An RNA-Binding Protein Secreted by a Bacterial Pathogen Modulates RIG-I Signaling

Abstract

RNA-binding proteins (RBPs) perform key cellular activities by controlling the function of bound RNAs. The widely held assumption that RBPs are strictly intracellular has been challenged by the discovery of secreted RBPs. However, extracellular RBPs have been described in eukaryotes, while secreted bacterial RBPs have not been reported. Here, we show that the bacterial pathogen Listeria monocytogenes secretes a small RBP that we named Zea. We show that Zea binds a subset of L. monocytogenes RNAs, causing their accumulation in the extracellular medium. Furthermore, during L. monocytogenes infection, Zea binds RIG-I, the non-self-RNA innate immunity sensor, potentiating interferon-β production. Mouse infection studies reveal that Zea affects L. monocytogenes virulence. Together, our results unveil that bacterial RNAs can be present extracellularly in association with RBPs, acting as "social RNAs" to trigger a host response during infection.

Keywords: bacteriophage A118; extracellular RNA; type I IFN.

Graphical abstract

Figure 1

Zea Is a Secreted Oligomeric Protein of L. monocytogenes (A) Syntheny analysis of the lmo2686/zea-containing genomic locus between L. monocytogenes and L. innocua. Arrows and stem and circle represent the transcriptional start sites (TSSs) and the transcriptional terminators, respectively. (B) Schematic representation and primary sequence of the Zea protein. The N-terminal signal peptide is highlighted in red. (C and D) Bacterial cytosol and culture medium from (C) L. monocytogenes WT and Δzea strains and from (D) WT and a FLAG-tagged Zea-overexpressing L. monocytogenes strain (zea^FLAG) were immunoblotted with the indicated antibodies (n = 2). (E) Ribbon diagram of hexameric Zea. (F) Electrostatic potential surface representation of hexameric Zea. (G) Immunoprecipitation (IP) of Zea with an anti-FLAG antibody from bacterial cytosol and culture medium from a L. monocytogenes strain co-overexpressing Zea^FLAG and Zea^HA (n = 2). Immunoblot of input and immunoprecipitated proteins were probed with an anti-FLAG and anti-HA antibodies. (H) ZeaFLAG elution profile from size exclusion gel chromatography (n = 2). (I) 280 nm (mAU) absorbance monitoring of a gel filtration profile of recombinant purified HisZea (green line; n = 2). The elution profile of protein markers is indicated with the orange line. Purified HisZea was analyzed by SDS-PAGE and Coomassie blue staining (top left-hand panel).

Figure 2

Zea Associates with RNA (A) Enrichment of Zea-bound RNAs (n) from bacterial cytosol and culture medium. Blue squares and red circles depict individual RNAs. The y axis shows the enrichment of the Zea-interacting RNAs relative to immunoprecipitation with IgG. (B) Expression of L. monocytogenes RNAs grown in BHI at stationary phase measured by tiling array compared with the enrichment of the Zea-bound RNAs. (C) Circular genome map of L. monocytogenes showing the position of the Zea-interacting RNAs. The first two circles from the inside show the genes encoded on the + (inner track) and – (outer track) strands, respectively. The positions of Zea-interacting small RNAs (rlis) are pointed at outside of the circular map. Dotted lines highlight the phage A118 locus. (D) Examples of normalized read coverage (reads per million) visualized by IGV from Zea and control (IgG) IP for a selection of phage A118 genes (blue arrows). Gene names marked in red show no significant enrichment in the Zea IP. (E) Heatmap showing the fold enrichment of phage A118 transcripts in the Zea IP compared to control IP in the bacterial cytosol and culture medium. (F) RIP-qPCR on RNAs isolated from Zea and control (IgG) immunoprecipitations in the bacterial cytosol (top) and culture medium (bottom). The enrichment of selected phage (lmo2282 to lmo2333) and control genes was calculated after normalization to the corresponding input fractions. Values represent means ± SEM, n = 3. †, not detected. Statistical significance (between the IP IgG and IP αZea) determined by two-tailed t test. See also Figure S2.

Figure 3

Zea Directly Binds RNA (A and B) Electrophoretic mobility gel shift assay with in vitro-transcribed 5′ end radiolabeled rli143 (n = 2) (A) and rli92 (n = 3) (B) in the presence of increasing concentration of HisZea, as indicated. (C and D) HisZea-rli143 (n = 2) (C) and HisZea-rli92 (n = 2) (D) complexes were incubated with increasing concentrations of the corresponding cold competitor RNA. (E) Immunoblotting (n = 3) of streptavidin affinity pull-down of in vitro-transcribed biotinylated transcripts in the presence of HisZea (left); quantification of Zea binding to rlis (right). Statistical significance determined by ANOVA with multiple testing against rli80.

