Poxviruses degrade host microRNAs through the catalytic subunit of their own poly(A) polymerase, VP55, which adds short nontemplated adenosine tails to argonaute-loaded guide strands and thereby marks them for cellular decay, while small RNAs carrying a 3 prime terminal 2 prime O-methyl group are spared.
Backes S, Shapiro JS, Sabin LR, Pham AM, Reyes I, Moss B, Cherry S, tenOever BR. Degradation of Host MicroRNAs by Poxvirus Poly(A) Polymerase Reveals Terminal RNA Methylation as a Protective Antiviral Mechanism. Cell Host & Microbe. 2012. Volume 12, issue 2, pages 200-210.
DOI 10.1016/j.chom.2012.05.019. PMID 22901540. PMCID PMC3782087.
Poxviruses degrade host microRNAs through the catalytic subunit of their own poly(A) polymerase, VP55, which adds short nontemplated adenosine tails to argonaute-loaded guide strands and thereby marks them for cellular decay, while small RNAs carrying a 3 prime terminal 2 prime O-methyl group are spared.
Poxviruses are large double-stranded DNA viruses that replicate entirely in the cytoplasm and transcribe genes with long 3 prime untranslated regions, so they might be expected to exploit host microRNAs the way several nuclear DNA viruses do, yet no poxvirus-encoded microRNA had been reported. The authors approached this puzzle from the small RNA side, sequencing the small RNA population of insect and mammalian cells during poxvirus infection. In Drosophila cells infected with vaccinia virus and in tiger moth cells infected with their natural entomopoxvirus, mature microRNAs acquired nontemplated 3 prime adenosines and were lost, while endogenous small interfering RNAs were not. The same happened in mammalian cells, where a recombinant vaccinia virus expressing miR-124 produced heterogeneous tailed species and simultaneously eliminated endogenous miR-93. Inhibitor experiments placed the responsible activity in the early phase of infection, and knockdown and reconstitution identified VP55, the catalytic subunit of the viral poly(A) polymerase, as both necessary and sufficient. Synthetic microRNA carrying seven adenosines was degraded in uninfected cells, indicating that the decay step is cellular. Tailed microRNAs remained associated with argonaute 2, placing the modification after strand selection, and a guide strand bearing a 3 prime 2 prime O-methyl group escaped tailing entirely. Restoring abundant microRNAs during infection roughly halved virus production, which the authors read as evidence that this degradation benefits the virus.
Plants, worms and insects use small interfering RNAs derived from the pathogen to restrict virus infection, while chordates respond largely through nucleic acid sensing and type I interferon. The paper frames its starting question as an anomaly in that landscape. Several nuclear DNA viruses and retroviruses encode their own microRNAs to reshape the host transcriptome, and the generalisation in the field at the time was that this strategy belongs to pathogens with large DNA genomes and sustained replication cycles. Poxviruses fit that description but had not been reported to use the pathway, which could not be explained by cytoplasmic replication because the laboratory and others had already shown that cytoplasmic RNA viruses can be engineered to generate functional microRNAs through noncanonical processing. Separately, 2 prime O-methylation was known to protect small RNAs from 3 prime uridylation-dependent decay and from exonucleolytic trimming, and methylation of the messenger RNA cap had recently been implicated in distinguishing host from viral RNA. These two lines had not been connected to poxvirus biology.
Why do poxviruses not appear to exploit the host microRNA pathway, and does poxvirus infection instead act against host small RNAs.
The study runs the same question through two host contexts and then reduces it to a single viral protein. Insect cells were used first because they carry both microRNA and small interfering RNA pathways, which allows the two classes to be compared within one sample, and because a genuine insect poxvirus pair exists to control for the artificial nature of infecting Drosophila cells with a mammalian virus. Small RNA deep sequencing at nucleotide resolution was the primary discovery readout, since the modification in question is the addition of a few nontemplated bases and would be invisible to abundance-only methods, and it was paired with northern blot so that mobility shift and abundance loss could be seen together. In mammalian cells the authors built a recombinant vaccinia virus carrying the miR-124 locus, which supplies a microRNA that is absent from the cells used and therefore reports on virus-derived processing and modification separately from the endogenous pool. Temporal placement used cycloheximide and cytosine arabinoside to restrict the infection to distinct gene expression classes. Attribution to VP55 used loss of function by small interfering RNA during infection and gain of function by transfection of the tagged subunits alone, with VP39 as the internal specificity control. Chemically defined mimetics carrying zero, one or seven adenosines separated the tagging step from the degradation step, and a 2 prime O-methylated guide strand tested the protection hypothesis directly. Argonaute 2 immunoprecipitation positioned the modification relative to strand selection, and reconstitution of the dominant microRNAs during infection tested whether the degradation matters for virus yield.
The biochemical chain is supported at each step. VP55, an early and virion-associated protein, adds a short nontemplated poly(A) tract of roughly two to nine adenosines to the guide strand of a microRNA after strand selection, and the tailed guide is then removed by host degradative machinery, since a chemically synthesised tailed microRNA is destroyed in the absence of any viral protein. A 3 prime terminal 2 prime O-methyl group blocks the adenylation step and thereby the entire pathway.
