Drosha acts in the cytoplasm during infection and restricts positive-strand RNA viruses by binding structured RNA rather than by producing small interfering RNAs
lab-led
Noncanonical cytoplasmic processing of viral microRNAs
lab-ledDiscovery
Insertion of a primary microRNA locus into the exclusively cytoplasmic Sindbis virus genome yields mature, functional miR-124 through a Dicer-dependent but microprocessor- and Exportin-5-independent route, defining a cytoplasmic hairpin-processing activity in vertebrate cells that the authors term a virtron.
- Senior authors
- Benjamin R. tenOever
- Correspondence
- Benjamin R. tenOever
Research areas & themes
Citation
Shapiro JS, Varble A, Pham AM, tenOever BR. Noncanonical cytoplasmic processing of viral microRNAs. RNA. 2010. Volume 16, issue 11, pages 2068-2074.
DOI 10.1261/rna.2303610. PMID 20841420. PMCID PMC2957047.
One-sentence contribution
Insertion of a primary microRNA locus into the exclusively cytoplasmic Sindbis virus genome yields mature, functional miR-124 through a Dicer-dependent but microprocessor- and Exportin-5-independent route, defining a cytoplasmic hairpin-processing activity in vertebrate cells that the authors term a virtron.
Executive summary
RNA interference serves as a primary antiviral defence in plants and many invertebrates, while vertebrates are generally understood to rely on interferon-driven innate immunity, with small RNAs confined to microRNA regulation of host transcripts. Canonical microRNA maturation begins in the nucleus with Drosha and DGCR8, requires Exportin-5 for nuclear export, and finishes with Dicer in the cytoplasm. Whether a hairpin that never enters the nucleus can be processed in vertebrate cells, and whether the resulting small RNA can act against the virus that produced it, was unresolved. To test this, the authors grafted the murine mmu-miR-124-2 primary microRNA locus into an additional nonessential subgenomic transcript of Sindbis virus, an alphavirus that replicates only in the cytoplasm. The recombinant virus generated abundant precursor and mature miR-124 at levels comparable to plasmid-based overexpression, and did so without detectable cost to viral protein accumulation or cell infectivity at high multiplicity. Genetic dissection placed the activity downstream of Dicer but showed no requirement for DGCR8, Exportin-5, or type I interferon signalling. The virus-derived miR-124 silenced a sensor transcript bearing complementary target sites and imposed roughly two logs of Dicer-dependent, interferon-independent attenuation on the virus at low multiplicity. The work identifies a cytoplasmic route to functional small RNAs in vertebrate cells and leaves the molecular identity of the processing step open.
Scientific context
The field distinguished two established routes to a pre-microRNA. Canonical primary microRNAs are cropped in the nucleus by the Drosha and DGCR8 microprocessor, exported by Exportin-5, and diced in the cytoplasm. Mirtrons bypass Drosha by arising from very short introns that are spliced, debranched, and refolded, but they remain nuclear in origin and Exportin-5 dependent. Separately, vertebrates were thought to have replaced RNA interference as an antiviral system with pattern recognition receptors and type I interferon. Two observations kept the question open. Artificial microRNA target sites engineered into viral genomes restrict replication in vertebrate cells, and deep sequencing had recovered small RNAs mapping to cytoplasmic RNA viruses, raising the possibility that cells can cleave viral RNA structures directly. The paper positions itself against this background by supplying both the hairpin substrate and its target within a single cytoplasmic virus, removing the need for an RNA-dependent RNA polymerase to generate double-stranded RNA.
Central question
Can a primary microRNA hairpin delivered by a virus that never enters the nucleus be processed into a mature, functional small RNA in vertebrate cells, and if so, which components of the canonical microRNA machinery does that processing require and can the resulting small RNA restrict the virus that encodes it?
