The Salmonella enterica Serotype Typhi Vi Capsular Antigen Is Expressed after the Bacterium Enters the Ileal Mucosa

S. Typhi strain Ty2 was obtained from the American Type Tissue Culture Collection (ATCC 19430; Manassas, VA). Strains and plasmids used in this study are listed in Table 1.

Strains were cultured aerobically in Luria-Bertani broth containing 300 mM NaCl (LBH) for optimal Vi suppression. Strains were cultured in SOB medium (Luria Bertani [LB] broth containing 10 mM NaCl) for optimal induction of Vi antigen expression. Media was supplemented with appropriate antibiotics at the following concentrations: carbenicillin, 100 mg/liter, and kanamycin, 100 mg/liter.

For T84 cell infection experiments, each strain was grown overnight at 37°C with shaking in LBH with the appropriate antibiotics. The next day, a 1:1,000 dilution of overnight culture was made, and bacteria were grown to an optical density at 600 nm (OD₆₀₀) of 1.0 to 1.5 (late log phase) for optimal Vi suppression without antibiotics. Bacteria were added at a concentration of 10⁷ CFU/well to tissue culture plates.

For inoculation of bovine ligated ileal loops, each strain was grown overnight at 37°C with shaking in 4 ml of LBH with the appropriate antibiotics. A volume of 0.04 ml of overnight culture was used for inoculation of 4 ml of LBH without antibiotics, and bacteria were grown to an OD₆₀₀ of 1.0 to 1.5 (late log phase).

A region 1.2 kb upstream of ompR was PCR amplified using FW (5′-GACGGTTCGTGTTCCAGAGCAG-3′) and RV (5′-CTCTTGCATTGTCTGTACTCC-3′), and a 1.0-kb region downstream of ompR was PCR amplified using primers FW (5′-CGCCGTATGGTGGAAGAAG-3′) and RV (5′-ACCTGGATGCTGCCTGCCTG-3′). Both the downstream and upstream flanking regions were cloned into pCR 2.1 (Invitrogen, Carlsbad, CA). Subsequently, the upstream fragment was subcloned into the BglII/SalI site of pGP704, giving rise to pQT13. The downstream fragment was cloned into pQT13 using the XbaI/SmaI site, giving rise to pQT17. The kanamycin cassette (1.5 kb) was excised from pKIXX and inserted in the SalI site in pQT17 to give rise to pQT19. The entire construct was confirmed via nucleotide sequencing. Using the suicide plasmid pQT19, alleles were introduced into S. Typhi Ty2 with standard allelic exchange methodologies. Colonies were screened for kanamycin resistance and carbenicillin sensitivity. A single kanamycin-resistant and carbenicillin-sensitive colony was further analyzed for deletion of the ompR gene using Southern blotting with an ompR-specific DNA probe and PCR analysis. The ompR deletion was confirmed phenotypically by agglutination with rabbit anti-Vi serum (Difco, Detroit, MI). The strain was designated QT74.

For complementation, the ompR gene was amplified using FW (5′-TGCCAGCCATCAGCGGGGGCTT-3′) and RV (5′-GCCCTGATGAATCTCGGTCAG-3′) primers. The PCR fragment was cloned into pBluescript KS using EcoRV/EcoRI, creating pQT50. The plasmid was then electroporated into QT74, giving rise to QT174(pQT50). Expression of the Vi antigen by QT174(pQT50) was investigated by agglutination with rabbit anti-Vi serum (Difco, Detroit, MI).

We constructed a Ty2 strain carrying a high-copy-number plasmid encoding the viaB promoter region (P_tviA) fused to a mutant variant of the gfp gene (egfp). The promoter of the viaB locus was PCR amplified using the primers FW (5′-TATACCATGGGAAGTCTCCTTATGCTGAAA-3′) and RV (5′-TATAGTCGACGCAGTCACGCACCATC-3′) flanked with restriction sites SalI and NcoI, respectively (italics indicate restriction enzyme sequences used). The resulting 600-bp PCR fragment was cloned into pCR2.1 (Invitrogen, Carlsbad, CA), and the fragment was confirmed by nucleotide sequencing. The fragment was subcloned into pEGFP, using the NcoI and SalI multiple cloning site, giving rise to plasmid pQT27. The P_tviA::egfp fusion was excised using SalI and EcoRI and subsequently subcloned into pBluescript SK+ (Stratagene, La Jolla, CA), giving rise to plasmid pQT28. The plasmid was electroporated into Ty2 to produce strain Ty2(pQT28). To confirm that Vi expression is modulated by osmolarity in vitro, Ty2(pQT28) was grown in culture containing different salt concentrations, and expression of the P_tviA::egfp reporter fusion was investigated by flow cytometry.

