Salmonella Induces the cGAS-STING-Dependent Type I Interferon Response in Murine Macrophages by Triggering mtDNA Release

Abstract

Salmonella enterica serovar Typhimurium (S. Typhimurium) elicited strong innate immune responses in macrophages. To activate innate immunity, pattern recognition receptors (PRRs) in host cells can recognize highly conserved pathogen-associated molecular patterns (PAMPs). Here, we showed that S. Typhimurium induced a robust type I interferon (IFN) response in murine macrophages. Exposure of macrophages to S. Typhimurium activated a Toll-like receptor 4 (TLR4)-dependent type I IFN response. Next, we showed that type I IFN and IFN-stimulated genes (ISGs) were elicited in a TBK1-IFN-dependent manner. Furthermore, cytosolic DNA sensor cyclic GMP-AMP synthase (cGAS) and immune adaptor protein stimulator of interferon genes (STING) were also required for the induction of type I IFN response during infection. Intriguingly, S. Typhimurium infection triggered mitochondrial DNA (mtDNA) release into the cytosol to activate the type I IFN response. In addition, we also showed that bacterial DNA was enriched in cGAS during infection, which may contribute to cGAS activation. Finally, we showed that cGAS and STING deficient mice and cells were more susceptible to S. Typhimurium infection, signifying the critical role of the cGAS-STING pathway in host defense against S. Typhimurium infection. In conclusion, in addition to TLR4-dependent innate immune response, we demonstrated that S. Typhimurium induced the type I IFN response in a cGAS-STING-dependent manner and the S. Typhimurium-induced mtDNA release was important for the induction of type I IFN. This study elucidated a new mechanism by which bacterial pathogen activated the cGAS-STING pathway and also characterized the important role of cGAS-STING during S. Typhimurium infection. IMPORTANCE As one of the most common foodborne transmitted zoonotic pathogens, S. Typhimurium infection causes diarrheal disease in humans and animals. S. Typhimurium infection has been implicated as an inducer for the type I interferon (IFN) response in macrophages, but the mechanisms are not fully understood. In this study, we reported that in addition to TLR4-dependent response, the cytosolic surveillance pathway (CSP) cGAS-STING is also required for the activation of type I IFN response during S. Typhimurium infection. We further showed that the infection of S. Typhimurium triggered mtDNA release into the cytosol, which induces the type I IFN response. In addition, physical interactions between cGAS and S. Typhimurium DNA have been identified in the context of infection. Importantly, we also provided convincing in vivo and in vitro evidence that the cGAS-STING pathway was potently implicated in the host defense against S. Typhimurium infection. Together, we uncovered a mechanism by which type I IFN response is elicited during S. Typhimurium infection in murine macrophages in an mtDNA-cGAS-STING-dependent manner.

Keywords: STING; Salmonella; cGAS; interferon; mtDNA; type I interferon.

FIG 1

S. Typhimurium infection triggers proinflammatory cytokine expression in murine macrophages. (A) qRT-PCR analysis of gene expression in mouse peritoneal macrophages (PMs) uninfected (mock) or infected with S. Typhimurium for 2, 4, or 8 h at an MOI of 10 (n = 3). (B) qRT-PCR analysis of gene expression in mouse PMs uninfected (mock) or infected with S. Typhimurium for 8 h at different MOIs (n = 3). Data in (A) and (B) were normalized to uninfected control (mock, set as 1), Actin was used as the housekeeping gene. Error bars represent ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 2

