Influenza A virus reassigns its own polymerase toward genome synthesis using a small RNA copied from its genome termini
lab-led
Influenza A virus-generated small RNAs regulate the switch from transcription to replication
lab-ledDiscovery
Influenza A virus produces a family of 22 to 27 nucleotide small viral RNAs corresponding to the 5 prime end of each genomic segment, which accumulate as the polymerase shifts toward genome synthesis and whose inhibition selectively depletes genomic RNA without comparably affecting messenger or complementary RNA.
- Senior authors
- Benjamin R. tenOever
- Correspondence
- Benjamin R. tenOever
Research areas & themes
Influenza Genome Regulation and ReplicationSmall Viral RNAs and the Transcription to Replication Switch
Citation
Perez JT, Varble A, Sachidanandam R, Zlatev I, Manoharan M, García-Sastre A, tenOever BR. Influenza A virus-generated small RNAs regulate the switch from transcription to replication. Proceedings of the National Academy of Sciences. 2010. Volume 107, issue 25, pages 11525-11530.
DOI 10.1073/pnas.1001984107. PMID 20534471. PMCID PMC2895093.
One-sentence contribution
Influenza A virus produces a family of 22 to 27 nucleotide small viral RNAs corresponding to the 5 prime end of each genomic segment, which accumulate as the polymerase shifts toward genome synthesis and whose inhibition selectively depletes genomic RNA without comparably affecting messenger or complementary RNA.
Executive summary
Influenza A virus uses a single RNA-dependent RNA polymerase for two incompatible jobs, primer-dependent messenger RNA synthesis early in infection and primer-independent genome synthesis later, and the molecular event that reassigns the enzyme between these modes had not been identified. Transcription requires the polymerase to remain bound in cis to the 5 prime end of the template, which is also what forces the stuttering that generates the polyadenylated tail, so genome synthesis appears to demand a way of reconstituting the double-stranded promoter without that steric constraint. Working from the premise that a small RNA could supply this function, the authors deep sequenced the sub-40 nucleotide RNA fraction of infected lung epithelial cells and found a discrete species, termed svRNA, matching the 5 prime terminus of each of the eight genomic segments. svRNA accumulates around twelve hours after infection, coincident with the measured rise in genomic RNA and after viral protein is already detectable. It is produced by multiple subtypes, in eggs and in cells from three host species, and it is not induced by an unrelated virus or by type I interferon. Reconstitution from plasmids showed a requirement for the three polymerase subunits, nucleoprotein and segment 8, with NEP/NS2 partially restoring production. svRNA co-precipitates only with the assembled polymerase. Locked nucleic acid inhibitors directed at one segment reduced that segment's genomic RNA and viral titres while leaving the other segments intact.
Scientific context
The influenza A virus genome consists of eight negative-sense segments whose conserved 5 prime and 3 prime noncoding ends form a panhandle or corkscrew structure recognised by the polymerase. During transcription the polymerase stays associated with the 5 prime end of the template, and that association both primes elongation and produces the steric hindrance that makes the enzyme stutter over the uracil tract to generate a polyadenylated messenger RNA. Genome synthesis proceeds through a full-length complementary RNA intermediate and is primer independent, which requires the polymerase to read through the 5 prime noncoding region rather than remain locked onto it. The paper states the resulting paradox directly, that the virus must keep the genomic ends associated throughout replication while also allowing complete synthesis across those ends, and that current models fail to reconcile this. Previous work had identified factors correlated with the switch, including complementary RNA stability, nucleotide pools and the soluble pools of nucleoprotein and polymerase, but the authors write that a true underlying mechanism remained elusive. Small RNAs were already known to serve viral life cycles in other systems, which motivated looking for one here.
Central question
Does influenza A virus generate a small RNA species, and if so does that species participate in the reassignment of the viral polymerase from transcriptase to replicase activity.
Experimental strategy
The design moves from unbiased discovery to requirement testing to functional perturbation. Discovery used SOLiD small RNA deep sequencing of the sub-40 nucleotide fraction from A549 cells infected with A/PR/8/34, which makes no assumption about the sequence or the structure of what might be present. Expression was then characterised by northern blot with a pan-specific probe across an infection time course, set against viral protein by western blot and against genomic RNA accumulation by a strand-selective quantitative PCR, so that the timing of svRNA could be placed relative to the transcription to replication transition rather than merely to infection. Generality was tested across three subtypes in embryonated eggs and across human, canine and murine cells, and specificity was tested against vesicular stomatitis virus and type I interferon to exclude an induced host species. Biogenesis requirements were addressed by transfecting the eight bidirectional segment plasmids and withdrawing one at a time, which separates the contribution of individual viral products from infection as a whole, with NS1 and NEP/NS2 supplied back individually to resolve the segment 8 requirement. Physical association was tested by Flag immunoprecipitation of individual polymerase subunits, the reconstituted trimer, nucleoprotein and NS1 against a synthetic 5 prime triphosphorylated svRNA mimetic. Function was interrogated with locked nucleic acid antisense oligonucleotides against individual segment svRNAs, read out by primer extension that resolves messenger, complementary and genomic RNA separately, which is the readout that can distinguish an effect on transcription from an effect on replication.
