Essential role of long non-coding RNAs in de novo chromatin modifications: the genomic address code hypothesis.

The epigenome, i.e., the whole of chromatin modifications, is transferred from mother to daughter cells during cell differentiation. When de novo chromatin modifications (establishment or erasure of, respectively, new or pre-existing DNA methylations and/or histone modifications) are made in a daughter cell, however, it has a different epigenome than its mother cell. Although de novo chromatin modification is an important event that comprises elementary processes of cell differentiation, its molecular mechanism remains poorly understood. We argue, in this letter, that a key to solving this problem lies in understanding the role of long non-coding RNAs (lncRNAs), a type of RNA that is becoming increasingly prominent in epigenetic studies. Many studies show that lncRNAs form ribonucleoprotein complexes in the nucleus and are involved in chromatin modifications. However, chromatin-modifying enzymes lack the information about genomic positions on which they act. It is known, on the other hand, that a single-stranded RNA in general can bind to a double-stranded DNA to form a triple helix. If each lncRNA forms a ribonucleoprotein complex with chromatin-modifying enzymes on one hand and, at the same time, a triple helix with a genomic region based on its specific nucleotide sequence on the other hand, it can induce de novo chromatin modifications at specific sites. Thus, the great variety of lncRNAs can be explained by the requirement for the diversity of "genomic address codes" specific to their cognate genomic regions where de novo chromatin modifications take place.