All Salmonella mutants were derived from S. Choleraesuis strain 6150 isolated from porcine feces (Isolated in Schierack Lab, Institute of Biotechnology, Faculty of Environment and Natural Sciences, Brandenburg University of Technology Cottbus-Senftenberg, Großenhainer Str. 57, 01968, Senftenberg, Germany). The bacterial strains and plasmids used in this study are listed in Table 1.

The deletion mutant was generated according to the Datsenko-Wanner method with minor modifications (Datsenko and Wanner, 2000). Briefly, electro-competent bacteria were transformed with pKD46 plasmid, grown at 30°C for 2 h with shaking and plated on agar with ampicillin (100 μg/ml) for 24 h at 30°C. Next, bacteria bearing pKD46 plasmid were used for preparation of electro-competent cells according to protocol described in Khetrapal et al. (2015) and transformed with PCR products containing chloramphenicol resistance cassette (Cam) and fimH gene homologous extensions. Bacteria, in which Cam resistance cassette was successfully introduced into fimH gene locus were selected on LB agar plates with Cam (final concentration of Cam: 10 μg/ml) and later successful introduction of resistance cassete into fimH gene locus was confirmed by colony PCR using following primers: fimH_Sal_For (ATCCAGTGGGGAGAGGG), fimH_Sal_Rev (GAGTTGGCCTGACTCAGC),O1535 fimH_Sal_qPCR_rev (CTTTGCCGCTGTAGCTAATGG) and C1 (Datsenko and Wanner, 2000). In the next step, Cam resistance cassette was removed using pCP20 plasmid and successful removal of Cam resistance cassete from fimH gene locus was confirmed by colony PCR and Sanger sequencing.

FimH gene sequences were amplified with PCR using the following primers: FimHpACYCfor- ACATGGATCCTTGACAATTAATCATCGGCTCGTATAATGTGTGGAG GAGGACAGCTATGAAAATATACTCAGCGCTATTG and FimHpACYCrev- ACATGGATCCTTAATCATAATCGACTCGTAGATAG. The resulting DNA fragments were amplified with Phusion High-Fidelity DNA Polymerase (Thermo- Scientific) according to the manufacturer's protocol. Briefly, 50 μl of the reaction mix contained: genomic DNA isolated from S. Choleraesuis isolate no. 1656/04 (endogenous CFimH) or S. Choleraesuis 1204 (endogenous C63FimH) or S. Enteritidis isolate no. 327 (endogenous EFimH) (1 μl), Phusion HF buffer (Thermo Scientific), MgCl₂ (at final concentration of 1.5 mM), 1 U of Phusion High-Fidelity DNA Polymerase (Thermo Scientific), 0.2 μM of each primer and 0.2 mM dNTPs mix. PCR was started with initial denaturation (30 s, 98°C), then 35 cycles of denaturation (5 s, 98°C), annealing (15 s, 58°C), elongation (40 s, 72°C) and final elongation (72°C, 5 min). PCR products were cloned into the pACYC177 plasmid. FimH protein sequence variants used in this study are presented in Table 2.

Salmonella Choleraesuis ΔfimH was electroporated with pACYC177/C, pACYC177/C63, or pACYC177/E plus empty pACYC177 plasmid as negative control. The resulting Salmonella strains were named SCΔfimH/C, SCΔfimH/C63 (S. Choleraesuis FimH mutant V63G), SCΔfimH/E, and SCΔfimH/Empty, respectively.

To measure the binding abilities of Salmonella mutants, we applied 100 μl of bacteria (CFU 5 × 10⁷/ml) to each well of 96-well plates coated with polyclonal rabbit anti-FimH antibody (Kisiela et al., 2005) and glycoprotein standards and then incubated them for 2 h at room temperature followed by three washing steps with PBS. Attached bacteria were fixed with 50 μl of 4% PFA (paraformaldehyde) in PBS for 30 min at 4°C, washed three times with 100 μl of PBS and stained with 50 μl of propidium iodide (10 μg/ml in ddH₂O) for 15 min at room temperature. Plates were read using the VideoScan module (Rodiger et al., 2013; Schierack et al., 2015). Three independent experiments with three repetitions for each dilution were prepared and measured (Figure S1).

