An inefficient splice site paces influenza gene expression, and the same processing step is a vulnerability across the orthomyxovirus family
lab-led
Influenza A Virus Utilizes Suboptimal Splicing to Coordinate the Timing of Infection
lab-ledDiscovery
The inefficient 5 prime splice site of influenza A virus segment 8 functions as a timing device, causing the nuclear export protein to accumulate slowly as a minor product of abundant NS1 transcription, with both raising and lowering that rate attenuating the virus through mistimed ribonucleoprotein export.
- Senior authors
- Benjamin R. tenOever
- Correspondence
- Benjamin R. tenOever
Research areas & themes
Influenza Genome Regulation and ReplicationSplicing and Temporal Control of Segment 8NS1 and Interferon Antagonism
Citation
Chua MA, Schmid S, Perez JT, Langlois RA, tenOever BR. Influenza A Virus Utilizes Suboptimal Splicing to Coordinate the Timing of Infection. Cell Reports. 2013. Volume 3, issue 1, pages 23-29.
DOI 10.1016/j.celrep.2012.12.010. PMID 23333274. PMCID PMC3563938.
One-sentence contribution
The inefficient 5 prime splice site of influenza A virus segment 8 functions as a timing device, causing the nuclear export protein to accumulate slowly as a minor product of abundant NS1 transcription, with both raising and lowering that rate attenuating the virus through mistimed ribonucleoprotein export.
Executive summary
Influenza A virus carries only ten major proteins on eight segments, each driven by a comparable promoter, yet it must order nuclear entry, genome replication, nuclear export and assembly in time. Segment 8 is bicistronic, producing the interferon antagonist NS1 as the unspliced transcript and the nuclear export protein NEP through a weak 5 prime splice site that captures only ten to fifteen percent of segment-derived messenger RNA. The authors asked whether that inefficiency is functional. Using microRNA target sites inserted into the segment, they silenced NS1 by more than ninety percent without touching NEP and found little effect on replication or on host antiviral gene induction in cell lines, in primary cells or in mice. Manipulating NEP in either direction told a different story. Lowering NEP, either by decoupling it from splicing with a 2A recoding site or by small interfering RNA, reduced titres. Raising NEP, either by supplying an extra copy from a replication-incompetent vector or by optimising the splice site so that NEP became the dominant product, reduced titres by about two orders of magnitude in culture and almost abolished replication in mice. Imaging showed premature cytoplasmic appearance of nucleoprotein when NEP accumulated early and delayed export when NEP was scarce.
Scientific context
Segment 8 of influenza A virus is one of two segments that undergo splicing and encodes both NS1 and NEP. NS1 is described in the prior literature as the dominant antagonist of the cellular response to infection, and viruses lacking it are attenuated by more than four logs in animals while retaining wild-type virulence in hosts without antiviral signalling. NEP is a 121 amino acid protein required for production of replication-competent virus, with its best-characterised role in exporting nucleoprotein-associated genomic RNA from the nucleus through association with the matrix protein M1, and disruption of its nuclear export signal abolishes replication. NEP had also been implicated by earlier work in controlling complementary RNA levels, in host tropism and in the synthesis of influenza small viral RNA. What had not been resolved was why the virus makes NEP through a splice site that works poorly, producing only ten to fifteen percent of segment 8 messenger RNA as the spliced form, and whether the relative amounts of the two segment 8 products matter for coordinating the infection.
Central question
Is the balance between NS1 and NEP, set by the efficiency of the segment 8 splice site, a requirement for coordinating the influenza A virus life cycle, or is the inefficiency of that splice site incidental.