Experimental strategy
The design rests on compartmental logic. Sindbis virus clone TE12Q carries a duplicated subgenomic promoter and a unique cloning site, allowing the mmu-miR-124-2 locus to be carried as an extra subgenomic message without disturbing the structural or nonstructural genes. Because Sindbis replication and transcription are strictly cytoplasmic, any mature miR-124 recovered from infected cells must either reflect nuclear trafficking of the viral transcript or a cytoplasmic processing route. miR-124 was chosen because it is neuron-restricted and therefore effectively absent from fibroblasts, giving a clean background, and because its canonical processing and its behaviour in engineered viruses had been characterised previously. Requirement for each biogenesis component was then tested by genetics rather than inhibitors, using fibroblasts lacking Dicer, DGCR8, or IFNAR1, and by pooled siRNA depletion of Exportin-5 in human fibroblasts. Function was assessed with a green fluorescent protein reporter carrying tandem miR-124 target sites in its untranslated region, and antiviral consequence was assessed by low multiplicity multicycle growth curves in the same panel of knockout fibroblasts, so that any attenuation could be assigned to Dicer rather than to interferon or to steric effects on the polymerase.
Key findings
-
Recombinant Sindbis virus carrying the pri-miR-124-2 locus in an extra subgenomic transcript, designated SV124, was rescued to stock titres comparable to the parental virus and produced abundant approximately 60-nucleotide precursor and approximately 20-nucleotide mature miR-124 in human fibroblasts from four hours post-infection through thirty-six hours, at levels comparable to a plasmid-based miR-124 expression construct (Figure 1B). Sindbis core protein levels and near-complete cell infectivity were equivalent between the parental and recombinant viruses at high multiplicity (Figures 1B and 1C). The authors read the absence of a replicative defect as indicating that a viral protein prevents formation of the genomic hairpin or that the viral polymerase resolves the structure during replication, and this interpretation is not tested directly.
-
Production of virus-derived miR-124 required Dicer. In Dicer-deficient murine embryonic fibroblasts both the Sindbis-derived miR-124 and the endogenous miR-93 were lost, while wild-type fibroblasts supported robust synthesis (Figure 2A).
-
Depletion of Exportin-5 by pooled siRNA in human fibroblasts, confirmed by loss of Exportin-5 protein, did not change the level of Sindbis-derived miR-124 (Figure 2B). The authors take this as evidence that the small RNA does not pass through a nuclear intermediate.
-
Cloning and sequencing of more than fifty individual small RNA products from SV124-infected cells showed that about half matched endogenous mature miR-124 exactly and about 44 percent carried some degree of 3 prime heterogeneity (Figure 2C). The authors note that this degree of heterogeneity is comparable to that reported for the mirtrons miR-344, miR-668 and miR-702.
-
Synthesis of Sindbis-derived miR-124 proceeded normally in DGCR8-deficient fibroblasts, whereas endogenous miR-93 was lost in those cells, and it also proceeded normally in fibroblasts lacking a functional type I interferon receptor (Figure 3A). Processing is therefore microprocessor-independent and does not depend on interferon signalling.
-
The virus-derived small RNA was functional. A green fluorescent protein reporter bearing tandem miR-124 target sites was silenced by SV124 infection, and the general reduction in host protein synthesis characteristic of alphavirus infection was enhanced in SV124 relative to parental Sindbis (Figure 3B). The authors interpret the reporter result as post-transcriptional gene silencing by virus-produced miR-124.
-
At low multiplicity, SV124 was attenuated by approximately two logs relative to parental Sindbis in wild-type fibroblasts at forty-eight hours (p equals 0.008). The difference was not significant in Dicer-deficient cells (p equals 0.164) and was reduced to roughly one log but still significant in Ifnar1-deficient cells (p equals 0.015) (Figure 4A). The authors read this as Dicer-dependent, largely interferon-independent self-restriction rather than steric hindrance of the polymerase or increased pathogen-associated molecular pattern production.
-
Supplying miR-124 in trans from a plasmid reduced SV124 core protein by 5.8-fold while leaving parental Sindbis core protein unaffected (Figure 4B), showing sequence-specific targeting of the recombinant virus.