Intestinal epithelial IPEC-J2 cells from newborn porcine (Schierack et al., 2006) were cultured in DMEM supplemented with 10% fetal calf serum (FCS, Invitrogen, Carlsbad, USA), 2 mM glutamine and antibiotics. The mouse small intestine cell line (ICE-1) was a kind gift from InScreenex (InSCREENeX, Braunschweig, Germany) (Schwerk et al., 2013). Cells were cultured in a defined medium supplied by the manufacturer.

For the generation of a vector expressing calreticulin, initially, IRES sequence derived from the pWP1 vector (kindly provided by Dr. D. Trono, École Polytechnique Fédérale de Lausanne, Switzerland) and a DNA cassette containing puromycin N-acetyl-transferase (PAC) cDNA from a pPUR vector (Clontech, Terra Bella Avenue Mountain View, CA, USA) were cloned into the pRRL-cPPT-CMV-X2-PRE-SIN vector. To obtain porcine CRT cDNA sequence, we isolated total RNA from IPEC-J2 cells, and a cDNA library was generated with a transcriptor first strand cDNA synthesis kit using an oligo (dT) primer (Roche). CRT cDNA was amplified by PCR with the primers 5′GAACGCGTATGCTGCTCCCAGTCCC′ and 5′GACTCGAGCTACAGCTCATCCTTGGCC3′ containing restriction sites for MluI and Xho1, respectively, and was cloned into the pRRL-CMV-IRES-PURO vector. This vector was named pRRL-CALR-IRES-PURO.

To produce virus, we used lentiviral vectors derived from the human immunodeficiency virus (HIV-1). Polyethylenimine-mediated (Sigma-Aldrich, Buchs, Switzerland) transient co-transfection was mediated using packaging LentiX 293T cells (Clontech, Cat. No. 632180) at 50–60% confluence with 20 μg of pRRL-CMV-CALR-IRES-PURO or pRRL-CMV-IRES-PURO, 10 μg of pMDL-g/p-RRE, 5 μg of pRSV-REV, and 5 μg of pMk-VSVG (D. Trono, École Polytechnique Fédérale de Lausanne, Switzerland). Viral supernatant was collected 48 h after transfection and was then clarified through a 0.45 μm pore size filter (Millipore, Billerica, MA, USA). Immediately afterwards, the virus-containing supernatant was concentrated 100X on an Amicon Ultra-15K:100.000 (Millipore, Billerica, MA, USA).

The pLVTHM/PURO-shCRT silencing vector was obtained essentially as described previously (Solatycka et al., 2012). pLVTHM/GFP lentiviral vector was kindly provided as part of the lentivirus system by Dr D. Trono (Ecole Polytechnique Fédérale de Lausanne, Switzerland). Briefly, shRNA/CRT targeted the GGATTGAATCCAAACACAAAC fragment of the CRT mRNA (NP_001167604). The DNA cassette consisted of the H1 promoter, and the shDNA/CRT sequence was cloned into the pLVTHM/PURO vector. To obtain the pLVTHM/PURO vector, we replaced GFP cDNA by the DNA cassette containing the puromycin N-acetyl-transferase (PAC) cDNA from a pPUR vector (Clontech, Terra Bella Avenue Mountain View, CA, USA). Lentivirus-induced CRT silencing was obtained essentially as described above with the exception of the lentivectors used: 20 μg of pLVTHM/PURO-shCRT or pLVTHM/PURO co-transfected with the packaging plasmids, 15 μg of psPAX2 and 5 μg of pMD2.G (kindly provided by Dr D. Trono, Ecole Polytechnique Federale de Lausanne, Switzerland), into the LentiX 293T cell line.

IPEC-J2 or ICE-1 cells were infected with concentrated virus stock (both silencing and expressing) by centrifugation (2,460 × g) at 23°C for 2.5 h. After overnight incubation, the medium was replaced, and after the next 48 h, cells were selected in 1 μg/ml of puromycin.

