RNA silencing

Moving targets

Article metrics

Viruses have evolved several strategies to attack plants, but the plants keep hitting back. So the viruses have upped the ante by stopping the plants' immune response from spreading to uninfected tissues.

Plants have developed a variety of weapons in their battle against viral invaders, one of the most elegant of which is 'RNA silencing'. The beauty of this immune response lies in its ability to adapt to all sorts of different viruses, because its specificity is dictated by the sequence of the viral genome itself. Another clever feature is that a signal that triggers silencing is not restricted to individual plant cells, but can spread from the site of infection, generating a response in more distant tissues. But it seems that viruses do not admit defeat easily. Both viruses and these plant-wide ('systemic') silencing signals spread from cell to cell through channels that traverse the cell walls. The spread of the viruses is dependent on viral 'movement proteins'. Writing in Cell1, Voinnet and colleagues show that these movement proteins may also debilitate the systemic arm of the silencing response.

RNA silencing can be triggered in plants by replicating viruses, double-stranded RNA molecules and foreign genes (transgenes) that allow the production of high levels of normal or aberrant messenger RNAs2. Silencing induced in this way by viruses limits the accumulation of the inducing virus, and can also confer immunity in the plant against closely related viruses.

The best model for how RNA silencing works3 suggests that double-stranded segments within an RNA target, such as a viral RNA, are recognized by a plant RNA- digesting enzyme (a 'dsRNase') that contains regions for binding to and untwisting double-stranded RNA. This nuclease cleaves the target RNA to produce segments 21–23 nucleotides in length. These small RNAs associate with the dsRNase or another RNA-degrading nuclease, and confer specificity in further RNA-degradation reactions because they can pair up with complementary single-stranded target RNA.

Viral genomes come in DNA or RNA forms; viruses with RNA genomes may be strong inducers of RNA silencing because double-stranded RNA is formed, at least temporarily, during replication of the genome. The silencing mechanism may halt the replication of the virus in the infected area. Moreover, plant tissues far from the infected site also show specificity in silencing the viral RNAs4. The most plausible explanation for this is that a mobile signal — containing nucleic acids corresponding to the RNA target — spreads from cell to cell.

One strategy that viruses deploy to overcome the silencing response is to move rapidly throughout the plant before silencing can shut them down. Such systemic spread of the virus requires virus-encoded movement proteins. These come in a variety of forms5: some are enzymes, while others have structural roles. Some of them localize to plasmodesmata (the channels connecting plant cells), affect the ability of these channels to open and close, and stimulate transport of themselves and nucleic acids (the viral genome) to the neighbouring cell. In many cases, the functions of the movement protein are split among several proteins. For example, the RNA virus potato virus X (PVX) has three movement proteins — p25, p12 and p8. The conventional view is that these proteins coordinate the movement of the virus to plasmodesmata, the alteration of the channels to accommodate the virus, and transit of the virus through the channels.

This model has merit but — given the finding by Voinnet et al. 1 that p25 is a suppressor of systemic RNA silencing — it may be too limited. The p25 protein is not the first silencing suppressor in a plant virus to be identified6,7,8. But the fact that it is also a movement protein indicates that viral movement might depend both on interactions between movement proteins and plasmodesmata, and on suppression of the plant's silencing mechanism.

Two of Voinnet et al.'s results1 are worth mentioning in more detail here. First, p25 specifically inhibits the production of the systemic silencing signal. This was shown using plants engineered to express a transgene encoding green fluorescent protein (GFP). GFP is a convenient molecule for these sorts of studies because if the GFP transgene is silenced the plants do not glow bright green under ultraviolet light. In Voinnet et al.'s study, silencing of this transgene — both locally and plant-wide — was induced by localized injection of either a PVX–GFP 'recombinant' virus or a highly expressed GFP gene. But when p25 was injected into the lower leaves of the plants simultaneously with either inducer, silencing was blocked in the upper leaves of the plants.

