What might the systemic signal be? The evidence is still a bit indirect, but favors the idea that the signal is an RNA, be it double or single stranded.
1. Baulcombe’s lab reported that a promoter is not required to initiate gene silencing by a gene fragment, although there is data from other laboratories suggesting that transcription is necessary. Curiously enough, non-overlapping gene fragments can induce cross-silencing. That is, a gene fragment can confer silencing on a viral vector containing a non-overlapping gene fragment. This suggests that the silencing is mediated by an interaction between the introduced sequence and either the resident gene itself or its transcript. Since there is ample evidence that silencing is sequence specific, the logical intermediary is the gene transcript that overlaps the two fragments.
2. So it is currently hypothesized that the signal is either an RNA degradation product or double-stranded RNA. Peter Waterhouse’s lab recently reported that transgenes comprising inverted repeats of gene sequences are particularly efficient silencing signals, but only if they have a promoter. They therefore suggest that double-stranded RNA might be the trigger.
3. And there are other clues. It has long been known that plants encode RNA-dependent RNA polymerases and their presence has been a continuing mystery. A recent exciting discovery in Neurospora is that RNA “quelling”, which strongly resembles post-transcriptional gene silencing in plants, requires a gene with homology to RNA-dependent RNA polymerase.
4. Unfortunately, the most straightforward test of the hypothesis, the induction of systemic silencing with double-stranded RNA, has not yet been successful in plants, although double-stranded RNA induces gene-specific silencing in C. elegans and in Drosophila.