There is some excitement in ecology and evolution that epigenetic variation could speed up plant adaptation to novel environments in the face of rapid global change. However, it is difficult to distinguish between adaptation based on standing genetic vs. epigenetic variation. Furthermore, it is unclear how epigenetic variation could accumulate or be maintained in the absence of genetic variation. We used a novel reduced-representation bi-sulfite sequencing method to assess DNA-sequence and -methylation differences between populations of five grassland species selected in monoculture or mixture (1). Due to limitations of this method, we also used whole-genome analysis to test for signs of epigenetic adaptation within two genetically homozygous lines of Arabidopsis thaliana subject to selection for increased dispersal ability in dynamic landscapes (2). Finally, we imposed environmental stress to offspring of a single homozygous Arabidopsis plant and used artificial selection for small vs. large plants in 150 replicate populations over three generations to test if selectable epigenetic variation would accumulate over this short time period.
Results/Conclusions
We found that multiple species showed rapid evolutionary changes in both methylation and DNA sequences in response to diversity-driven selection (1). But epigenetic differences were likely caused by genetic effects, including effects of genes not included in the reduced-representation sequencing. Nevertheless, in the case of selection within homozygous Arabidopsis lines (2), we could demonstrate heritable phenotypic changes towards increased dispersal in response to selection, in the absence of correlated genetic changes. These phenotypic changes could be linked to changes in methylation patterns and expression levels in genes related to flowering. We suggest that this epigenetic selection response was possible due to reminiscent epigenetic variation after an old hybridization event between different genotypes, whereby the gene(s) causing the epigenetic variation had been lost in the course of repeated selfing and thus epigenetic marks were lost after diverse numbers of generations. We consider it less likely that epigenetic variation accumulated during the selection in dynamic landscapes, because we were unable to detect any effects of artificial selection among the large number of offspring of the single homozygous plant in (3), after three generations. Our results suggest that, even though rates of epimutations are much higher than rates of genetic mutations, it still requires large populations and many generations to accumulate an amount of standing epigenetic variation that can play a role in plant adaptation to rapid environmental change.