To confirm the phenotype of ompR as a regulator of Vi expression, pQT28 was electroporated into QT74, giving rise to QT74(pQT28). This strain was grown under different osmolarity conditions in broth and subjected to flow cytometric analysis.

T84 cells are a human epithelial cell line of colon carcinoma cells that can be polarized in Transwell plates. T84 cells were seeded at 5 × 10⁵ cells/well, and transepithelial resistance was monitored daily. At a transepithelial resistance of 500 to 1,500 Ω (28), bacteria were added at approximately 10⁷ CFU/well (multiplicity of infection was approximately 10:1) to the apical compartment of polarized T84 cells for 1 h at 37°C in 5% CO₂ to allow for invasion. After 1 h, the supernatant from each well was removed. To compare bacterial gene expression inside and outside mammalian cells by monitoring expression of the P_tviA::egfp reporter using flow cytometry, gentamicin was not added (4). Instead, prewarmed Dulbecco's modified Eagle's medium (DMEM)-F12 medium was added to the apical side, and cells were incubated for 2 h at 37°C in 5% CO₂. After this incubation period, the supernatant was collected, and extracellular bacteria were harvested via centrifugation. The monolayer was then washed three times with ice-cold phosphate-buffered saline (PBS), and T84 cells were lysed with 1% Triton X-100 on ice for 10 min to collect intracellular bacteria.

Samples for flow cytometry were prepared as described previously (4). To characterize Ty2(pQT28) and QT74(pQT28), 1 ml of bacteria culture was harvested by centrifugation at 6,000 × g for 5 min. Bacterial cells were washed twice with PBS to remove LB media. Bacterial pellets were resuspended in 1 ml of 4% paraformaldehyde and incubated for 30 min at room temperature. Fluorescence was measured by using a FACSCalibur system (Becton Dickinson, Franklin Lakes, NJ) and the Cell quest software provided by the supplier. The bacterial population was gated by using side and forward scatter parameters and then analyzed for green fluorescence (fluorescein isothiocyanate [FITC] channel).

For infection of polarized T84 cells with Ty2(pQT28), the flow cytometry results were pooled from experimental infections of two separate Transwell plates. The bacterial population was gated with side and forward scatter parameters using samples from a Ty2(pQT28) grown under Vi-inducing conditions (SOB medium), and uninfected T84 polarized cells were treated under the same conditions as infected cells in order to set the region of fluorescence. The results were expressed as the percentage of the population which was fluorescing.

Four milk-fed 3- to 4-week-old calves were obtained from a Texas A&M University cattle herd. Calves were tested for Salmonella infection by fecal swabs, enriched in tetraiodothionate broth, and plated on XLT4. To perform the ligated ileal loop experiments, calves were anesthetized using propofol induction and isoflurane maintenance for the duration of the experiment (1). A right-flank laparotomy was performed, the ileum was exposed, and loops with lengths ranging from 6 to 9 cm were ligated, leaving 1-cm loops between them. The loops were inoculated by intraluminal injection of a 4-ml suspension containing approximately 1 × 10⁹ CFU/ml of S. Typhi strains Ty2 or QT74 grown in LBH. Loops injected with LBH served as a negative control. Loops were excised at 2 and 8 h postinfection to (i) collect the fluid that has accumulated in the lumen, (ii) harvest extracellular bacteria from 6.0-mm biopsy-punched tissue samples of the mucosa/submucosa for RNA extraction, and (iii) collect tissues for histopathology/frozen optimal cutting temperature (OCT) sections. The frozen sections were later sectioned at 10 μm thickness for fluorescence immunohistochemistry and stored at −80°C.