Experimental strategy
The design isolates each segment 8 product and moves its abundance up and down by independent means, so that a shared phenotype cannot be attributed to any one manipulation. NS1 was reduced without disturbing NEP by inserting tandem perfect target sites for host microRNAs into the intergenic region of a modified NS vector, which places silencing under the control of the cell's own microRNA content rather than requiring a mutation in the coding sequence. Two microRNAs with different expression patterns were used, one ubiquitous and one restricted to haematopoietic cells, and a scrambled insert served as the isogenic control, which lets the same virus be silenced or not depending on the cell type it enters. Cells lacking Dicer1 confirmed that the effect depended on the microRNA pathway. NEP was lowered in two ways, by expressing it from a 2A ribosome recoding site instead of as a splice product and by small interfering RNA during wild-type infection, and raised in two ways, by supplying an extra copy from a replication-incompetent influenza-like vector grafted into segment 4 and by mutating the 5 prime splice site so that NEP became dominant and NS1 a readthrough product. Because splice site optimisation also changes an amino acid, a parental virus carrying the same substitution with the native splice site was built as the matched control. Replication was read by multicycle growth curves in culture and by titres from intranasally infected mice, host response by interferon-regulated gene expression, and the proposed timing defect by immunofluorescence of nucleoprotein localisation at early time points.
Key findings
- Insertion of perfect microRNA target sites silenced NS1 at both protein and messenger RNA level without changing nucleoprotein or NEP, with silencing matching the microRNA content of the cell, complete in haematopoietic cells for both targeted viruses and restricted to the ubiquitous microRNA target in non-haematopoietic cells (Figures 1A and 1B, Figure S1C). Cells lacking Dicer1 restored NS1 in all three viruses, establishing that the effect requires microRNA processing (Figure 1C).
- Despite more than ninety percent loss of NS1, growth curves in lung epithelial cells showed the silenced viruses reaching titres comparable to the scrambled control and to around ten to the seventh plaque forming units per millilitre, while a virus deleted for NS1 was substantially impaired (Figure 1D). Induction of the interferon-regulated gene Mx1 was unchanged in the silenced viruses and elevated only for the deletion virus (Figure S1D).
- The same result held in primary lung fibroblasts and bone marrow-derived macrophages, where NS1 fell in proportion to endogenous microRNA levels without affecting nucleoprotein (Figures 2A and 2B), and in mice, where silencing was confirmed in lungs of both interferon-competent and signalling-deficient animals and titres and interferon-regulated gene induction did not differ across viruses (Figures 2C to 2E, Figures S2A and S2B). The authors read this as low levels of NS1 being sufficient to antagonise the host response, which they describe as surprising.
- A virus expressing NEP from a 2A recoding site produced adequate NS1 but reduced NEP and lost roughly two logs of progeny despite normal synthesis of structural protein (Figures 3A and 3B). Independently, small interfering RNAs reducing NEP during wild-type infection produced about one log of attenuation at forty-eight hours with a reported p value of 0.01 (Figures S3B and S3C).
- A replication-incompetent influenza-like vector carrying an extra NEP copy in place of hemagglutinin raised NEP substantially, and co-infection of this vector with wild-type virus caused approximately two logs of attenuation relative to the scrambled vector (Figures 3C and 3D). Excess NEP is therefore as damaging as insufficient NEP.
- Optimising the segment 8 5 prime splice site inverted the expression ratio, reducing NS1 and raising NEP, confirmed by immunoblot and by RT-PCR of the spliced products (Figure 4A, Figure S4A). The optimised virus was attenuated by about two logs relative to the matched parental virus carrying the same amino acid change (Figure 4B), and the attenuation persisted in an NS1-complementing cell line, which the authors take as evidence that the defect is caused by excess NEP rather than by the accompanying loss of NS1 (Figure 4C, Figures S4B to S4D). In mice the optimised virus never exceeded one hundred plaque forming units per millilitre while infection was confirmed by quantitative RT-PCR for nucleoprotein (Figure 4D, Figure S4E).
- Nucleoprotein staining in infected lung epithelial cells appeared in the cytoplasm as early as five hours after infection with the splice-optimised virus while remaining nuclear with the parental virus, and at seven hours every optimised-virus infected cell showed cytoplasmic signal against roughly half of parental-virus infected cells (Figure 4E). The virus with reduced NEP showed the converse, a delay in export relative to wild type (Figure S4F).
