The Role of Topologically Associating Domains for Developmental Gene Regulation - ⁠ A Systematic Functional Analysis at the Sox9 and Shh Loci

Abstract

Precise spatiotemporal gene expression during embryonic developmental is controlled by cis- regulatory elements (CREs) such as enhancers and promoters. Their physical chromatin proximity is correlated with active transcription and thought to be restricted to topologically associated domains (TADs) that help establish interactions between CREs and limit inappropriate contacts. Accordingly, TADs frequently overlap with gene regulatory landscapes, in which are contained diverse enhancers that transmit their activity across the domain towards their target promoter. Large structural variants reorganizing TADs were shown to cause gene misexpression and disease thereby linking gene regulation to chromatin structure. Recently, several studies revealed controversial results questioning the importance of TADs for transcriptional control. Acute depletion of CTCF and other architectural proteins in vitro led to loss of TAD structures with surprisingly modest effects on gene expression. However, the cytotoxicity of such depletion assays hindered analysis of more complex gene regulatory scenarios and their effect during development. This study specifically addresses the connection between TADs and developmental gene regulation through two projects using the murine limb as a model system. First, we took advantage of the Sox9/Kcnj2-locus that is subdivided into two adjacent TADs with distinct expression patterns of Sox9 and Kcnj2. The systematic deletion of individual CTCF binding sites at the TAD boundary and within the TAD resulted in gradual fusion of the neighboring domains without major effects on gene expression. TAD rearrangement by TAD-spanning inversions and repositioning of the boundary, however, redirected the regulatory activity and resulted in pathogenic gene misexpression. Thus, TAD structures may not be essential for developmental gene regulation, yet CTCF-dependent rearrangement of TADs can lead to the redirection of enhancer–promoter contacts and gene misexpression. In the second project, we studied how enhancer position relative to its TAD influences the function of an individual enhancer at the Shh-locus. Therefore, we repositioned the Shh-limb enhancer ZRS to five alternative locations inside and outside of its TAD. As expected, the enhancer lost all function in the positions outside of the Shh-TAD. Interestingly, the new positions inside the TAD also displayed decreased enhancer activity, albeit to varying degrees. Further analysis suggests that CTCF likely functions in some positions as a facilitator of enhancer-promoter contacts, while insulating short-range contacts in others. Ultimately, the ZRS is only able to ectopically activate some genes if repositioned to novel TADs, displaying strong enhancer-promoter selectivity. In summary, the results demonstrate that TADs provide robustness and precision to gene regulation, guiding enhancer-promoter interaction without being essential. The findings in this work build a basis for future studies aiming to understand enhancer-promoter interaction and can help in contextualizing potential disease-causing mutations disrupting TADs.