Influenza A virus reassigns its own polymerase toward genome synthesis using a small RNA copied from its genome termini
lab-led
A Small-RNA Enhancer of Viral Polymerase Activity
lab-ledDiscovery
Influenza A virus small viral RNAs are shown to be synthesized from the complementary RNA intermediate, to load into the RNA binding cleft of the polymerase PA subunit, and to act there as segment-specific allosteric enhancers of full-length genome synthesis.
- 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, Zlatev I, Aggarwal S, Subramanian S, Sachidanandam R, Kim B, Manoharan M, tenOever BR. A Small-RNA Enhancer of Viral Polymerase Activity. Journal of Virology. 2012. Volume 86, Issue 24, pages 13475-13485.
DOI 10.1128/jvi.02295-12. PMID 23035211. PMCID PMC3503082.
One-sentence contribution
Influenza A virus small viral RNAs are shown to be synthesized from the complementary RNA intermediate, to load into the RNA binding cleft of the polymerase PA subunit, and to act there as segment-specific allosteric enhancers of full-length genome synthesis.
Executive summary
Influenza A virus carries a heterotrimeric RNA-dependent RNA polymerase that must perform two different jobs on the same eight templates, first transcribing capped and polyadenylated messenger RNA and later replicating full-length genome segments. How the enzyme switches between these modes, and how it keeps the eight segments in balance so that complete genomes can be packaged, had remained unsettled. Earlier work from this laboratory had identified small viral RNAs of roughly 22 nucleotides derived from the 5-prime ends of the genomic segments and had associated them with the transcription-to-replication switch, but their origin, their location, their binding site on the polymerase and their mode of action were unknown.
The study combines a plasmid-based polymerase reconstitution system, subcellular fractionation of infected human alveolar epithelial cells, deep sequencing of purified virions, immunoprecipitation of tagged polymerase subunits with synthetic 5-prime triphosphate small viral RNAs, structure-guided mutagenesis of the PA subunit, cell-free polymerase assays with purified recombinant enzyme, and a recombinant virus engineered to lose small viral RNA from one segment.
The small RNAs are shown to be templated from the complementary RNA intermediate, to remain nuclear rather than being packaged, to require the PB1 and PA heterodimer and in particular the basic residue R566 of PA for binding, and to enhance full-length genome synthesis in vitro without needing a free 3-prime hydroxyl. A virus unable to make small viral RNA from one segment loses genome synthesis for that segment alone and retains normal messenger RNA production.
Scientific context
The influenza A virus polymerase enters the nucleus tethered to the partially complementary 3-prime and 5-prime ends of each negative-sense segment, which fold into the panhandle or corkscrew promoter. From that template it first synthesizes capped viral messenger RNA using host-derived cap primers and stuttering at a poly-uridine track to add a poly-adenosine tail, and it later shifts to synthesizing full-length complementary RNA and progeny genomic RNA. The paper states that models for this switch have generally invoked external factors such as newly made polymerase acting in trans, changing concentrations of nucleoprotein, nucleotides or substrates, or simple stabilization of the complementary RNA intermediate, and that the precise contribution of each remains unknown. It further states that little was understood about how a replicase-competent polymerase maintains stoichiometric balance across the eight segments.
Prior work from the same laboratory reported influenza-derived small viral RNAs mapping to the 5-prime ends of each genomic segment and implicated them in the transcription-to-replication switch, and independent work reported that the relative concentrations of template sources bias the polymerase toward replicase mode. What those small RNAs bound, where they acted, and by what mechanism they acted were open questions entering this study.
Central question
By what molecular mechanism do influenza A virus small viral RNAs modulate the activity of the viral polymerase, and does that mechanism account both for the transition to genome replication and for the maintenance of balanced amounts of the eight genomic segments?
Experimental strategy
The design separates the question into origin, location, binding site and function, and answers each with a system in which the relevant variables can be set independently.
Origin and cofactor requirements are approached with a bidirectional eight-plasmid expression system in which individual segments can be omitted, and then with a reduced polymerase reconstitution assay containing a single template of defined polarity. Supplying a genomic-sense template or a complementary-sense template directly is what allows the nuclear export protein requirement to be placed upstream or downstream of complementary RNA synthesis. Truncations of the nuclear export protein separate its nuclear export function from its effect on RNA synthesis.