Mechanistic model
The study establishes that the quantity of NEP, and the rate at which it accumulates, determines the outcome of infection, and it establishes that the segment 8 splice site sets that rate. It does not establish the molecular step at which mistimed export becomes lethal to the infection, and the export timing data are correlative with respect to the titre phenotype.
The model the authors propose is that the virus uses a deliberately poor 5 prime splice site as a molecular timer. Because all eight segments carry comparable promoters, the virus cannot delay a protein simply by transcribing its segment later. Instead it overproduces the NS1 transcript, from which NEP is generated as a minor spliced byproduct, so that NEP concentration rises slowly and crosses the threshold for ribonucleoprotein export only after replication has proceeded. On this reading the apparent overproduction of NS1 is not waste but the mechanism by which NEP accumulation is paced, and coupling the timer to NS1 rather than to a polymerase subunit, nucleoprotein or a surface protein is advantageous because excess NS1 is tolerated while excess of the others would perturb replication or assembly. The authors also note, citing their own and others' earlier work, that the transcription to replication switch is likely triggered by NEP accumulation, and they present their NEP manipulation data as corroborating that deregulated NEP produces aberrant replication. That link is offered as consistent with the data rather than tested here.
Conceptual or technical advance
The work reframes an apparent inefficiency in viral gene expression as a regulatory device, and it does so by making the splice site itself the experimental variable rather than the protein it produces. It also separates the two functions carried on one segment in a way that had not previously been possible in a replication-competent virus, using host microRNA targeting to knock down NS1 while leaving NEP intact, which yields the unexpected result that the interferon antagonist is produced far in excess of what is needed for its antagonist function, at least late in infection. The set of five complementary reagents built here, the microRNA-targeted viruses, the 2A recoded virus, the NEP-bearing replication-incompetent vector, the splice-optimised virus and its matched parental control, provides a way to titrate segment 8 output in either direction and is reusable for other questions about influenza gene expression stoichiometry.
Relationship to the broader research program
This paper continues a line in the corpus on how influenza A virus times its own life cycle without a transcriptional clock. It cites the laboratory's earlier work identifying influenza small viral RNA and the requirement for NEP in its production, and its discussion places NEP accumulation as the likely trigger for the switch from transcription to replication that the earlier paper addressed, which ties the two studies to one question approached from opposite ends. It also draws on the laboratory's ongoing use of microRNA targeting to control viral gene expression in vivo, a technique the group had applied to influenza in earlier work and which is used here as a precision knockdown rather than as an attenuation strategy. Reading the 2010 and 2013 papers together suggests a picture in which a single segment 8 product sits at the centre of both genome synthesis regulation and export timing, which is a category 3 synthesis emerging from the pair rather than a claim either paper makes alone.
Related publications
- Perez and colleagues, 2010, predecessor, from the same laboratory. Identified influenza small viral RNA and showed a requirement for NEP in its production, and is cited here in the discussion of what triggers the transcription to replication switch.
- Varble and colleagues, 2010, and Langlois and colleagues, 2012, methodological foundation, from the same laboratory. Supply the modified NS vector and the microRNA-targeting approach used to silence NS1 in a replication-competent virus.
- Perez and colleagues, 2009, methodological foundation, from the same laboratory. Cited among the demonstrations that influenza A virus can be placed under host microRNA control.
- Manicassamy and colleagues, 2010, methodological foundation, from another laboratory. Source of the 2A recoding strategy used to uncouple NEP expression from splicing.
- García-Sastre and colleagues, 1998, and Egorov and colleagues, 1998, predecessor, from other laboratories. Establish the attenuation of NS1-deficient influenza virus that frames the expectation this paper tests against.
- Robb and colleagues, 2009, predecessor, from another laboratory. Cited for the regulatory role of NEP in transcription and replication of the influenza genome.