Second, although p25 arrested systemic silencing by either inducer, it inhibited localized silencing induced by the injected GFP gene but not by the replicating PVX–GFP recombinant virus. This surprising result is explained if there are two avenues by which localized silencing can be triggered, only one of which leads to systemic silencing (Fig. 1). One branch is activated by both the injected GFP gene (or rather, its mRNA) and the PVX–GFP viral RNA; this branch is necessary for production of the systemic signal, and is inhibited by p25. Given that both local and systemic silencing are arrested when the injected GFP gene is used together with p25, it seems likely that p25 inhibits this pathway before the silencing signal is even produced. The other branch results only in local silencing; it is activated only by replicating PVX–GFP virus, is not suppressed by p25, and does not lead to production of the systemic signal. So, for replicating PVX–GFP virus, systemic silencing occurs through the p25-sensitive branch, but local silencing occurs through the p25-insensitive pathway.

Figure 1: Plants may have a two-pronged strategy for silencing viruses, as suggested by Voinnet et al.1.

a, The 'weak-inducer' branch. RNAs with limited double-stranded structure — such as those encoded by foreign genes (transgenes) or resulting from replication of a viral genome — feed into this pathway. Silencing depends on a cellular RNA-dependent RNA polymerase (RdRP)-like protein (such as SDE1/SGS2). SDE1/SGS2 produces double-stranded RNA using the single-stranded RNA as a template. This branch also results in the production of a systemic silencing signal and is inhibited by a movement protein, p25, from the potato virus X. b, The 'strong-inducer' branch. Replication of the genome of an RNA virus, using a viral RdRp, produces some replication intermediates with double-stranded-RNA structure. This pathway is not inhibited by p25 and does not lead to systemic silencing. c, Both pathways lead to localized silencing. A 'dsRNase' enzyme cuts up the double-stranded RNAs. The resulting small RNAs associate with the dsRNase or another RNA-degrading nuclease, which cleaves target RNAs that pair up with the small RNAs.

Why do these two pathways exist? It is likely that there are differences in the silence-inducing RNAs formed by the GFP gene and the replicating virus. Expression of the injected GFP gene yields a functional mRNA that is predicted to adopt limited double-stranded RNA structure — such molecules are probably 'weak' inducers of silencing3. Such weak inducers are predicted to be processed through the p25-sensitive, systemic-signal-producing pathway. This pathway may require a cellular RNA-dependent RNA- polymerase-like enzyme, such as SDE1/ SGS2, which is necessary for RNA silencing induced by transgenes but not by several replicating viruses9,10. Replication of the PVX–GFP virus, on the other hand, yields both single-stranded RNA with limited double-stranded structure, and replication intermediates with more double-stranded-RNA structure. The single-stranded viral RNA may enter the 'weak-inducer' pathway. In contrast, the double-stranded RNA formed during replication might be a 'strong' inducer that serves directly as a substrate for the dsRNase in a p25-insensitive manner.

The details of most key steps in this systemic signalling pathway remain to be determined. But one needs to look no further than the viruses to find new entry points into this fast-moving field.


  1. 1

    Voinnet, O., Lederer, C. & Baulcombe, D. C. Cell 103, 157– 167 (2000).

  2. 2

    Meins, F. Plant Mol. Biol. 43, 261–273 (2000).

  3. 3

    Bass, B. L. Cell 101, 235–238 ( 2000).

  4. 4

    Fagard, M. & Vaucheret, H. Plant Mol. Biol. 43 , 285–293 (2000).

  5. 5

    Lazarowitz, S. G. & Beachy, R. N. Plant Cell 11, 535–548 (1999).

  6. 6

    Anandalakshmi, R. et al. Proc. Natl Acad. Sci. USA 95, 13079 –13084 (1998).

  7. 7

    Brigneti, G. et al. EMBO J. 17, 6739–6746 (1998).

  8. 8

    Kasschau, K. D. & Carrington, J. C. Cell 95, 461–470 (1998).

  9. 9

    Dalmay, T. et al. Cell 101, 543–553 (2000).

  10. 10

    Mourrain, P. et al. Cell 101, 533–542 (2000).

Download references

Author information

Correspondence to James C. Carrington.

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.