The second decade of 3C technologies: detailed insights into nuclear organization

Abstract

The relevance of three-dimensional (3D) genome organization for transcriptional regulation and thereby for cellular fate at large is now widely accepted. Our understanding of the fascinating architecture underlying this function is based on microscopy studies as well as the chromosome conformation capture (3C) methods, which entered the stage at the beginning of the millennium. The first decade of 3C methods rendered unprecedented insights into genome topology. Here, we provide an update of developments and discoveries made over the more recent years. As we discuss, established and newly developed experimental and computational methods enabled identification of novel, functionally important chromosome structures. Regulatory and architectural chromatin loops throughout the genome are being cataloged and compared between cell types, revealing tissue invariant and developmentally dynamic loops. Architectural proteins shaping the genome were disclosed, and their mode of action is being uncovered. We explain how more detailed insights into the 3D genome increase our understanding of transcriptional regulation in development and misregulation in disease. Finally, to help researchers in choosing the approach best tailored for their specific research question, we explain the differences and commonalities between the various 3C-derived methods.

Keywords: 3C technology; 3D genome; CTCF; chromatin loops; long-range gene regulation; transcription.

Figure 1.

Hierarchical genome organization. Schematic representation of the organization of the 3D genome into A (blue) and B (red) compartments and topologically associated domains (TADs), which are composed of several sub-TADs (depicted here as spheres), which in turn harbor several chromatin loops. Panels below the respective schematics depict how these structures are perceived in Hi-C (the “checkerboard” pattern for compartments; TADs, sub-TADs, and loops as detected in the interaction matrix) and 4C (chromatin contact plot showing loops). Note, however, that domains can also be appreciated in 4C data. Arrowheads indicate loops detected in tissue 1. Bars: compartment map, 10 Mb; tissue comparison Hi-C map, 100 kb. The bottom left Hi-C panel was created using the Juicebox software (Rao et al. 2014). The top right Hi-C panel is reprinted from Krijger et al. (2016).

Figure 2.

Overview of the established and newly developed 3C-derived methods. Schematics illustrate the experimental steps specific or common to the different methods. (*) DNase Hi-C has been combined with target enrichment, rendering it a “many versus all” method such as targeted chromatin capture (T2C), capture Hi-C (Chi-C), HiCap, and Capture-C. (†) HiCap differs from Chi-C mostly by employing a four-cutter instead of a six-cutter for the restriction digest. (‡) Ligation may be performed under diluted conditions (i.e., in solution) or within the intact nucleus (in situ Hi-C).

Figure 3.

CTCF binding polarity determines chromatin looping. Convergently oriented CTCF-binding sites are found at the base of chromatin loops and recruit the additional architectural protein cohesin. Motif inversion using CRISPR impedes looping, with cohesin recruitment being unaltered. Gene expression can also be affected. (Reprinted from de Wit et al. 2015.)

Figure 4.

Manipulating chromatin looping in vivo. In both murine and human adult erythroblasts, the transcription cofactor Ldb1 can be recruited to the developmentally silenced embryonic or fetal β-globin promoter via a zinc finger (ZF). This results in increased interaction with the LCR at the expense of the adult globin genes and concomitant changes in gene expression. (Reprinted from Deng et al. 2014 with permission from Elsevier.)

Figure 5.

Visualizing 4C data. (A) Especially when interested in local contact profiles, a coverage plot representation of the normalized 4C data may be chosen. Data are from the same locus as the 4C plot in Figure 1. The panel depicts an overlay of contact profiles from two tissues (depicted in blue and orange). An overlay is shown as dark red. (B) Alternatively and when depicting long-range contacts, a “spider plot” or arachnogram can be employed. Contacts from the viewpoint to other regions on the cis chromosome are depicted in brown. Black lines within the chromosome represent genes.

Figure 6.

PE-SCAn to combine Hi-C and ChIP-seq data. A PE-SCAn (de Wit et al. 2013) analyzes whether sequences bound by a chromatin factor of interest show preferential long-distance (i.e., TAD-crossing) interactions. The depicted example is reprinted from Krijger et al. (2016) and shows the relationship of Sox2-binding profiles obtained in mESCs and mouse NPCs (neural progenitor cells) and the chromatin interactions detected in NSCs and iPSCs (induced pluripotent stem cells). Note that the Sox2-binding sites detected in ESCs exclusively cluster in iPSCs, whereas the binding sites found in NPCs cluster in NSCs.