFU/ml. β-Actin was used as an internal control. Proteins were loaded at 15 μg/well. Blots are representative of at least three individual experiments. (B) Microscopic analysis of Hsp60 (green) on the surfaces of uninfected or L. monocytogenes-infected (10⁶ CFU/ml [MOI of 10:1] and 10⁸ CFU/ml [MOI of 1,000:1]) Caco-2 cells. Caco-2 cells were labeled with anti-Hsp60 antibody and an FITC-conjugated secondary antibody. Propidium iodide was used to stain dead cells (red). Samples were viewed on a Leica fluorescence microscope with a 40× objective. Bars, 10 μm. (C) Confocal microscopic analysis of intracellular Hsp60 (FITC; green) expression in Caco-2 cells left uninfected (top panels) or infected with 10⁶ CFU/ml L. monocytogenes (bottom panels). The tight junction protein ZO1 was labeled (Cy5; red) for the purpose of visualizing cell borders. Yellow highlighting indicates areas where Hsp60 and ZO-1 are colocalized (right panels). Hsp60 expression was more abundant in L. monocytogenes-infected cells than in uninfected cells.

L. monocytogenes infection at a low dosage influences Hsp60 expression.Recent reports indicate that bacterial and viral infections can induce a host heat shock response (4, 39, 62). We sought to determine whether L. monocytogenes infection might influence Hsp60 expression in intestinal epithelial cells by infecting Caco-2 cells with various doses (10⁴, 10⁶, and 10⁸ cells) of L. monocytogenes and monitoring protein expression. Infection with each dose of L. monocytogenes increased hsp60 transcript levels, as analyzed by RT-PCR (data not shown), as well as levels of Hsp60 protein in plasma membrane and intracellular protein fractions (Fig. 6A). However, the greatest increase in Hsp60 was in the plasma membrane following infection with 10⁴ or 10⁶ cells of L. monocytogenes (Fig. 6A). Fluorescence microscopy confirmed increased levels of surface (Fig. 6B) and intracellular (Fig. 6C) Hsp60 and demonstrated that colocalization of Hsp60 with Caco-2 cell tight junctions increased following L. monocytogenes infection (Fig. 6C).

To determine whether Hsp60 expression occurred specifically in response to L. monocytogenes or in response to general bacterial factors, we exposed Caco-2 cells to 10⁶ CFU/ml of the L. monocytogenes WT or lap strain or nonpathogenic L. innocua and monitored Hsp60 expression by Western blotting (Fig. 7A) and flow cytometry (Fig. 7B). The Hsp60 level increased in both intracellular (62% increase) and membrane (80% increase) fractions following infection with L. monocytogenes but was not dependent upon LAP expression, since the lap mutant had similar levels to those of the WT (Fig. 7A), and did not increase after L. innocua exposure. These data indicate that the Hsp60 response may be specific to L. monocytogenes. Furthermore, no change in Hsp60 expression was observed when bacterial cells were separated from Caco-2 cells by use of a 0.4-μm filter (indirect exposure) (Fig. 7), suggesting that direct bacterial contact with epithelial cells, irrespective of LAP expression, is required and that diffusible microbial factors may not induce Hsp60 expression.

• Open in new tab
• Download powerpoint

FIG. 7.

Hsp60 expression in Caco-2 cells following direct versus indirect exposure to Listeria. (A) Immunoblot analysis of Hsp60 levels in Caco-2 plasma membrane (Mem-Hsp60) and intracellular (Int-Hsp60) fractions following direct or indirect (via separation by a 0.4-μm-pore-size filter) exposure to the L. monocytogenes (Lm) WT, the L. monocytogenes lap strain, or L. innocua at 10⁶ CFU/ml (MOI, 10:1). β-Actin was used as an internal control. Blots are representative of three individual experiments. Proteins were loaded at 15 μg/well. Lanes marked with an open box indicate increased Hsp60 expression compared to that of controls. (B) Flow cytometric analysis of Hsp60 expression on the surfaces of Caco-2 cells following direct or indirect exposure to L. monocytogenes WT or lap strain. Data are averages for 3 experiments, with treatments run in quadruplicate (n = 12), and are presented as fold changes in the number of Caco-2 cells expressing surface Hsp60 compared to uninfected cells. Data are presented with SEM. Lowercase letters (a and b) represent significant differences (P < 0.05) in surface Hsp60 expression in Caco-2 cells due to direct versus indirect bacterial exposure.

