Argues that chordates, unlike plants, nematodes and arthropods, do not use small RNAs as antiviral defence, and that this absence of interplay between host microRNAs and RNA viruses is what leaves the microRNA machinery available for engineering viral tropism and small RNA delivery.
tenOever BR. RNA viruses and the host microRNA machinery. Nature Reviews Microbiology, 2013, volume 11, issue 3, pages 169-180.
DOI 10.1038/nrmicro2971. PMID 23411862.
Argues that chordates, unlike plants, nematodes and arthropods, do not use small RNAs as antiviral defence, and that this absence of interplay between host microRNAs and RNA viruses is what leaves the microRNA machinery available for engineering viral tropism and small RNA delivery.
Small RNA silencing systems are found in all three domains of life, but their uses differ. Plants, nematodes and arthropods generate virus-derived interfering RNAs and use them as antiviral immunity, whereas chordates rely on a protein-based system built on pattern recognition receptors, type I interferons and interferon-stimulated genes. This review sets out the case that chordate microRNAs do not serve an antiviral role, organizing the argument around three constraints. Copy number is the first, since an abundance threshold of roughly 100 copies per cell is needed for a microRNA to act on a host transcript, and a higher threshold would be needed to act on the far more numerous viral transcripts. Silencing capacity is the second, since genome-encoded microRNAs bind with partial complementarity and repress by less than twofold, which prevents the catalytic cleavage that virus-derived interfering RNAs achieve. Kinetics is the third, since most viral life cycles are shorter than the time needed to transcribe, process and load a new microRNA against proteins with half-lives beyond ten hours. Work from several laboratories profiling small RNAs in infected cells found no evidence of virus-derived small RNAs in vertebrates, and reported exceptions among herpesviruses, poxviruses and adenovirus are treated as a distinct set of cases. The practical argument follows from the biological one. Because viruses leave the host small RNA landscape intact, target sites for tissue-restricted or ubiquitous microRNAs can be grafted into viral genomes to dictate tropism, and viruses can be engineered to produce artificial microRNAs.
By 2013 it was established that small RNAs silence gene expression across all three domains of life, with clustered regularly interspaced short palindromic repeat RNAs restricting phage in bacteria and archaea, PIWI-interacting RNAs restricting transposable elements and foreign nucleic acid in animal germlines, and virus-derived interfering RNAs constituting the antiviral innate immune system in nematodes, arthropods and plants. MicroRNAs are the eukaryote-conserved class, encoded in the genome, processed from polymerase II transcripts by the Drosha and DGCR8 microprocessor and then by Dicer, and loaded into RISC where they fine-tune endogenous transcripts. Because chordates possess this machinery, it had been assumed that they would also generate an RNA-based antiviral defence and that such a defence might reinforce the interferon response. The review notes that the evidence for this expectation did not materialise. It records that small RNA profiling of virus-infected cells by other laboratories found no virus-derived small RNAs in vertebrate cells, and that an independent survey across hepatitis C virus, poliovirus, dengue virus, vesicular stomatitis virus and West Nile virus recovered small RNAs only at very low abundance, Dicer independent and without silencing capacity. The unresolved question the review addresses is not whether such a defence exists, but why it does not, and what follows from its absence.
Do host microRNAs contribute to antiviral defence in chordates, and if the answer is no, what does the resulting absence of interplay between RNA viruses and the host small RNA machinery permit in the way of engineered virological tools?
Not applicable in the experimental sense, since this is a single-author review rather than a primary research report. The argumentative strategy is worth stating, because it determines what the article can and cannot support. The review does not present new data. It assembles published findings from many laboratories and organizes them around a quantitative argument. Rather than asking whether any microRNA has ever been reported to affect a virus, it asks what abundance, complementarity and timing a small RNA would need in order to matter within the life cycle of an acute RNA virus, and then evaluates the literature against those thresholds. This framing is used to set aside a body of reports that measure microRNA induction as fold change rather than copies per cell, on the grounds that a large fold induction from a negligible baseline is not physiologically meaningful. The second half of the review inverts the same logic, treating the absence of viral engagement with the microRNA machinery as an engineering resource rather than as a negative result.
This is a review, so the entries below separate what the tenOever laboratory reported from what other laboratories reported, and both from the interpretive claims the review itself advances.
Claims advanced by the review itself, presented as argument rather than as new data. Three constraints are proposed to explain why chordate microRNAs cannot function in antiviral defence. Copy number, with a conservative estimate of roughly 100 copies per cell for activity on a host transcript, probably an order of magnitude higher for a viral target, against the observation that the 20 to 30 most abundant microRNAs make up more than 90 percent of the microRNA content of a cell. Silencing capacity, since partial complementarity gives less than twofold repression and precludes the catalytic cleavage and turnover that perfect complementarity permits. Kinetics, since most viral life cycles run under 12 hours while average protein half-life under stress exceeds 10 hours, leaving only newly transcribed targets and short-lived proteins within reach.