Location is approached by nuclear and cytoplasmic fractionation of infected A549 cells across a time course, and independently by deep sequencing small RNAs from sucrose-purified virions against small RNAs from infected cells, which tests packaging rather than steady-state distribution.
The binding site is approached with immunoprecipitation of individually tagged polymerase subunits in the presence of chemically synthesized 5-prime triphosphate small viral RNA, then with subunit truncations to find the minimal binding unit, then with point mutation of the conserved basic residues lining a reported cleft in the PA crystal structure.
Function is approached in a cell-free reaction with purified trimeric polymerase, an in vitro transcribed complementary RNA minireplicon that cannot itself generate small viral RNA, and synthetic small viral RNA variants including a 13-nucleotide form and forms blocked at the 3-prime hydroxyl, which distinguishes a primer from an allosteric effector. Finally, a recombinant virus carrying a poly-uridine track substitution in the neuraminidase segment tests the contribution during infection while leaving the other seven segments as internal controls.
Key findings
-
Omitting individual segments from the bidirectional plasmid system showed that loss of the polymerase subunits, of nucleoprotein, or of segment 8 abolished small viral RNA production, whereas loss of hemagglutinin, neuraminidase or the matrix proteins did not, and complementation identified the nuclear export protein rather than nonstructural protein 1 as the segment 8 component required (Figure 1A). The authors interpret this as a requirement for a fully elongation-competent polymerase together with the nuclear export protein.
-
In the polymerase reconstitution assay, increasing nuclear export protein enhanced complementary RNA synthesis from a genomic-sense template in a dose-dependent manner, and small viral RNA levels tracked with it (Figure 1B). Supplying a complementary-sense template directly made small viral RNA production independent of the nuclear export protein (Figure 1C). The observation is the loss of dependence on the alternate template. The interpretation offered is that the complementary RNA serves as the template for small viral RNA and that the nuclear export protein requirement reflects only its role in generating enough of that intermediate.
-
Nuclear export protein truncations that lacked the nuclear export sequence still drove complementary RNA synthesis and small viral RNA production, whereas truncations retaining the nuclear export sequence but lacking an intact C-terminal domain did neither (Figure 2). The effect on RNA synthesis is therefore separable from nuclear export activity. The authors note this corroborates earlier findings of Robb and colleagues.
-
Subcellular fractionation of infected A549 cells detected small viral RNA predominantly in the nuclear fraction from 8 hours post-infection onward and throughout infection, while nucleoprotein genomic RNA accumulated in the cytoplasm and peaked by 16 hours (Figure 3A). Modest cross-fraction contamination was visible on the blot and is acknowledged in the text.
-
Deep sequencing of sucrose-purified virions found that small RNA reads accounted for about 1.8 percent of segment 1 reads from virions against about 28 percent from infected cells (Figure 3B). The interpretation is that small viral RNAs are largely not packaged, supporting a nuclear site of action.
-
Immunoprecipitation of tagged subunits with synthetic 5-prime triphosphate small viral RNA showed association with the intact trimer and with the PB1 and PA heterodimer, and no association with the green fluorescent protein control (Figure 4A). Truncation mapping identified the PB1 interaction domain of PA, residues 155 to 716, and the catalytic domain of PB1, residues 1 to 500, as the minimal requirement, while the PB2 interaction domain of PB1 and the PA endonuclease domain were dispensable (Figure 4B).
-
Point mutation of R566 in the reported basic cleft of PA strongly reduced small viral RNA binding, K328 and K539 had minimal effects individually, and double and triple combinations substantially reduced binding (Figure 4C). The authors additionally report, as data not shown, that cleft mutants were incapable of both transcription and replication in reconstitution assays, which they read as evidence that the cleft matters for overall polymerase activity and not only for this interaction.
-
In cell-free reactions with purified trimeric polymerase and a complementary RNA minireplicon, little full-length product was made with buffer alone or with increasing scrambled RNA, whereas equimolar synthetic 22-nucleotide small viral RNA produced full-length product (Figure 5A). The first 13 nucleotides were sufficient, and small viral RNAs blocked at the 3-prime hydroxyl still supported full-length synthesis while abolishing aberrant 30 to 40 nucleotide products (Figure 5B). Because extension from the oligonucleotide is chemically prevented, the enhancement cannot be primer-mediated, and the authors conclude the effect is allosteric and is carried by the conserved 5-prime region.