-
Sequence and free-energy analysis of the two viral RNA species indicated that the negative-strand genome carries a perfect miR-124 target with a substantially more favourable hybridisation free energy than the corresponding site on the positive-strand genome, which lacks a seed match longer than six nucleotides (Figure 4C). The authors propose on this basis that targeting occurs on the negative-strand genome, which is inference from sequence rather than a demonstrated site of cleavage.
Mechanistic model
The study does not establish the molecular mechanism of processing, and the authors say so directly, noting that how a capped and polyadenylated cytoplasmic transcript of roughly 500 nucleotides is converted into a precursor hairpin remains unknown and that future work must define it. What the data constrain is the boundary of the pathway. The activity operates on a substrate that is never nuclear, it requires Dicer, it does not require DGCR8 or Exportin-5, and it does not require type I interferon signalling. The product enters the silencing machinery, because a complementary reporter is repressed and because the virus encoding the hairpin is attenuated in a Dicer-dependent manner. The authors group these features under the term virtron and point out that while Dicer dependence with microprocessor independence resembles mirtrons, the Exportin-5 independence, the overall GC content, and the hairpin end structures do not match previously characterised endogenous mirtrons. What the data do not constrain is the identity of the enzyme or complex that liberates the precursor from the subgenomic transcript, whether an endogenous cellular RNA would be handled the same way, and whether the attenuation reflects cleavage of the negative-strand genome as proposed or another route to reduced replication.
Conceptual or technical advance
The work makes a cytoplasmic route to mature microRNA in vertebrate cells experimentally visible and supplies a genetic framework for isolating it, since the substrate can be delivered at will by a cytoplasmic virus and each canonical biogenesis component can be removed independently. It also provides a design principle for engineered viral vectors, showing that a primary microRNA locus can be carried in an alphavirus subgenomic message with little cost to replication at high multiplicity while yielding a functional small RNA, and that the same arrangement produces measurable sequence-specific self-restriction at low multiplicity. The separation of Dicer-dependent restriction from interferon-dependent restriction gives a way to ask how much small RNA activity persists in vertebrate cells independently of the interferon system.
Relationship to the broader research program
The paper belongs to a line of work in the laboratory on engineering RNA viruses to carry and produce small RNAs, and on the use of microRNA targeting as a tool for species-specific or tissue-specific attenuation of virus replication. The immediately preceding work from the group established that influenza A virus can be engineered to synthesise microRNAs and that microRNA target insertion attenuates influenza A virus in a species-specific way. The present study extends that framework from a segmented negative-sense virus that replicates in the nucleus to a positive-sense virus that is strictly cytoplasmic, which is what makes the compartmental argument possible. The recurring question the paper feeds is whether small RNA activity in vertebrates retains any antiviral role alongside interferon. The authors are cautious on this point, arguing that because virtrons entail self-targeting they are unlikely to be genuine viral products and would more plausibly be cellular by-products of infection, and stating that establishing whether they constitute an antiviral pathway requires substantially more work.
Related publications
- Varble et al. 2010 (PNAS), methodological foundation. Established engineered RNA viral synthesis of microRNAs and provided the pri-miR-124-2 construct and the GFP_124t reporter used here.
- Perez et al. 2009 (Nature Biotechnology), predecessor. Established microRNA-mediated species-specific attenuation of influenza A virus, the targeting logic that the self-targeting experiments here build on, and supplied constructs cited in the methods.
- Perez et al. 2010 (PNAS), companion. Influenza A virus-generated small RNAs and their regulatory role, from the same laboratory in the same period, and the source of the small RNA Northern protocol used here.
- tenOever 2009 (Discovery Medicine), review or synthesis. The author's own framing of microRNA-based manipulation of viral therapeutics, cited in the introduction as the basis for the vestigial-pathway argument.