RNA-seq analysis reveals the induction of type I IFN response in S. Typhimurium-infected murine macrophages. (A) Volcano plot showing gene expression analysis of PMs uninfected (mock) or infected with S. Typhimurium for 8 h at an MOI of 10. The x-axis shows a fold change in gene expression by calculating log₂(infected RPKM/mock RPKM) and the y-axis shows statistical significance (log₁₀[Q-value]). Downregulated genes are plotted on the left (blue) and upregulated genes are on the right (red). Each treatment had 3 biological replicates. RPKM: reads per kilobase of exon model per Million mapped reads. Differentially expressed genes were those with a false discovery rate (FDR) cutoff of 0.05 and a fold change of ≥±2. (B and C) qRT-PCR analysis of gene expression of downregulated genes (B) and upregulated (C) genes in mouse PMs uninfected (mock) or infected with S. Typhimurium for 8 h at an MOI of 10 (n = 3). (D) Heatmap of RNA-seq analysis. PMs were uninfected (mock) or infected with S. Typhimurium for 8 h at an MOI of 10. Heatmap was made by calculating log₂(infected RPKM/mock RPKM). Numbers 1, 2, and 3 indicate 3 biological replicates. (E) Ingenuity pathway analysis of upregulated gene expression changes in PMs infected with S. Typhimurium. (F) ELISA analysis of IFN-β production in mouse PMs uninfected (mock) or infected with S. Typhimurium for 2, 4, 6, 8, or 10 h at an MOI of 10 (n = 3). (G) qRT-PCR analysis of gene expression in C57BL/6 (WT) or TLR4^−/− mouse-derived PMs uninfected (mock) or infected with S. Typhimurium for 2, 4, and 8 h at an MOI of 10 (n = 3). (H) ELISA analysis of IFN-β production in the supernatant from C57BL/6 (WT) or TLR4^−/− mouse-derived PMs uninfected (mock) or infected with S. Typhimurium for 6 h at an MOI of 10 (n = 3). Data in (B) and (C) were normalized to uninfected control (mock, set as 1), Data in (G) were normalized to mock-infected C57BL/6 (WT) PMs, mock-infected TLR4^−/− mouse-derived PMs, respectively (mock, both set as 1). Actin was used as the housekeeping gene. Error bars represent ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 3

TBK1-IFN axis is required for the S. Typhimurium-induced type I IFN response. (A) Immunoblot analysis of protein expression in Raw264.7 cells untreated (con) or pretreated with BX795 (2 μM) for 2 h and uninfected (mock) or infected with S. Typhimurium for 8 h at an MOI of 100. (B) qRT-PCR analysis of gene expression in PMs untreated (con) or pretreated with BX795 (2 μM) for 2 h and uninfected (mock) or infected with S. Typhimurium for 8 h at an MOI of 10 (n = 3). (C) qRT-PCR analysis of gene expression in C57BL/6 or IFNAR1^−/− mouse-derived PMs uninfected (mock) or infected with S. Typhimurium for 2, 4, or 8 h at an MOI of 10 (n = 3). Data in (B) were normalized to untreated, mock-infected control and BX795-treated, mock-infected control, respectively (mock, both set as 1, not shown). Data in (C) were normalized to mock-infected C57BL/6 PMs and mock-infected IFNAR1^−/− PMs, respectively (mock, both set as 1). Actin was used as the housekeeping gene. Error bars represent ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 4

cGAS-STING contributes to S. Typhimurium-induced type I IFN response. (A) ELISA analysis of IFN-β production in the supernatant from C57BL/6 (WT), cGAS^−/− or STING^−/− mouse-derived PMs uninfected (mock) or infected with S. Typhimurium for 6 h at an MOI of 10 (n = 3). (B) and (C) qRT-PCR analysis of gene expression in C57BL/6 (WT), cGAS^−/− or STING^−/− mouse-derived PMs uninfected (mock) or infected with S. Typhimurium for 8 h at an MOI of 10 (n = 3). Data in (B) and (C) were normalized to mock-infected C57BL/6 (WT) PMs, mock-infected cGAS^−/− PMs, and mock-infected STING^−/− PMs, respectively (mock, all set as 1). Actin was used as the housekeeping gene. Error bars represent ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 5

Activation of STING potentiates S. Typhimurium-induced type I IFN response. (A) Immunoblot analysis of protein expression in Raw264.7 untreated (con) or pretreated with DMXAA (100 μg/mL) for 12 h and uninfected (mock) or infected with S. Typhimurium for 8 h at an MOI of 100. (B) qRT-PCR analysis of gene expression in PMs untreated (con) or pretreated with DMXAA (100 μg/mL) for 12 h and uninfected (mock) or infected with S. Typhimurium for 8 h at an MOI of 10 (n = 3). (C) PMs were untreated or lipofectamine-transfected with 5 μM 2′3′-cGAMP (cGAMP) for 4 h and then uninfected (mock) or infected with S. Typhimurium at an MOI of 10. At 6 h postinfection, qRT-PCR analysis of gene expression (n = 3). Data in (B) were normalized to untreated, mock-infected control (mock, set as 1, not shown). Data in (C) were normalized to untreated, mock-infected control (mock, set as 1). Actin was used as the housekeeping gene. Error bars represent ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 6

