The Typhoid Toxin Produced by the Nontyphoidal Salmonella enterica Serotype Javiana Is Required for Induction of a DNA Damage Response In Vitro and Systemic Spread In Vivo

Abstract

The Salmonella cytolethal distending toxin (S-CDT), first described as the "typhoid toxin" in Salmonella enterica subsp. enterica serotype Typhi, induces DNA damage in eukaryotic cells. Recent studies have shown that more than 40 nontyphoidal Salmonella (NTS) serotypes carry genes that encode S-CDT, yet very little is known about the activity, function, and role of S-CDT in NTS. Here we show that deletion of genes encoding the binding subunit (pltB) and a bacteriophage muramidase predicted to play a role in toxin export (ttsA) does not abolish toxin activity in the S-CDT-positive NTS Salmonella enterica subsp. enterica serotype Javiana. However, S. Javiana strains harboring deletions of both pltB and its homolog artB, had a complete loss of S-CDT activity, suggesting that S. Javiana carries genes encoding two variants of the binding subunit. S-CDT-mediated DNA damage, as determined by phosphorylation of histone 2AX (H2AX), producing phosphorylated H2AX (γH2AX), was restricted to epithelial cells in S and G₂/M phases of the cell cycle and did not result in apoptosis or cell death. Compared to mice infected with a ΔcdtB strain, mice infected with wild-type S. Javiana had significantly higher levels of S. Javiana in the liver, but not in the spleen, ileum, or cecum. Overall, we show that production of active S-CDT by NTS serotype S. Javiana requires different genes (cdtB, pltA, and either pltB or artB) for expression of biologically active toxin than those reported for S-CDT production by S. Typhi (cdtB, pltA, pltB, and ttsA). However, as in S. Typhi, NTS S-CDT influences the outcome of infection both in vitro and in vivoIMPORTANCE Nontyphoidal Salmonella (NTS) are a major cause of bacterial food-borne illness worldwide; however, our understanding of virulence mechanisms that determine the outcome and severity of nontyphoidal salmonellosis is incompletely understood. Here we show that S-CDT produced by NTS plays a significant role in the outcome of infection both in vitro and in vivo, highlighting S-CDT as an important virulence factor for nontyphoidal Salmonella serotypes. Our data also contribute novel information about the function of S-CDT, as S-CDT-mediated DNA damage occurs only during certain phases of the cell cycle, and the resulting damage does not induce cell death as assessed using a propidium iodide exclusion assay. Importantly, our data support that, despite having genetically similar S-CDT operons, NTS serotype S. Javiana has different genetic requirements than S. Typhi, for the production and export of active S-CDT.

Keywords: DNA damage; Salmonella; nontyphoidal; typhoid toxin.

FIG 1

Infection with wild-type S. Javiana activates a DNA damage response and results in an accumulation of cells in the G₂/M phase. HIEC-6 cells were infected with S. Javiana for 48 h prior to immunofluorescence staining or flow cytometry analyses. Cells treated with the topoisomerase inhibitor etoposide (at a final concentration of 2 µM) for 24 h served as a positive control. Data are from three independent experiments. (A) Representative histograms showing the cell cycle progression of HIEC-6 cells infected with wild-type (WT) and ΔcdtB S. Javiana strains; an untreated control is included to show a normal cell cycle progression, and cells treated with 2 µM etoposide demonstrate the accumulation of cells in G₂/M phase due to DNA damage. (B) Quantification of cell cycle analyses shown in panel A. Histogram bars that do not share letters within a given cell cycle phase group (i.e., G₁, S, and G₂/M) are significantly different (P < 0.05). (C) Representative images of cells infected with S. Javiana strains (antibody stain shown in green) and positive controls (treated with 2 µM etoposide) and negative controls (untreated). The nuclei of HIEC-6 cells were stained with DAPI (blue). Scale bars represent 20 microns. (D) Quantification of the proportions of cells with at least four 53BP1 foci (shown in yellow) that colocalized with γH2AX foci (shown in red). Treatments that do not share letters have significantly different (P < 0.05) proportions of cells with 53BP1 foci that colocalized with γH2AX foci. P values were corrected for multiple comparisons using the Tukey honestly significant difference (HSD) correction method. Values in panels B and D are means ± standard errors of the means (error bars).

FIG 2

pltB, ttsA, and STY1887 are not required for S-CDT-mediated DNA damage in HIEC-6 cells. HIEC-6 cells were infected with S. Javiana strains harboring single-gene deletions in the S-CDT islet. Immunofluorescence staining was performed for DNA damage response foci γH2AX and 53BP1. (A) Organization of the S-CDT islet in S. Javiana. Genes shown in blue represent genes encoding S-CDT subunits; genes shown in white are contained within the islet but do not encode protein products that constitute part of the S-CDT holotoxin. (B) Representative images of cells infected with wild-type S. Javiana and S-CDT single gene deletion strains (both colored green in the merged image); HIEC-6 cell nuclei are shown in blue. (C) Quantification of the proportions of cells with at least four 53BP1 foci that colocalized with γH2AX foci. Treatments that do not share letters have significantly different (P < 0.05) proportions of cells with 53BP1 foci that colocalized with γH2AX foci. Data represent three independent experiments. P values were corrected, to account for multiple testing, using Tukey’s HSD test. Error bars represent standard errors of the means.

