Description
The human genome contains approximately 27,700 CpG islands (CGIs). Most are associated with promoters and their DNA is nearly always unmethylated. By contrast, CGIs lying within the bodies of genes usually become methylated during differentiation and development. CGIs also normally become methylated at X-inactivated and imprinted genes and abnormally methylated in genome rearrangements and in malignancy. In such circumstances, methylation of CGIs is often associated with RNA transcripts reading through these elements but the relationship of this RNA to methylation of CGIs is not clear. Here we investigated a previously described form of a-thalassemia caused by a genome rearrangement leading to abnormal transcription and DNA methylation of the CGI at the promoter of the a-globin gene. We show that transcription per se is responsible for DNMT3B-mediated methylation of the globin CGI, and that this is a general mechanism responsible for methylation of most intragenic CpG islands. Overall design: CapSeq was performed on day 7 in vitro differentiated EBs containing the human gene sequence of RHBDF1 with (RHBDF1+P; chr16:47,861-63,210, hg18) or without (RHBDF1-P; chr16:47,911-60,819, hg18) its promoter in the a recombination mediated cassette exchange (RMCE) system established within the mouse a-globin locus (Lynch et al., 2012, DOI: 10.1038/emboj.2011.399 ) to map transcription initiation sites within the transgene. Please note that the Cap-seq methods captures the 5' end of any short RNA that was Capped, capturing both coding and non-coding RNA.