Tissue samples were fixed in formalin, processed according to standard procedures for paraffin embedding, sectioned at 5 μm, and stained with hematoxylin and eosin. Histopathology slides were examined blindly by two board-certified veterinary pathologists.

Frozen sections of ileal tissue inoculated with S. Typhi Ty2 grown in LBH and loops inoculated with LBH (negative control) were fixed in methanol. Samples were incubated with a primary anti-Vi rabbit antibody (1:250 dilution) (Difco, Detroit, MI) overnight in 4°C. The next day, samples were incubated for 2 h at room temperature in a humidified chamber with secondary anti-rabbit conjugated to Alexa Fluor 549 (1:250 dilution) (Molecular Probes, Eugene, OR) in the dark. All antibodies were diluted in antibody dilution buffer (1% bovine serum albumin [BSA] in 0.02 M PBS-Tween). Slides were washed three times for 10 min each in 0.3% Tween-PBS. A drop of Prolong antifade DAPI (4′,6-diamidino-2-phenylindole) solution (Invitrogen, Carlsbad, CA) was added to each slide. A coverslip was placed over the top, and slides were allowed to dry overnight at room temperature. The next day, slides were stored at 4°C until ready to view. Samples were viewed and photographed with an Olympus IX-70 camera.

For analysis of intracellular and extracellular S. Typhi gene expression in vitro, bacteria were collected and stored in ethanol/phenol mRNA stop solution until RNA was extracted with a hot acid phenol-chloroform assay. For analysis of intracellular and extracellular S. Typhi gene expression during infection of bovine ileal loops, extracellular bacteria were harvested from the lumen and frozen on dry ice. Tissue samples were frozen on dry ice in TRI Reagent (Molecular Research Center, Cincinnati, OH) at the site of surgery until further processing for RNA extraction. After RNA extraction from ileal tissue, samples were processed using the microbe enrichment kit (Ambion, Austin, TX) in order to isolate bacterial RNA. All samples were DNase treated (Ambion DNA-free kit). Subsequently, 1,000 ng of each sample was reversely transcribed in a 50-μl volume (TaqMan reverse transcription reagents; Applied Biosystems, Foster City, CA), and 4 μl of cDNA was used for each real-time (RT)-PCR. RT-PCR with primers was used to detect the tviB gene in bacterial mRNA in order to measure Vi expression and rpoD, the internal control gene, in S. Typhi-infected T84 polarized cells, using SYBR green (Applied Biosystems, Foster City, CA) and the ABI 7500 real-time PCR system. The rpoD gene has been shown to have no significant variation of expression in either S. Typhi or S. Typhimurium inside macrophages (8, 9) and T84 polarized cells (this study). RNA samples were analyzed for the prgH and fliC genes to monitor the expression levels of the Salmonella pathogenicity island 1 (SPI-1) type III secretion system (T3SS) and flagellin, respectively. For each run, the calculated threshold cycle (C_T) was normalized to the C_T of the rpoD gene amplified from the corresponding sample, and the data were analyzed using the comparative C_T method (Applied Biosystems, Foster City, CA). Levels of S. Typhi gene expression in T84 polarized cells were calculated relative to the inoculum of S. Typhi Ty2 culture grown in LBH. The levels of S. Typhi gene expression in calves were calculated for each loop inoculated with S. Typhi Ty2 relative to a loop inoculated with the Vi-negative S. Typhi ompR mutant (QT74) collected at the same time point from the same animal. A list of genes analyzed in this study with respective primers is provided in Table 2.

For statistical analysis of ratios (i.e., increases in S. Typhi gene expression or data, expressed as percentages), data were transformed logarithmically prior to performance of statistical analysis. A parametric test (paired Student's t test for T84 polarized cell samples and for calf ligated loop samples) was used to calculate whether differences were statistically significant.

GFP (green fluorescent protein) has been previously applied to study host pathogen interactions and Salmonella gene expression (4, 5, 46, 47). We first wanted to determine whether a plasmid-based reporter construct (pQT28) containing the viaB promoter (P_tviA) fused to egfp would be responsive to changes in osmolarity when introduced into S. Typhi isolate Ty2. Ty2(pQT28) was grown under Vi-inducing (SOB medium) and Vi-suppressing (LBH medium) conditions. As a negative control, SOB medium was inoculated with the Ty2 parent strain.

