Over the past few years there has been heightened interest in understanding the ramifications of 'epigenetic' changes to genomes. Such alterations are heritable, non-random and intricate, and occur in nearly all organisms. But they do not directly change DNA sequences. Instead, epigenetic processes include the reversible attachment of certain chemical groups to genomic DNA or to the proteins that package it into a tidy, compact form known as chromatin.
These modifications are essential for gene expression and silencing, and the inactivation of one of the two X chromosomes in female mammals. They are also involved in the development of cancer, and possibly ageing. And, as highlighted on page 212 of this issue1 and in Genes and Development2, they are also required for the regulation of plant transposable elements — the 'jumping genes' discovered by Barbara McClintock. Miura and co-workers1 and Singer and colleagues2 have dramatic evidence that defects in epigenetic modifications lead to transposable elements (transposons) hopping about the genome of thale cress, Arabidopsis thaliana, much more often than usual.
In essence, transposons are discrete segments of DNA that can move from one site in the genome to another within a cell. Genome-sequencing projects have revealed that these mobile elements are likely to occur at a high frequency, and in many forms, in eukaryotic (non-bacterial) organisms, including Arabidopsis3.
Transposons are typically covered with methyl groups, so it has been proposed that this epigenetic modification evolved, at least in part, as a defence mechanism to limit the spread of such 'junk' DNA4. In the delicate balance between the damage that parasitic elements such as these can wreak in their quest to proliferate, and the occasional benefit that transposon movement may afford the host, methylation would provide a flexible means of regulating transposon jumping. Indeed, long-term studies of maize have shown that transposon activity inversely correlates with methylation, which can be altered during different stages of development5, 6. In general, methylation is likely to work by preventing genes from being expressed (transcribed). So, another advantage of methylating transposons would be to reduce the pointless expression of RNA (so-called transcriptional noise) from transposon-encoded genes, most of which are riddled with mutations and so cannot function7.
The results of Miura et al.1 and Singer et al.2 provide striking support for the genome-defence model4. They show that at least two distinct kinds of transposon in Arabidopsis are activated and start jumping specifically in the context of reduced DNA methylation (Fig. 1). Both groups used ddm1 (for 'decrease in DNA methylation') mutant plants, which have genome-wide decreases in methylation. The normal DDM1 protein is similar to some chromatin-remodelling proteins8, so the mutant plants may also have altered chromatin structures, another type of epigenetic modification.
Figure 1: Possible causes and consequences of uncontrolled movement of transposable elements ('jumping genes').
Red and green boxes represent silenced and active transposons, respectively. a, A transposable element in the genome is typically silenced by methylation (addition of CH3 groups) and probably by repressive chromatin structures — certain modifications of the proteins that bundle up the genome (purple ovals). b, When these controls are disrupted — for example, in plants with the ddm1 mutation1, 2 — the transposon is expressed (represented by a wavy line) and jumps about the genome at a much higher rate. c, Possible consequences of the transposon jumping into new sites elsewhere in the genome, such as into or nearby other genes (open rectangles).
High resolution image and legend (66K)Using a predictive bioinformatics approach, Singer et al.2 identified a family of 22 transposons — related to maize transposons called Mutator elements — scattered throughout the Arabidopsis genome. At least one of these transposons readily hopped about the genome of ddm1 mutants. By contrast, Miura et al.1 started by investigating a morphological abnormality that arose spontaneously in an inbred lineage of ddm1 plants. They showed that this defect was caused by a transposon hopping into, and disrupting, a gene encoding a key protein involved in biosynthesis. The transposon was from the CACTA family, so called because the ends of these transposons have the DNA sequence CACTA. Miura et al. found three other related transposons in Arabidopsis, and at least two of the four jumped about avidly in ddm1 plants.
Neither group found appreciable transposon movement in normal plants, indicating that methylation efficiently suppresses these mobile DNA elements. Moreover, a previous study found that the transcription of a different transposon is activated in ddm1 plants9. Similarly, the expression of mouse transposons is activated in methylation-deficient animals10, 11. Collectively, these studies give considerable support to the genome-defence model4. Another implication is that loss of methylation of transposons may also result in increased transcriptional noise, regardless of the consequences for transposon movement.
Like any good experiment, these studies1, 2 raise further questions, some specific and others broader. One is whether the ddm1 mutation has effects on transposons — other than increased transcription — that might contribute to their increased mobility. For example, might it alter the chromatin structure at the transposons' donor or target sites? The biochemical functions of the normal DDM1 gene are not yet certain, so the basis for methylation changes in ddm1 mutant plants remains unclear. Another question is whether mutations in other Arabidopsis genes that alter methylation or chromatin structure will have similar consequences, or perhaps affect only particular subsections of the genome or classes of mobile elements.
Certainly, the new work1, 2 will allow the generation of a practical resource, consisting of transposons sprinkled across the genome. This will be useful for investigating the possible consequences of transposon mobilization (Fig. 1), including gene disruption1, altered gene transcription, chromosomal rearrangements and epigenetic instability12.
It remains to be seen whether other Arabidopsis transposons are mobilized in methylation-deficient mutants. And it will be interesting to find out whether transposons in animal cells are mobilized by disrupting methylation patterns or chromatin structures. If so, this finding might give pause to those contemplating treating patients with drugs that perturb genome-wide epigenetic processes, such as chemotherapeutic drugs intended to 'unsilence' aberrantly hypermethylated and repressed genes in cancers.
These studies open new vistas on transposons and epigenetic controls, which together have shaped eukaryotic genomes through evolution, and provide a useful way of studying jumping genes. They also highlight the destabilizing effects of perturbations in methylation and chromatin structure, illustrating how changes in epigenetic regulation can spiral out of control towards genomic instability.
