Skip to main content
Premium Trial:

Request an Annual Quote

New Studies Uncover Antiviral RNAi in Mammals


While plants and invertebrates have long been known to use RNA interference for viral defense, it was unclear whether mammals did the same. Two new papers published in Science this month, however, reveal that at least some mammalian cells use the gene-silencing mechanisms to fend off viruses.

Although there has been evidence of an RNAi-based anti-viral system in mammals, its existence has been a topic of much debate.

For instance, mammalian viral mRNAs are known to be susceptible to siRNA-induced RNAi, and virus-derived RNAs have been identified in mammalian cells infected by RNA viruses. In addition, certain mammalian viral proteins are capable of suppressing insect and plant RNAi, as well as artificially induced RNAi in mammalian cells, a team from the University of California, Riverside wrote in the first paper.

Yet in some infected somatic cells, viral dsRNA triggers a potent and non-specific interferon response that some speculate may have replaced antiviral RNAi activity, Swiss Federal Institute of Technology Zurich scientists noted in the second report. Meanwhile, the viral suppressor of RNAi, or VSR, -like activities displayed by some mammalian viral proteins have yet to be validated in authentic virus-expression contexts.

"Third, diverse virus-infected mammalian cell types accumulate virus-derived small RNAs but these have unspecified functions and lack the biochemical features, size, and distribution patterns of plant and invertebrate viral siRNAs," they added.

To address these uncertainties, the UC Riverside researchers looked to nodamura virus, or NoV, which is mosquito-transmissible and highly virulent to suckling mice and hamsters. Importantly, NoV belongs to the same bipartite positive-strand RNA virus genus as the insect pathogen Flock house virus, which in Drosophila requires expression of the dsRNA-binding protein VSR protein B2 to inhibit Dicer processing of dsRNA viral replication intermediates into siRNAs.

"Because the B2 ortholog of NoV exhibits similar in vitro VSR activities and suppresses experimental RNAi in mammalian cells, we reasoned that use of NoV-delta-B2, a B2-deficient mutant of NoV, to challenge baby hamster kidney 21 cells might facilitate detection of mammalian viral siRNAs," they wrote in their paper.

In two independent studies, the investigators compared deep sequencing profiles of 18- to 28-nucleotide long small RNAs from the kidney cells following inoculation with either NoV or NoV-delta-B2. In those infected with NoV, virus-derived sRNAs were "highly abundant," but showed an overwhelming bias for positive strands, displayed none of the size preference expected for Dicer products, and appeared to breakdown products from the positive-stand viral RNAs.

The virus-derived sRNAs from NoV-delta-B2-infected cells, in contrast, were "much less abundant and exhibited reduced positive-strand bias," according to the UC Riverside team. Meanwhile, around 77 percent of the total negative-strand viral sRNA reads in both libraries were in the 21- to 23-nucleotide size range, similar to Dicer-dependent cellular microRNAs.

NoV-delta-B2 was also found to maintain infection at low levels in the kidney cells, although higher accumulation levels of NoV-delta-B2 were restored in cells engineered with either NoV B2 or a viral protein known to suppress experimental RNAi in mammalian cells.

Together, these findings indicate that NoV-delta-B2 is defective only in RNAi suppression and that the RNAi response induced by NoV-delta-B2, which is characterized by the production of viral siRNAs, has "potent antiviral activity" in the hamster kidney cells, the UC Riverside researchers stated in their paper.

Looking to validate their findings in vivo, the team infected 7-day-old mice with NoV, which is lethal, or NoV-delta-B2. Both viruses spread similarly from the injected abdominal cavity to the fore and hind limbs 24 hours after inoculation, and showed similar accumulation levels.

By day four, NoV RNA levels, while comparable to ribosomal RNA, were more than 1,000 times that of NoV-delta-B2. "Accordingly, unlike the 100 percent mortality observed 5 days post–NoV infection, suckling mice challenged by NoV-delta-B2 remained healthy for the duration of the experiment, up to 4 weeks postinoculation," the UC Riverside researchers wrote.

Additional experimentation indicated that the rapid in vivo clearance of NoV-delta-B2 was not mediated by one of the known innate antiviral pathways. Interestingly, a NoV mutant carrying a single mutation known to abolish B2's VSR activity was nonvirulant at NoV-delta-B2 in the suckling mice and was progressively cleared.

"Thus, in vivo infection and virulence of NoV require the RNAi suppressor activity of B2," according to the Science paper.