When Ty2(pQT28) was grown under conditions repressing Vi antigen expression (LBH medium), the peak fluorescence intensity was comparable to that of an S. Typhi strain lacking the reporter construct (Ty2). However, growth of Ty2(pQT28) under conditions inducing expression of the Vi antigen (SOB medium) resulted in marked induction of the P_tviA::egfp reporter construct, as indicated by a 10-fold increase in the peak fluorescence intensity (Fig. 1). These results were consistent with previous reports that expression of the Vi antigen of S. Typhi is affected by osmolarity (51) and validated the use of the plasmid-based P_tviA::egfp reporter construct.

To investigate whether expression of the Vi antigen changes when S. Typhi transits from the intestinal lumen into tissue, we used intestinal model epithelia as an in vitro model. Polarized T84 intestinal epithelial cells were infected with S. Typhi strain Ty2(pQT28) grown under high osmolarity (LBH) to mimic the conditions in the intestinal lumen. The P_tviA activity was monitored prior and subsequent to invasion of epithelial cells using flow cytometry. The peak fluorescence intensity was greater for Ty2(pQT28) associated with the T84 monolayer than for extracellular bacteria recovered from the apical compartment (Fig. 2). Furthermore, a larger percentage of cell-associated bacteria expressed fluorescence (53.8%) than did extracellular bacteria in the apical supernatant (22.9%).

To further examine Vi expression during infection of T84 cells, we performed real-time PCR analysis. Polarized T84 cells were infected with S. Typhi Ty2 grown in LBH medium, and for each experiment, bacterial RNA was pooled from infections of 10 separate Transwells. We monitored expression of the tviB gene, the fliC gene, and genes encoded by the SPI-1 T3SS (prgH and iagA). The tviB gene was expressed at a significantly higher level (P = 0.0059) by cell-associated bacteria than by extracellular bacteria (Fig. 3). The fliC, prgH, and iagA genes were expressed at significantly higher levels by extracellular bacteria than by cell-associated bacteria. These data were consistent with previous reports indicating that the invasion-associated SPI-1 genes and the genes involved in flagellar biosynthesis are downregulated following invasion (10, 11, 12, 14, 17, 40, 41).

The transcript levels of tviB in extracellular and cell-associated bacteria were expressed as fold increase over those measured in the inoculum culture, which was grown under Vi-suppressing conditions (300 mM NaCl). The induction of tviB expression seen in extracellular bacteria may be due to the use of tissue culture medium containing 150 mM NaCl, an osmolarity that induces Vi expression. Nonetheless, experiments with model epithelia indicated that the levels of Vi expression were significantly higher in cell-associated bacteria than in extracellular bacteria.

A deletion in ompR was previously found to inhibit Vi synthesis and result in a negative Vi slide agglutination reaction (35). A deletion of the ompR gene was constructed in S. Typhi strain Ty2 (QT74), and the mutation was verified by Southern blotting (Fig. 4A). The inability of the S. Typhi ompR mutant to express the Vi antigen was verified by slide agglutination using Vi antiserum (data not shown). A plasmid carrying the intact ompR gene (plasmid pQT50) was introduced into the S. Typhi ompR mutant (QT74) by electroporation. Expression of the Vi capsular antigen in QT74(pQT50) was restored with slide agglutination using Vi antiserum (data not shown). Complementation with a low-copy-number plasmid (pWSK29) carrying the ompR gene did not restore Vi production when detected by slide agglutination (data not shown). This finding was consistent with prior ompR complementation studies (35).

Vi expression by the ompR mutant was further characterized by introducing plasmid pQT28 (carrying an egfp reporter gene fused to the promoter of the viaB region). The ompR mutant [QT74(pQT28)] and S. Typhi wild type [Ty2(pQT28)] were grown under conditions inducing Vi capsule expression, and expression of the P_tviA::egfp reporter construct was analyzed by flow cytometry. The fluorescence intensities emitted by S. Typhi strain Ty2 (negative control) (data not shown) and strain QT74(pQT28) grown under Vi-inducing conditions (SOB broth) were similar. In contrast, Ty2(pQT28) (positive control) grown under Vi-inducing conditions (SOB broth) exhibited a 10-fold increased peak fluorescence intensity compared to that of QT74(pQT28) (Fig. 4B). These data were consistent with previous reports that OmpR is a positive regulator required for viaB expression (35).