Key findings
- Deep sequencing of the sub-40 nucleotide RNA from infected A549 cells at twelve hours returned over four million unique reads, more than ninety percent of which were cellular microRNAs. Roughly thirty percent of the influenza-derived sequences in that fraction were enriched at the 5 prime end of the genomic RNA rather than distributed across the genome as breakdown products were. These svRNAs were 22 to 27 nucleotides, unique to each segment beyond position 14 to 16, and terminated three to four bases short of the polyU tract (Figure 1, Tables S1 to S3).
- By northern blot with a pan-specific probe, svRNA became detectable around twelve hours after infection and continued to accumulate to thirty-six hours (Figure 2A). NS1 protein was already visible by eight hours, so protein expression preceded svRNA detection (Figure 2B), and strand-selective quantitative PCR placed the rise in genomic RNA at twelve hours, concurrent with svRNA appearance (Figure 2C). The concurrence is a temporal correlation and the paper does not establish which event is upstream.
- svRNA was absent from cells treated with vesicular stomatitis virus or type I interferon, indicating it is not a virus-inducible or interferon-inducible host small RNA (Figure 2D). It was produced by H1N1, H3N2 and H5N1 in embryonated eggs, with levels tracking the extent of replication (Figure 2E), and by H1N1 infection of human, canine and murine fibroblasts (Figure 2F).
- Transfection of all eight bidirectional segment plasmids produced svRNA in the absence of infection. Withdrawing segment 1, 2, 3 or 8 abolished it, withdrawing segment 5 abolished the mature species while leaving fragments of approximately 32 and 42 nucleotides, withdrawing segment 7 reduced it, and withdrawing segments 4 or 6 had no significant effect (Figure 3A). Supplying NEP/NS2 back to a segment 1 to 7 transfection gave partial rescue while NS1 gave none (Figure 3B). The authors read this as the polymerase, nucleoprotein and NEP/NS2 constituting the minimal machinery for svRNA production.
- A synthetic 5 prime triphosphorylated svRNA mimetic co-precipitated with the reconstituted three-subunit polymerase but not with PB1, PB2 or PA individually, nor with nucleoprotein or NS1 (Figure 3C). Phosphatase treatment confirmed that the bound species carried a 5 prime triphosphate, alongside a less abundant species lacking it (Figure 3D).
- Despite carrying an exposed 5 prime triphosphate, transfected svRNA did not induce IRF3 phosphorylation or interferon beta, whereas polyinosinic-polycytidylic acid and an NS1 RNA-binding mutant virus did (Figure S3B). The authors take this as consistent with published length and structure requirements for cytosolic helicase recognition rather than as a direct test of receptor engagement.
- A locked nucleic acid complementary to segment 4 svRNA had little effect on HA messenger RNA or complementary RNA but substantially reduced HA genomic RNA by primer extension (Figure 4A). Nucleoprotein and NS1 protein were unaffected and HA protein modestly reduced (Figure 4B), while supernatant transferred to MDCK cells produced a marked loss of HA, NP and NS1 signal, which the authors attribute to a packaging defect arising from loss of segment 4 genomic RNA (Figure 4C).
- Inhibitors against NA or NS svRNA left HA transcripts unchanged, indicating that the effect of the segment 4 inhibitor is segment restricted (Figure 5A). Prolonged segment 4 inhibition removed HA protein without affecting NP or NS1 (Figure 5B) and reduced infectious titres by approximately eighty percent (Figure 5C).
Mechanistic model
The study does not establish a definitive mechanism. It identifies a viral small RNA, places its appearance at the time of the transcription to replication transition, shows that it binds the assembled polymerase, and shows that blocking it selectively depletes genomic RNA. It does not show that svRNA acts on the polymerase to cause the switch, and it does not determine how svRNA is made.
The model the authors propose is that the replicase form of the polymerase is an NEP/NS2-containing complex loaded with svRNA, while the transcriptase form lacks svRNA and therefore remains bound in cis to the 5 prime template end, which permits the stuttering that generates the polyadenylated tail. Because inhibition acts segment by segment, they note that the model implies eight distinct replicase complexes distinguished by which svRNA is bound, with the bound svRNA acting as a guide that could reconstitute the promoter in trans and free the genomic ends. This is author interpretation and is presented as such in the discussion.