Generated cell lines were named IPEC-PURO (control line for overexpression), IPEC-CRT (overexpression), ICE-PURO (control line for ectopic expression), ICE-CRT (ectopic expression), IPEC-shPURO (control line for silencing), and IPEC-shCRT (silencing). Growth rate of generated cell lines was tested by SRB assay.

The cells (1 × 10⁴) were grown in 96-well plates (Sarstedt, Germany) in complete DMEM. After every 24 h, they were fixed by the addition of 200 μl of 10% trichloroacetic acid for 30 min at 4°C, washed with water, and dried. Fixed cells (after 24, 48, 72, and 96 h) were incubated with 0.4% sulforhodamine B (SRB, Sigma) in 1% acetic acid for 20 min in RT. After washing with 1% acetic acid, the protein-bound dye was extracted with 10 mM Tris for 1 h at 37°C. The absorbance at 562 nm was measured in plate reader “Quanta” (Bio-tek). Data were presented as the means ± SD from three independent assays.

Total RNA was isolated from 10⁶ cells using RNeasy plus mini Kit (Qiagen) according to the manufacturer's recommendations. First-strand cDNA was synthesized using the iScript (Bio-Rad). The relative amounts of mRNA were quantified by qPCR using the CFX thermocycler (Bio-Rad) and SYBR-green (Life Technologies). Actb was used as the reference gene, and all primers used are listed in Table 3. The comparative Cq method was used for relative quantification of gene expression. All experiments were performed at least four times, and triplicate samples were analyzed in each experiment to confirm the accuracy and reproducibility of the qPCR. Fold change was measured over unstimulated cells set at 1.

The cell lysates were obtained by solubilising cells in RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 0.1% SDS, 1% IGEPAL Ca-630, 0.5% sodium deoxycholate, pH 8.0) containing protease inhibitor cocktail (Sigma) for 30 min at room temperature (RT). The amounts of protein were determined with a bicinchoninic acid protein assay kit (Sigma). Isolation of cell surface proteins was performed using a Pierce® Cell Surface Protein Isolation Kit (Thermo Scientific) according to the manufacturer's instructions.

ConA affinity chromatography was performed as described by Soper et al. (Soper and Aird, 2007).

The presence of porcine CRT in cell lysates was confirmed by Western blotting using primary rabbit monoclonal antibody directed against CRT (Cell Signaling, D3E6, XP® Rabbit mAb #12238) and horseradish peroxidase (HRP)-conjugated goat anti-rabbit immunoglobulins (DAKO, #00077731).

The binding of the FimH proteins to cell lysates was analyzed by far-Western blotting. Cell lysates or purified cell surface proteins separated by SDS-PAGE in 12% gel were transferred to nitrocellulose (Bio-Rad), blocked overnight with 3% BSA in TBST (TBS supplemented with 0.1% Triton X-100) and incubated with FimH proteins (0.1 mg/ml) for 4 h at RT. Immunodetection of FimH proteins was performed using primary rabbit polyclonal antibodies against FimH (Kisiela et al., 2005) and horseradish peroxidase (HRP)-conjugated goat anti-rabbit immunoglobulins (DAKO, #00077731).

The samples (300 μg of total protein) were lysed in extraction buffer (7 M urea, 2 M thiourea, 4% w/v CHAPS, 2% Brij, 2% w/v 3-10 ampholytes, 40 mM Tris, and 100 mM DTT) and precipitated with ice-cold acetone for 2 h at −20°C. The first dimension was run using pH 3–7, 17 cm NL ReadyStrip™ IPG Strips (Bio-Rad). Prior to the second dimension SDS-PAGE, focused IPG strips were reduced with DTT (Sigma) in equilibration buffer (7 M urea, 2 M thiourea, 4% w/v CHAPS, 0.5% w/v 3–10 ampholytes, and 100 mM DTT) and alkylated with iodoacetamide (2.5% w/v). After equilibration, strips were placed on the top of 10% SDS polyacrylamide gels and held with molten 0.5% (w/v) agarose. After 2-DE separation, gels were either visualized with colloidal Coomassie brilliant blue G-250 according to the Hoving protocol (Westermeier, 2006) or transblotted onto nitrocellulose membranes. Selected spots were cut out from the gel and digested with trypsin using the All-in-One Trypsin Digestion Kit (Sigma).

