RIG-I detects mRNA of intracellular Salmonella enterica serovar Typhimurium during bacterial infection

Abstract

The cytoplasmic helicase RIG-I is an established sensor for viral 5'-triphosphorylated RNA species. Recently, RIG-I was also implicated in the detection of intracellular bacteria. However, little is known about the host cell specificity of this process and the bacterial pathogen-associated molecular pattern (PAMP) that activates RIG-I. Here we show that RNA of Salmonella enterica serovar Typhimurium activates production of beta interferon in a RIG-I-dependent fashion only in nonphagocytic cells. In phagocytic cells, RIG-I is obsolete for detection of Salmonella infection. We further demonstrate that Salmonella mRNA reaches the cytoplasm during infection and is thus accessible for RIG-I. The results from next-generation sequencing analysis of RIG-I-associated RNA suggest that coding bacterial mRNAs represent the activating PAMP. IMPORTANCE S. Typhimurium is a major food-borne pathogen. After fecal-oral transmission, it can infect epithelial cells in the gut as well as immune cells (mainly macrophages, dendritic cells, and M cells). The innate host immune system relies on a growing number of sensors that detect pathogen-associated molecular patterns (PAMPs) to launch a first broad-spectrum response to invading pathogens. Successful detection of a given pathogen depends on colocalization of host sensors and PAMPs as well as potential countermeasures of the pathogen during infection. RIG-I-like helicases were mainly associated with detection of RNA viruses. Our work shows that S. Typhimurium is detected by RIG-I during infection specifically in nonimmune cells.

FIG 1

RIG-I is essential for detection of Salmonella infection in nonphagocytic cells but is obsolete for detection in macrophages. (A) RIG-I^+/+ and RIG^−/− murine embryonic fibroblasts (MEFs) were infected for 8 h with S. Typhimurium SL1344 at an MOI of 10 or with SeV at an MOI of 10. (B) MAVS^+/+ and MAVS^−/− MEFs were infected with SL1344 at an MOI of 10, SeV at an MOI of 10, or EMCV (encephalomyocarditis virus) at an MOI of 10 for 8 h. (C) TRIF/Myd88^+/+ and TRIF/Myd88^−/− MEFs were infected with SL1344 at an MOI of 10 for 8 h. (D) RIG-I^+/+ and RIG^−/− BMDM were infected with SL1344 at an MOI of 10 or with SeV at an MOI of 10 or were treated with 2 µg/ml LPS for 4 h. Data representing average n-fold expression of IFN-β mRNA over 18S rRNA levels of three independent biological samples each measured in technical triplicate experiments ± standard deviations (SD) are depicted. (E) MAVS^+/+ and MAVS^−/− BMDM were infected with SL1344 at an MOI of 10 or with SeV at an MOI of 10 or were treated with 2 µg/ml LPS for 4 h. (F) TRIF/Myd88^+/+ and TRIF/Myd88^−/− BMDM were infected with SL1344 at an MOI of 10 or with SeV at an MOI of 10 or were treated with 2 µg/ml LPS for 4 h. (G) RIG-I^+/+ and RIG^−/− BMDM and TRIF/Myd88^+/+ and TRIF/Myd88^−/− BMDM were transfected with 1 µg SL1344 RNA for 6 h. (H) RIG-I^+/+ and RIG^−/− MEFs were infected for 8 h with L. monocytogenes at an MOI of 10. (G) RIG-I^+/+ and RIG^−/− MEFs were infected for 4 h with L. monocytogenes at an MOI of 10. Each column in this figure represents the average n-fold expression of IFN-β over 18S rRNA levels of three independent biological samples each measured in technical triplicate experiments ± SD as determined by specific qPCR.

