PerspectivePlant Science

Viruses Face a Double Defense by Plant Small RNAs

See allHide authors and affiliations

Science  07 Jul 2006:
Vol. 313, Issue 5783, pp. 54-55
DOI: 10.1126/science.1130818

All organisms, from bacteria to mammals, can be infected by viruses. Bacteria counter infection with endonucleases that cleave viral DNA. Mammals fight infection with antibodies and lymphocytes that are adapted to specific viral antigens. They also employ non-adaptive defenses, such as producing interferons that block viral replication and stimulate the host immune response, and launching an editing attack on viral RNA (adenosine deaminases acting on RNA). However, none of these defense mechanisms has been found in plants. Instead, plants appear to rely heavily on an adaptive RNA degradation system (most plant viruses have RNA genomes) similar to that mediating RNA interference (a gene silencing mechanism) in animals. On page 68 of this issue, Deleris et al. (1) reveal new molecular insights into how viruses both trigger and suppress this antiviral RNA silencing mechanism in host plants.

Plant antiviral defense.

Pathways directed by Dicer-like (DCL) enzymes interact with viral RNAs and are affected by viral suppressor proteins. Viruses are repressed mainly through DCL4. However, Turnip crinkle virus encodes a suppressor protein that inhibits DCL4 activity. In this situation, the 22-nucleotide siRNAs produced by DCL2 become the major directors of antiviral activity.

The plant antiviral response involves Dicers, enzymes similar to ribonuclease III, which cleave viral double-stranded RNA into smaller pieces (∼21 to 24 nucleotides in length) called short interfering RNA (siRNA) duplexes (see the figure). One strand of each duplex is incorporated into a large ribonucleoprotein complex called the RNA-induced silencing complex (RISC). Guided by the associated siRNA, RISC recognizes and destroys complementary target viral RNAs (2). In animals, and probably all multicellular eukaryotes, Dicer is also involved in the production of microRNAs, which are very similar to siRNAs but generated from endogenous, partially self-complementary precursor RNAs. These microRNAs direct RISCs to silence the expression of key genes during important developmental transitions (3). From the available complete genome sequences, it seems that whereas metazoans and fungi have only one or two Dicer genes, plants have multiplied and modified them over the past billion years to produce a basic set of at least four different Dicer-like (DCL) variants of the enzyme (4). In the model plant Arabidopsis thaliana, one DCL produces microRNAs and another makes siRNAs that direct chromatin modification. Deleris et al. show that the remaining two DCLs combine to provide an antiviral defense system that itself becomes a target of viral counterdefense mechanisms.

When a virus infects a plant, the partially self-complementary structures of its genomic RNA, double-stranded RNA intermediates of its replication, and/or double-stranded RNA produced from the viral RNA by host polymerases, become substrates for the production of siRNAs (5). These siRNAs guide RISCs against invading viral RNA. Unsurprisingly, most viruses have evolved ways to combat this defense pathway. Many viral genomes encode suppressor proteins that bind to the siRNA duplexes and prevent their incorporation into RISCs (6). Others produce suppressors that act directly on the enzymes or cofactors of this defense pathway (7).

Since the first detection of siRNAs in plants (8), their sizes and the relevance of the size differences have been debated. The current view is that healthy plants generate 21-nucleotide microRNAs and siRNAs from their own RNA, with DCL1 and DCL4, respectively, to regulate gene expression during development. DCL3 generates 24-nucleotide siRNAs to modify chromatin. But the size profile of viral siRNAs varies depending on the virus. Many viruses induce the production of mostly 21-nucleotide siRNAs. Others cause the plant to make a mix of 21- and 24-nucleotide siRNAs. And other viruses trigger the generation of predominantly 24-nucleotide or exclusively 22-nucleotide viral siRNA species. So, why are the different viruses treated differently by plants? Deleris et al. show that the answer has a lot to do with DCL redundancy and the action of viral suppressor proteins on DCL-directed activity.

Arabidopsis infected with Tobacco rattle virus (TRV) generates virus-specific 21- and 24-nucleotide siRNAs, and when infected by Turnip crinkle virus (TCV), it produces only 22-nucleotide siRNAs. However, these size profiles change in different Arabidopsis Dcl-mutant backgrounds. The TCV 22-nucleotide siRNAs produced by DCL2 direct destruction of the virus, whereas the TRV 21- and 24-nucleotide species are generated by DCL4 and DCL3, respectively, with only the 21-nucleotide siRNAs directing destruction of the virus. Why is this and how is it done? Deleris et al. show that TCV RNA would be targeted by DCL4-produced siRNAs if it were not for the virus expressing a suppressor protein that blocks DCL4 activity. The contrary situation exists for TRV. If DCL4 activity is eliminated from the plant by mutation, RISC is guided against TRV by 22-nucleotide siRNAs produced by DCL2. Both viruses replicate to greater levels in a double dcl2-dcl4 mutant, showing that even though the siRNAs from one of the DCLs predominates, both DCLs are producing siRNAs that restrict the virus. However, DCL4 appears to be the first line of antiviral defense, with DCL2-mediated activity coming to the fore when DCL4 is deactivated by a viral suppressor protein.

Have plants duplicated DCLs as a response to viral suppressors? Mammals have only one Dicer plus a suite of other viral defense systems, and it is currently debated whether their RNA interference pathway has an antiviral role (9). Insects have fewer antiviral defense systems than mammals but have two Dicers, and a number of insect viruses encode suppressors that inhibit these Dicers or their pathways (10, 11). Plants such as rice and poplar have even more DCLs than Arabidopsis, which are probably involved in antiviral defense (4). Other plant viruses, such as citrus tristeza, encode three different suppressor proteins to shut down the DCL-mediated defense responses (12). All this suggests that the RNA silencing-based antiviral defense pathway is increasingly important to eukaryotes that have fewer alternative antiviral protection systems, and that there is an escalating arms race between viruses and their plant hosts.

The study raises the question: How can a viral suppressor protein inhibit the activity of a specific DCL in the production of viral siRNAs? In this case, the answer may reside in the RNA-binding capacity of the viral suppressor protein P38, which can bind to long as well as short (21- to 26-nucleotides) double-stranded RNA (13). P38 may be competing with DCL4, blocking DCL4's access to long double-stranded viral RNA molecules. Alternatively, P38 might directly interact with DCL4, affecting its activity, but not that of DCL2.

The complexity of small RNA-mediated pathways in plants, and the roles within them played by each of the DCLs, is becoming ever more apparent. A major challenge will be to determine the subcellular localization of the pathway components and details of their trafficking to these locations. DCL1, DCL3, and DCL4 have all been detected exclusively in the nucleus (14, 15), which is consistent with the production of microRNAs and siRNAs from host transcripts. However, the demonstration by Deleris et al. that all of these enzymes have access to RNAs from viruses thought not to enter the nucleus suggests that either DCLs or viral RNA substrates are being trafficked between the cytoplasm and nucleus.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
View Abstract

Stay Connected to Science

Navigate This Article