Northern blot hybridization studies revealed an accumulation of discrete species of viral siRNAs — particularly 22-nucleotide long ones — in NoV-delta-B2-inoculated suckling mice, mirroring plant and invertebrate hosts after viral infection.

Deep sequencing of sRNAs from the mice indicated no size preference for NoV viral sRNAs, and those 22-nucleotide long RNA detected were not enriched for canonical siRNAs. In contrast, roughly 85 percent of NoV-delta-B2 sRNAs were between 21- and 23-nucleotides long, with 22 nucleotides as the predominant length.

In their study, the ETH-Z group focused on pluripotent mouse embryonic stem cells, or mESCs, which can survive the complete ablation of Dicer or Argonaute functions and support RNAi triggered by long dsRNA as they lack an interferon response.

Several mESC lines were infected with purified forms of encephalomyocarditis virus, or EMCV, a mammalian positive-sense single-stranded RNA picornavirus producing high levels of dsRNA within its eight-hour infection cycle.

All cells accumulated the EMCV-encoded VP1 capsid to different degrees. RNA was isolated from the line with the highest levels and subjected to deep sequencing in two separate experiments. Six hours after infection, 0.4 and 0.7 percent of total reads mapped the EMCV genome, of which 33 and 27 percent were in the 21- to 23-nucleotide long size range.

"For comparison, miR-134, miR-296, and miR-470, which functionally target the mESC pluripotency factors Nanog, Oct4, and Sox2, represented respectively 0.11 percent, 0.02 percent, and 0.05 percent of total reads," the Swiss researchers wrote in their Science paper. The remaining reads — a heterogenous mix in the 24- to 44-nucleotide long range — mapped "nearly exclusively" along the viral positive strand and were most likely viral breakdown products.

In contrast, 36 percent and 28 percent of the 21- to 23-nucleotide reads mapped to both viral strands within the first 200 nucleotides of the EMCV 5' UTR and thus exhibited a roughly 2-to-1 positive-to-negative strand ratio versus the approximately 10-to-1 ratio of all other reads.

"The symmetrical 5' and 3' EMCV reads mapped to the regions where dsRNA replication-intermediates initiate during positive- and negative-strand synthesis," the investigators noted. Similar to replication-intermediate-derived siRNAs in virus-infected plants and invertebrates, abundant positive and negative reads at the EMCV 5' end formed contiguous and perfectly complementary duplexes with 2-nucleotide 3' overhands. Additionally, all EMCV-derived 21- to 23-nucleotide reads defined a dominant, phase register initiated from the 5' end at a roughly 22-nucleotide periodicity, with complementary positive and negative strands offset by 2 nucleotides.

Northern analysis and experiments using Dicer knockout mESCs confirmed that EMCV-derived viral sRNAs were bona fide siRNAs.

In order to examine the antiviral activity of these siRNAs, the researchers turned to NoV, noting in their paper that the genetic rescue of VSR-dependent viruses in RNAi-compromised host cells is not possible with EMCV.

NoV and NoV-delta-B2 were titrated to similar levels in stable B2-expressing baby hamster kidney 21 cells, which were then used to infect mESCs. The latter accumulated "considerably less" than the former after three days, and only the NoV-delta-B2 infect was able to generate virus-derived 21- to 23-nucleotide deep sequencing reads.

"NoV-derived reads, heterogeneous in size, mapped mostly along the RNA1 positive strand," according to the ETH-Z group's paper. In contrast, those from NoV-delta-B2, which are almost exclusively in the 21-to-23 nucleotide range, derived mainly from the 5' and 3' ends of RNAi positive and negative strands — a feature that resembles the NoV-delta-B2 siRNA pattern in fruitflies.

Notably, NoV-delta-B2 5' end reads had a roughly 22-nucleotide periodicity and formed "contiguous, perfect duplexes" with 2- to 3-nucleotide 3' overhangs reminiscent of the Dicer-dependent EMCV siRNAs.

To examine the antiviral RNAi in NoV-infected mESCs, the team used a cell line in which Ago1, 2, 3, and 4 has been knocked out, and in which an ectopically expressed human Ago2 transgene can be shut off using tamoxifen.

Following tamoxifen administration, the cells were infected with either NoV or NoV-delta-B2. In two separate experiments, the NoV and NoVDB2 RNA1 levels were, respectively, approximately two and eight times as high in tamoxifen-induced cell as in untreated mESCs.

"Combined with those obtained with EMCV, the results demonstrate that antiviral RNAi operates in mammalian cells," they wrote.