Many mammalian genes are arranged in a bidirectional manner, sharing a common promoter and regulatory elements. This is especially true for promoters containing a CpG island, usually unmethylated and associated with an 'open' or accessible chromatin structure. In evolutionary terms, a primary function of genomic methylation is postulated to entail protection of the host genome from the disruption associated with activity of parasitic or transposable elements. These are usually epigenetically silenced following insertion into mammalian genomes, becoming sequence degenerate over time. Despite this, it is clear that many transposable element-derived DNAs have evaded host-mediated epigenetic silencing to remain expressed (domesticated) in mammalian genomes, several of which have demonstrated essential roles during mammalian development.
The current study provides evidence that many CpG island-associated promoters associated with single genes exhibit inherent bidirectionality, facilitating "hijack" by transposable elements to create novel antisense 'head-to-head' bidirectional gene pairs in the genome that facilitates escape from host-mediated epigenetic silencing. This is often associated with an increase in CpG island length and transcriptional activity in the antisense direction. From a list of over 60 predicted protein-coding genes derived from transposable elements in the human genome and 40 in the mouse, we have found that a significant proportion are orientated in a bidirectional manner with CpG associated regulatory regions.
These data strongly suggest that the selective force that shields endogenous CpG-containing promoter from epigenetic silencing can extend to exogenous foreign DNA elements inserted in close proximity in the antisense orientation, with resulting transcription and maintenance of sequence integrity of such elements in the host genome. Over time, this may result in "domestication" of such elements to provide novel cellular and developmental functions.