Fig. 1: Tracking TAD anchors in living human cells.

a, TetOx96 and CuOx150 repeat arrays were inserted at TAD anchors in HCT116 cells and visualized using TetR-splitGFPx16 (GFP11 and GFP1_10) and CymR-2xHalo^JFX646, respectively. Multimerized GFP11 fragments are not shown for clarity. Antibiotic cassettes (Bsd-TK and Neo) were removed by Cre (through loxP sites) and FLP (through FRT sites) recombinases to avoid interference of transcription with loop extrusion. A Capture Micro-C contact map corresponding to TAD L2 is shown. b, Cells were imaged without or with auxin treatment, which leads to RAD21 depletion. The 3D distance (d) between the two fluorescent reporters was computed as function of time. In absence of auxin, the chromatin region between the two anchors is in one of three states: open (no loop), extruding (that is, containing one or more DNA loop(s)) or closed (the two anchors are maintained in direct contact). c, Capture Micro-C maps of cells left untreated or treated with auxin for 3 h. Green and magenta spots indicate the genomic locations of the inserted repeat arrays. The genomic distance between TAD anchors is indicated below each Micro-C map. All Micro-C maps show a 1,125-kb region, at 2-kb resolution. The T1 Micro-C map was used to illustrate the Adjacent locus. The white space in the L1 map corresponds to a region that was not covered by capture probes. d, Live-cell images of L1 TAD anchors at t = 0 min. The arrows indicate spots that were identified as not replicated. Example image from 1 of 30 image time series. Scale bars = 5 µm. e, Time-lapse images of the region inside the dotted white box in d. Scale bars = 1 µm. f, Time series of 3D anchor–anchor distances of the L1 TAD in untreated cells (red, corresponding to images in d and e) or auxin-treated cells (blue) and the adjacent locus (gray). g, CDF of 3D anchor–anchor distances. All images are maximum-intensity projections of 20 z-stacks. CDF, cumulative distribution function; a.u., arbitrary units.