Two elements are interpretation rather than demonstration. The first is where in the complex the reaction occurs. The argonaute 2 immunoprecipitation shows that tailed species are argonaute associated and that star strands are not tailed, and the authors infer from this that VP55 acts through an interaction with argonaute or another component of the mature silencing complex. No direct interaction between VP55 and argonaute is shown. The second is the purpose of the activity. The authors offer two evolutionary readings, evasion of microRNA-mediated restriction arising from the long 3 prime untranslated regions of poxvirus transcripts or from indirect control of host factors, and evasion of a more general small RNA-based restriction of the kind present in insects. They then propose that 2 prime O-methylation may have evolved, at least in part, as a cellular countermeasure protecting antipathogen small RNAs. This last proposal is explicitly speculative in the paper and rests on comparative arguments about the distribution of the HEN1 methyltransferase rather than on experiment.
The work supplies a concrete reason why a family of large DNA viruses does not use host microRNAs, namely that it destroys them, which converts an absence in the literature into a testable enzymatic activity with a named subunit. It also assigns a second function to a protein previously characterised for its role in viral messenger RNA maturation, so that the VP55 and VP39 heterodimer is described as carrying both replication functions and evasion functions. By showing that a 3 prime terminal methyl group determines whether a small RNA is a substrate, the study places terminal RNA methylation alongside cap methylation as a mark distinguishing RNA that is protected from RNA that is not, and makes that distinction experimentally accessible using defined synthetic substrates. The engineered poxvirus expressing a heterologous microRNA is itself a reusable reagent for separating virus-derived from host-derived small RNA in infected cells.
The paper belongs to a sustained line of work in this corpus on what happens to small RNA biology during virus infection and on the use of engineered viruses that express microRNAs as experimental instruments. It cites the laboratory's own earlier demonstrations that cytoplasmic RNA viruses can be built to generate functional microRNAs through noncanonical processing, and the recombinant vaccinia virus here extends that toolkit to a DNA virus. It also connects to the laboratory's interest in why chordates rely on interferon rather than RNA interference, since the finding that a virus spends an enzyme on destroying host microRNAs argues that those small RNAs impose a cost on the virus. Setting this paper beside the laboratory's other small RNA work, the recurring move is to treat small RNAs as both a subject of viral antagonism and a means of interrogating infection, which is a category 3 synthesis drawn from reading these papers together rather than a claim made here.
The authors state plainly that the physiological requirement for VP55 in microRNA degradation cannot be isolated, because VP55 is also required for viral replication and the two activities cannot be separated with the reagents available. The functional test of whether microRNAs restrict the virus therefore relies on reconstituting a pool of four microRNA families rather than on a virus lacking the tailing activity, and it produced an approximately two-fold effect on virus yield that the authors explicitly describe as possibly direct or indirect. All infection work is in cultured cells, and the inference that a fifty percent loss of extracellular enveloped virus would attenuate the virus in an animal is extrapolation from earlier literature rather than a result reported here. The argonaute 2 immunoprecipitation establishes association, not the site or partner of the enzymatic reaction. The protection experiment tests one 2 prime O-methylated guide strand in one sequence context. The evolutionary claims about the origin of 2 prime O-methylation as an antiviral countermeasure are presented in the discussion as speculation supported by comparative genomic arguments from other laboratories, and no evolutionary experiment is performed. Vaccinia virus infection of Drosophila cells is abortive and arrests before intermediate gene expression, which the authors acknowledge and address by adding the natural entomopoxvirus pairing, but the insect data are otherwise from a non-natural host. Finally, the sequencing-based claim of unbiased degradation rests on the microRNA populations of the specific cell types profiled.
Poxviruses carry an enzyme that adds short adenosine tails to host microRNAs, marking them for destruction. Small RNAs with a methylated end escape this fate.
Poxviruses have never been shown to use host microRNAs, and this work suggests why. In insect and mammalian cells, poxvirus infection adds a few nontemplated adenosines to mature microRNAs, after which the cell's own machinery destroys them. The responsible enzyme is VP55, the catalytic subunit of the viral poly(A) polymerase, which is both necessary and sufficient. Small RNAs carrying a methyl group at their 3 prime end are not modified and survive.
Poxviruses replicate in the cytoplasm and transcribe genes with long 3 prime untranslated regions, yet none had been reported to encode microRNAs. Sequencing small RNAs during vaccinia virus infection of Drosophila cells, entomopoxvirus infection of its natural moth host, and vaccinia infection of mammalian cells showed that mature microRNAs acquire short nontemplated adenosine tracts and then disappear, while endogenous small interfering RNAs do not. Chemical inhibitors placed the activity in early infection, and knockdown together with reconstitution identified VP55, the catalytic subunit of the viral poly(A) polymerase, as necessary and sufficient. A synthetic microRNA bearing seven adenosines was destroyed in uninfected cells, showing that the nuclease is cellular. Tailed guides remained argonaute associated, placing the modification after strand selection, and a 2 prime O-methylated guide was fully protected. Restoring the dominant microRNAs during infection halved virus production. The authors propose that terminal methylation evolved partly to shield small RNAs from this class of attack.
lab-led for Backes 2014, Benitez 2015 on engineered RNA interference, Aguado 2015 and the tenOever 2013 review, co-led with the Cherry laboratory for Backes 2012, and collaborative for Cullen 2013, a Minireview led by Bryan Cullen with Sara Cherry and tenOever