Reported by other laboratories. Small RNA profiling of virus-infected cells found no evidence of virus-derived interfering RNA production in vertebrates, in work the review credits to the Pfeffer and colleagues studies that identified herpesvirus-encoded microRNAs. An independent survey by Parameswaran and colleagues across six RNA viruses recovered small RNAs only at very low concentrations, Dicer independent and without silencing capacity.
Reported by other laboratories. Exceptions to the general rule cluster among DNA viruses. Herpesviruses encode their own microRNAs and antagonize specific host microRNAs. Poxvirus infection leads to degradation of host microRNAs. Adenovirus expresses the 160 nucleotide non-coding RNA VA1, which interferes with small RNA export and RISC loading, work attributed to the Cullen laboratory and others. Hepatitis C virus uses the liver-specific miR-122 to stabilize its genome and mask the 5 prime end from host nucleases, work attributed to the Sarnow laboratory and colleagues. The review classifies these as viral piracy rather than as host defence.
Reported by other laboratories. The engineering of microRNA target sites into viral genomes to control tropism was introduced in lentiviral vectors using the haematopoietic-restricted miR-142, attributed to Brown and colleagues in the Naldini laboratory, and extended to live-attenuated vaccine design in poliovirus by Barnes and colleagues in the Andino laboratory using let-7a and the neuron-specific miR-124. Application to flaviviruses is attributed to the Pletnev laboratory, and to oncolytic vectors to several groups.
Reported by the tenOever laboratory. The same target-site strategy was adapted to influenza A virus to produce a vaccine candidate that is attenuated in mammals but grows to wild-type titres in fertilized eggs, by exploiting a ubiquitous microRNA absent from the allantoic membrane, addressing attenuation and a manufacturing constraint at once, reported by Perez and colleagues.
Reported by the tenOever laboratory. MicroRNA targeting was used as a tool to assign a requirement to a cell type rather than to attenuate. Engineering four miR-142 target sites into dengue virus blocked spread in vivo, which was read as evidence that replication in haematopoietic cells is required for dissemination, reported by Pham, Langlois and tenOever. Targeting influenza A virus in antigen-presenting cells while leaving lung epithelial replication intact led to the conclusion that antigen-presenting cells acquire viral antigen by engulfing non-haematopoietic cells or debris, reported by Langlois and colleagues.
Reported by the tenOever laboratory, with independent corroboration. RNA viruses of both nuclear and cytoplasmic origin can be engineered to produce functional microRNAs without self-cleavage or attenuation, reported by Varble and colleagues and by Shapiro and colleagues from the tenOever laboratory, and by Rouha, Thurner and Mandl independently. In vivo delivery of cytoplasmic RNA virus-derived microRNAs was reported by Langlois, Shapiro, Pham and tenOever. Mechanistic analysis of how hairpins in a cytoplasmic genome are processed, attributed to Shapiro and colleagues from the same laboratory, was interpreted as evidence for a non-nuclear microprocessor activity forming during infection.
Reported by other laboratories. The main liabilities of vector-delivered small RNAs were identified as pathogen-associated molecular pattern generation, since polymerase I or III transcripts carry a 5 prime triphosphate and hairpins present cytoplasmic double-stranded RNA, and saturation of the export and RISC machinery, with fatal shRNA oversaturation in mice attributed to Grimm and colleagues and export bottlenecks noted as resembling the adenovirus VA1 phenotype. Artificial microRNAs delivered as microprocessor substrates are presented as the least toxic route.
Escape from engineered targeting is described as less frequent than anticipated, with no escape mutants recoverable under ubiquitous let-7a pressure in poliovirus, single-nucleotide mismatches but retained silencing under neuron-restricted miR-124 pressure, and outright excision of the target region in dengue virus. The generalization drawn, that multiple targets placed at separate genomic locations mitigate excision, is the review's own interpretation rather than a tested rule.
The review does not establish a mechanism, since it presents no new data. It advances an evolutionary and quantitative account, and the status of that account should be read as author interpretation throughout.
The account runs as follows. Chordates traded an RNA-based antiviral system for a protein-based one built on interferon, and the two appear to have been mutually exclusive, a point the review states remains a mystery. The article notes as supporting circumstance that small interfering RNAs and short hairpin RNAs can themselves induce type I interferon, and that the very features marking viral RNA as foreign, double-stranded character and an exposed 5 prime triphosphate, are also the products of the RNA-dependent RNA polymerase activity that amplifies virus-derived interfering RNAs in other phyla. Because chordate microRNA sequences are fixed in the genome, they are easily evaded, and the three constraints of copy number, silencing capacity and kinetics then explain why they were not retained for defence. The corollary, which is the engineering premise of the second half of the article, is that RNA viruses of chordates experience no selective pressure to disrupt the microRNA machinery and therefore leave it intact and available.
