Edwardsiella tarda-Induced cytotoxicity depends on its type III secretion system and flagellin

Abstract

Many Gram-negative bacteria utilize a type III secretion system (T3SS) to translocate virulence proteins into host cells to cause diseases. In responding to infection, macrophages detect some of the translocated proteins to activate caspase-1-mediated cell death, called pyroptosis, and secretion of proinflammatory cytokines to control the infection. Edwardsiella tarda is a Gram-negative enteric pathogen that causes hemorrhagic septicemia in fish and both gastrointestinal and extraintestinal infections in humans. In this study, we report that the T3SS of E. tarda facilitates its survival and replication in murine bone marrow-derived macrophages, and E. tarda infection triggers pyroptosis of infected macrophages from mice and fish and increased secretion of the cytokine interleukin 1β in a T3SS-dependent manner. Deletion of the flagellin gene fliC of E. tarda results in decreased cytotoxicity for infected macrophages and does not attenuate its virulence in a fish model of infection, whereas upregulated expression of FliC in the fliC mutant strain reduces its virulence. We propose that the host controls E. tarda infection partially by detecting FliC translocated by the T3SS, whereas the bacteria downregulate the expression of FliC to evade innate immunity.

FIG 1

Survival and replication of E. tarda PPD130/91 in C57BL/6 BMDMs. (A) E. tarda PPD130/91 fails to replicate in BMDMs at 37°C, but T3SS-dependent replication is observed at 30°C. BMDMs were infected with opsonized bacteria at an MOI of ∼2.5. Intracellular bacteria were determined by CFU counting at the indicated times after infection. Experiments were performed in quadruplicate wells for each infection, and the data are from one representative infection. *, P < 0.05 (relative to the 1-h value). (B) Microscopic analysis of E. tarda PPD130/91 replication in BMDMs at 30°C. BMDMs were infected with bacteria expressing GFP and fixed at different time points. Cells were labeled with rhodamine phalloidin and subjected to confocal microscopy. At least 300 infected cells per infection were counted for the number of intracellular bacteria. Data are means ± standard errors of the means (SEM) from three independent experiments.

FIG 2

T3SS-dependent cytotoxicity in BMDMs and J774A.1 cells. (A) Confocal micrographs of PI uptake by BMDMs. BMDMs were infected with GFP-expressing strains for 2 h and stained with PI in Opti-MEM without fixation. Green, bacteria; red, PI-positive BMDMs. Bar, 50 μm. (B) Quantification of PI-positive infected cells. More than 300 infected cells for each infection were counted. (C) T3SS-dependent cytotoxicity of E. tarda PPD130/91 in BMDMs at 30°C determined by LDH assay. BMDMs were infected with different bacterial strains, and the culture supernatants were collected and processed for the LDH release assay. (D) T3SS dependent cytotoxicity of E. tarda PPD130/91 in J774A.1 at 30°C. (E) T3SS-dependent cytotoxicity of E. tarda PPD130/91 in BMDMs at 37°C. Samples were collected 1 h after uptake for analysis. Means ± SEM from three experiments are shown. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 3

Caspase-1 of BMDMs is activated during E. tarda infection. (A) Effect of the caspase-1 inhibitor Ac-YVAD-AOM on the release of LDH. BMDMs were pretreated with 317 μM Ac-YVAD-AOM or dimethyl sulfoxide (DMSO) for 90 min and then infected with the wild type for 2 h. Culture supernatants were collected to determine LDH release. ***, P < 0.001. (B) E. tarda-induced activation of caspase-1. BMDMs were either mock infected or infected with indicated strains for 2 h. Cell culture supernatants were collected, and cell monolayers were lysed with 4× SDS sample buffer. Membrane was probed with anti-caspase-1 (p20) mouse monoclonal antibody. β-Actin was used to show the same loading of the cell lysates.

FIG 4

The E. tarda T3SS is required for secretion of IL-1β but not TNF-α by BMDMs. BMDMs were infected with different strains for 6.5 h, and supernatants were collected for ELISA. Supernatant from a noninfected cell culture was used as a control. Values are means ± standard deviations (SD) for triplicate cultures. **, P < 0.01; ***, P < 0.001. (A) Secretion of IL-1β. (B) Secretion of TNF-α.

FIG 5

E. tarda T3SS and FliC are required for cytotoxicity for fish leukocytes. (A) Fish leukocytes isolated from head kidney were left noninfected or were infected at an MOI of 5 with different strains. After 1 h and 2 h, the levels of LDH released were analyzed. Values are means ± SD for 8 replicate cultures. ***, P < 0.001. (B) The release of LDH by fish leukocytes is inhibited by Ac-YVAD-AOM. Primary leukocytes were pretreated with 317 μM Ac-YVAD-AOM or DMSO for 90 min and then infected with the wild type or the esaN mutant for 2 h. Culture supernatants were collected to determine LDH release. ***, P < 0.001. (C) FliC2HA is secreted into the culture supernatant in a T3SS-dependent manner. Bacterial strains carrying pACYC-fliC2HA were grown in TSB for preparing protein samples from bacterial pellets (P) or culture supernatants (S). Membranes were probed with anti-HA for FliC2HA, anti-DnaK (a bacterial cytosolic marker), and anti-EvpC (a T6SS protein, which served as an internal loading control for different strains). DnaK was not detected in any culture supernatants, showing that detection of FliC2HA is not due to contamination from bacterial cells.

FIG 6

Competitive index analysis. Five naive blue gourami fish were used for each group. Fish were injected intramuscularly with mixtures of equal numbers of cells of the indicated strains and the wild-type strain and sacrificed 72 h after injection. CIs from livers for individual fish are presented, and the mean ± SD is also shown. Student's t test was used to calculate the P value with the hypothetical mean of 1.0 (*, P < 0.05).