-
Replacing the poly-adenosine track with a poly-uridine track in a complementary RNA template abolished small viral RNA production at 24 and 48 hours in the reconstitution assay (Figure 6A and 6B), indicating that the canonical track is needed for small viral RNA generation from the promoter.
-
A recombinant virus carrying the poly-uridine neuraminidase segment made less total small viral RNA (Figure 6C), retained normal neuraminidase messenger RNA (Figure 6B and 6D), and failed to synthesize neuraminidase genomic RNA robustly while hemagglutinin, nucleoprotein and nonstructural segment genomic RNA levels were unchanged relative to wild type (Figure 6E). The observation is a segment-restricted genome synthesis defect with messenger RNA intact. The interpretation is that small viral RNA acts specifically on replication of its own segment and thereby contributes to segment balance.
Mechanistic model
The model the authors advance, shown schematically in Figure 7, runs as follows. The incoming polymerase, bound to the genomic promoter and not yet occupied by small viral RNA, works predominantly as a transcriptase using host caps as primers. As nuclear export protein accumulates from viral transcripts, complementary RNA is synthesized and stabilized. Nascent complementary RNA then serves as the template for small viral RNA. The small viral RNA loads into the basic RNA binding cleft of PA, and the loaded polymerase becomes competent to engage its cognate complementary RNA template and synthesize full-length genomic RNA. Because the 3-prime complementarity of each small viral RNA differs, loading would create segment-specific replication-competent polymerase populations, which the authors propose is how stoichiometric balance across the eight segments is maintained.
Several parts of this model are directly demonstrated and others are proposed. Directly demonstrated are the binding of small viral RNA to the PB1 and PA heterodimer, the dependence of that binding on PA basic residues including R566, the sufficiency of synthetic small viral RNA to promote full-length synthesis from a complementary RNA template with purified enzyme, the fact that a blocked 3-prime hydroxyl does not remove that activity, and the segment-restricted genome synthesis defect in the engineered virus. The allosteric character of the effect is an inference from the blocked 3-prime hydroxyl experiment and from the sufficiency of 13 nucleotides, which excludes priming but does not itself show a conformational change. No structure of a small viral RNA bound to the polymerase is presented, and the cleft is identified from a previously reported crystal structure of a PA and PB1 fragment rather than from a complex with this RNA. The proposal that one polymerase complex is dedicated to one segment, and the claim that this is how segment balance is set, is author interpretation consistent with the single-segment defect but not established by a direct measurement of polymerase occupancy. The conclusion that the nuclear export protein requirement reflects nothing more than complementary RNA availability is also an interpretation, supported by the template-swap and truncation experiments.
Conceptual or technical advance
The work places a small RNA inside the catalytic machinery of a viral RNA-dependent RNA polymerase and assigns it a defined binding surface and a non-priming mode of action. That makes the transcription-to-replication switch addressable as a ligand-occupancy problem rather than only as a question of protein and nucleotide concentrations. It also supplies a reagent-level toolkit for testing the idea further, namely chemically synthesized 5-prime triphosphate small viral RNAs including 3-prime blocked forms, a purified-enzyme reaction in which de novo small RNA production is suppressed, PA cleft point mutants, and a recombinant virus in which small viral RNA can be removed from a single segment while leaving the others intact. The single-segment virus in particular turns segment balance into something that can be perturbed one segment at a time.
The paper states in its discussion that this finding is the first example of a small RNA capable of controlling RNA-dependent RNA polymerase activity. That is a priority claim made by the authors.
Relationship to the broader research program
The study extends a line of work in this laboratory on small RNAs generated during virus infection and on what those RNAs do to the infected cell and to the virus itself. It builds directly on the earlier identification of influenza small viral RNAs by Perez and colleagues and connects to the laboratory's broader interest in whether small RNA species encountered during infection are regulatory or incidental. Reading this alongside the laboratory's work on mammalian small RNA silencing during infection is category 3 synthesis, and it would need the other records in the corpus to support it, so it is flagged here rather than asserted.
The engineered poly-uridine virus also belongs to a recurring methodological thread in which recombinant influenza is used as an instrument, with a defined genetic change installed to remove one regulatory element while leaving the rest of the genome as an internal control.
Related publications
- Perez and colleagues 2010, Proceedings of the National Academy of Sciences, identification of influenza A virus-generated small RNAs regulating the switch from transcription to replication. Relationship predecessor. Cited in this paper as the source of the original identification and of the methods for small viral RNA detection.