After tryptic digestion, all samples were acidified with formic acid (FA) and desalted on a 0.6 μl ZipTip C18 (Millipore) according to the manufacturer instructions. Peptides were eluted using 50% acetonitrile with 0.1% FA in water. Afterwards, 1 μl of each sample was mixed with 1 μl of saturated α-cyano-4-hydroxycinnamic acid (4-HCCA) matrix in 50% acetonitrile in water with 0.1% FA, and 1 μl of the sample/matrix solution was spotted onto the MALDI steel plate. Peptides were measured by matrix-assisted laser desorption/ionization-time of flight/time of flight mass spectrometry (MALDI-TOF-TOF-MS/MS) on an Ultraflex Extreme MALDI-TOF-TOF mass spectrometer (Bruker Daltonics, Bremen, Germany). The instrument was operated in positive ion reflection mode with an acceleration voltage of 25 kV. Spectra were acquired in the range of 700–3,500 Da with an average of 2,500 laser shots. Calibration of the MALDI-TOF mass spectrometer was performed with the peptide calibration standard (Bruker Daltonics, Bremen, Germany). All spectra were acquired by flexControl and then evaluated by flexAnalysis 3.4.7.76.0 software (Bruker Daltonics, Bremen, Germany).

Calibrated spectra were submitted to database searches using the MASCOT search engine (Matrix Science; http://www.matrixscience.com/). The search was performed with up to 1 trypsin miss-cleavages, carbamidomethylation of cysteine as a fixed modification, and oxidation of methionine as a potential modification. Peptide mass tolerance was set to ±0.45 Da, and MS/MS fragment ion tolerance was set to ±0.1 Da. A MASCOT search of all data was made against the SwissProt protein database restricted to mammalian taxonomy (66345 sequences).

Recombinant FimH adhesins were obtained as described in our previous paper (Grzymajlo et al., 2013) and named CFimH (active FimH variant from S. Choleraesuis), EFimH (active FimH variant from S. Enteritidis), and C63FimH (inactive FimH variant from S. Choleraesuis). Recombinant porcine calreticulin was obtained from CUSABIO (USA/China).

The binding of FimH adhesins to recombinant porcine CRT immobilized on CM5 sensor chips (GE Healthcare) was analyzed by SPR using a BIAcore T200. CRT immobilization level corresponds to approximately 150 RU. All binding experiments were carried out at 25°C with a 30 μl min⁻¹ flow rate. To determine FimH affinity to CRT, 100 μl of five different concentrations (8, 4, 2, 1, and 0.5 μM) of FimH variant solutions, as well as a sample buffer blank, were passed over the ligand-immobilized chip surface (association phase) for 2 min, followed by 10 min of dissociation with running buffer. The same samples were passed over a control chip surface without immobilized ligand. Three replicates of each analyte concentration were injected. The resulting sensorgrams were obtained first by subtracting the buffer blank from curves recorded for the interactions of FimH proteins with CRT. Then, the curves recorded when adhesins were passed over the blank surface were subtracted from such sensorgrams. The equilibrium constants were determined using BIAevaluation 3.1 software. For global fitting, a 1:1 Langmuir binding model with an included mass transport step was applied based on criteria provided by the BIAevaluation handbook.

For the invasion assay, Salmonella, after the fifth passage, were washed in PBS, resuspended in cell culture medium and adjusted by dilution to provide an MOI (multiplicity of infection) of 10:1 bacteria to host cells in culture wells of a 6-well plate. Confluent monolayers were infected for 1 h, followed by an additional hour of incubation with culture media containing 50 μg/ml of gentamicin (Sigma). After incubation, cells were washed three times with PBS and lysed with 0.1% Triton X-100 (Sigma) for 10 min. Bacterial suspensions were serially diluted with PBS, plated on LB-agar plates, and incubated overnight at 37°C, and the colonies were then counted to calculate the CFU.

For the adhesion assay, bacteria and cells were grown as described for the invasion assay. Infection was performed at an MOI of 50:1 bacteria to host cells. After 2 h of infection, cells were washed and plated as described for invasion assay. The percentage of dead cells was measured using bromophenol blue. Each assay was conducted in triplicate and was independently repeated at least three times.

