RNA as a signalling molecule - Intercellular Communication in Plants - page 58

intramolecular base-pairing (Sijen et al. , 1996; Stam et al. , 1997). Moreover, trans-
genes that were deliberately engineered to produce dsRNA caused a very high inci-
dence of co-suppression in tomato and Arabidopsis (Hamilton et al. , 1998; Chuang
&Meyerowitz, 2000), leading to the suggestion that, as in animals, dsRNA was a
critical molecule in the development of PTGS in plants.
These findings, however, did not explain how PTGS was triggered in transgenic
lines expressing single copy, sense transgenes with no intrinsic potential to form
dsRNA, such as the co-suppressed petunias described above. It was proposed that
some single-stranded transcripts produced in these silenced lines had a distinctive,
'aberrant feature' that triggered their conversion into dsRNA through the action of
an elusive plant-encoded RNA-dependent RNA polymerase (RdRp) (Dougherty &
Parks, 1995; Wassenegger & Pelissier, 1998), a suggestion supported by the fact
that RdRp activities had been indeed previously detected in plants (Schiebel et al. ,
1993a,b, 1998). This hypothesis was subsequently confirmed by the identification
of several genes that were specifically required for this 'sense transgene PTGS'
in Arabidopsis. Among those genes, SDE1/SGS2 was found to encode a putative
RNA-dependent RNA polymerase and SDE3 a putative RNA helicase (Mourrain
et al. , 2000; Dalmay et al. , 2001). The current model is that the combined action
of SDE1 and of SDE3 produces dsRNA de novo using transgene-derived, single-
stranded transcripts as templates. The resulting dsRNA then triggers PTGS (Beclin
et al. , 2002). A similar scenario probably accounts for the initial stages of quelling
in Neurospora , since a SDE1 homologue, QDE1, was also found to be required
for this process (Cogoni & Macino, 1999). At present, the presumed 'aberrant
feature' that distinguishes RdRp templates from other cellular sense transcripts
during the initiation of PTGS and quelling remains unclear. It may include the
lack of a polyA tail or inadequate subcellular localization. Transgene dosage has
been often positively correlated with the onset of co-suppression in plants (Pang
et al. , 1996) and this can also be accommodated with the 'aberrant RNA model',
since higher levels of transcription will increase the chances of such RNA being
synthesized and detected in the cell.
3.3.1.3 Short interfering (si)RNAs are the specificity determinants
of RNA silencing
Whether introduced exogenously, produced from inverted repeat constructs or syn-
thesized by cellular RNA-dependent RNA polymerases (RdRps), the dsRNA trigger
of RNAi and PTGS induces degradation of cytoplasmic RNAs that are identical in
sequence. Therefore, both PTGS and RNAi must involve a trans -acting specificity
determinant with a nucleic acid component, whose nature was first investigated in
plants. Since several cases of co-suppression were triggered by sense transcripts and
targeted against sense mRNAs, it was likely that the nucleic acid involved was an
antisense RNA. However, many studies of co-suppressed lines failed to provide ev-
idence for the involvement of a high-molecular-weight antisense RNA (Dougherty
et al. , 1994; Kunz et al. , 1996). It was therefore reasoned that, perhaps, this elusive
antisense molecule was in fact too short to be detected by conventional procedures.