S. Typhimurium triggers mtDNA release to induce the type I IFN response. (A) Fluorescence microscopy analysis of mitochondrial membrane potential by assessing TMRE in Raw264.7 cells uninfected (mock) or infected with S. Typhimurium for 2 h at an MOI of 100. Scale bar: 20 μm. (B) The average integrated density of TMRE fluorescence. The average density was obtained from 20 individual cells for different conditions. 3 independent experiments were performed. (C) Immunoblot analysis of protein expression in different fractions in Raw264.7 cells uninfected (mock) or infected with S. Typhimurium for 8 h at an MOI of 10 or 100. Cells were treated with Tunicamycin (4 μg/mL) for 8 h as a positive-control for mtDNA release. (D) qPCR analysis of mtDNA in the cytosolic fraction of Raw264.7 cells uninfected (mock) or infected with S. Typhimurium for 8 h at an MOI of 10 or 100 (n = 3). (E) qRT-PCR analysis of gene expression in EB-treated and untreated (con) Raw264.7 cells uninfected (mock) or infected with S. Typhimurium for 8 h at an MOI of 100 (n = 3). (F) qRT-PCR analysis of gene expression in PMs pretreated with Mito Tempo (100 μM) or NecroX-5 (40 μM) for 2 h and uninfected (mock) or infected with S. Typhimurium for 6 h at an MOI of 10 (n = 3). Data in (D) were normalized to mock-infected control (con, set as 1). Data in (E) and (F) were normalized to untreated, mock-infected control and pretreated, mock-infected control, respectively (mock, all set as 1). Actin was used as the housekeeping gene. Error bars represent ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 7

Bacterial DNA binds to cGAS during S. Typhimurium infection. (A) qRT-PCR analysis of gene expression in PMs untreated (con) or pretreated with BafA1 (200 nM) or chymostatin (100 μM) for 2 h and uninfected (mock) or infected with S. Typhimurium for 8 h at an MOI of 10 (n = 3). (B) qPCR analysis of S. Typhimurium-derived sequences (dnaE and sciK) in the cytosolic fraction of PMs uninfected (mock) or infected with S. Typhimurium for 8 h at an MOI of 100 (n = 3). ND, not detected. (C) Raw264.7 cells were uninfected or infected with S. Typhimurium for 8 h at an MOI of 100, then immunoprecipitated with IgG or cGAS antibody. qPCR of S. Typhimurium-derived sequences (sciK and dnaE) from DNA isolated from IPs (n = 3). Data in (A) were normalized to untreated, mock-infected control and BafA1-treated, mock-infected, and chymostatin-treated, mock-infected, respectively (mock, all set as 1, not shown). Quantities in (C) were normalized to inputs. Actin was used as the housekeeping gene. Error bars represent ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 8

cGAS-STING contributes to host defense against S. Typhimurium infection. (A) and (B) C57BL/6, cGAS^−/−, STING^−/− mice were intragastrically inoculated with 1 × 10⁸ CFU S. Typhimurium. (A) The survival rate of the mice was determined (C57BL/6, n = 10; STING^−/−, n = 10). (B) Homogenates of the liver and spleen were plated to determine the bacterial CFU counts per gram in the indicated organs at 120 h postinfection (n = 6). (C) Fluorescence microscopy analysis of C57BL/6 or STING^−/− mouse-derived PMs infected with S. Typhimurium for 2 h or 4 h at an MOI of 10. Hoechst33342: nuclei, GFP, S. Typhimurium carry a GFP expressing plasmid. Scale bar: 75 μm. The ratio of GFP fluorescence integrated density to Hoechst33342 fluorescence integrated density is on the right (n = 3). (D) qRT-PCR analysis of S. Typhimurium 16S rRNA level in PMs untreated (con) or pretreated with DMXAA (100 μg/mL) for 12 h, or pretreated lipofectamine-transfected with 5 μM 2′3′-cGAMP (cGAMP) for 4 h and infected with S. Typhimurium for 8 h at an MOI of 100 (n = 3). Data in (D) were normalized to untreated S. Typhimurium-infected control (S. T, set as 1). Actin was used as the housekeeping gene. Error bars represent ± SEM. *, P < 0.05; **, P < 0.01.

FIG 9

S. Typhimurium triggers the cGAS-STING-dependent type I IFN response by inducing mtDNA release. In murine macrophages, S. Typhimurium infection led to the release of mtDNA into cytosol that possibly engages cytosolic DNA sensor cGAS. The activated cGAS led to the activation of immune adaptor protein STING and phosphorylation of kinase TBK1, which activates the transcription factor IRF3. Translocation of IRF3 resulted in transcriptional induction of the type I IFN response.