FIG 2

Salmonella RNA induces IFN-β expression upon transfection. (A) 293T-FF reporter cells were transfected with 1 µg of mock-treated or RNase A-treated total RNA from A549 cells infected for 8 h with influenza A/Viet Nam/1203/2004 HAlo virus (IAV) or S. Typhimurium (SL1344) at an MOI of 5. Each column represents the mean increase in n-fold relative luminescence units (RLU) compared to the average results from mock-infected cells at 24 h posttransfection of three independent biological samples each measured in technical triplicate experiments ± SD. (B) A549 control (shCTRL) and A549 shRIG-I cells were transfected with 1 µg of total RNA from S. Typhimurium (SL1344). Each column represents the average n-fold expression of IFN-β over 18S rRNA levels of three independent biological samples each measured in technical triplicate experiments ± SD. (C) Western blot analysis of total protein lysates of A549 cells stably expressing shRNAs against RIG-I (shRIG-I) or control shRNAs (sh ctrl). Cells were stimulated for 8 h with 100 U IFN-α or left untreated. Membranes were incubated with anti-RIG-I (upper panel) or anti-β-actin (lower panel).

FIG 3

Salmonella Typhimurium RNA is accessible in the cytoplasm (A) MEFs were infected with S. Typhimurium WT-GFP at an MOI of 1 or with S. Typhimurium ΔinvG at the indicated MOI for 8 h. Values represent the average n-fold expression level of IFN-β mRNA over 18S rRNA levels of three independent biological samples each measured in technical triplicate experiments ± SD as determined by qPCR. (B) Intracellular titers of MEF infected with S. Typhimurium WT-GFP and ΔinvG at an MOI of 5, as determined by serial dilution of cell lysates. Each column represents the average titers of three independent biological samples each measured in technical triplicate experiments ± SD. (C) HeLa cells transfected with pCAGGS-RIG-I-mCherry (red) were infected 48 h posttransfection with S. Typhimurium (green) at an MOI of 10 for 8 h. Z-stacked pictures were taken at a distance of 0.5 µm, covering the whole cell body to exclude colocalization of RIG-I and S. Typhimurium signal. (D to G) HeLa cells constitutively expressing MS2-GFP (green) (D and E) were infected with WT S. Typhimurium SL1344 (F) or SL1344 expressing MS2-aptamer RNA (G) (a magnification of the indicated area is depicted on the right side). Cytoplasmic complexes of MS2-GFP are indicated with white arrows; Salmonella is indicated in red. (H) Quantification of MS2-GFP complexes in the cytoplasm of 40 randomly chosen SL1344-positive cells.

FIG 4

Salmonella coding RNAs bind to RIG-I and activate IFN-β expression in a 5′-triphosphate-dependent fashion (A) Column fractionation was used to eliminate RNAs smaller than 30 or smaller than 500 bp. A549 cells were transfected with 1 µg or 4 µg of total RNA and equal volumes of fractionated RNAs. Average n-fold expression levels of IFN-β mRNA over 18S rRNA levels of three independent biological samples each measured in technical triplicate experiments ± SD are depicted. n.s., not significant. (B) Concentrations of RNAs after column fractionation were compared to total RNA as determined by spectrometry. (C) Relative representations (%) of reads of coding (black) and noncoding (white) RNAs from total bacterial RNA (Total), GFP pulldown RNA (GFP), and RIG-I pulldown RNA groups. n.d. (gray) represents RNAs that could not be classified as coding or noncoding. (D) A549 cells were transfected with 40 ng of pulldown RNA for 16 h. Average n-fold expression levels of IFN-β mRNA over 18S rRNA levels of three independent biological samples each measured in technical triplicate experiments ± are depicted. (E) In vitro-transcribed RNA was transfected into RIG-I^+/+, RIG^−/−, or reconstituted RIG^−/− MEF. Average n-fold expression levels of IFN-β mRNA over 18S rRNA levels of three independent biological samples each measured in technical triplicate experiments ± SD are depicted. Significance is indicated in reference to no-enzyme control samples. (F) Enrichment of marker mRNA from Salmonella-infected A549 cells (MOI of 10, 8 h) after pull down with RIG-I-specific antibody over pull down with control IgG antibody. Equal amounts of pulldown RNA were reverse transcribed using gene-specific primers for marker mRNA. Average n-fold levels over control IgG pull down levels from three independent biological samples each measured in technical triplicate experiments ±SD are depicted.