RNase L limits host and viral protein synthesis via inhibition of mRNA export

Abstract

RNase L is widely thought to limit viral protein synthesis by cleaving host rRNA and viral mRNA, resulting in translation arrest and viral mRNA degradation. Here, we show that the mRNAs of dengue virus and influenza A virus largely escape RNase L-mediated mRNA decay, and this permits viral protein production. However, activation of RNase L arrests nuclear mRNA export, which strongly inhibits influenza A virus protein synthesis and reduces cytokine production. The heterogeneous and temporal nature of the mRNA export block in individual cells permits sufficient production of antiviral cytokines from transcriptionally induced host mRNAs. This defines RNase L-mediated arrest of mRNA export as a key antiviral shutoff and cytokine regulatory pathway.

Fig. 1. DENV mRNAs escape RNase L–mediated mRNA decay and are translated.

(A) smFISH for DENV mRNA and GAPDH mRNA and immunofluorescence for DENV NS3 protein in WT and RL-KO A549 cells 48 hours after infection with DENV (MOI = 0.1). Images are from a single Z plane. Orange arrow: RNase L not activated (GAPDH mRNA not degraded); green arrow: RNase L activated (GAPDH mRNA degraded); blue arrow: uninfected cells (neither DENV mRNA nor NS3 detected). (B) Quantification of GAPDH mRNA in WT and RL-KO cells. WT cells in which GAPDH mRNA levels were lower than the lowest level observed in RL-KO cells were designated as RNase L active (RL+). WT cells with GAPDH mRNA levels above that threshold were designated RNase L inactive (RL–). (C) Quantification of DENV mRNA foci. (D) Scatterplot of DENV mRNA foci and DENV NS3 immunofluorescence. (E) Similar to (B) and (C) but quantifying the integrated density (a.u., arbitrary units) of DENV NS3 immunofluorescence as represented in (A).

Fig. 2. RNase L inhibits nuclear export of interferon (IFN) mRNAs in response to DENV infection.

(A) smFISH for IFN-λ1 mRNA, GAPDH mRNA, and DENV mRNA in WT and RL-KO A549 cells 48 hours after infection with DENV (MOI = 0.1). Orange arrow: DENV-positive but RNase L not activated (GAPDH mRNA not degraded); green arrow: DENV positive, RNase L activated (GAPDH mRNA degraded), and IFN-λ1 mRNA localized to cytoplasm; white arrow: DENV positive, RNase L activated, and IFN-λ1 mRNA retained in the nucleus. (B) Scatterplot of GAPDH mRNA and IFN-λ1 mRNA in individual cells. (C) Quantification of smFISH foci in WT and RL-KO cells infected with DENV and that induced IFN-λ1 mRNA expression. (D) Scatterplot of the number of smFISH foci of IFN-λ1 mRNA localized to the nucleus or cytoplasm in individual WT and RL-KO cells as represented in (A).

Fig. 3. RNase L inhibits nuclear export of IFN mRNAs in response to dsRNA.

(A) smFISH for IFN-β and GAPDH mRNAs in WT and RL-KO A549 cells 8 hours after poly(I:C) lipofection. Solid lines mark the cell boundary of WT cells since it is difficult to view cell boundaries with nuclear-retained IFN mRNA. (B) Quantification of IFN-β smFISH foci, as represented in (A), in individual cells (dots). (C) Scatterplot of the nuclear and cytoplasmic IFN-β foci in individual WT and RL-KO cells, as represented in (A). (D) smFISH for IFN-β and GAPDH mRNAs in RL-KO cells expressing RNase L or RNase L-R667A catalytic mutant via lentiviral transduction after poly(I:C) lipofection. (E) Immunoblot of whole cell (W), nuclear (N), and cytoplasmic (C) fractions from WT and RL-KO cells 6 hours after poly(I:C) used for RT-qPCR. (F) qRT-PCR quantification of IFN-β mRNA from nuclear-cytoplasmic fractions as shown in (E). Bars represent the average 2^−ΔCt (nuclear-cytoplasm) ± SD (n = 4). Dots represent independent experiments. The P value was determined using Student’s t test. (G) Immunofluorescence for NPC and smFISH for IFN-β mRNA in a WT cell displaying nuclear retention of IFN-β mRNA.

Fig. 4. RNase L–mediated inhibition of mRNA export inhibits influenza virus protein synthesis.

(A) smFISH for IAV NA and NS1 mRNAs 7 hours after infection (MOI = 0.5). Cells were transfected with poly(I:C) 1 hour after infection. GAPDH mRNA degradation is a marker for RNase L activation. Additional images are shown in fig. S5 (B and C). White arrows, nuclear retention of NA and NS1; blue arrows, nuclear retention of NA but not NS1. (B) Scatterplot of nuclear intensity and cytoplasmic intensity of NA or NS1 mRNAs from individual WT cells with active RNase L (RL+; loss of GAPDH mRNA) or inactive RNase L (RL−; contain GAPDH mRNA) and RL-KO cells. (C) Similar to (A) but smFISH for NS1 mRNA and GAPDH mRNA and immunofluorescence staining of NS1 protein. Green arrow: RL−, cytoplasmic NS1 mRNA+; orange arrow: RL+, cytoplasmic NS1 mRNA+; white arrow: RL+, nuclear NS1 mRNA retention+. (D) Cytoplasmic staining intensity of NS1 mRNA in RL− WT cells, RL+ WT cells displaying either NS1 mRNA nuclear retention (export block+) or cytoplasmic localization (export block−), and RL-KO cells. WT cells with export block were defined as having a nuclear-to-cytoplasmic 8/NS1 mRNA intensity ratio that was higher than the maximum ratio observed in RL-KO cells for each replicate (fig. S6C). (E) Similar to (D) but plotting immunofluorescent intensity of NS1 protein. Statistical significance (**P < 0.005) was determined by t test analysis.

Fig. 5. RNase L–mediated mRNA export block contributes to inhibition of antiviral cytokine production.

(A) ELISAs for IFN-β and IFN-λ1. Markers represent the mean SD from at least six replicates. Statistical significance (*P < 0.05; **P < 0.005; ***P < 0.0005) was determined by t test analysis. (B) WT A549 cells displaying varying degrees of nuclear retention of IFN-β or IFN-λ1 mRNA after poly(I:C). The number and percentage of smFISH foci localized to the nucleus are shown in the top right corners. Nuclei were determined by DAPI staining (not shown for space and clarity) and are outlined. Additional images are shown in figs. S8 and S9. (C) The percentage of cells displaying 70% or greater IFN-β or IFN-λ1 smFISH foci in the cytoplasm derived from fig. S9D. (D) RT-qPCR for IFN-β mRNA in WT, RL-KO, PKR-KO, and RL/PKR-dKO cell lines 6 hours after poly(I:C). Bars represent the average cycle threshold (Ct) ± SD. Each dot represents a Ct value from an individual experiment. (E) Quantification of IFN-β via ELISA 12 hours after poly(I:C). (F) Fold change in IFN-β secretion via ELISA 12 hours after poly(I:C) with or without cotreatment with ISRIB. (G) WT and RL-KO cells stained with DAPI to visualize apoptotic cells (white arrows). The graph below shows the mean percentage and SD of apoptotic cells. (H) Immunoblot analysis of PARP and caspase-mediated cleavage fragment of PARP (PARPc). The graph below shows the mean ± SD of PARPc/PARP ratio. n.s., not significant.