Biogenesis is left open. The authors consider cleavage of the genomic RNA by the polymerase and synthesis from the complementary RNA intermediate, argue against cleavage because it would destroy the template, and note that synthesis from complementary RNA would require a stochastic svRNA-independent round of complementary RNA production. They connect this second possibility to prior work on complementary RNA stability as the trigger for the switch. Neither route is demonstrated here.
Conceptual or technical advance
The work makes an RNA species a candidate participant in a step of the influenza life cycle that had been described only in terms of proteins, nucleotide pools and RNA stability, and it supplies the reagents that make that participation testable, namely the pan-specific and segment-specific detection probes, the synthetic 5 prime triphosphorylated mimetics, and the segment-specific locked nucleic acid inhibitors. Because the svRNA sequences derive from the conserved noncoding ends, the observation that the species is produced across subtypes and host species raises the possibility of inhibitors that are not strain restricted, which the authors name as a potential therapeutic direction rather than demonstrate.
Relationship to the broader research program
The paper sits at the intersection of two recurring interests in this corpus, the temporal control of influenza gene expression and the use of small RNA biology as both a subject and a tool. The finding that the virus times a transition using an RNA species anticipates later work from the same laboratory on how influenza schedules its own gene expression through the splicing of segment 8, and the deep sequencing and small RNA northern blot methods used here recur across the laboratory's subsequent small RNA work. Reading this alongside later papers in the corpus, the framing of viral noncoding RNA as a regulatory rather than merely incidental product is a thread that continues, which is a category 3 synthesis observation drawn from setting this paper next to Chua and colleagues in 2013 and the laboratory's later small RNA studies rather than something stated here.
Related publications
- Chua and colleagues, 2013, follow-up. The later paper on suboptimal splicing of the influenza segment 8 transcript addresses the same question of how the virus times its life cycle, and it cites this work when discussing genomic RNA accumulation and premature ribonucleoprotein export.
- Robb and colleagues, 2009, predecessor, from another laboratory. Cited here as the report that NEP/NS2 regulates transcription and replication of the influenza genome, which is the basis for interpreting the segment 8 requirement.
- Vreede and colleagues, 2004 and 2008, predecessor, from another laboratory. Cited as the prior models in which complementary RNA stabilisation and nucleoside triphosphate concentration govern the switch, against which the svRNA model is positioned.
Limitations and boundaries
The function of svRNA is inferred from an antisense inhibition experiment, and the segment 4 inhibitor is complementary to a sequence that is also present at the 5 prime terminus of the segment 4 genomic RNA itself, so the reported experiments do not formally separate depletion of the free small RNA from engagement of the template end. No experiment reconstitutes replicase activity with and without svRNA in vitro, so the proposed role in converting the polymerase between activities is not directly tested. The coincidence in timing between svRNA accumulation and the rise in genomic RNA is correlative. The biogenesis route is unresolved and the authors say so. The size of the species is also unresolved, since the sequencing data indicate predominant lengths of 25 and 27 nucleotides while the northern blots indicate approximately 22 and 25 nucleotides, a discrepancy the authors describe as difficult to explain and attribute tentatively to the 5 prime triphosphate. The absence of IRF3 phosphorylation is evidence that transfected svRNA does not trigger this pathway under the conditions used and is not a general claim about innate immune invisibility. All work is in cultured cells and embryonated eggs, with no animal infection model, and the titre reduction of approximately eighty percent was measured in a single-cycle and short growth curve setting rather than in vivo.
Audience summaries
25 words
Influenza makes short RNAs copying the start of each genome segment. They appear as the virus begins copying its genome, and blocking them reduces genome production.
75 words
Influenza A virus must switch its single polymerase from making messenger RNA to copying its genome, and how it does so was unclear. Sequencing small RNAs from infected cells revealed a 22 to 27 nucleotide species matching the start of each genome segment. It appears when genome copying begins, binds the assembled polymerase, and requires the polymerase, nucleoprotein and NEP/NS2 to be made. Blocking one segment's small RNA reduces that segment's genome and viral output.
150 words
Influenza A virus uses one RNA-dependent RNA polymerase for primer-dependent messenger RNA synthesis and for primer-independent genome synthesis, and the transition between the two had no established mechanism. Deep sequencing of small RNAs from infected A549 cells identified a discrete 22 to 27 nucleotide species, svRNA, corresponding to the 5 prime end of each of the eight genomic segments. It accumulates from roughly twelve hours after infection, coincident with the measured rise in genomic RNA and after viral protein appears. It is produced across H1N1, H3N2 and H5N1 and across human, canine, murine and avian systems, and is not induced by an unrelated virus or by interferon. Production requires the polymerase subunits, nucleoprotein and segment 8, with NEP/NS2 partially sufficient, and svRNA associates only with the assembled polymerase. Segment-specific antisense inhibition depletes that segment's genomic RNA while sparing messenger and complementary RNA. The authors propose an svRNA-loaded replicase, which the data support but do not establish.