TADs and Their Borders: Free Movement or Building a Wall?

Abstract

The tridimensional (3D) organization of mammalian genomes combines structures from different length scales. Within this organization, Topologically Associating Domains (TADs) are visible in Hi-C heat maps at the sub-megabase scale. The integrity of TADs is important for correct gene expression, but in a context-dependent and variable manner. The correct structure and function of TADs require the binding of the CTCF protein at both borders, which appears to block an active and dynamic mechanism of 'Cohesin-mediated loop extrusion'. As a result, mammalian TADs appear as so-called 'loop domains' in Hi-C data, which are the focus of this review. Moreover, we present a reanalysis of TADs from three "golden-standard" mammalian Hi-C data sets. Despite the prominent presence of TADs in Hi-C heat maps from all studies, we find consistently that regions within these domains are only moderately insulated from their surroundings. Moreover, single-cell Hi-C and superresolution microscopy have revealed that the structure of TADs and the position of their borders can vary from cell-to-cell. The function of TADs as units of gene regulation may thus require additional aspects, potentially incorporating the mechanism of loop extrusion as well. Recent developments in single-cell and multi-contact genomics and super-resolution microscopy assays will be instrumental to link TAD formation and structure to their function in transcriptional regulation.

Loop extrusion as a mechanism for DNA Double-Strand Breaks repair foci formation

“…In agreement, H2AX, MDC1 and 53BP1 were significantly more enriched within the damaged TADs compared to neighboring TADs (Fig. S1b), although spreading through boundaries was observed to some extent, in agreement with the rather moderate insulation properties of these domains and boundaries 11 . (-DSB), the histone H1 and Ubiquitin (FK2) ChIP-seq log2 ratio in damaged versus undamaged condition (+DSB/-DSB) and the γH2AX, MDC1 and 53BP1 ChIP-seq signals after DSB induction (+DSB) as indicated.…”

Lessons from the analysis of TAD boundary deletions in normal population

“…OMIM genes, DEV-genes and disease causing-genes were found to locate preferentially closer to the undeleted boundary compared to other genes. This was consistent with the fact that TAD insulating capacity is relatively weak (Chang et al 2020). By contrast, deletion of TAD boundaries in individuals without observable phenotype, probably indicate that the expression of genes within U-TADs was less dependent on the insulator abilities of the TAD boundaries, explaining the shape of the genes distribution along U-TADs and the lower enrichment in diseases-causing genes.…”

“…We also extracted TAD boundaries coordinates from 36 HI-C datasets available online, providing TAD-boundary data in different regulatory contexts: immortalized cell lines (GM12878, HMEC, HUVEC, IMR90, K562, KBM7, NHEK, A549, Caki2, G401, LNCaP, NCIH460, PANC1, RPMI7951, SJCRH30, SKMEL5, SKNDZ, SKNMC, T470); human embryonic stemcells derived lineages (H1-ESC, H1-MES, H1-MSC, H1-NPC, H1-TRO) and human tissues (aorta, liver, thymus and heart left ventricle, cerebral cortex, adrenal gland, bladder, small bowel, lung, psoas muscle, pancreas, spleen and heart right ventricle). Considering TADs in various cellular and tissue contexts was important, since single-cell Hi-C assays showed that TAD structure could be variable in a same tissue (Chang et al 2020). We collected genomic coordinates of 307,430 benign and likely benign copy number deletions from various public databases (ClinVar, DGV, DECIPHER and Phenotype in Humans using Ensembl Resources, GnomAD and from control cases of previously published series).…”

“…These structures are characterized by high level of intra-domain chromatin interactions contrasting with low level of inter-domain chromatin interactions (Szabo et al 2019). On Hi-C heat maps, TADs appear as pyramid-shaped structures, arising from chromatin-contact enrichment and separated by chromatin-contact depletion zones, referred as TAD-boundaries (Chang et al 2020). The boundaries of TADs are specific structures playing the role of insulators, i.e.…”

Epigenomic characterization of latent HIV infection identifies latency regulating transcription factors

“…CTCF is a large, ubiquitously expressed regulator of gene expression with an 11 zinc finger central domain and N and C termini that interact with numerous protein partners (Ghirlando and Felsenfeld, 2016). The human genome contains 15,000-40,000 CTCF binding sites (Ghirlando and Felsenfeld, 2016; Holwerda and de Laat, 2013), and CTCF homodimerization drives the formation of large topologically-associated DNA domains (TADs) by mediating chromatin loop structures (Chang et al, 2020). These loops serve to insulate domains of transcriptional regulation by disconnecting genes from nearby enhancer elements, and by restricting lateral spread of domains of both inhibitory and activating histone modifications (Hansen et al, 2019;Saldaña-Meyer et al, 2019).…”