The RNA Exosome Syncs IAV-RNAPII Transcription to Promote Viral Ribogenesis and Infectivity

Abstract

The nuclear RNA exosome is an essential multi-subunit complex that controls RNA homeostasis. Congenital mutations in RNA exosome genes are associated with neurodegenerative diseases. Little is known about the role of the RNA exosome in the cellular response to pathogens. Here, using NGS and human and mouse genetics, we show that influenza A virus (IAV) ribogenesis and growth are suppressed by impaired RNA exosome activity. Mechanistically, the nuclear RNA exosome coordinates the initial steps of viral transcription with RNAPII at host promoters. The viral polymerase complex co-opts the nuclear RNA exosome complex and cellular RNAs en route to 3' end degradation. Exosome deficiency uncouples chromatin targeting of the viral polymerase complex and the formation of cellular:viral RNA hybrids, which are essential RNA intermediates that license transcription of antisense genomic viral RNAs. Our results suggest that evolutionary arms races have shaped the cellular RNA quality control machinery.

Keywords: Influenza virus polymerase; RNA chimeras; RNA exosome; RNA hybrids; RNA surveillance; RNAPII elongation; epigenetics; host-pathogen interactions; neurodegeneration; non-coding RNA.

Figure 1.. Proteomics of Influenza Polymerase Subunit PA Reveal Interactions with RNA Exosome

(A) IAV polymerase acidic protein (PA) from (left) H1N1 (California 2009) or (right) H3N2 (Wyoming 2003) with C-terminal 2xStrep tags were expressed in A549 cells and subjected to affinity purification and mass spectrometry (AP-MS) in biological triplicate with and without interferon pre-treatment. Human proteins (prey) significantly enriched after PA purification (orange - bait) relative to vector and GFP-2xStrep control samples are depicted (yellow - without interferon, blue – with interferon, green – both with and without interferon). Known exosome interacting partners are boxed in blue. (B) Heatmap depicting the relative protein abundance of bait proteins and exosome subunits across biological triplicates (without interferon) in the AP-MS samples. (C) Significantly enriched protein complexes from the CORUM database are annotated with pertinent complex members circled and labeled in red in the above networks. (D) A549 cells were infected with a PA-FLAG-tagged IAV for 10 hours. Endogenous EXOSC10 was immunoprecipitated from cell lysate. Co-immunoprecipitated proteins were detected by immunoblot. See also Figure S1 and Table S1

Figure 2.. The RNA Exosome Enhances Influenza Polymerase Activity and Viral Biogenesis

(A-C) qPCR of Influenza viral mRNA (A), RNA (B), and host infection-induced gene mRNA levels (C) in human A549 cells infected with the A/Puerto Rico/8/1934 (PR8) H1N1 strain at the indicated time points. Cells were transfected with siRNAs targeting EXOSC3 or DIS3, a non-targeting siRNA (siCTRL), or left un-transfected (No si). (D) mRNA expression levels for the targets of the siRNAs used above. (E) IAV minigenome replicon assay in A549 cells transfected with siRNAs as used above. The viral polymerase transcribed firefly luciferase levels were normalized to co-transfected renilla luciferase levels. (F) Renilla luciferase expression in A549 cells infected with the IAV PR8-luc at the indicated time points post-infection (single cycle growth – left; multi cycle growth – right). Cells were transfected with siRNAs as used above. (G) Viral replication growth curves (single cycle – top; multi cycle – bottom) assayed from supernatants harvested from A549 cells infected with PR8 at the indicated time points post-infection. A549 cells were first transfected with siRNAs as used above. Statistical analyses between datasets were performed with a two-tailed Student’s t-test, adjusted with a Holm-Bonferroni test for sequential comparisons. For all panels, *p<.05, **p<.005, and ***p<.0005. Error bars indicate SD from triplicate experiments. See also Figure S2

Figure 3.. Patient-Derived Cells with EXOSC3 Mutation (Asp132Ala) Suppress Viral Ribogenesis and Growth

(A-C) qPCR of IAV viral mRNA (A), RNA (B), and host infection-induced gene mRNA levels (C) in primary dermal fibroblasts isolated from patients with pontocerebellar hypoplasia type I (affected; homozygous mutation – Asp132Ala) or family members (carrier; heterozygous mutation). Cells were infected with the PR8 strain at the indicated time points. (D) mRNA expression levels for EXOSC3. (E) IAV minigenome replicon assay in patient primary dermal fibroblasts. Viral polymerase transcribed firefly luciferase levels were normalized to co-transfected renilla luciferase levels. (F) Renilla luciferase expression in patient primary dermal fibroblasts infected with the IAV PR8-luc at the indicated time points post-infection (single cycle growth – left; multi cycle growth – right). (G) Viral replication growth curve assayed from supernatants harvested from patient primary dermal infected with IAV PR8 at the indicated time points post-infection. Statistical analyses between datasets were performed with a two-tailed Student’s t-test, adjusted with a Holm-Bonferroni test for sequential comparisons. For all panels, *p<.05, **p<.005, and ***p<.0005. Error bars indicate SD from triplicate experiments. See also Figure S3

