The secRNome of Listeria monocytogenes Harbors Small Noncoding RNAs That Are Potent Inducers of Beta Interferon

Abstract

Cellular sensing of bacterial RNA is increasingly recognized as a determinant of host-pathogen interactions. The intracellular pathogen Listeria monocytogenes induces high levels of type I interferons (alpha/beta interferons [IFN-α/β]) to create a growth-permissive microenvironment during infection. We previously demonstrated that RNAs secreted by L. monocytogenes (comprising the secRNome) are potent inducers of IFN-β. We determined the composition and diversity of the members of the secRNome and found that they are uniquely enriched for noncoding small RNAs (sRNAs). Testing of individual sRNAs for their ability to induce IFN revealed several sRNAs with this property. We examined ril32, an intracellularly expressed sRNA that is highly conserved for the species L. monocytogenes and that was the most potent inducer of IFN-β expression of all the sRNAs tested in this study, in more detail. The rli32-induced IFN-β response is RIG-I (retinoic acid inducible gene I) dependent, and cells primed with rli32 inhibit influenza virus replication. We determined the rli32 motif required for IFN induction. rli32 overproduction promotes intracellular bacterial growth, and a mutant lacking rli32 is restricted for intracellular growth in macrophages. rli32-overproducing bacteria are resistant to H₂O₂ and exhibit both increased catalase activity and changes in the cell envelope. Comparative transcriptome sequencing (RNA-Seq) analysis indicated that ril32 regulates expression of the lhrC locus, previously shown to be involved in cell envelope stress. Inhibition of IFN-β signaling by ruxolitinib reduced rli32-dependent intracellular bacterial growth, indicating a link between induction of the interferon system and bacterial physiology. rli32 is, to the best of our knowledge, the first secreted individual bacterial sRNA known to trigger the induction of the type I IFN response.IMPORTANCE Interferons are potent and broadly acting cytokines that stimulate cellular responses to nucleic acids of unusual structures or locations. While protective when induced following viral infections, the induction of interferons is detrimental to the host during L. monocytogenes infection. Here, we identify specific sRNAs, secreted by the bacterium, with the capacity to induce type I IFN. Further analysis of the most potent sRNA, rli32, links the ability to induce RIG-I-dependent induction of the type I IFN response to the intracellular growth properties of the bacterium. Our findings emphasize the significance of released RNA for Listeria infection and shed light on a compartmental strategy used by an intracellular pathogen to modulate host responses to its advantage.

Keywords: Listeria monocytogenes; secreted RNA; type I IFN.

FIG 1

secRNome of L. monocytogenes. (A) The secRNome of L. monocytogenes comprises two subcompartments: “naked” RNA (sec-RNA) and RNA associated with membrane vesicles (MVs), shed in the medium by bacteria. Cytosolic RNA pertains to the RNA inside the bacterial cell. Lm, wild-type strain. (B) IFN-β induction by sec-RNA and MV-RNA. Data represent levels of induction of IFN-β expression upon transfection of sec-RNA, MV-RNA, and cytosolic RNA in BMDM (100 ng/10⁶ cells). Transfection of sec-RNA and MV-RNA highly induced IFN-β expression (70-fold and 20-fold, respectively) compared to cytosolic RNA. Treatment of the three RNA fractions with CIAP (calf intestinal alkaline phosphatase) significantly reduced IFN-β expression levels in sec-RNA and MV-RNA. Cells treated with a transfection reagent (Lipofectamine) only were used as the control. Data are presented as means ± standard deviations (SD) of results from three experiments (ns, nonsignificant; *, P < 0.05; **, P < 0.01). T.R., transfection reagent only; None, untransfected. (C) The transcriptome of the secRNome. Data represent percentages of total reads of cytosolic RNA, MV-RNA, and sec-RNA mapped to the L. monocytogenes genome. Data represent results from three independent experiments (experiments I, II, and III). (D) Heat map of 30 sRNAs differentially expressed in sec-RNA and MV-RNA compared to cytosolic RNA of L. monocytogenes. The 20 sRNA candidates selected for in vitro transcription are indicated with arrows. Data represent means of results from three independent experiments.

FIG 2

Individual transfection of the IVT sRNAs in BMDM cells revealed rli32, rli48, rli99, and rli100 as potent inducers of IFN-β. (A) IFN-β induction in BMDM cells by sRNAs detected in sec-RNA and MV-RNA. BMDM cells were transfected with different IVT sRNA molecules (50 ng/10⁶ cells). The IFN-β induction was assayed 6 h posttransfection by qRT-PCR. Data are presented as means ± SD of results from three experiments. Cells treated with the transfection reagent (Lipofectamine) only were used as the control. The results identified candidate motifs in rli32 essential for IFN-β induction. (B) Folding prediction of three candidate motifs of rli32 by the use of RNAfold ViennaRNA software. (C) The three motifs were transcribed in vitro and individually transfected in BMDM. Motif 2 showed the highest level of IFN-β induction, that level is similar to the level seen with the complete rli32 sequence. Data are presented as means ± SD of results from three experiments (ns, nonsignificant; **, P < 0.01; ***, P < 0.001).