Confluent monolayers of tested cells in a 96-well plate were incubated for 1 h with a monoclonal antibody against CRT (D3E6 XP®Rabbit mAb #12238), an antibody against CRT (Antibodies Online, Rabbit polyconal Ab #0805) that does not recognize porcine CRT at a 1:50 dilution, or with a FimH adhesin (CFimH or EFimH) with a final concentration of 0.1 mg/ml. Cells were incubated at 20°C for 1 h to allow for antibody or protein binding; then, SCΔfimH/C bacteria (MOI 50:1) were added to quadruplicate wells, and the adhesion assay procedure was applied.

Cells were grown to confluence in 6-well plates, detached with Accutase solution (Biowest), washed with staining buffer (PBS buffer supplemented with 1% bovine serum albumin and 0.002% NaN₃) and incubated with or without 1 μl of calreticulin (D3E6) XP®Rabbit mAb #12238 for 2 h at 4°C. After the incubation, the cells were washed with staining buffer and incubated with 0.125 μg of fluorochrome-conjugated secondary antibody, PE Donkey anti-rabbit IgG (minimal x-reactivity) antibody (Biologend, clone Poly4064) or with 0.125 μg Alexa Fluor®647 Donkey anti-rabbit IgG (minimal x-reactivity) antibody (Biologend, Clone Poly4064), for 30 min at 4°C. Afterwards, cells were washed twice with staining buffer and fixed in 1% buffered formaldehyde. The samples were then immediately analyzed with a FACSCalibur II cytometer (BD Pharmingen). All analysis was carried out with Weasel 3.0.2 software (http://en.bio-soft.net/other/WEASEL.html).

Cells (5 × 10³) were grown on 18 mm glass coverslips for 24 h in a humidified cell culture chamber (37°C, 5% CO₂) and subsequently fixed in 2% paraformaldehyde at RT for 15 min. Afterwards, the cells were briefly rinsed in PBS and permeabilized with PBS containing 0.1% Triton X-100 for 5 min at 4°C. Permeabilization solution was next aspirated, and cells were blocked with FCS (20 min, RT) and further incubated overnight with primary anti-calreticulin antibody (D3E6; dilution 1:50) at 4°C. Cells were rinsed with PBS and incubated for 45 min at RT with the Alexa Fluor®633 anti-rabbit secondary antibody (Thermo Fischer Scientific; dilution 1:400) in the dark, and nuclei were stained with mounting medium containing DAPI (Abcam, UK). Images were acquired on an LSM 510 META microscope (Carl Zeiss, GmbH, Germany) using a PLAN-APOCHROMAT 63x/1.4OIL DICM27 objective. Image acquisition was performed using ZEN 2009 Light Edition software.

All statistical calculations were performed in GraphPad (GraphPad Software, Inc., La Jolla, CA, USA). The shape of the data distribution was assessed with the Shapiro normality test and analysis of quantile-normal plots. Mann-Whitney test (nonparametric), Student's t-test (parametric) or one-way ANOVA with post hoc Dunn's multiple comparison test were performed based on the shape of data distribution. All data collected in this study were obtained from at least three independent experiments for each condition. A p-value of less than 0.05 was considered statistically significant. Data are presented as the means ± standard deviation (SD).

To identify the host receptor for FimH protein, we applied two independent pathways (Figure 1A). First, IPEC-J2 cell lysates were subjected to 2D electrophoresis and either stained with Coomassie brilliant blue (Figure 1B) or transferred onto a nitrocellulose membrane. The presence of protein bound by S. Choleraesuis FimH was analyzed by far-Western blotting using recombinant S. Choleraesuis FimH (CFimH) protein. As CFimH adhesin bound to a protein spot with a molecular mass approximately 55 kDa and pI in a range between 4 and 6 (Figure 1C), all spots from the Coomassie-stained 2D gel with molecular masses between 35 and 70 kD and pI between 4 and 6 were cut out, digested with trypsin and analyzed by MALDI-MS and peptide mass fingerprinting.