- Shapiro, Varble, Pham and tenOever 2010, RNA, noncanonical cytoplasmic processing of viral microRNAs. Relationship methodological foundation. Cited for the small RNA library preparation used here.
- Perez and colleagues 2009, Nature Biotechnology, microRNA-mediated species-specific attenuation of influenza A virus. Relationship predecessor. Cited for the infection procedures used here.
- Robb, Smith, Vreede and Fodor 2009, Journal of General Virology, on regulation of transcription and replication by the nuclear export protein. Relationship predecessor from another laboratory. The truncation results here are explicitly described as corroborating it.
Limitations and boundaries
The cell-based work is confined to two human cell lines, A549 and HEK293, and to influenza A/Puerto Rico/8/34. The in vitro polymerase reactions use a purified trimer from an avian isolate, A/chicken/Nanchang/3-120/01, so the cell-free and infection systems are not matched for strain. No animal work is included, and no claim is made about pathogenesis or transmission.
The cell-free system lacks nucleoprotein, which the authors note prevents de novo small viral RNA production and also means the reaction does not reproduce the ribonucleoprotein context in which the polymerase normally works. The authors themselves caution that the ability of a short synthetic small viral RNA to promote full-length synthesis in vitro may not reflect a system the virus would use in a cell, since such promiscuity could be disadvantageous.
The allosteric mechanism is inferred rather than observed. There is no structure of the complex, no direct measurement of a conformational change, and no binding affinity determination. The PA cleft mutants are reported to be dead for both transcription and replication, which means the cleft mutations cannot cleanly separate small viral RNA function from general polymerase function in a functional assay, and this control is presented as data not shown.
The segment-specificity conclusion rests on one engineered segment, neuraminidase, altered by a poly-uridine track substitution that is known from prior work to affect messenger RNA export as well, so the manipulation is not exclusive to small viral RNA production even though messenger RNA levels for that segment were intact here. The fractionation result showed acknowledged cross-contamination between fractions, and the packaging conclusion rests on relative read proportions from purified virus stocks rather than on a direct assay of virion content.
Audience summaries
25 words
Influenza makes short RNAs from its own genome copies. These dock in a pocket of the viral polymerase and switch it toward copying that specific segment.
75 words
Influenza A virus must switch its polymerase from making messenger RNA to copying its eight genome segments, and it must copy them in balanced amounts. This study shows that short viral RNAs, made from the positive-sense copy of the genome, bind a basic cleft in the polymerase PA subunit and boost full-length genome synthesis without acting as primers. Removing this small RNA from one segment blocks genome synthesis for that segment alone.
150 words
The influenza A virus polymerase transcribes and replicates the same eight templates, and the control of that switch has been attributed variously to new polymerase acting in trans, to nucleoprotein or nucleotide concentrations, and to stabilization of the complementary RNA intermediate. Working in human cell lines, in reconstituted polymerase assays and with purified enzyme, this study shows that the small viral RNAs previously found at the 5-prime ends of the segments are templated from the complementary RNA intermediate, remain nuclear and are largely excluded from virions, and bind the polymerase through the PB1 and PA heterodimer, with the PA basic residue R566 most important. Synthetic small viral RNA promotes full-length genome synthesis from a complementary RNA template even when its 3-prime hydroxyl is blocked, which excludes priming and is read as allosteric enhancement. A recombinant virus lacking small viral RNA from the neuraminidase segment loses genome synthesis for that segment alone.
Discoveries supported by this paper
Documented publication relationships
- Degradation of Host MicroRNAs by Poxvirus Poly(A) Polymerase Reveals Terminal RNA Methylation as a Protective Antiviral Mechanism — conceptual extension.
- Engineered RNA viral synthesis of microRNAs — methodological foundation.
- Noncanonical cytoplasmic processing of viral microRNAs — methodological foundation.
- MicroRNA-mediated species-specific attenuation of influenza A virus — predecessor.
- Drosha as an interferon-independent antiviral factor — methodological foundation.
- Reduced Nucleoprotein Availability Impairs Negative-Sense RNA Virus Replication and Promotes Host Recognition — predecessor.
- The Host Factor ANP32A Is Required for Influenza A Virus vRNA and cRNA Synthesis — conceptual extension.