Abstract
Abstract 4585
Post-translational histone modifications influence expression by creating a chromatin environment which is conducive to or inhibitory of transcription. Modifications such as trimethylation of histone H3 lysine 4 and acetylation of histone H3 lysine 9 are generally associated with euchromatin and gene activation, while modifications such as trimethylation of histone H3 lysine 27 are associated with regions of heterochromatin and/or gene repression. Monomethyl histone H3 lysine 27 (H3K27me1) is a poorly studied post-translational histone modification for which variable associations with mRNA expression have been observed. Initially, H3K27me1 was localized to areas of pericentric heterochromain and was thought to be a marker of gene repression. Later reports described H3K27me1 enrichment throughout the body of actively transcribing genes (Vakoc C et al. MCB 26:9185, 2006; Wang Z, 40:897 Nat Genet, 2008). Some reports describe selective depletion of H3K27Me1 at promoters and transcription start sites (TSS), implying that depletion of H3K27me1 at the TSS is necessary for active transcription, (Vakoc C et al.), while others have associated increased enrichment for H3K27me1 at the promoter with increased levels of mRNA expression (Barski A et al. Cell 129:823, 2007). We hypothesize that the relationship between H3K27me1 occupancy and gene expression varies depending on both the cell-type and the location in the gene (i.e. promoter, transcription start site (TSS) and body of the gene) and that varying H3K27me1 levels in each of these locations is associated with alterations in the level of mRNA expression. To assess the association of H3K27me1 level with mRNA expression, H3K27me1 binding was determined using chromatin immunoprecipitation on microarray analysis (ChIP-chip) and correlated with mRNA levels determined using Illumina human expression arrays. ChIP-chip was performed using an antibody specific for H3K27me1 and the resulting DNA applied the to a custom designed genomic tiling NimbleGen microarray containing over 100 erythroid expressed genes and 10-100kb of flanking DNA for each locus. Probes were tiled with 10-100bp spacing, typically ∼65bp. Regions of repetitive DNA were excluded. ChIP-chip was performed in erythroid (K562) and non-erythroid (SY5Y-neural, RD-muscle) cells and the pattern of H3K27me1 enrichment assessed. mRNA transcript analyses were performed using Illumina human V6-2 expression arrays and quantitative real time RT-PCR. H3K27me1 levels at the promoter (-1000 to +1), in the 200bp surrounding the TSS (-100 to +100), and over the body of the gene were correlated with the level of mRNA expression. Increasing levels of H3K27me1 over the body of the gene lead correlated with increased levels of gene expression (R2=0.6122), while the amount of H3K27me1 at the promoter (-1000 to +1) had no correlation with gene expression (R2=0.2769). In agreement with Vakoc et al., decreased enrichment for H3K27Me1 at the TSS (-100 to +100) correlated with increased levels of mRNA expression. This is in sharp contrast to H3K4Me3, which accumulates at the start site of active genes. H3K27me1 has not been studied in detail in transcriptionally silent genes. Interestingly, genes without H3K27Me1 enrichment had no expression, implying that H3K27me1 is a marker of active transcription. Patterns of H3K27Me1 enrichment were cell-type specific. For example, in erythroid (K562) cells, the beta-globin locus was highly enriched for H3K27me1. This enrichment was not present in non-erythroid cells (RD, SY5Y). Finally, H3K27Me1 may also mark enhancers in a cell type-specific manner. For example, in the well studied HS2 enhancer in the beta-globin LCR, there is significant enrichment for H3K27me1 in K562 cells, but not in SY5Y or RD cells. These data indicate that modulation of chromatin architecture by monomethylation of histone 3 lysine 27 influences the level of gene expression in erythroid and non-erythroid cells.
Disclosures:
No relevant conflicts of interest to declare.
Roles of GPRC5 family proteins: focusing on GPRC5B and lipid-mediated signalling