Alternatively, IPEC-J2 cell lysates were subjected to affinity chromatography on Concanavalin A (ConA), and the fraction enriched in ConA-binding glycoproteins was subjected to SDS-PAGE and far-Western blotting with CFimH (Figures 1D,E). In agreement with the results presented above, a protein band with a molecular mass approximately 55 kDa was detected by CFimH adhesin; thus, protein bands of apparent molecular mass between 35 and 70 kDa were cut from the gel, digested with trypsin and analyzed by MALDI-MS and peptide mass fingerprinting. The results from both pathways were combined, and the analysis has provided a probability-based Mowse score for the best matching protein hits, and using the MASCOT analysis tool, the receptor has been identified as calreticulin.

To confirm MS data and show that IPEC-J2 cells express calreticulin, we analyzed cell lysate and the membrane fraction of IPEC-J2 cells by Western blotting with an anti-calreticulin antibody (Figure 1F). In addition, flow cytometry analysis revealed that CRT is expressed on the surface of IPEC-J2 cells (Figure 1G). To further prove that CRT is a receptor for S. Choleraesuis FimH adhesin, we also tested whether preincubation of IPEC-J2 cells with anti-calreticulin antibody could inhibit adhesion of S. Choleraesuis to porcine cells. It was found that anti-calreticulin antibody significantly (p < 0.01) inhibited binding of the SCΔfimH/C Salmonella strain to IPEC-J2 cells (Figure 1H, Table 1).

To show that calreticulin is a specific receptor for S. Choleraesuis FimH adhesin, we studied its interaction with variants of FimH proteins by far-Western blotting and surface plasmon resonance (SPR). It was found that, in static conditions, only active S. Choleraesuis FimH adhesin (CFimH) was able to bind to recombinant porcine CRT in contrast to inactive FimH (C63FimH) from S. Choleraesuis and active S. Enteritidis FimH (EFimH) (Figure 2A). It is worth to mention, that we observed three bands in far-western blot interaction between CFimH and CRT, probably because of different glycoforms of the recombinant CRT (see Figure S5). These data were further confirmed by direct kinetic analysis using SPR. For each allelic variant, analysis was carried out at different concentrations of recombinant FimH proteins (0.5, 1, 2, 4, and 8 μM; Figure 2B). According to SPR data, CFimH bound significantly stronger (KD = 6 × 10⁻⁸) to CRT in comparison with that of C63FimH and EFimH (K_D = 2 × 10⁻⁷) (Figure 2C).

To verify this interaction on the cell line model, we tested whether the purified FimH could block the Salmonella-IPEC-J2 cell interaction and lead to decreased bacterial attachment. Cells were plated and incubated with CFimH or EFimH and then challenged with our standard adhesion test with SCΔfimH/C strain (Figure 2D). It was found that only active CFimH strongly inhibited the interaction between bacteria and tested cells (p < 0.001), indicating that porcine calreticulin is a specific receptor for S. Choleraesuis FimH adhesin.

To support our hypothesis that porcine calreticulin is a specific receptor for an allelic variant of FimH adhesin from S. Choleraesuis, the adhesion and invasion of S. Choleraesuis isogenic strains expressing different variants of FimH adhesin (Table 1, 2) to porcine IPEC-J2 cells with different CRT expression levels were studied. To generate a specific loss-of-function phenotype, IPEC-J2 cells expressing CRT were transduced with the pLVTHM/shCRT construct to inhibit expression of the receptor. Such cells, named IPEC-shCRT, were characterized by a strongly decreased level of calreticulin mRNA (Figure 3A) and highly decreased binding of anti-calreticulin antibodies to cell lysates in comparison to control IPEC-shPURO cells obtained after transduction of IPEC-J2 cells with the pLVTHM/PURO vector (Figure 3B). To create a gain-of-function cellular model, we transduced IPEC-J2 cells with the pRRL-PURO-CRT-IRES expression vector, and the resulting cell population, which overexpressed calreticulin at the mRNA and protein levels (Figures 3D,E), was named IPEC-CRT. Control IPEC-PURO cells with basal CRT expression level were obtained by transduction of IPEC-J2 cells with the pRRL-PURO-IRES-vector. The membrane level of CRT was decreased in the silencing model and upregulated in the overexpression model as presented by flow cytometry plots (Figures 3C,F). The cells growth rate do not change after genetic modifications (Figure S2).

