Predicting modulaters of TE-driven chromatin 3D structure (HiC)

DNA is tightly packaged in the human nucleus and is wrapped in complex three-dimensional (3D) structures that have been implicated in regulatory processes However, no comprehensive model describes the formation of the 3D genome. At least 40% of the human genome consists of fossilized inactive transposable element (TE) sequences. A role for TEs in 3D genome structure has been suggested by several studies that illustrate how TEs are involved in 3D genome formation. However mechanisms that mediate the formation of TE-mediated 3D contacts is lacking. As TEs are rich in TF binding sites it seems likely that TFs bound to TEs are responsible for forming the 3D genome structure. We used the comprehensive TF binding data available in human and mouse pluripotent stem cells (PSCs), coupled with HiC data to explore the role of TFs bound to TEs in 3D genome organization. Based on these computational predictions we divide TFs into three main classes, those that utilize TEs to drive 3D genome formation, those that are neutral, and a third class that breaks 3D contacts at specific TEs. We then experimentally validate four proteins, and show that SMARCA5 and MAFK are involved in promoting chromatin contacts at TEs, whilst E2F6 and KDM1A are disruptive. Human pluripotent stem cells were transfected with shRNAs targetting SMARCA4, MAFK, E2F6 or KDM1A, and LUC (as a control) and subjected to RNA-seq to measure gene expression changes and HiC to determine chromsome contacts. All experiments were performed in biological duplicate