Figure 4

Zea Controls the Abundance of Its Target RNAs in the Culture Medium (A–D) qPCR analysis on RNA extracted from the culture medium of different L. monocytogenes strains (as indicated) for (A) selected phage and control genes, (B) the lma-monocin locus, (C) rli143 in L. monocytogenes, and (D) rli143 in L. innocua. The relative abundance was calculated after normalization to the WT sample. Values represent means ± SEM, n = 3. †, not detected. Statistical significance determined by unpaired ANOVA with multiple testing against WT. See also Figures S4 and S5.

Figure 5

Zea Regulates L. monocytogenes Virulence (A and B) BALB/c mice were inoculated intravenously with L. monocytogenes EGD-e (WT) or the zea-deleted strain (Δzea). After 48 h and 72 h post-infection, livers (A) and spleens (B) were recovered and CFUs assessed by serial dilution and plating. The number of bacteria in each organ is expressed as log₁₀ CFUs. The lines denote the means ± SEM, n = 2. Statistical significance determined by two-tailed t test.

Figure 6

Zea Interacts with RIG-I and Modulates a RIG-I Dependent IFN Response (A) qPCR analysis of IFN-β, IFN-γ, and interleukin 8 (IL8) (n = 3) expression in response to infection with WT and zea⁺L. monocytogenes in LoVo cells (left); qPCR analysis of interleukin 8 (IL8) expression in response to infection with WT and Lmo1656⁺L. monocytogenes in LoVo cells infected as above (right). The relative expression was calculated after normalization to (1) the GAPDH as a housekeeping gene and (2) to the WT sample. †, not detected. Statistical significance determined by two-tailed t test. (B) qPCR analysis of IFN-β expression in response to infection with WT and zea⁺L. monocytogenes in LoVo cells transfected with control siRNA (ctrl siRNA) or with RIG-I targeting siRNA (RIG-I siRNA) and infected as above (n = 3). Values represent means ± SEM. Statistical significance determined by two-tailed t test. (C) Representative confocal images of LoVo cells transfected with FLAG-tagged Zea (top) or FLAG-tagged mCherry (bottom). The co-localization between Zea and RIG-I was assessed with a line scan (white line) whose fluorescence intensity is plotted in red for ZeaFLAG and in green for RIG-I. Top right insets: magnification of the region in which the line scan was performed. Scale bars, 10 μm. (D) Representative coIP between FLAG-tagged Zea and Strep-tagged RIG-I (left, n = 2). Immunopurified ZeaFLAG treated (+ RNaseA) or not (−RNaseA) with RNase was incubated with a cell lysate from HEK293 cells stably expressing Strep-tagged RIG-I (Sanchez David et al., 2016). Quantification of Zea-bound RIG-I in presence or absence of RNase (right). Statistical significance determined by two-tailed t test. (E) Representative coIP between FLAG-tagged Zea and Strep-tagged RIG-I (n = 2). LoVo cells were co-transfected with the plasmids encoding FLAG-tagged Zea and Strep-tagged RIG-I and ZeaFLAG was then immunoprecipitated and treated (+ RNaseA) or not (−RNaseA) with RNase before elution with an anti-FLAG peptide. (F) The immunostimulatory activity of Zea-interacting small RNAs was assessed by transfection into ISRE reporter cells lines (Lucas-Hourani et al., 2013). Values represent means ± SEM, n = 3. Firefly luciferase activity was normalized to mock-transfected cells. HMW (high molecular weight), LMW (low molecular weight), and 5′3P (5′ triphosphate-RNA) were used as positive controls. An mCherry RNA fragment served as a negative control. Statistical significance determined by two-tailed t test. (G) The immunostimulatory activity of the Zea protein was assessed by transfection of a Zea-encoding plasmid (zea) into the ISRE reporter cells line (Lucas-Hourani et al., 2013). Values represent means ± SEM, n = 3. Statistical significance determined by two-tailed t test. See also Figures S6 and S7.

Figure 7

Proposed Model for Zea-Mediated Regulation of IFN Response (A and B) During infection, L. monocytogenes secretes Zea and RNA into (A) the cytoplasm, which then assemble to form (B) a ribonucleoprotein complex. (A′) In alternative, Zea can be directly secreted in an RNA-bound form. Zea-containing ribonucleoprotein complexes then associate to RIG-I, triggering a signaling cascade which would then activate the type I IFN response.