Induction and Suppression of RNA Silencing by an Animal Virus

See allHide authors and affiliations

Science  17 May 2002:
Vol. 296, Issue 5571, pp. 1319-1321
DOI: 10.1126/science.1070948


RNA silencing is a sequence-specific RNA degradation mechanism that is operational in plants and animals. Here, we show that flock house virus (FHV) is both an initiator and a target of RNA silencing in Drosophila host cells and that FHV infection requires suppression of RNA silencing by an FHV-encoded protein, B2. These findings establish RNA silencing as an adaptive antiviral defense in animal cells. B2 also inhibits RNA silencing in transgenic plants, providing evidence for a conserved RNA silencing pathway in the plant and animal kingdoms.

Posttranscriptional gene silencing, quelling, and RNA interference (RNAi) are mechanistically related RNA silencing processes that destroy RNA in a sequence-specific manner (1, 2). Available data show that double-stranded RNA (dsRNA) serves as the initial trigger of RNA silencing and, after recognition, is processed by the Dicer RNase into short fragments of 21 nucleotides (nt) in length. These short interfering RNAs (siRNAs) are then incorporated into a dsRNA-induced silencing complex (RISC) to guide cycles of specific RNA degradation (1, 2). Here, we report that RNA silencing plays a natural antiviral role in animal cells, as has been established in plants (3, 4).

We focused on the flock house virus (FHV) because itsB2 gene (see fig. S1) shares key features, but not sequence similarity, with the plant cucumoviral 2b gene (5), which encodes a known group of silencing suppressors (6, 7). Both open reading frame (ORF) 2b and B2 overlap the carboxyl terminal region and occupy the +1 reading frame of the ORF encoding the viral RNA-dependent RNA polymerase and are translated in vivo by a subgenomic mRNA (5).

The FHV B2 protein indeed exhibited a potent silencing-suppression activity (Fig. 1) in theAgrobacterium co-infiltration assay (8), established in transgenic plants that express green fluorescent protein (GFP). Transient B2 expression prevented RNA silencing of the GFP transgene, leading to a strong and prolonged green fluorescence examined under ultraviolet (UV) illumination (Fig. 1, left), similar to suppression by the cucumoviral 2b proteins (9) (Fig. 1, right). In contrast, a broad red fluorescent zone surrounding the infiltrated patch (Fig. 1, middle) became clearly visible 6 days after infiltration, when the co-infiltrated transgene directed translation of neither 2b nor B2.

Figure 1

Cross-kingdom suppression of RNA silencing in plants by an animal viral protein. The GFP-expressingNicotiana benthamiana leaves were co-infiltrated with a mixture of two Agrobacterium tumefaciens strains, as described (8, 9). One directs the expression of GFP and thereby induces GFP RNA silencing, and the other simultaneously expresses the FHV-encoded B2 (left leaf), B1 (middle leaf), or the plant cucumoviral 2b (right leaf). The leaves were detached and photographed under UV illumination 6 days after infiltration. GFP silencing is visualized in the middle leaf as a bright red color zone surrounding the infiltrated patch caused by chlorophyll fluorescence.

RNA blot hybridizations confirmed that expression of either protein was associated with high accumulation levels of the GFP mRNA (see fig. S2). In addition, the GFP-specific siRNAs, a hallmark of RNA silencing (10), remained at extremely low levels in the leaves where there was expression of either B2 or 2b (see fig. S2). We further demonstrated that B2 was able to functionally substitute for 2b of cucumber mosaic virus (CMV) in whole plant infections (see methods sections of online material), as found previously for a CMV 2b homolog (11). B2 suppression of RNA silencing in plants explains why FHV is able to overcome the RNA silencing defense and establish systemic infections in transgenic plants that express a plant viral protein that facilitates virus cell-to-cell movement (12).

Our finding that an FHV-encoded protein suppresses RNA silencing in plants suggests a role for RNA silencing in FHV infections of animal hosts. FHV belongs to the Nodaviridae family, members of which naturally infect vertebrate and invertebrate hosts, and Drosophila cells support complete infection cycles of FHV (13). We found that infection of DrosophilaS2 cells with FHV virions resulted in a rapid appearance of the FHV-specific siRNAs of both positive (Fig. 2A) and negative polarities. Accumulation of the siRNAs trailed that of FHV genomic and subgenomic RNAs (Fig. 2C), which suggests that the decreased accumulation of FHV RNAs at later stages of FHV infection (14) (Fig. 2B) may be caused by an FHV-specific RNA silencing.

Figure 2

Induction and suppression of RNA silencing in Drosophila by FHV. (A toC) A time course analysis is shown on the accumulation of FHV siRNAs (A) and RNAs 1 to 3 (B) in S2 cells infected with FHV virions, and densitometry measurements of the accumulation levels of FHV RNA1 and siRNA are shown in (C). An RNA marker of 22 nt in length transcribed in vitro was loaded in the right lanes (A). (D) Accumulation of FHV RNAs in S2 cells transfected with pRNA1 or pRNA1-ΔB2, with or without dsRNA. dsRNA corresponding to mRNA of cyclin E (cycE), GFP, and two fly Dicer (DCR) genes; to the 5′ and 3′-terminal 1000 nt of the AGO2 mRNA (16, 17,23); or to the 3′-terminal 500 nt of FHV RNA1; and a B2-expressing plasmid that was co-transfected into S2 cells are indicated above each lane.

