SummaryCo.suppreseion of the pigmentation gene chelcone synthese (chs) in Petunia hybrida by chs transgenes leads to white or variegated flowers and is characterized by a reduction in steady-state mRNA levels. To determine the level at which suppression occurs different petunia transformants were analysed containing CaMV-35S RNA promoter-driven hybrid genes consisting of the ~glucuronidase gene (uidA) linked to the full-length chsA cDNA, the 5' half or to the 3' half. With these transgenes one out of 12-15 primary transformants showed suppression of the transgenes and of the resident chs genes throughout the flower or in sectors. The reduction in steady-state chs mRNA was not the result of a transcriptional inactivation event. As determined by nuclear run-on experiments, transcription of suppressed chs genes was similar to that of nonsuppressed genes. This indicates that co.suppression occurs post-transcriptionally. Among individual transformants the transgenes were transcribed at different levels but neither a high nor a low level correlated with a particular degree or pattern of suppression. Surprisingly, even a promoterless chs transgene construct was found to suppress the endogenous chs genes in three out of 15 transformanta. It remains, however, unknown whether or not transcription of the transgene locus is required to induce co-suppression. The data suggest that properties of the chs transgene locus other than the expression level are important for inducing cosuppression. The possible role of antisense RNA, which was detected in all transformants, ectopic pairing and the structure of the integrated T-DNAs in the mechanism of the selective increase in chs RNA turnover are discussed.

Two distinct gene-silencing phenomena are observed in plants: transcriptional gene silencing (TGS), which involves decreased RNA synthesis because of promoter methylation, and posttranscriptional gene silencing (PTGS), which involves sequence-specific RNA degradation. PTGS is induced by deliberate [1-4] or fortuitous production (R.v.B., unpublished data) of double-stranded RNA (dsRNA). TGS could be the result of DNA pairing [5], but could also be the result of dsRNA, as was shown by the dsRNA-induced inactivation of a transgenic promoter [6]. Here, we show that when targeting flower pigmentation genes in Petunia, transgenes expressing dsRNA can induce PTGS when coding sequences are used and TGS when promoter sequences are taken. For both types of silencing, small RNA species are found, which are thought to be dsRNA decay products [7] and determine the sequence specificity of the silencing process [8, 9]. Furthermore, silencing is accompanied by the methylation of DNA sequences that are homologous to dsRNA. DNA methylation is assumed to be essential for regulating TGS and important for reinforcing PTGS [10]. Therefore, we conclude that TGS and PTGS are mechanistically related. In addition, we show that dsRNA-induced TGS provides an efficient tool to generate gene knockouts, because not only does the TGS of a PTGS-inducing transgene fully revert the PTGS phenotype, but also an endogenous gene can be transcriptionally silenced by dsRNA corresponding to its promoter.

SummaryTo induce post.transcriptional silencing of flower pigmentation genes by homologous sense transgenes in transgenic petunias, it is not necessary for the transgenes to be highly transcribed. Even promoterless transgenes can induce silencing. Here it is shown that in these cases silencing is mediated by multimeric transgene/T-DNA loci in which the T-DNAs are arranged as inverted repeats (IRs). With the transgene constructs used, monomeric T-DNA loci are unable to confer silencing even though they modulate IR-induced silencing. IRs with the silencing sequences proximal to the centre (IR c) induce a more severe silencing than IRs with these sequences distal to the centre (IR,). Somatic reversion of silencing, as observed in a side branch of one of the chalcone synthase (Chs) transformants, was associated with a deletion of the IR locus from L1 cells, the meristematic cell layer that expresses the endogenous Chs genes in the flower corolla. Taken together, these data indicate that the post-transcriptional silencing mechanism can be activated by inverted transgene repeats. It is also shown that a silent IR UidA-ChsA locus silences the expression of a monomeric 35S promoter-driven UidA-ChsA transgene only in corollas where the endogenous Chs genes are highly transcribed. These results are consistent with a model in which an IR, by virtue of its palindromic sequence organization, is able to promote the production of aberrant RNAs from the endogenous homologs as a result of ectopic pairing.

The expression of transgenic proteins is often low and unstable over time, a problem that may be due to integration of the transgene in repressed chromatin. We developed a screening technology to identify genetic elements that efficiently counteract chromatin-associated repression. When these elements were used to flank a transgene, we observed a substantial increase in the number of mammalian cell colonies that expressed the transgenic protein. Expression of the shielded transgene was, in a copy number-dependent fashion, substantially higher than the expression of unprotected transgenes. Also, protein production remained stable over an extended time period. The DNA elements are small, not exceeding 2,100 base pairs (bp), and they are highly conserved between human and mouse, at both the functional and sequence levels. Our results demonstrate the existence of a class of genetic elements that can readily be applied to more efficient transgenic protein production in mammalian cells.

Chalcone synthase (CHS) catalyzes the first step in the biosynthesis of flavonoids that function in flower pigmentation, protection against stress, and induction of nodulation. The petunia genome contains eight complete chs genes, of which four are differentially expressed in floral tissues and UV-light-induced seedlings. The 5[prime]-flanking regions of these four chs genes were fused to the [beta]-glucuronidase (GUS) reporter gene and introduced into petunia plants by Agrobacterium-mediated transformation. We show that expression of each construct is identical to the expression of the authentic chs gene, implying that the differences in expression pattern between these chs genes are caused at least in part by their promoters. Histochemical analyses of GUS expression show that chs promoters are not only active in pigmented cell types (epidermal cells of the flower corolla and tube and [sub] epidermal cells of the flower stem) but also in a number of unpigmented cell types (mesophylic cells of the corolla, several cell types in the ovary and the seed coat). Comparison of chs-GUS expression and flavonoid accumulation patterns in anthers suggests that intercellular transport of flavonoids and enzymes occurs in this organ. Analysis of the flavonoids accumulated in tissues from mutant lines shows that only a subset of the genes that control flavonoid biosynthesis in the flower operates in the ovary and seed. This implies that (genetic) control of flavonoid biosynthesis is highly tissue specific.