Vi-suppressing conditions (300 mM NaCl) encountered in the intestinal lumen (13, 30, 45) are difficult to mimic using tissue culture models, which rely on the use of tissue culture media containing 150 mM NaCl (Fig. 3). To further study whether Vi expression is affected by the transition from the intestinal lumen (300 mM NaCl) into tissue (150 mM NaCl), we wanted to use an in vivo model suited for investigating the interaction of bacteria with intestinal tissue within hours after infection. Although cattle are resistant to oral infection, the bovine ligated ileal loop model has been implemented successfully to study early events during the interaction of S. Typhi with the intestinal mucosa (39, 40). In order to study the expression of the Vi antigen in vivo, we inoculated calf ligated ileal loops with S. Typhi strain Ty2 grown under high-osmolarity conditions (300 mM NaCl containing LB medium). Loops inoculated with the ompR mutant (QT74) served as a control, lacking expression of the Vi capsule while carrying an intact tviABCDE vexABCDE operon. To illustrate Vi expression, bacterial gene expression for each loop inoculated with the S. Typhi wild type (Ty2) was expressed as log fold increase over expression in the S. Typhi ompR mutant (QT74), collected at the same time point.

We performed real-time PCR analysis to investigate the expression of the Vi antigen by monitoring expression of the tviB gene. In addition, expression of fliC and invasion genes (prgH and iagA) was assessed in calf ileal tissue following invasion at 2 and 8 hours after infection. At 2 and 8 h, expression of tviB among intracellular bacteria within calf tissue was significantly higher than that in the inoculum, with P values of 0.001 and 0.013, respectively (Fig. 5). The level of tviB expression was also higher at 8 h than at 2 h, although this difference was not statistically significant (P = 0.06). Although no statistical significance was obtained, there was a trend that fliC, prgH, and iagA were downregulated in tissue at 2 h and 8 h compared to expression levels detected in the ompR mutant.

To directly detect Vi production within tissue, frozen sections of calf intestinal tissue collected at 2 h and 8 h after infection were labeled with primary anti-Vi antibody and secondary anti-rabbit goat antibody conjugated to Alexa Fluor 594 (orange-red fluorescence), and tissue was counterstained with DAPI nuclear stain (blue fluorescence). As a negative control, tissue collected at corresponding time points from loops inoculated with sterile medium (mock infection) was labeled by the same procedure. No Vi antigen production was detected in the mock-infected ileal tissue (Fig. 6D and H). Ileal loops infected with S. Typhi strain Ty2 revealed Vi-expressing bacteria along the intestinal villus tips 2 h after infection, with bacterial penetration of the intestinal epithelial barrier and entry into the lamina propria (Fig. 6B and C). Fluorescent staining of bovine Peyer's patch sections revealed large numbers of Vi-expressing bacteria in the mantle and in the area between germinal centers (Fig. 6F and G). At 8 h, our findings were similar, but the numbers of invasive bacteria expressing the capsule within the tissue were elevated. These findings were consistent with the concept that production of the Vi antigen occurs during bacterial passage through the intestinal epithelium.

The histopathology of the same sections of ileum and Peyer's patches (Fig. 6A and E, respectively) revealed moderate-to-severe changes in the epithelium infrastructure characterized by epithelial detachment, blunting, and erosion, which agrees with previous studies that demonstrate the induction of invasion genes and the destruction of the intestinal epithelium with S. Typhi under high-osmolarity growth conditions (12, 51). Only mild increases in neutrophilic counts were detected in the lamina propria and submucosa of the ileum, whereas the numbers of macrophages were markedly increased (data not shown). Sections of bovine Peyer's patches illustrated mild-to-moderate increases in neutrophil counts and markedly increased numbers of mononuclear cells (data not shown). The lack of infiltrating neutrophils and the presence of a predominate mononuclear cell infiltrate within the intestine are characteristic of typhoid fever infections (24, 32, 33).