The Energetics and Physiological Impact of Cohesin Extrusion

Abstract

Cohesin extrusion is thought to play a central role in establishing the architecture of mammalian genomes. However, extrusion has not been visualized in vivo, and thus, its functional impact and energetics are unknown. Using ultra-deep Hi-C, we show that loop domains form by a process that requires cohesin ATPases. Once formed, however, loops and compartments are maintained for hours without energy input. Strikingly, without ATP, we observe the emergence of hundreds of CTCF-independent loops that link regulatory DNA. We also identify architectural "stripes," where a loop anchor interacts with entire domains at high frequency. Stripes often tether super-enhancers to cognate promoters, and in B cells, they facilitate Igh transcription and recombination. Stripe anchors represent major hotspots for topoisomerase-mediated lesions, which promote chromosomal translocations and cancer. In plasmacytomas, stripes can deregulate Igh-translocated oncogenes. We propose that higher organisms have coopted cohesin extrusion to enhance transcription and recombination, with implications for tumor development.

Keywords: CTCF; DNA damage; Nipbl; chromosomal translocations; class switching; cohesin; loop extrusion; nuclear architecture; topoisomerase II.

Figure 1. Cohesin translocation requires ATP and Smc ATPases

(A) Schematics representing the extrusion model of nuclear topology. Cohesin rings are loaded at Nipbl+ sites, extrude along DNA, and halt at CTCF+ anchor sites in the convergent orientation. (B) Venn diagram showing the distribution of Rad21 ChIP-Seq peaks relative to CTCF and Nipbl. (C) ChIP-Seq tracks of CTCF, Nipbl, and Rad21 in the presence and absence of ATP in activated B cells. (D) Histograms showing global Rad21 (left) or CTCF (right) occupancy at Nipbl+ loading sites in oligomycin-treated (black) and untreated (red) activated B cells. (E) ChIP-Seq profiles of transduced Smc3-biotag WT and ATPase mutants in activated B cells. (F) Occupancy of Smc3 WT and mutants at loading and anchor sites.

Figure 2. Creation of loop domains but not their maintenance requires ATP

(A) APA analysis of HCT-116 cells, where Rad21 is temporarily degraded and allowed to recover in the absence of ATP, transcription, or replication. (B) Box and APA plots showing Hi-C signals at loops or intra-domain interactions under various conditions relative to 24h activated B cells. Conditions were: G₀, ATP-, transcription-, or replication-inhibited B cells. Data are represented as mean ± SEM. (C) Compartments in HCT-116 and activated B cells ± oligomycin. HCT cells were also ± Auxin.

Figure 3. Characterization of architectural stripes

(A) Examples of stripes (black arrowheads) in Bcl6 and Basp1 domains in activated B cells. CTCF and Nipbl tracks are shown below. “a” single-stripe domain; “b” double-stripe domain; “c” conventional loop domain. (B) Bar graph illustrates the percentage of stripe anchors associated with CTCF binding motifs. The pie chart shows the percentage of motifs facing the stripe. Data are represented as mean ± SEM. (C) Violin plot shows insulation scores for stripe and stripe partner anchors. (D) Upper schematic depicts stripe formation at the Basp1 domain based on the extrusion model. Composite graph shows Nipbl profiles at loop domains displaying double (blue), left (red), or right stripes (green), or at no stripes (grey). (E) Hi-C matrices show the loss of stripes in ZF9-11 cells at loci where stripe anchors fail to recruit CTCF. (F) Molecular dynamic simulations of Hi-C data where cohesin is loaded near the left (first domain) or right anchor CTCF anchors (third domain).

Figure 4. Stripes form preferentially at SEs and play a role in transcription and recombination of Ig genes

(A) Composite graph showing SEs at loop domains carrying left (red) or right (green) stripes. (B) Contact matrix showing stripes at the Bach2 SE locus. (C) A nearly 3Mb-size stripe covers the Igh locus in B cells. Igh SE elements (hs3a, 1–2, 3b, and 4), Nipbl, and CTCF are shown. Arrowheads denote the orientation of CTCF motifs. CTCF tracks for WT, SA^−/−, and SA^partial mutants are shown. (D) Box plot shows Hi-C signals at loops (grey) and active regulatory elements linked by stripes (red) or not associated with stripes (white). Interactions were normalized based on non-regulatory DNA Hi-C signals in the same domain. p for all differences were 2.2e-16. (E) Hi-C² analysis of the Igh locus in SA^+/+ and SA^−/− CH12 B cells. Black arrowhead shows loss of stripe signals. (F) Igα germline transcripts in SA^+/+, SA^partial and SA^−/− activated cells. (G) Representative IgA recombination in WT and SA^−/− cells. All data are represented as mean ± SEM.

Figure 5. A role for loop extrusion in tumor development

(A) Karyotype analysis of plasmacytoma 7134 carrying a t(12;15) translocation. (B) Schematic depicts the translocated domain. (C) Hi-C² analysis of t(12;15) in SA^+/+ and SA^−/− cells. (D) Transcriptional analysis of genes upstream and downstream of the translocation breakpoint. (E) Proliferation of SA^+/+ and SA^−/− PCT cells. (F) Distribution of TOP2B induced damage in activated B cells relative to Rad21 and CTCF occupancy at loop anchors. (G) DNA damage profiles at double (red line), single (green line) or nostripe (back line) domains. (H) Extent of DNA damage in domains with no stripes (0) or with stripes up to 300Kb or 500Kb in length. Data are represented as mean ± SEM.

Figure 6. Loops extrusion untangles regulatory DNA hubs

(A–B) Examples of enhancer-enhancer loop formation in the absence of cohesin extrusion in ATP-depleted B cells relative to activated B cells (72h). (C) Bar graph characterizes gained and lost loops in oligomycin-treated cells relative to CTCF binding, motif orientation, and overlap with active regulatory DNA. (D) ChIP-Seq tracks showing promoters and enhancers (H3K4me1^highH3K4me3^low), Nipbl, Rad21, and CTCF.

Figure 7. Transcriptional regulation by loop extrusion

Based on the new and our recent findings (Kieffer-Kwon et al., 2017), we propose two mechanisms whereby nuclear architecture enhances transcription. First, as B cells enter the cell cycle, ATP synthesis increases, fueling cohesin ATPase-driven extrusion. Loop domains are thus constrained and intra-domain interactions (including regulatory DNA interactions) are increased “non-specifically”. Second, at sites with extensive loop extrusion, stripe contacts can directly tether promoters to enhancers, including SEs by a reeling in mechanism. As shown in Figure 6, this activity may disengage enhancer-enhancer loops and realigns them into functional interactions involving promoters near stripe anchors. In consequence of extensive extrusion, stripe anchors are hotspots for topoisomerase-induced damage.