FIG 3

The presence of either artB or pltB is essential for S-CDT-mediated intoxication. Normal human intestinal epithelial HIEC-6 cells were infected with S. Javiana strains harboring either single gene deletions in artB or artAB or double deletions in pltB and either artB or artAB. At 48 hpi, cells were stained for the DNA damage response foci γH2AX and 53BP1. (A) Schematic of the gene deletions for ΔartB and ΔartAB strains. (B) Representative immunofluorescence images of HIEC-6 cells infected with S. Javiana deletion strains. S. Javiana cells (green) and HIEC-6 cell nuclei (blue) are shown. (C) Quantification of the proportions of cells with colocalized foci of at least four 53BP1 foci and γH2AX foci. Treatments that do not share letters are significantly different (P < 0.05). Results are from three independent experiments. P values were adjusted using Tukey’s HSD test to correct for multiple-comparison testing. Error bars represent standard errors of the means.

FIG 4

Key amino acid residues necessary for PltB binding to sugar moieties are also present in ArtB. (A) Alignments of translated amino acid sequences of PltB and ArtB from S. Javiana strain CFSAN0001992 and PltB from S. Typhi strain CT-18. Key residues that are conserved are shown on light blue background. Residues that differ between S. Typhi PltB and S. Javiana ArtB or PltB (24) are shown on dark blue background. Amino acid residues conserved in all three sequences (59) are indicated by an asterisk below the sequence alignment. (B) Predicted 3D structure of PltB and ArtB based on translated amino acid sequences of pltB and artB extracted from S. Javiana strain CFSAN0001992, generated using SWISS-MODEL online software (60).

FIG 5

S-CDT-mediated DNA damage primarily occurs in the S and G₂/M phases of the cell cycle of HIEC-6 human intestinal epithelial cells. HIEC-6 cells infected with S. Javiana wild-type (WT) and ΔcdtB strains for 48 h were stained to detect γH2AX and DNA content to determine the specific phases of the cell cycle during which S-CDT-mediated DNA damage occurred. (A) Representative histograms of DNA cell cycle analysis of HIEC-6 cells (gray) with γH2AX-positive cells overlaid (blue). (B) Quantification of γH2AX-positive HIEC-6 cells by cell cycle phase. Results are from two independent experiments. Treatments that do not share letters represent significantly different mean proportions (P < 0.05) of cells positive for a given cell cycle phase (e.g., proportions of γH2AX-positive cells in G₁ were compared for the four treatment groups). P values were corrected for multiple comparisons using Tukey’s HSD test. Error bars represent standard errors of the means.

FIG 6

Infection of HIEC-6 human intestinal epithelial cells with S. Javiana does not induce apoptosis or cell death regardless of S-CDT status. HIEC-6 cells were infected with wild-type or ΔcdtB S. Javiana strains for 48 h or treated with 50 µM H₂O₂ for 8 h (for a positive control). Cells were stained with annexin V conjugated to Alexa Fluor 488 (AF488); PI was used as a live/dead stain. (A) Representative images of density plots of annexin V and PI staining of HIEC-6 cells infected with S. Javiana, treated with H₂O₂, or untreated. (B and C) Quantification of the proportions of PI-positive cells (B) and annexin V-positive cells (C) shown in panel A. Treatments that do not share letters have significantly different (P < 0.05) proportions of cells positive for either annexin V or PI staining. Results in panels B and C are from four independent experiments. Error bars represent standard errors of the means.

FIG 7

Infection of C56BL/6 mice with WT S. Javiana results in a higher bacterial load in the liver, but not in the spleen, ileum, or cecum, compared to infection with a ΔcdtB strain. Eight-week-old female C56BL/6 mice were orally inoculated with 1 × 10⁹ CFU S. Javiana wild-type or ΔcdtB strain (five mice in each group). At 48 hpi, the mice were euthanized, and tissues were harvested for bacteriologic and histological examination. (A) S. Javiana cells were enumerated from the liver, spleen, ileum, and cecum. The average bacterial load was calculated for each tissue. The limit of detection (LOD) for the average bacterial load for all tissues was 5 CFU. Differences in S. Javiana recovered from each tissue were compared using the nonparametric Kruskal-Wallis test. (B and C) Sections of the distal ileum, cecum, and proximal colon were fixed in buffered formaldehyde, embedded in paraffin, sectioned, and stained with hematoxylin and eosin. The levels of polymorphonuclear leukocytes (PMN) in the cecum (Ce) and proximal colon (Co) were scored on a scale of 0 to 4 (0 representing no PMNs and 4 representing large numbers of PMNs); cryptitis was assessed using the same scale. The degree of edema was scored from 0 (no edema) to 3 (severe edema). (C) The pathology score for typhlocolitis represents the sum of the scores (maximum score is 15) in panel B. Statistical differences in the counts of mice in each category of pathology score were assessed using the nonparametric Kruskal-Wallis test. Each symbol represents the value for an individual mouse. Black bars represent the median bacterial load (A) or median pathology scores (B and C) for mice infected with each S. Javiana strain. Values that are significantly different (P < 0.05) are indicated by a bar and asterisk. Only statistical associations with P < 0.05 are shown.