Endonucleolytic RNA cleavage drives changes in gene expression during the innate immune response

Figures

• 

  Figure 1.. RNase L activation leads to transcriptional changes through ZAK, JNK and p38. A Schematics of RNA-seq experiments. A549 cells were transfected with 2–5A and RNA-seq was performed as described in Materials and Methods. B Volcano plot showing upregulation of example proinflammatory cytokines and transcription factors during 2–5A treatment. Differentially expressed genes define as padjusted value <0.05, log2fold change >1. C Gene Ontology analysis of molecular functions shows that upregulated genes are enriched in genes related to transcription activities. 5 out of 9 categories with the highest Normalized Enrichment Scores are shown where padjusted <0.05. Categories not shown had highly overlapping gene sets with those that are highlighted. D Western blots show that transcription factors FOSB and JUN are increased due to 2–5A activation. Different FOSB isoforms are marked with black arrows. E Western blots show JNK and p38 are activated in A549 cells after 2–5A treatment. Different JNK isoforms are marked with black arrows. ZAKα is hyperphosphorylated during RNase L activation. Upper bar shows phosphorylated ZAKα in a Phos-Tag gel (shifted from its expected molecular weight). Lower…

• 

  Figure 2.. OAS-RNase L and other dsRNA pathways act together to induce the proinflammatory response. A Schematics of RNA-seq experiments. A549 cells were transfected with poly I:C (leading to activation of MDA5, RIG-I, and PKR, besides OAS). B Venn-diagram showing the overlap of DE genes in cells treated with poly I:C in WT (magenta), 2–5A in WT (yellow) and poly I:C in RNASEL KO cells (green, abbreviated to KO). C Correlation between differential expression changes in 2–5A (activation of RNase L) vs poly I:C (activation of all dsRNA pathways including RNase L) treated cells. Blue and red dots represent JNK/p38 MAPK activation-related or IFN stimulated genes (as defined in Materials and Methods), differentially expressed in 2–5A or poly I:C treated cells, respectively. Black dot represents GADD34. D Violin plot representation of genes related to JNK and p38 MAPK activation and interferon stimulation. Statistical testing (paired t-tests) showed significant differences between data sets as indicated (p value <0.001, marked by ***). E Effects of activating the OAS-RNase L pathway on phosphorylation of JNK, p38, eIF2α and IRF3 assayed by Western blotting. ZAKα is hyperphosphorylated during…

• 

  Figure 3.. Global RNA cleavage in the cell leads to transcriptional changes. A Schematics of RNA-seq experiments. RNASEL KO A549 cells were electroporated with RNase A or BSA (control) and RNA-seq experiments were performed as described in Materials and Methods. B Venn-diagram showing the number of upregulated genes (defined as padjusted value <0.05 and log2fold changes >1) in RNase A electroporated samples (purple) and 2–5A treated cells (yellow). C Comparison of RNA log2 fold changes between RNase L activated vs RNase A electroporated samples. Blue dots are representing transcripts related to JNK and p38 activation. D Violin plot representation of those genes related to JNK and p38 activation (N = 54, as defined in Materials and Methods) in RNase A electroporated cells. RNASEL KO is abbreviated to KO. Statistical analysis (paired t-test) showed that the log2 fold changes between RNase A electroporated and 2–5A activated samples were not significantly different (padjusted value >0.05, marked by n.s.). E JNK, p38 and ZAKα are activated due to RNase A treatment, as shown by western blots. Different JNK isoforms are marked with black arrows. Upper bar shows phosphorylated ZAKα in a Phos-Tag…

• 

  Figure 4.. RNA fragmentation leads to alternative ORF (altORF) translation. A Schematics of ribosome profiling experiments. RNASEL KO A549 cells were electroporated with RNase A or BSA (control) and ribosome profiling experiments were performed as described in Materials and Methods. B Ribosome profiling reads in the 3’UTR of the ACTG1 gene model shows the signature of altORF translation in RNase A electroporated and RNase L activated A549 cells, but not in the controls (−2–5A and BSA). C Increased 3’UTR:ORF ratios indicate higher fraction of translation in 3’UTRs when active RNases are present in the cell, but not in controls (−2–5A and BSA). D Normalized average ribosome footprint occupancy (metagene plot) around the stop codon of main ORFs (left panel) reveals increased proportion of ribosome footprints in the 3’ UTRs (right panel, zoom in). E Increased 3’UTR:ORF ratios indicate altORF translation during thapsigargin treatment and activation of the UPR, but not in untreated cells. In all panels, data shown for RNase L activation (+2–5A) were obtained from 21.

• 

  Figure 5.. RNase L activation influences translational control and effects of poly I:C treatment. A. Comparison of changes in RNA-seq vs ribosome profiling under conditions of 2–5A treated WT cells (left), poly I:C treated WT cells (center), and poly I:C treated RNASEL KO A549 cells (right). To emphasize the highest and most significant changes, genes were filtered for the following criteria: log2 fold changes >4 (either RNA-seq or ribosome profiling), padjusted <0.01, normalized mean counts across the samples >50 (see Materials and Methods for details). Dotted diagonal lines represent TE changes >2 fold. Orange circles represent genes encoded by the mitochondria. B. TEs from individual replicates of representative genes with downregulated (MUC5AC and MUC5B) or upregulated (CXCL3 or CXCL8) translation were computed. TEs calculated as described in Materials and Methods. RNASEL KO is abbreviated to KO in all panels.

• 

  Figure 6.. RNase L modulates elements of the Integrated Stress Response. A Bar plots representing TEs of GADD34 mRNA during poly I:C treatment in WT and RNASEL KO cells. These data show that TE changes are greater in RNASEL KO cells during the ISR as compared to respective controls. B Ribosome profiling tracks of gene models for GADD34 (PPP1R15A) in WT and RNASEL KO cells during poly I:C treatment suggest a change in 5’UTR vs main ORF translation ratio in RNASEL KO cells during the ISR as compared to WT cells and their respective controls. C 5’UTR:main ORF ratios of ribosome profiling data in WT and RNASEL KO cells with and without poly I:C treatment were computed for individual replicates. Data show this value is more strongly reduced upon poly I:C treatment in cells where RNase L is absent. D. Western blot for GADD34 shows increased levels of protein in poly I:C treated RNASEL KO cells, reflecting both the lack of mRNA degradation and enhanced TE in cells where RNase L is absent.

• 

  Figure 7.. IFIT mRNAs undergo changes in translation that depend on RNase L. A uORFs were identified on IFIT1 mRNAs based on the increase in ribosomes density at alternative start codons (CUG, green markings) in poly I:C treated cells. Red marking shows the start codon of the main ORF. B TEs of IFIT1–3 mRNAs are decreased in RNase L deficient cells (comparing WT and RNASEL KO poly I:C treated cells) consistently across replicates. C IFIT1–3 mRNAs were found to have less 5’UTR relative to main ORF translation in poly I:C treated RNASEL KO (abbreviated to KO) cells as compared to WT cells. D Schematics of how eIF2α phosphorylation and activation of RNase L affects gene expression in the cell. While both mechanisms lead to reduced gene expression (less translation when eIF2α is phosphorylated and less overall mRNA when RNase L is activated), eIF2α phosphorylation increases translation of GADD34 but RNase L reduces its translation. Similarly, RNase L reduces IFIT mRNA translation. It may also be increased via uORF regulation under conditions of eIF2α phosphorylation (marked by asterisk since further confirmation is needed).