Abstract
The mammalian genome is coiled, compacted and compartmentalized into complex non-random three-dimensional chromatin loops in the nucleus1–3. At the core of chromatin loop formation is CCCTC-binding factor (CTCF), also described as a “weaver of the genome”45. Anchored by CTCF, chromatin loops are proposed to form through a loop extrusion process6, organising themselves into gene neighbourhoods2 that harbour insulated enhancer-promoter domains, restricting enhancer activities to genes within loops, and insulating genes from promiscuous interactions outside of loops2,7–9. Studies targeting CTCF binding site deletions at gene neighbourhood boundaries result in localised gene expression dysregulation8,10–12, and global CTCF depletion recently showed CTCF to be crucial for higher hierarchical chromatin organisation of topologically associating domains (TADs)13. However, the role for CTCF in maintaining sub-TAD CTCF gene neighbourhoods and how gene transcription is affected by CTCF loss remains unclear. In particular, how CTCF gene neighbourhoods govern genome-wide enhancer-promoter interactions require clarification. Here, we took an in vivo approach to assess the global dissolution of CTCF anchored structures in mouse cardiomyocyte-specific Ctcf-knockout (Ctcf-KO), and uncovered large-scale ectopic de novo Enhancer-Promoter (E-P) interactions. In vivo cardiomyocyte-specific Ctcf-KO leads to a heart failure phenotype14, but our analysis integrates genome-wide transcription dysregulation with aberrant E-P interactions in context of CTCF-loop structures, identifying how genes engage their E-P interactions, requiring CTCF looping for their maintenance. Our study points to a mammalian genome that possesses a strong propensity towards spontaneous E-P interactions in vivo, resulting in a diseased transcriptional state, manifest as organ failure. This work solidifies the role of CTCF as the central player for specifying global E-P connections.