In most eukaryotes methylation of histone H3 on lysine 9 (H3K9me) is the key post-translational modification required for the assembly of constitutive heterochromatin at centromeres and other chromosomal regions. H3K9me is bound by the chromodomain proteins HP1/Swi6 and the Suv39/Clr4 H3K9 methyltransferase itself suggesting that, once established, H3K9me might act as an epigenetic mark that can transmit the chromatin state independently of the initiator signal. However, it has not been demonstrated that H3K9me does indeed act as an epigenetic mark. Fission yeast represents an excellent system to address this question since S. pombe lacks DNA methylation and H3K9me is catalysed by the unique, non-essential H3K9 methyltransferase Clr4. To determine whether H3K9me carries epigenetic properties it is important to uncouple H3K9me from genomic domains that have the intrinsic ability to recruit the heterochromatin machinery. One way to solve this problem is to isolate H3K9me from its original context and investigate whether at an ectopic site H3K9me can self-propagate through cell division. To accomplish this, we tethered regulatable TetR-Clr4 fusion protein at euchromatic loci in fission yeast. This resulted in the assembly of an extensive domain of H3K9me-dependent heterochromatin that is rapidly disassembled following TetR-Clr4 release. Strikingly, the inactivation of Epe1, a putative histone demethylase, is sufficient to maintain the silent H3K9me-dependent heterochromatin at the tethering sites through mitotic and meiotic cell divisions in absence of TetR-Clr4. These results indicate that H3K9me acts as an epigenetic mark to maintain heterochromatin domains; however, a regulatory mechanism dependent on Epe1 exists to actively remove H3K9me and thus prevent heterochromatin from being transmitted when assembled at inappropriate regions of the genome.