FIG 3

rli32-triggered IFN-β induction is RIG-I dependent. (A) Contribution of cytosolic receptors RIG-I and MDA5 to the recognition of rli32. Transfection of RIG-I^−/− and MDA5^−/− deficient HEK293 cells with rli32 revealed that the triggering of IFN-β expression by rli32 is RIG-I dependent. The absence of RIG-I abolished the expression of IFN-β almost completely. Treatment of rli32 with CIAP led to significantly reduced IFN-β expression after transfection in HEK293 cells. Cells treated with the transfection reagent (Lipofectamine) only were used as the control. WT, wild type. (B) Conformational switch of RIG-I. A549 cells transfected with ssrS, rli32, and 3pRNA, as a positive control, were lysed and treated with trypsin. Trypsin digestion of lysates from rli32-transfected and 3pRNA-transfected A549 cells resulted in the emergence of a protease-resistant RIG-I fragment, whereas the ssrS transfection resulted in rapid degradation of RIG-I. (C) Quantification of the differences in the band intensities of RIG-I following trypsin digestion depicted as fold changes. (D) Dose-dependent inhibition of influenza A virus replication by rli32. HEK293 cells were transfected with three different concentrations of rli32 and ssrS (10, 50, and 100 ng) 24 h prior to infection with A/PR/8/34 (H1N1).Transfection of 100 ng rli32 diminished the virus titer by as much as 95%, while transfection of only 10 ng resulted in a reduction of 75%. On the other hand, the pretreatment with the weak IFN-β inducer ssrS did not inhibit virus proliferation. T.R., transfection reagent; Cont., untreated control). (E) Contribution of the cytosolic receptors RIG-I and MDA5 to the rli32-triggered antiviral response. rli32 was transfected into RIG-I^−/− and MDA5^−/− deficient HEK293 cells, and the respective supernatants were used for virus titer assay. The virus titer was unaffected in HEK293 cells deficient for RIG-I, whereas the pretreatment of MDA5^−/− deficient HEK293 cells with rli32 led to a strong decrease of the virus titer. Data are presented as means ± SD of results from three experiments (ns, nonsignificant; *, P < 0.05; **, P < 0.01; ***, P < 0.001).

FIG 4

Impact of rli32 deletion and overproduction on IFN-β induction. (A) sec-RNA, MV-RNA, and cytosolic RNA were isolated from wild-type L. monocytogenes, strain Lm-rli32 (L. monocytogenes rli32-overproducing strain), and strain Lm-Δrli32 (L. monocytogenes Δrli32 deletion mutant). Transfection of the three RNA fractions (sec-RNA, MV-RNA, and cytosolic RNA) isolated from Lm-rli32 induced a higher IFN-β response in macrophages than those from L. monocytogenes and Lm-Δrli32. SecRNA and MV-RNA isolated from the strain lacking rli32 showed a significant decrease in the IFN-β response. Cells treated with the transfection reagent (Lipofectamine) only were used as the control. (B) Addition of membrane vesicles (MVs) isolated from the supernatant of the Lm-rli32 strain to macrophages showed that those MVs were more potent in inducing an IFN-β response than the MVs from the wild-type strain or strain Lm-Δrli32. Data are presented as means ± SD of results from three experiments (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

FIG 5

Rli32 affects the intracellular growth of L. monocytogenes. Macrophages were infected with the parental strain (Lm), rli32-overproducing strain (Lm-rli32), and Δrli32 deletion mutant (Lm-Δrli32). (A) The number of intracellularly grown bacteria (quantified as CFUs) was counted on agar plates following lysis of the macrophages at 4, 8, and 24 h p.i. The data from time points 8 and 24 h p.i. revealed increased growth of Lm-rli32 over that of the wild type. The level of growth of Lm-Δrli32 was reduced compared to that of the parental strain at 24 h p.i. Data are presented as means ± SD of results from three experiments. (B) Amounts of rli32 transcripts in the cytosol of macrophages infected with the Lm, Lm-rli32, and Lm-Δrli32 strains. The cytosolic host RNA was isolated 8 h p.i. and reverse transcribed in cDNA, followed by amplification using qRT-PCR. The bands were separated on an agarose gel. As the control, the amount of the host reference gene Rplp in the cytosolic host RNA was determined. (C and D) Induction and repression of IFN-β response during infection with strains Lm-rli32 and Lm-Δrli32. Expression of IFN-β in macrophages infected with strains Lm, Lm-rli32, and Lm-Δrli32 was determined at 8 h p.i. (C) and 24 h p.i. (D). Data are presented as means ± SD of results from three experiments (ns, nonsignificant; *, P < 0.05; **, P < 0.01; ***, P < 0.001). (E to G) Heat map representing ratios of differences in the compositions of sRNAs in (E) sec-RNA, (F) MV-RNA, and (G) cytosolic RNA in strains Lm-rli32 and Lm-Δrli32 in comparison to strain Lm.

FIG 6

Physiological consequences of rli32 for the growth of L. monocytogenes. Bacterial growth in the presence of cefuroxime and H₂O₂ was assayed. (A and B) Impact of (A) cefuroxime (Cef) (4 μg/ml) and (B) H₂O₂ (0.15%) on the growth of strains Lm, Lm-Δrli32, and Lm-rli32. (C) Impact of the IFN signaling inhibitor ruxolitinib (Rux) (1 μM and 10 μM) on the growth (24 h p.i.) of strains Lm, Lm-rli32, and Lm-Δrli32 in macrophages. Data are presented as means ± SD of results from three experiments (ns, nonsignificant; **, P < 0.01).