Limitations and boundaries
The link between mistimed ribonucleoprotein export and loss of infectivity is inferred rather than demonstrated. Nucleoprotein localisation is scored by immunofluorescence at two early time points, and no experiment shows that restoring correct export timing rescues the titre defect. The splice-optimised virus necessarily carries an amino acid substitution, which is controlled by a matched parental virus and by an NS1-complementing cell line but not by a construct that raises NEP without altering the segment 8 sequence. Excess NEP was delivered in one experiment by a replication-incompetent vector during co-infection at high multiplicity, a setting that differs from natural infection. The conclusion that low NS1 suffices for antagonism is bounded by the readouts used, which are interferon-regulated gene induction and virus titre in the cells, primary cultures and mouse strain tested, and by the fact that silencing removes more than ninety percent of NS1 rather than all of it, so residual NS1 activity cannot be excluded as the explanation. The animal work is a sublethal intranasal mouse model, and the study does not address transmission, pathology or other host species. The proposed connection between NEP accumulation and the transcription to replication switch is drawn from prior literature and is not tested here.
Audience summaries
25 words
Influenza splices one of its genes badly on purpose. The resulting slow build-up of an export protein times the infection, and correcting the splicing cripples the virus.
75 words
Influenza A virus segment 8 makes the interferon antagonist NS1 as its main product and the nuclear export protein NEP through a weak splice site. Silencing NS1 by over ninety percent barely affected the virus, in cells or in mice. Changing NEP in either direction crippled it. Making the splice site efficient raised NEP, sent genome complexes to the cytoplasm hours early, and cost two logs of growth. The poor splice site is a timer.
150 words
Influenza A virus has eight segments with comparable promoters and no obvious way to schedule its life cycle. Segment 8 encodes the interferon antagonist NS1 as the unspliced transcript and the nuclear export protein NEP through a 5 prime splice site used only ten to fifteen percent of the time. Inserting host microRNA target sites allowed NS1 to be reduced by more than ninety percent without touching NEP, and this had little effect on replication or on interferon-regulated gene induction in cell lines, primary cells or mice. Reducing NEP by a 2A recoding site or by small interfering RNA lowered titres, and raising NEP by an extra gene copy or by optimising the splice site lowered them by about two logs in culture and nearly abolished replication in mice. Nucleoprotein reached the cytoplasm hours early when NEP was abundant. The authors propose that inefficient splicing serves as a molecular timer pacing NEP accumulation.
Discoveries supported by this paper
Documented publication relationships
- Influenza A virus-generated small RNAs regulate the switch from transcription to replication — follow-up.
- Engineered RNA viral synthesis of microRNAs — methodological foundation.
- MicroRNA-based strategy to mitigate the risk of gain-of-function influenza studies — companion.
- The Mammalian Response to Virus Infection Is Independent of Small RNA Silencing — predecessor.
- A Versatile RNA Vector for Delivery of Coding and Noncoding RNAs — methodological foundation.
- Influenza A Virus Transmission Bottlenecks Are Defined by Infection Route and Recipient Host — predecessor.
- Engineered Mammalian RNAi Can Elicit Antiviral Protection that Negates the Requirement for the Interferon Response — methodological foundation.
- In Vivo RNAi Screening Identifies MDA5 as a Significant Contributor to the Cellular Defense against Influenza A Virus — predecessor.
- Synthetic Virology: Building Viruses to Better Understand Them — methodological foundation.
- Archaeal Kink-Turn Binding Protein Mediates Inhibition of Orthomyxovirus Splicing Biology — methodological foundation.
Pathogens
Technologies
Reverse genetics and virus rescueImmunofluorescence and confocal microscopySmall interfering RNA and short hairpin knockdownMicroRNA target site insertion into viral genomesSite-directed and alanine scanning mutagenesisRoute-controlled in vivo infection and deliveryMulticycle growth curves2A peptide recoding of viral segmentsReplication-incompetent influenza vector