A central tenet of evolution is that small changes in an organism’s genome can be passed on to subsequent generations. Generally, we accept that this happens through the DNA sequences: small, random mutations are inherited by offspring. Indeed, many inherited characteristics, such as fruit color, flower shape, body size, or the direction a snail’s shell whorls are encoded in genes, but they do not always obey the simple laws of Mendelian inheritance. While transposable elements, extragenomic DNA, and—as was the case for the hawkweed that tormented Gregor Mendel himself—parthenogenesis can explain some of these anomalies, recently the spotlight has fallen on another type of inheritance altogether—epigenetic modifications.
Frank Johannes at the University of Groningen in The Netherlands has been trying to understand the intricacies of epigenetic inheritance—specifically, how methylation of DNA bases can contribute to the inheritance of particular characteristics in Arabidopsis. “People are beginning to speculate,” he says. “ ‘ Wait a minute. What’s going on in nature? Does this contribute significantly to adaptation?’ ”
But it’s hard to tell in most natural populations whether inheritance is due to DNA sequence variation or epigenetic changes. “We cannot delineate these two causes very well,” Johannes says. So with his collaborator, Fabrice Roux at the University of Science and Technology in Lille, France, he has been studying a large population of Arabidopsis plants with disrupted methylation patterns. The plants were derived from two Arabidopsis parents with essentially identical genomes, but with one having a mutated DDM1 DNA methylation gene. DDM1 is required for normal methylation—the conversion of cytosine, in cytosine-guanine pairs in the DNA, into 5-methylcytosine—and its mutation reduces genomic methylation by 70 percent.
A team headed by Vincent Colot, now at the École Normale Supérieure in Paris, backcrossed the first generation offspring and selected progeny that were homozygous for the wild type DDM1 gene; in other words, with fully functional methylation machinery. They propagated the plants through a further six rounds of inbreeding, creating “epigenetic recombinant inbred lines” (epiRILs), which carried a mosaic of the parental epigenome. When Roux grew them in a common garden in northern France to subject the almost 6,000 plants to “realistic” ecological selection, they found that the epiRILs yielded plants with distinctly different phenotypes despite being effectively genetically identical.
The segregation and heritability of these traits—which included flowering time and plant height—mirrored those found in naturally divergent Arabidopsis populations, in which phenotypic variation represents adaptations to different environmental conditions. But natural populations have had thousands of years to generate these variations: the epiRILs managed to do it in just eight generations. Andrew Hudson at the University of Edinburgh says there is a clear implication that “DNA methylation and epigenetic changes are important in evolution.”
Johannes explains that there are at least two processes that can influence the epigenome: point mutations in genes that control methylation such as DDM1 that create an additional layer of variation; and environmental impacts that can influence the methylation state, which can then be inherited. New variations of plants, perhaps better adapted to a change in environment, could therefore arise much more quickly than previously thought. Research in this area is “still correlative but nevertheless very interesting,” Hudson says.
But epigenetic changes are not typically as stable as changes in DNA sequence. Some stretches of DNA do remain unmethylated for at least ten generations, Johannes says, but other sequences revert to their “wild type” methylation state due to random fluctuations, or reversion brought about by small RNAs that try to correct the defects. It may be that epigenetic changes could be reinforced by mutations in the DNA, making them stable and heritable in the conventional way. Indeed, some of the sequences affected by the DDM1 mutation are likely to be associated with the mobilization of transposable elements, which would result in immediate—and heritable—DNA sequence changes.
Another complication arises because some traits, particularly those associated with seed production, don’t seem to dabble in epigenetic inheritance. Johannes speculates that there might be an “obscure epigenetic editing process going on” that repairs disadvantageous epigenetic states for crucial genes. “You can imagine,” he says, that there’s “some sort of rescue mechanism,” particularly for gene networks that control a process as important as seed production.
In collaboration with Colot, Johannes hopes to answer such questions by performing genome-wide measurements to establish exactly which genetic elements in the genome are affected by epigenetic perturbation. He already has measurements of genome-wide methylation from more than a hundred individual plants, and is performing what he thinks is the first genome-wide epigenetic linkage study—the epigenetic equivalent of genome-wide association studies of human disease. In parallel, he wants to mathematically model the epigenetic effects and incorporate them into population genetics models, to understand dynamic inheritance patterns that cannot be explained by purely Mendelian genetics.
A Hidden Jewel refers to an article, published in a specialist journal, which has been evaluated in Faculty of 1000, a post-publication peer review service of the Science Navigation Group. Read the evaluation of Johannes’ article.
Correction (10/04/2011): The original version of this article incorrectly stated that Fabrice Roux created the epiRILs used to demonstrate the effects of methylation on the phenotypic variation that can arise in just 8 generations within a population of Arabidopsis plants grown in the same environment. In fact, a team led by Vincent Colot created the plants, and the story has been corrected to reflect this. The Scientist regrets the error.