Intragenic CpG Islands and Their Impact on Gene Regulation

Abstract

The mammalian genome is depleted in CG dinucleotides, except at protected regions where they cluster as CpG islands (CGIs). CGIs are gene regulatory hubs and serve as transcription initiation sites and are as expected, associated with gene promoters. Advances in genomic annotations demonstrate that a quarter of CGIs are found within genes. Such intragenic regions are repressive environments, so it is surprising that CGIs reside here and even more surprising that some resist repression and are transcriptionally active within a gene. Hence, intragenic CGI positioning within genes is not arbitrary and is instead, selected for. As a wealth of recent studies demonstrate, intragenic CGIs are embedded within genes and consequently, influence 'host' gene mRNA isoform length and expand transcriptome diversity.

Keywords: CpG island (CGI); DNA methylation; alternative polyadenylation (APA); epigenetics; mRNA processing; orphan CpG-Islands; polyadenylation.

FIGURE 1

Definitions and states of CpG islands. (A) Depiction of the typical CpGs in the mammalian genome which are DNA methylated in isolation but are devoid of this mark in the CGI context. CGIs were originally defined bioinformatically. (B) Schematic demonstrating CGIs locations were biochemically determined. MBD and CXXC proteins fixed to a Sepharose column allowed purification of DNA methylated and unmethylated CGIs in mammals. (C) Representation of CGIs across mammalian genomes and a table, summarising the proportion and methylation status of CGIs as reported by Illingworth et al. (2010) in mouse and human across TSS, Intragenic and Intergenic regions. Total CGI numbers in mouse = 23,021, human = 25,495 (D) Summary of the main states of CGIs across the genome. Their active state is associated with binding of transcription factors (TF) and subsequent RNA Polymerase II (Pol II) binding. Repressed states of CGIs are through combinations of DNA methylation, H3K4me3 and H3K27me3.

FIGURE 2

Schematics of how iCGIs impact gene regulation mechanisms. (A) Transcription through a ‘weak’ iCGI can silence it, depositing H3K36me3 and DNA methylation at the iCGI. (B) However, if the iCGI exhibits strong transcriptional activity, it can lead to transcriptional interference. This can result in events akin to those at the (C) H13/Mcts2 locus, that exhibits allele-specific PAS usage. Usage of the PAS is highlighted in yellow. (D) Similar mechanisms have been found other iCGIs. Alternatively, and in some cases, simultaneously, (E) the iCGI can act as a promoter itself, highlighted in blue, for either the host gene itself (gene X) or for a different ‘nested’ gene (gene Y).