Overview of the i3C protocol. Living cells are harvested in a close‐to‐physiological buffer (PB; step 1); intact nuclei isolated by mild NP‐40 treatment (step 2); chromatin digested using ApoI or NlaIII, nuclei spun to release unattached chromatin (step 3); and leave cut chromatin bound to the nuclear substructure (step 4). Then, ligation takes places in situ, and DNA is isolated (step 5).
Percentage of total cell chromatin present at the different steps of the procedure (± SD; n = 2).
Relative contribution of the different HUVEC ChromHMM features in each i3C fraction.
i4C‐seq (blue shades) and conventional 4C (gray shades) were performed side by side in HUVECs, using ApoI and the SAMD4A TSS as a viewpoint (triangle); profiles from two replicates are overlaid. The browser view shows interactions in the ˜1 Mbp around SAMD4A. The zoom‐in shows interactions in the SAMD4A TAD (gray rectangle). Strong (red) and intermediate (brown) interactions called by fourSig, RefSeq gene models, and ENCODE ChIP‐seq data are shown below.
Figure 2.Native interactions are confined by TAD boundaries and describe prelooping
i4C‐seq was performed in HUVECs using NlaIII and the TSSs of BMP4, CDKN3, CNIH, and SAMD4A as viewpoints (triangles). Interactions are shown aligned to TAD boundaries (gray rectangles; from Dixon et al, 2012) and HUVEC ENCODE ChIP‐seq data (below). Prelooping of the SAMD4A and BMP4 TNF‐responsive TSSs to enhancers is indicated (orange lines).
Figure 3.TALE‐iD verifies native looping at the human ZFPM2 locus
An overview of TALE‐iD. A construct encoding a TALE DNA‐binding domain that targets an active enhancer in the ZFPM2 first intron is fused to a bacterial Dam methylase and introduced into K562 cells. Cells are harvested 48 h after transfection; genomic DNA is isolated and digested with DpnI to reveal sites specifically methylated by the Dam activity. Finally, qPCR using primers flanking different DpnI sites is used as readout.
i4C performed in K562 cells using ApoI and the ZFPM2 TSS as a viewpoint (triangle). i4C interaction in the 458‐kbp ZFPM2 locus is shown, and the enhancer targeted by the TALE‐iD construct is indicated (red triangle). K562 ENCODE ChIP‐seq data are also shown below.
qPCR readout at different DpnI sites. DpnI sites at the ZFPM2 promoter (p1–p4) and enhancer (e1–e3; positions in panel B) were targeted in qPCRs after restriction digest. Bar plots show log2‐fold enrichment of cut sites (1/ΔΔCt) over background DpnI cutting levels in untransfected K562 cells. Regions c1–c4 serve as controls; region m1 (an enhancer shown to interact with the TSS by i4C) is also methylated as part of a multi‐loop structure. The same DpnI sites were also tested in transfections involving a construct encoding either a non‐targeting (“scrambled”) TALE domain or the targeting domain fused to an inactive Dam protein (“ΔDam”). *P < 0.05; two‐tailed unpaired Student's t‐test (n = 3). The bars and error bars denote mean ± SEM.
Figure 4.3D organization of a 2.8‐Mbp human locus analyzed by iT2C/conventional T2C
Interaction maps from conventional T2C (left) and iT2C (middle, right) in the 2.8 Mbp around SAMD4A on chromosome 14. Magnifications show interactions at increasingly higher resolution. Bottom: HUVEC ENCODE ChIP‐seq data are aligned to interactions mapped at 1.5‐kbp resolution in the 250 kbp around SAMD4A.
PE‐SCAN graphs (see de Wit et al, 2013) show the enrichment of iT2C interactions (± 5 kbp) for CTCF (gray), H3K27ac (pink), and H3K27me3 (blue), while H3K9me3 (brown) that is absent from this region serves as a control.