To investigate how the calreticulin expression level affects bacterial attachment and invasion, we used our gain-of-function (IPEC-CRT) and loss-of-function (IPEC-shCRT) cell models. Isogenic strains of S. Choleraesuis that express non-binding (SCΔfimH/C63) and high-binding (SCΔfimH/C) FimH variants from S. Choleraesuis and FimH variant from S. Enteritidis (SCΔfimH/E), as well as S. Choleraesuis without FimH expression (SCΔfimH/Empty), were compared. This experiment revealed that only SCΔfimH/C binds significantly better (p < 0.01) to IPEC-CRT cells (Figure 4A). Thus, under these conditions, Salmonella expressing active FimH from S. Choleraesuis is able to adhere better thanks to a higher amount of CRT expressed on the cell surface. When we compared adhesion of all strains to IPEC-shCRT cells representing model with decreased CRT expression (Figure 4B), the pattern was generally the same. Together with lower expression of calreticulin, the SCΔfimH/C strain binds to IPEC-shCRT cells significantly weaker (p < 0.05) than to control cells. Taking these two models into account, we found that the expression level of calreticulin correlates with the level of adhered SCΔfimH/C, which indicates that the receptor is involved in mediating infection specifically for this variant of FimH. When bacterial internalization was examined (Figures 4C,D), again only the isogenic strain SCΔfimH/C invaded IPEC-CRT cells significantly better (p < 0.01) and IPEC-shCRT significantly weaker (p < 0.05) than that of control IPEC cells. However, the invasion level was low and therefore its biological significance still needs to be proven. As expected, strains producing FimH variants that are able to bind to oligomannosidic structures (SCΔfimH/C, SCΔfimH/E) bind to both epithelial cell variants up to 8-fold higher than that of strains with inactive FimH and up to 20-times higher than that of the FimH knockout strain.

We next wanted to evaluate whether the role of calreticulin as a FimH-dependent host-specific receptor was restricted to IPEC-J2 cells by testing epithelial intestinal cells from other species. Transduction of ICE-1 cells of mouse origin, with the pRRL-PURO-CRT-IRES expression vector results in strong neoexpression of porcine calreticulin on protein level (Figure S3). When we perform adhesion and invasion tests using mouse intestinal epithelial cells with neoexpression of porcine calreticulin (ICE-CRT) with two isogenic strains, SCΔfimH/C63 and SCΔfimH/C, profound differences were found in their binding patterns. Strain SCΔfimH/C binds and invades ICE-CRT cells with porcine calreticulin expression significantly better (p < 0.05) than those of control ICE-PURO cells (Figures 4E,F). Importantly, binding of S. Choleraesuis isogenic variants to cells of mouse origin was significantly weaker (p < 0.05) when compared with binding to porcine cells, again highlighting the host specificity of such an interaction. Taken together, the presence of porcine calreticulin correlates with adhesive and invasive properties of S. Choleraesuis with active FimH-variant, which supports the host specificity hypothesis of this interaction.

When we investigated the membrane CRT level 4 h after Salmonella infection using flow cytometry, noticeable differences were found between SCΔfimH/C63 and SCΔfimH/C infected cells. In cells infected with the active strain, a significantly higher number of cells expressing surface calreticulin was observed when compared to the number of inactive strain infected or uninfected cells (Figure 5A). Interesting differences in CRT localisation between SCΔfimH/C63 and SCΔfimH/C infected cells were found using confocal microscopy. When compared to uninfected cells as well as to cells infected with inactive strain, IPEC-J2 cells infected with SCΔfimH/C present higher surface expression of calreticulin. Furthermore, surface CRT on such cells was organized in dot aggregates (Figure 5B and Figure S4), which can be a potential entry site for pathogens. Taken together, these results suggest that there is an association between infection of Salmonella Choleraesuis with active FimH and surface calreticulin presence and organization.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.