Figure 4.. Conditional ablation of Exosc3 attenuates Viral Polymerase Activity

(A) Splenic B cells were isolated and expanded from two mouse types: COIN/+ (conditional heterozygous for Exosc3 depletion), and COIN/COIN (conditional homozygous for Exosc3 depletion). Cells were treated with tamoxifen (+TAM) or DMSO (−TAM) for 3 days to induce the deletion of Exosc3 and inversion of GFP. Top: Fluorescence-activated cell sorting (FACS) displaying GFP conversion with and without tamoxifen treatment at 6 hours post-infection. Bottom: FACS displaying GFP conversion against population scatter. (B) qPCR of IAV viral mRNA levels in B cells infected with PR8 before (mock) or 6 hours post-infection (infected), along with the control of mRNA expression level for the efficiency of tamoxifen-induced depletion of Exosc3. Statistical analyses (B) between datasets were performed with a two-tailed Student’s t-test, adjusted with a Holm-Bonferroni test for sequential comparisons. For panel B, *p<.05, **p<.005, and ***p<.0005. Error bars indicate SD from duplicate experiments. (C) Heat map of average expression changes of positive and negative-stranded viral RNAs in murine splenic B cells in the presence or absence of tamoxifen, and at 6 hours post-infection. Expression changes (left) reflect the log2 fold-change in expression between homozygous (COIN/COIN) and heterozygous (COIN/+) Exosc3 depletion conditions. Corresponding eBayes adjusted and FDR-corrected P-values are shown on the right. Levels of positive-sense viral mRNA are overall an order of magnitude lower than negative-sense viral RNA in this directional RNA-Seq experiment (regardless of the TAM status). It is conceivable that this moderated the extent of the changes we see in vmRNA compared to vRNA levels. (D) Heat map of average expression changes of differentially expressed cellular genes (FDR q<0.05) between COIN/+ and COIN/COIN in the presence or absence of tamoxifen, and at 6 hours post-infection. (E) Top-enriched GO biological process categories among the differentially regulated genes from D - key to right. (F) Heat map of average expression changes for differentially expressed genes in E annotated with the “Immune response” GO biological process term. See also Figure S4 and Table S2

Figure 5.. Synthesis of host:viral chimeric RNAs is dependent on the Nuclear RNA Exosome

(A-B) qPCR of host mRNA and IAV mRNA (A), and RNA levels (B) in A549 cells infected with PR8 at the indicated time points. Cells were transfected with a siRNA targeting EXOSC3 or a non-targeting siRNA (siCTRL) and treated, at the time of infection, with cycloheximide (CHX) or DMSO. (C) Schematic of how the synthesis of each IAV mRNA requires the formation of a host:viral RNA hybrid followed by polymerization to generate the (+) strand from the viral (−) strand template. The plus strand (viral mRNA) is then a chimeric RNA formed by a 9-16mer of cellular RNA (see panel D for size distribution). (D) Length distribution of cellular transcript fragments (snatched RNA) found at the 5’ ends of viral mRNA in PR8 infected A549 cells at 4 hours post-infection. (E) Heat map of average expression changes of positive-stranded host:viral chimeric mRNAs in A549 cells transfected with siRNAs targeting DIS3 or EXOSC10, or a non-targeting siRNA (siCTRL). Cells were infected with PR8 for 4 hours and expression changes reflect the fold-change ratio between either siDIS3 or siEXOSC10 and siCTRL conditions. (F) Box plots of the log2 fold-change in expression for different transcript biotypes in uninfected A549 cells transfected with siRNAs as used in E, compared to siCTRL. PROMPTs were defined as transcripts identified between 1 and 2.5 kb upstream of annotated transcription start sites (TSS). Antisense TSS transcripts were identified in a 500 bp region upstream of annotated TSS features. (G) Left: Cap-snatch per-million events across cellular gene categories in A549 cells (from panel E). Duplicates are shown. Right: Proportional distribution of cap-snatch origins according to cellular transcript biotype. (H) Top: RNA antisense purification (RAP) experimental strategy for IAV NS1 segment purification. Bottom: RAP-qPCR of host:influenza virus chimeric mRNA levels in A549 cells infected with PR8 before (mock) or 6 hours post-infection (infected). Cells were transfected with a siRNA targeting EXOSC3 or a non-targeting siRNA (siCTRL). At time of collection, cells were cross-linked (+) or left native (−). Primers spanning the internal region of the viral segment and the 5’ or internal portion of the indicated PROMPTs were used to amplify chimeric RNAs. Statistical analyses between datasets were performed with a two-tailed Student’s t-test, adjusted with a Holm-Bonferroni test for sequential comparisons. For panels A-B, H, *p<.05, **p<.005, and ***p<.0005. Error bars indicate SD from triplicate (A-B) and duplicate experiments (D-H). See also Figure S5

Figure 6.. Models for the co-transcriptional interference of influenza virus with RNAPII and the RNA Exosome

(A) Model displaying IAV RNA and mRNA kinetics over the time of an infection in exosome proficient (gray) and deficient (blue) cells. (B) IAV cap-snatching model in exosome proficient cells. IAV polymerase (vPOL) accesses host ncRNA and mRNA in kinetic competition with RNAPII- and RNA exosome-dependent capbinding complex (CBC) dynamics, which control productive elongation of mRNA and co-transcriptional maturation at ncRNA loci. (C) IAV cap-snatching model in exosome deficient cells. IAV polymerase (vPOL) access to host ncRNA and mRNA is impaired because of reduced targeting of the viral polymerase to promoters and by defects in promoter-proximal RNAPII activity. This results in stabilization of ncRNA and TSS-RNA and a concomitant reduction of chimeric transcripts. See also Figure S6