Abstract
In many plant species, a subset of transcribed genes are characterized by strictly CG-context DNA methylation, referred to as gene body methylation (gbM). The mechanisms that establish gbM are unclear, yet flowering plant species naturally without gbM lack the DNA methyltransferase, CMT3, which maintains CHG (H = A, C, or T) and not CG methylation at constitutive heterochromatin. Here, we identify the mechanistic basis for gbM establishment by expressing CMT3 in a species naturally lacking CMT3. CMT3 expression reconstituted gbM through a progression of de novo CHG methylation on expressed genes, followed by the accumulation of CG methylation that could be inherited even following loss of the CMT3 transgene. Thus, gbM likely originates from the simultaneous targeting of loci by pathways that promote euchromatin and heterochromatin, which primes genes for the formation of stably inherited epimutations in the form of CG DNA methylation.