Chromatin accessibility classification of TAD boundaries discloses new architectural proteins

Abstract

Abstract 3D genome organization is crucial to modulate gene expression. Topologically associating domain boundaries isolate genes and their regulatory elements within the same topological neighborhood, avoiding crosstalk between regulatory elements. Perturbation of domain boundaries causes aberrant genomic contacts and gene expression misregulation. CTCF and the cohesin complex are critical to form boundaries through loop extrusion. However, we still lack a complete understanding of what makes a boundary more effective at insulating genomic contacts than others. Here, we experimentally classified boundaries according to their chromatin accessibility as a proxy of protein occupancy in K562 human cells. Highly accessible boundaries are occupied by more proteins including CTCF, are more robust contact insulators, have a more conserved CTCF DNA-binding motif, and are more conserved across cell types. By exploring proteins enriched at boundaries with different accessibility, we found that CTCF and cohesin bind together with REST forming one module, while ZNF316 and EMSY form another, both occupying boundaries very frequently. CRISPR–Cas9 deletion of the ZNF316/EMSY binding region at a robust CTCF-free boundary increases inter-domain contacts, demonstrating that this module contributes to chromatin insulation. Our results emphasize the importance of protein combination and abundance in boundary strength and identify ZNF316/EMSY as a novel chromatin insulator module.