Sainsbury Laboratory, John lnnes Centre, Norwich, UK

Background and objectives
Post-transcriptional gene silencing (PTGS) is manifest as the reduction in steady-state levels of specific RNAs after the introduction of homologous sequences in the plant genome. Recent findings that viruses can both initiate [1] and be the target of [2] PTGS, even in a non-transgenic background [3], led to the suggestion that PTGS is a mechanism by which plants recognize and reject 'foreign' nucleic acids. The challenge now is to find out what makes a nucleic acid foreign to the plant. Until now, this issue has been difficult to address experimentally because PTGS is inconsistently activated in transgenic plants and is often influenced by complex and poorly understood chromosomal features affecting the transgenic loci [4] . However, we have developed a new approach in which systemic PTGS can be consistently induced by localized, transient expression of a foreign silencer sequence that can be precisely customized. The silencer sequence is provided to the plant cells either by agro-infiltration or by particle bombardment. We refer to this method as systemic induced gene silencing (SIGS). Using transgenic Nicotiana benthamiana carrying the GFP reporter gene as a target of SIGS, we discovered that there is a systemic, sequence-specific signal of gene silencing involved in PTGS [5].

Results and conclusions
Here we show that SIGS can be interpreted as a three-step phenomenon. In the initiation step, a silencer sequence is transferred to a discrete number of cells (less than 10 cells are sufficient to trigger systemic silencing) and leads to the production of a highly sequence-specific signal of gene silencing. The silencer sequence needs to share homology with the coding region of the target gene, but does not need to be transcribed, and can be as short as 200 bp. We show that the signalling molecule is translocated from these cells to the rest of the plant within 3 days. The propagation step, in which the signal is moving in systemic parts of the plant, is independent of the tissues where the initiation took place, and does not require the continuous presence of the target gene. We show that the systemic movement of the signal through the plant is rapid, bidirectional and graft-transmissible. Both long-distance phloem transport as well as local cell-to-cell movement involving plasmodesmata gating are recruited. In the suppression step, cells that receive the signal exhibit PTGS of the target gene and, as a consequence, are resistant to a virus sharing sequence homology with this target gene. The suppression effect is dynamically maintained during the whole plant development and involves self-regeneration of the signalling molecule.

Others have also confirmed production of a systemic signal of gene silencing but, depending on the transgene involved, targeted against GUS or nitrate reductase genes [6]. We are currently attempting to characterize the signalling molecule. Considering its sequence specificity, we infer that the signal is an RNA molecule that moves in the plant through the channels involved in virus or viroid movement.

1. Kumagai MH, Donson J, Delia-Cioppa G et al., 1995. Proceedings of the National Academy of Sciences, USA 92, 1679-1683.
2. English JJ, Mueller E, Baulcombe DC, 1996. Plant Cell 8, 179-188.
3. Ratcliff F, Harrison BD, Baulcombe DC, 1997. Science 276, 1558-1560.
4. Mlynarova L, Jansen RC, Conner AJ et al., 1995. Plant Cell 7, 599-609.
5. Voinnet 0, Baulcombe DC, 1997. Nature 389, 553.
6. Palauqui JC, Eimayan T, Poilien JM, Vaucheret H, 1997. EMBO Journal 16, 4738-4745.