The review is explicit that the counter-evidence is not settled. Reports that the double-stranded RNA-binding proteins of influenza A virus, vaccinia virus and Ebola virus suppress RNA silencing are described as provocative but difficult to weigh, since the same proteins mask pattern recognition and since viruses lacking them are attenuated in wild-type cells but replicate normally in interferon-deficient cells irrespective of microRNA expression. A speculative proposal is also set out in a boxed section, that ubiquitous microRNAs might function as sequence restriction elements analogous to bacterial restriction enzymes rather than as regulators with physiological targets. The review labels this as a hypothesis and proposes growing viruses in microRNA-deficient cells as a way to test it.
The article supplies a framework rather than a result. Its contribution is to convert a negative observation, the absence of virus-derived small RNAs in chordates, into a set of quantitative criteria against which claims of antiviral microRNA activity can be judged, and to press the field toward reporting microRNA abundance as copies per cell rather than as fold change. It also gathers the engineering consequences into a single account, showing that the same fact about chordate biology underwrites two distinct classes of tool. The first is control of viral tropism by grafting target sites, which supports attenuation, vaccine manufacture and the assignment of replication requirements to specific cell lineages. The second is delivery, where viruses engineered to express artificial microRNAs address the solubility, stability and membrane-crossing problems that had limited small interfering RNA therapeutics, with the additional observation from the tenOever laboratory and others that cytoplasmic RNA viruses can serve this role without nuclear trafficking.
The review is written from the tenOever laboratory's own line of work and reads as a statement of its organizing premise, that engineered viruses are instruments for asking biological questions and not only objects of study. Several threads visible here recur across the corpus. Microbe-restricted and tissue-restricted microRNA targeting is used as a genetic tool for cell-type attribution rather than only for attenuation. Small RNA and interferon biology are treated as alternative solutions to the same problem rather than as cooperating systems. Viral genome engineering is treated as a general-purpose capability.
Category 3 synthesis, visible only when this article is set beside other papers in the corpus. The interferon-centred account of chordate antiviral defence that this review takes as given is the same system dissected experimentally in the 2007 Science report on IKKε, where the interferon-stimulated gene set was shown to be internally structured. The review's closing position, that virus design is a route to understanding, is stated in its own right in the later Cold Spring Harbor Perspectives article on synthetic virology. Establishing these continuities requires the other papers and is not supported by this article alone.
The article is a review and presents no new experimental data, so none of its claims are demonstrated here and all of them rest on the cited primary literature. Its central negative conclusion, that chordate microRNAs do not contribute to antiviral defence, is an argument from quantitative thresholds applied to published work rather than a direct test, and the review itself acknowledges contradictory reports that it sets aside on methodological grounds. The copy number threshold of roughly 100 copies per cell is described as a conservative estimate, and the extrapolation to a tenfold higher threshold for viral targets is reasoning rather than measurement. The scope is explicitly RNA viruses of chordates. DNA viruses are treated as exceptions and the review states that the herpesvirus, poxvirus and adenovirus cases are incompletely explained, with the question of why poxviruses are threatened by host microRNAs left open. Latent and chronic infections fall outside the kinetic argument by the review's own statement. The engineering sections describe results from a period when most applications had not reached clinical use, and the article flags unresolved problems, including competing endogenous RNA effects that may preclude the approach for persistent vectors, incompletely explained toxicity from hairpins that should not be pattern recognition receptor substrates, and the need to combine artificial microRNAs with adeno-associated virus vectors in future work. Priority and emphasis language in the source, including its characterization of two profiling studies and several annotated references as first demonstrations, belongs to the review and is not adopted here.
Chordates abandoned small RNA antiviral defence for interferon, leaving their microRNA machinery untouched by RNA viruses and available for engineering viral tropism and delivering designed small RNAs.
Plants and insects fight viruses with small RNAs, but chordates use interferon instead. This review argues that host microRNAs cannot serve antiviral roles in chordates, because they are too scarce, bind too weakly and act too slowly relative to a viral life cycle. Viruses therefore have no reason to disrupt the microRNA machinery, so it stays intact and can be exploited, by grafting target sites into viral genomes and by engineering viruses to deliver designed small RNAs.
Virus-derived interfering RNAs provide antiviral immunity in plants, nematodes and arthropods, while chordates rely on pattern recognition receptors and type I interferon. This review sets out why the chordate microRNA machinery does not substitute for the lost RNA-based defence, arguing from three constraints. Genome-encoded microRNAs rarely reach the abundance needed to act on abundant viral transcripts, partial complementarity limits repression to less than twofold and precludes catalytic cleavage, and the time required to transcribe, process and load a new microRNA exceeds the life cycle of most acute RNA viruses. Published small RNA profiling from several laboratories is consistent with this position, and reported counterexamples are largely confined to DNA viruses and read as viral piracy. The consequence is practical. Because RNA viruses leave the machinery intact, microRNA target sites can be grafted into viral genomes to control tropism and attenuation, and viruses can be engineered to deliver artificial microRNAs.
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