Caco-2 cells expressing high levels of Hsp60 are more susceptible to subsequent L. monocytogenes infection.Pathogen-induced Hsp60 expression is not unprecedented (4, 39); however, it is intriguing in the case of L. monocytogenes infection of intestinal epithelial cells, since Hsp60 is the receptor for LAP-mediated infection. We therefore designed a series of cell-based studies to test whether L. monocytogenes-induced Hsp60 expression would render host cells more susceptible to further LAP-mediated (secondary) infection. First, control (untransfected) Caco-2 monolayers were exposed to WT L. monocytogenes or to L. innocua in a primary infection; after a 3-h recovery period with gentamicin, cells were then subjected to secondary infection by L. monocytogenes. The number of L. monocytogenes organisms bound to Caco-2 cells resulting from the secondary infection was significantly greater (P = 0.032) following a primary L. monocytogenes exposure than in previously uninfected cells or in monolayers which were first exposed to L. innocua (Fig. 8A). These data suggest that initial infection with L. monocytogenes may render host cells more susceptible to further (secondary) infection by this pathogen, but this alone does not imply involvement of Hsp60 or LAP in susceptibility to secondary infection. Therefore, a second experiment was conducted with control-sh (transfected with shRNA noncoding vector)- and shRNA-Hsp60-transfected Caco-2 cells, where the primary infection microorganism was the WT and the secondary microorganism was the WT, lap, or ΔinlA strain (Fig. 8B). In control-sh-transfected Caco-2 cells, we observed significantly greater adhesion of the WT (3.3-fold increase) (P = 0.01) and the ΔinlA strain (4.4-fold increase) (P = 0.025), but not the lap mutant, following primary L. monocytogenes WT infection than that to cells which were previously uninfected (Fig. 8B). This indicates that primary infection rendered the Caco-2 cells more susceptible to LAP-mediated infection, in a manner which was independent of InlA. There was no apparent effect of primary infection on adhesion of any microorganism during secondary infection in shRNA-Hsp60-transfected Caco-2 cells (Fig. 8C).

• Open in new tab
• Download powerpoint

FIG. 8.

Influence of primary Listeria infection on Caco-2 cell susceptibility to subsequent (secondary) L. monocytogenes adhesion. (A) Secondary adhesion of L. monocytogenes WT to normal Caco-2 cells after primary infection with L. monocytogenes WT or L. innocua compared to that with previously uninfected cells. Data are averages for at least three independent experiments performed in quadruplicate (n ≥ 12) and are presented with SEM. Lowercase letters (a and b) represent significant differences (P < 0.05) in adhesion of bacterial strains. (B and C) Adhesion of L. monocytogenes WT, lap, lap⁺, ΔinlA, and inlA⁺ strains to control-sh-transfected Caco-2 cells (Caco-2 cells transfected with noncoding shRNA) (B) and to Hsp60 knockdown Caco-2 cells (shRNA Hsp60⁻) (C) following primary infection with L. monocytogenes WT for 1 h. After primary infection, cells were incubated for 3 h in gentamicin (50 μg/ml)-containing cell culture medium (recovery period) prior to secondary infection for 1 h with L. monocytogenes. Adhesion data are averages for at least three independent experiments performed in quadruplicate (n ≥ 12) and are presented with SEM. Bars marked with asterisks indicate significant differences in secondary bacterial adhesion between previously uninfected and infected Caco-2 monolayers (P < 0.05). Bars marked with lowercase letters (a to e) represent significant (P < 0.05) differences in LAP-mediated adhesion between bacterial strains.

Similarly, in transepithelial translocation studies, we observed an increase in LAP-mediated translocation in control-sh-transfected (Fig. 9A) but not shRNA-Hsp60-transfected (Fig. 9B) Caco-2 cell monolayers following primary L. monocytogenes WT infection. Despite differences observed in LAP-mediated translocation in control-sh-transfected Caco-2 cells, the level of translocation during secondary infection in both Caco-2 cell lines was higher for all microorganisms than that observed in earlier primary infection studies. This may be due to reductions in TEER which resulted from primary WT infection (data not shown). The secondary adhesion and transepithelial translocation findings emphasize the contribution of L. monocytogenes-induced host Hsp60 expression to subsequent LAP-mediated infectivity of L. monocytogenes.

• Open in new tab
• Download powerpoint

FIG. 9.

Influence of Listeria-induced Hsp60 expression on subsequent LAP-mediated translocation through Caco-2 monolayers. Translocation of L. monocytogenes WT, lap, lap⁺, ΔinlA, and inlA⁺ strains is shown for control-sh-transfected cells (Caco-2 cells transfected with noncoding shRNA) (A) and shRNA-Hsp60-transfected Caco-2 cells (B) following primary infection with L. monocytogenes WT. Caco-2 cell preparation and treatment were the same as those described in the legend to Fig. 7, except that the translocation assay during secondary infection was performed for 2 h. Translocation assays were repeated at least three times in quadruplicate wells (n ≥ 12), and data are presented with SEM. Bars marked with asterisks indicate significant differences in secondary bacterial adhesion between previously uninfected and infected Caco-2 monolayers (P < 0.05), while bars marked with lowercase letters (a to c) represent significant (P < 0.05) differences in LAP-mediated adhesion between bacterial strains.