Limitations and boundaries
The findings rest on immortalised fibroblast lines from human and mouse and on a single engineered alphavirus carrying a single microRNA locus, so the generality across cell types, primary cells, tissues, and other cytoplasmic viruses is untested. No animal work is included. The attenuation phenotype appeared only at low multiplicity, and the authors note explicitly that rapid replication at high multiplicity in immortalised fibroblasts may have masked it, which means the magnitude of any effect is regime-dependent. Exportin-5 was removed by pooled siRNA knockdown rather than by genetic deletion, so residual protein cannot be excluded even though the immunoblot showed effective loss. The proposed site of targeting on the negative-strand genome rests on free-energy and seed-match analysis, not on direct mapping of cleavage. The processing enzyme responsible for generating the precursor from the subgenomic transcript is not identified. The system supplies the hairpin artificially, so the study does not show that natural infections generate comparable structures, and the authors themselves frame natural virtrons as hypothetical. Finally, the silencing observed was described as reduced relative to canonical microRNA activity, which bounds how much regulatory weight the pathway can carry.
Audience summaries
25 words
A hairpin carried by a virus that stays in the cytoplasm still becomes a working microRNA, needing Dicer but not the nuclear microRNA machinery, and restricting the virus.
75 words
Vertebrate microRNAs are normally cut first in the nucleus and only finished in the cytoplasm. Putting a microRNA gene into Sindbis virus, which never enters the nucleus, still produced mature miR-124. The process required Dicer but not DGCR8, Exportin-5, or interferon signalling. The small RNA silenced a reporter and cut the virus back by about two logs in a Dicer-dependent way, indicating an uncharacterised cytoplasmic route to functional small RNAs.
150 words
Canonical microRNA biogenesis requires nuclear cropping by Drosha and DGCR8 followed by Exportin-5-dependent export and cytoplasmic dicing. Shapiro and colleagues inserted the mmu-miR-124-2 primary locus into an extra subgenomic transcript of Sindbis virus, an alphavirus restricted to the cytoplasm, and recovered abundant precursor and mature miR-124 at levels comparable to plasmid overexpression without measurable replicative cost at high multiplicity. Genetic tests in knockout fibroblasts placed the activity downstream of Dicer and showed no requirement for DGCR8, Exportin-5, or type I interferon signalling, while small RNA sequencing revealed 3 prime heterogeneity comparable to that of known mirtrons. The product was functional, silencing a target-bearing reporter and attenuating the virus by roughly two logs at low multiplicity in a Dicer-dependent, largely interferon-independent manner, with plasmid-supplied miR-124 reducing SV124 protein 5.8-fold. The authors name these products virtrons and state that the responsible processing mechanism remains undefined.
Discoveries supported by this paper
Documented publication relationships
- Engineered RNA viral synthesis of microRNAs — methodological foundation.
- MicroRNA-mediated species-specific attenuation of influenza A virus — predecessor.
- Degradation of Host MicroRNAs by Poxvirus Poly(A) Polymerase Reveals Terminal RNA Methylation as a Protective Antiviral Mechanism — methodological foundation.
- In Vivo Delivery of Cytoplasmic RNA Virus-derived miRNAs — predecessor.
- A Small-RNA Enhancer of Viral Polymerase Activity — methodological foundation.
- Evidence for a cytoplasmic microprocessor of pri-miRNAs — predecessor.
- RNA viruses and the host microRNA machinery — methodological foundation.
- An In Vivo RNAi Screening Approach to Identify Host Determinants of Virus Replication — methodological foundation.
- A Versatile RNA Vector for Delivery of Coding and Noncoding RNAs — predecessor.
- Drosha as an interferon-independent antiviral factor — predecessor.
- microRNA Function Is Limited to Cytokine Control in the Acute Response to Virus Infection — predecessor.
- RNase III nucleases from diverse kingdoms serve as antiviral effectors — methodological foundation.
- SARS-CoV-Encoded Small RNAs Contribute to Infection-Associated Lung Pathology — predecessor.