To investigate this possibility, we constructed a full-length FHV RNA1 cDNA clone (pRNA1) (see methods sections of online material), which, after transfection into S2 cells, directed RNA1 self-replication and transcription of RNA3 (15), the subgenomic mRNA for B2 (Fig. 2D, lane 2). We found that depleting the mRNA for Argonaute2 (AGO2) by RNAi, an essential component of the RISC complex (16), led to a pronounced increase (two- to threefold) in the accumulation of FHV RNAs 1 and 3 (Fig. 2D, lanes 6 to 8), whereas co-transfection of cyclin E or GFP dsRNAs with pRNA1 had minimal effect (Fig. 2D, lanes 4 and 5), indicating that a functional RNA-silencing pathway naturally restricted FHV accumulation in the host cells. Furthermore, co-transfection of pRNA1 with a dsRNA targeting the 3′-terminal 500 nucleotides of FHV RNA1 completely prevented the accumulation of intact FHV RNA1 in S2 cells (Fig. 2D, lane 3). These results collectively demonstrate that FHV is both an initiator and a target of RNA silencing in this animal host.

Further studies showed that B2 was essential for FHV accumulation in Drosophila cells, which is in contrast to a previous study carried out in nonhost mammalian cells (15). A B2-knockout mutant of FHV RNA1, referred to as RNA1-ΔB2 (see methods sections of online material), which contains the previously described point mutations (15) that converted the first and 58th codons of the B2 ORF into serine and stop codons, respectively, failed to accumulate to detectable levels after transfection into S2 cells (Fig. 2D, lanes 12 and 20). This defect was partially trans-complemented (up to 10% of the wild-type level) by co-transfection of a plasmid expressing either B2 (Fig. 2D, lanes 13 and 21) or a His-tagged B2 (Fig. 2D, lane 22). Expression of the His-tagged B2 from the co-transfected plasmid was detected in S2 cells by Western blot analysis using an antibody recognizing the His tag. Reverse transcription–polymerase chain reaction and sequencing revealed that the introduced mutations were stably maintained in the progeny FHV RNAs extracted from infected cells, indicating that B2 was indeed expressed from the co-transfected plasmid rather than from a revertant RNA1.

Accumulation of RNA1-ΔB2 in S2 cells was efficiently rescued, up to 50% of the wild-type level, by co-transfection with the AGO2 dsRNAs, either singly (Fig. 2D, lanes 14 and 15) or in combination (Fig. 2D, lane 16). However, co-transfection with dsRNAs targeting mRNAs of the two Drosophila Dicer genes (17) was not effective (Fig. 2D, lane 17) under the same conditions. This is possibly due to a more efficient mRNA depletion by RNAi for AGO2 (Fig. 2D, lanes 14 to 16) than for Dicer (16, 17), which is required to process the input dsRNA. Notably, the level of complementation by RNAi of AGO2 (Fig. 2D, lanes 14 to 16) was higher than that achieved by the B2-expressing plasmid (Fig. 2D, lane 13), although was still achieved less efficiently than B2 expression from wild-type RNA1 (Fig. 2D, lane 10). Therefore, in the absence of B2 expression, FHV RNAs 1 and 3 accumulated to substantial levels when the RISC complex was disrupted by AGO2 depletion. These data confirmed the previous finding (15) that B2 is not required for RNA1 self-replication and indicate that the essential function of B2 for FHV infection of the S2 host cells observed in this study was to suppress RNA silencing that targeted FHV RNAs for degradation. Thus, the same protein blocks RNA silencing in both animals and plants, providing the first experimental evidence for a highly conserved RNA silencing pathway in different kingdoms.

It is known that RNA silencing operates in animals, including mammals (1, 2, 18). In this work, we demonstrate that infection of Drosophila cells with an RNA virus triggers strong virus RNA silencing and that the same virus is equipped with an effective silencing suppressor essential for infection. These data provide direct evidence that RNA silencing naturally acts as an adaptive antiviral defense in animal cells. The specificity mechanism of this adaptive defense is based on nucleic acid base pairing between siRNA and its target RNA (1,2) and thus is distinct from cellular and humoral adaptive immunity based on peptide recognition (19). A prediction from our work is that heterologous sequences inserted into a replicating virus genome will lead to the production of a population of siRNAs capable of silencing other viral and cellular RNAs in trans that are homologous to the insert. Indeed, recent studies showed that viral sequences inserted in alphavirus vectors give rise to virus resistance in mosquitoes, which is dependent on the inserted RNA sequence rather than on its protein product (20, 21). It will be of interest to determine if RNA silencing also plays a role in observed protection against mammalian viruses, derived similarly from heterologous expression of RNA sequences from a replicating RNA virus vector (22).

  • * To whom correspondence should be addressed. E-mail: shou-wei.ding{at}


View Abstract

Navigate This Article