Sense- and antisense-mediated gene silencing in tobacco is inhibited by the same viral suppressors and is associated with accumulation of small RNAs.

Antisense-mediated gene silencing (ASGS) and posttranscriptional gene silencing (PTGS) with sense transgenes markedly reduce the steady-state mRNA levels of endogenous genes similar in transcribed sequence. RNase protection assays established that silencing in tobacco plants transformed with plant-defense-related class I sense and antisense chitinase (CHN) transgenes is at the posttranscriptional level. Infection of tobacco plants with cucumber mosaic virus strain FN and a necrotizing strain of potato virus Y, but not with potato virus X, effectively suppressed PTGS and ASGS of both the transgenes and homologous endogenes. This suggests that ASGS and PTGS share components associated with initiation and maintenance of the silent state. Small, ca. 25-nt RNAs (smRNA) of both polarities were associated with PTGS and ASGS in CHN transformants as reported for PTGS in other transgenic plants and for RNA interference in Drosophila. Similar results were obtained with an antisense class I beta-1,3-glucanase transformant showing that viral suppression and smRNAs are a more general feature of ASGS. Several current models hold that diverse signals lead to production of double-stranded RNAs, which are processed to smRNAs that then trigger PTGS. Our results provide direct evidence for mechanistic links between ASGS and PTGS and suggest that ASGS could join a common PTGS pathway at the double-stranded RNA step.

Beta-1, 3-glucanase and chitinase transgenes in hybrids show distinctive and independent patterns of posttranscriptional gene silencing.

Nicotiana sylvestris Speg. & Comes transformed with a tobacco class-I beta-1,3-glucanase (GLU I ) cDNA driven by CaMV 35S RNA expression signals exhibits posttranscriptional gene silencing (PTGS) which is triggered between the cotyledon and two-leaf stages of seedling development and is postmeiotically reset to the high-expressing state during seed development. The incidence of GLU I PTGS in sibling plants differed for the two different transformants tested and increased with the number of T-DNA loci. Comparison of host class-I and class-II beta-1,3-glucanase gene expression suggests that a similarity of 60-70% in the coding-region is required for PTGS of the homologous host genes. The GLU I transformants exhibited a spatial gradient in PTGS, in which expression of the silent phenotype gradually increased in successive leaves toward the bottom of the plant. In contrast, transformants carrying an unrelated tobacco class I chitinase (CHN I) cDNA in the same expression vector exhibited discontinuous patterns of PTGS with adjacent high-expressing and silent leaves. The GLU I- and CHN I-specific patterns were maintained in hybrids homozygous for both T-DNA's indicating that two different transgenes present in the same genome can exhibit independent and distinctive patterns of PTGS. This implies that the nature of the transgene rather than a general pre-pattern of competence for PTGS or propagation of the silent state are important for pattern determination.

Graft transmission of induced and spontaneous post-transcriptional silencing of chitinase genes.

Sense and antisense tobacco chitinase (CHN) transgenes, Luciferase-CHN transcriptional fusions, and promoterless CHN cDNAs were introduced biolistically into CHN transformants of tobacco that never exhibit spontaneous gene silencing. All of the constructs tested induced systemic silencing of the resident CHN transgene and endogenes. Nuclear run-on transcription assays showed that local introduction of additional gene copies triggers systemic post-transcriptional gene silencing (PTGS). Together, this provides evidence that additional transgene copies need not be either highly transcribed or produce sense transcripts to evoke production of systemic PTGS signals. CHN PTGS was transmitted by top grafting, but not by reciprocal grafting of mature stems or the exchange of tissue plugs. Thus, the commonly encountered difficulties in achieving graft-transmission could reflect the method used. Silencing in sense but not antisense transformants was transmitted by grafting to a high-expressing sense CHN scion suggesting that the elaboration of mobile signals may not be an essential feature of antisense-mediated gene silencing.

Stochastic and nonstochastic post-transcriptional silencing of chitinase and beta-1,3-glucanase genes involves increased RNA turnover-possible role for ribosome-independent RNA degradation.

Stochastic and nonstochastic post-transcriptional gene silencing (PTGS) in Nicotiana sylvestris plants carrying tobacco class I chitinase (CHN) and beta-1,3-glucanase transgenes differs in incidence, stability, and pattern of expression. Measurements with inhibitors of RNA synthesis (cordycepin, actinomycin D, and alpha-amanitin) showed that both forms of PTGS are associated with increased sequence-specific degradation of transcripts, suggesting that increased RNA turnover may be a general feature of PTGS. The protein synthesis inhibitors cycloheximide and verrucarin A did not inhibit degradation of CHN RNA targeted for PTGS, confirming that PTGS-related RNA degradation does not depend on ongoing protein synthesis. Because verrucarin A, unlike cycloheximide, dissociates mRNA from ribosomes, our results also suggest that ribosome-associated RNA degradation pathways may not be involved in CHN PTGS.

Silencing of transgenes introduced into leaves by agroinfiltration: a simple, rapid method for investigating sequence requirements for gene silencing.

Agroinfiltration--the infiltration of Agrobacterium tumefaciens into intact plant levels--provides a rapid and simple way of screening large numbers of transgene constructs for silencing in response to a resident transgene. Transgenic Nicotiana sylvestris plants homozygous for the tobacco class I chitinase A gene CHN48 under the control of the cauliflower mosaic virus 35S RNA promoter (P35S) show a high incidence of postranscriptional gene silencing. We forced suspensions of A. tumefaciens, carrying P35S-CHN48 in a binary Tiplasmid vector, into wild-type and transgenic N, sylvestris leaves with a blunt-tipped plastic syringe. The infiltrated CHN48 transgene was expressed in leaves transformed with the vector alone, but not in CHN48-transformed leaves showing the silent phenotype. In contrast, expression of a chimeric P35S-E. coli beta-glucuronidase gene (uidA) infiltrated into leaves was not affected by the presence of the CHN48 transgene stably integrated in the host genome. These results show that extra copies of CHN48 are silenced by resident, silent copies of the same gene and confirm that CHN48 silencing is not the result of promoter interactions. The results also suggest that silencing of the additional CHN48 copies does not require their integration into chromosomes.