Epigenetic silencing of plant transgenes

Epigenetics concerns stable but potentially reversible changes in gene expression that occur without a change in nucleotide sequence. Epigenetic controls on gene expression are essential for the normal development of plants and animals, and disruption of these controls can in some cases contribute to human disease. Our group was the first to report on a novel form of epigenetic gene silencing that involves interactions between homologous nucleic acid sequences in somatic cells. Variants of this type of gene silencing have subsequently been identified in diverse organisms and are currently under intense investigation by many groups around the world. Gene silencing effects that rely on recognition of nucleic acid sequence homology appear to reflect host defenses to parasitic sequences, e.g. transposable elements, viruses, viroids. These defenses have apparently been recruited during evolution to establish epigenetic control mechanisms and thus have been instrumental in the evolution of plants and vertebrates. Unwanted epigenetic silencing of transgenes in genetically engineered plant lines is a problem for the agricultural biotechnology industry. To understand the requirements for stable transgene expression in plants, we have carried out the most thorough molecular and cytogenetic analyses to date of numerous genetically well characterized transgene loci in tobacco. This work has led to the identification of several new plant repetitive DNA sequences, most notably sequences derived from a plant virus. Our study is the first showing frequent integration of viral DNA into a plant genome, thus demonstrating the contribution of these sequences to plant evolution.

In different cell types of plants and animals, only a subset of genes is actually active and expressed. The rest of the genetic information must be kept stably silent, otherwise development is impaired or disease can result. To study how genetically silent states are established an maintained, we use the small mustard plant, Arabidopsis as a model system. Arabidopsis appears to use many of the same mechanisms to silence genes as do animals. Therefore, we can learn things that might ultimately be important for human health by studying plants. Our group is interested in a special type of gene silencing, called homology-dependent gene silencing, in which silencing is triggered by interactions between nucleic acids (RNA, DNA) that have a similar sequence. We have uncovered an unanticipated regulatory role for RNA in inducing chemical changes in DNA that silence gene expression. The changes do not lead to alterations of the DNA sequence, but involve attaching methyl groups onto cytosines in DNA. This process is called RNA-directed DNA methylation. From our work on Arabidopsis mutants that are defective in RNA-directed DNA methylation, we are piecing together the pathway of this process. We are also studying the nature of the RNA signal that is able to elicit DNA methylation. We found out that the RNA must initially be double stranded, then processed into tiny RNAs, which we think actually provide the signal for methylation. To study another possible gene silencing mechanism involving DNA-DNA interactions, we have developed a system where we can visualize fluorescent-tagged genes under the fluorescent microscope. In a second aspect of our work, we analyze virus sequences that are part of plant chromosomes. Similarly to the human genome, which is at least 8% virus-derived, many plant genomes contain large amounts of so-called `endogenous viruses`, that is viral sequences that are integrated into plant DNA. We are studying how these sequences contribute to plant genome composition and evolution, and also how they might either make a plant resistant to the free form of the virus (through a gene silencing mechanism?) or, conversely, how normally silent, benign viral sequences can be reawakened to produce symptoms of virus infection. We study these viruses in tobacco and tomato.