Integrated Genome-Wide CTCF and CohesinSA1 Occupancy and Expression Analyses in Erythropoiesis

Abstract

Abstract 1305CTCF and cohesion are critical regulators of cellular growth, development and differentiation. CTCF has multiple functions including acting at gene promoters as a transcriptional activator or repressor, mediating long-range chromatin interactions, and acting as a chromatin insulator element. The cohesin complex is also multifunctional, participating in chromosome segregation during cell division, facilitating DNA-promoter interactions through cell-type specific DNA-looping, participating in DNA repair, and participating with CTCF in enhancer blocking. The cohesin complex is composed of 4 proteins Smc1, Smc3, Scc1, and either SA1 or SA2. The presence of SA1 or SA2 is mutually excusive, leading to 2 related, but distinct complexes, cohesinSA1 and cohesin.SA2. The SA1 component of the complex directly interacts with CTCF. To gain insight into how CTCF and cohesin regulate genes in erythroid development, chromatin immunoprecipitation coupled with high throughput sequencing (ChIP-seq) and mRNA transcriptome analyses were performed in human CD34+ hematopoietic stem and progenitor cells and cultured primary human erythroid (R3/R4 stage) cells, the results combined, and the interactomes compared. The MACS program identified 26,330 sites of CTCF and 23,396 sites of cohesinSA1 occupancy in CD34+ and 39,782 sites of CTCF and 33,497 sites of cohesinSA1 occupancy in erythroid cell chromatin (p<10e-5, fold enrichment>5). In CD34+ cells, the majority of CTCF and cohesinSA1 binding sites were located in intergenic regions (56 and 57%,) and introns (33 and 34%). In contrast, in erythroid cells, CTCF and cohesinSA1 binding had migrated to gene promoters (16% vs 2% and 24% vs 2%, respectively) with less binding in intergenic regions and introns. Sites of binding in erythroid cells were similar to that observed in fibroblasts, another differentiated cell-type. CTCF has sites of both cell-type specific and cell-type invariant binding. The Galaxy tool was utilized to compare sites of CTCF occupancy in 7 additional cell types. In CD34+ cells, only 5% sites of CTCF binding were CD34+ cell-type specific. In erythroid cells, 36% of CTCF binding sites were erythroid-specific. These unique sites were located primarily in enhancers and introns and were rarely seen in promoters. Refseq genes within 3kb of erythroid cell-specific CTCF sites were highly significantly enriched for the following GO terms: induction of apoptosis by extracellular signals, cytoskeleton organization, cellular response to stress, and macromolecule catabolic process. In both cell types, RefSeq genes within 3kb of an invariant CTCF site were consistently expressed at lower levels c.f. genes within 3kb of CD34+- or erythroid cell-specific CTCF sites. Analyzing CTCF-cohesinSA1 co-occupancy, there were 17,755 sites of CTCF and cohesinSA1 co-occupancy in CD34+ cells, accounting for 75% of CTCF sites and 67% of cohesinSA1 sites. In erythroid cells, 19,933 sites of occupancy were shared between CTCF and cohesinSA1, representing 50% of CTCF sites and 60% of cohesinSA1 sites. Finally, it has been suggested that CTCF marks chromatin domains in a cell-type specific manner. To determine whether CTCF and cohesinSA1 are present at domain boundaries in erythropoiesis, ChIP-seq for H3K27me3, a repressive chromatin mark, was performed. Chromatin domains were predicted using the Rseg program. 9,480 and 18,511 H3K27me3 chromatin domains were identified in CD34+ and erythroid cells, respectively, with average domain lengths of 31kb in CD34+ and 28kb in erythroid cells. There were 692 and 2,096 CTCF sites that marked domain boundaries in CD34+ and erythroid cells, respectively. These CTCF sites were cell-type specific, as only 75 of these CTCF sites were shared between CD34+ and erythroid cells. In both cell types, the majority of CTCF sites marking domain boundaries were found in distal intergenic regions and introns. CohesinSA1 was also frequently found at domain boundaries, present at 566 and 1830 domain boundaries in CD34+ and erythroid cells, respectively. Co-localization of CTCF with cohesinSA1 at domain boundaries was also common, with 66% of CTCF sites and 58% of CTCF sites binding both CTCF and cohesionSA1 in CD34+ and erythroid cells, respectively. These data indicate that CTCF and cohesin have multiple roles in regulating gene expression in erythropoiesis. Disclosures:No relevant conflicts of interest to declare.