Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

The fundamental role of chromatin loop extrusion in physiological V(D)J recombination


The RAG endonuclease initiates Igh V(D)J assembly in B cell progenitors by joining D segments to JH segments, before joining upstream VH segments to DJH intermediates1. In mouse progenitor B cells, the CTCF-binding element (CBE)-anchored chromatin loop domain2 at the 3′ end of Igh contains an internal subdomain that spans the 5′ CBE anchor (IGCR1)3, the DH segments, and a RAG-bound recombination centre (RC)4. The RC comprises the JH-proximal D segment (DQ52), four JH segments, and the intronic enhancer (iEμ)5. Robust RAG-mediated cleavage is restricted to paired V(D)J segments flanked by complementary recombination signal sequences (12RSS and 23RSS)6. D segments are flanked downstream and upstream by 12RSSs that mediate deletional joining with convergently oriented JH-23RSSs and VH-23RSSs, respectively6. Despite 12/23 compatibility, inversional D-to-JH joining via upstream D-12RSSs is rare7,8. Plasmid-based assays have attributed the lack of inversional D-to-JH joining to sequence-based preference for downstream D-12RSSs9, as opposed to putative linear scanning mechanisms10,11. As RAG linearly scans convergent CBE-anchored chromatin loops4,12,13,14, potentially formed by cohesin-mediated loop extrusion15,16,17,18, we revisited its scanning role. Here we show that the chromosomal orientation of JH-23RSS programs RC-bound RAG to linearly scan upstream chromatin in the 3′ Igh subdomain for convergently oriented D-12RSSs and, thereby, to mediate deletional joining of all D segments except RC-based DQ52, which joins by a diffusion-related mechanism. In a DQ52-based RC, formed in the absence of JH segments, RAG bound by the downstream DQ52-RSS scans the downstream constant region exon-containing 3′ Igh subdomain, in which scanning can be impeded by targeted binding of nuclease-dead Cas9, by transcription through repetitive Igh switch sequences, and by the 3′ Igh CBE-based loop anchor. Each scanning impediment focally increases RAG activity on potential substrate sequences within the impeded region. High-resolution mapping of chromatin interactions in the RC reveals that such focal RAG targeting is associated with corresponding impediments to the loop extrusion process that drives chromatin past RC-bound RAG.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Fig. 1: Role of RSS-based and RAG scanning mechanisms in DFL16.1 deletional joining.
Fig. 2: Mechanism of orientation-biased D-to-JH joining of seven Ds between DFL16.1 and DQ52.
Fig. 3: Binding of dCas9 impedes downstream RAG scanning and associated loop formation.
Fig. 4: Active transcription across Sγ2b impedes loop extrusion-mediated RAG scanning.

Data availability

HTGTS V(D)J-seq, Hi-C, 3C-HTGTS, GRO-seq and ChIP–seq sequencing data reported in this study have been deposited in the GEO database under the accession number GSE130224. Specifically, HTGTS V(D)J-seq data are deposited in the GEO database under the accession number GSE130216 and are related to Figs. 1e–h, 2b–d, 3a–c, e, 4c, Extended Data Fig. 2d, e, g, h, 3e, f, 4a, c, d, 5a–c, 6c, 7d, e, 10b, and Supplementary Tables 1, 2. Hi-C data are deposited in the GEO database under accession number GSE134543 and are related to Extended Data Fig. 8a. 3C-HTGTS data are deposited in the GEO database under accession number GSE130214 and are related to Figs. 3f, 4d and Extended Data Figs. 8b, 9q, r, 10c. GRO-seq data are deposited in the GEO database under accession number GSE130215 and are related to Figs. 3d, 4b and Extended Data Fig. 4e, 6d, 7f, 10a. ChIP–seq data are deposited in the GEO database under the accession number GSE130213 and are related to Extended Data Figs. 8c, d, 10d.

Code availability

HTGTS V(D)J-seq and 3C-HTGTS data were processed through a published pipeline available at Code for Hi-C data processing is available at GRO-seq and ChIP–seq were aligned to the mm9 genome with Bowtie2 v.2.2.8 (, processed using SAMtools v.1.8 ( and graphs were generated using the RSeQC tool v.2.6 (


  1. 1.

    Alt, F. W. et al. Ordered rearrangement of immunoglobulin heavy chain variable region segments. EMBO J. 3, 1209–1219 (1984).

    CAS  PubMed  PubMed Central  Google Scholar 

  2. 2.

    Rao, S. S. P. et al. A 3D map of the human genome at kilobase resolution reveals principles of chromatin looping. Cell 159, 1665–1680 (2014).

    CAS  PubMed  PubMed Central  Google Scholar 

  3. 3.

    Guo, C. et al. CTCF-binding elements mediate control of V(D)J recombination. Nature 477, 424–430 (2011).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  4. 4.

    Jain, S., Ba, Z., Zhang, Y., Dai, H. Q. & Alt, F. W. CTCF-binding elements mediate accessibility of RAG substrates during chromatin scanning. Cell 174, 102–116 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  5. 5.

    Teng, G. & Schatz, D. G. Regulation and evolution of the RAG recombinase. Adv. Immunol. 128, 1–39 (2015).

    CAS  PubMed  Google Scholar 

  6. 6.

    Alt, F. W., Zhang, Y., Meng, F.-L., Guo, C. & Schwer, B. Mechanisms of programmed DNA lesions and genomic instability in the immune system. Cell 152, 417–429 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  7. 7.

    Sollbach, A. E. & Wu, G. E. Inversions produced during V(D)J rearrangement at IgH, the immunoglobulin heavy-chain locus. Mol. Cell. Biol. 15, 671–681 (1995).

    CAS  PubMed  PubMed Central  Google Scholar 

  8. 8.

    Bolland, D. J. et al. Two mutually exclusive local chromatin states drive efficient V(D)J recombination. Cell Rep. 15, 2475–2487 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

  9. 9.

    Gauss, G. H. & Lieber, M. R. The basis for the mechanistic bias for deletional over inversional V(D)J recombination. Genes Dev. 6, 1553–1561 (1992).

    CAS  PubMed  Google Scholar 

  10. 10.

    Wood, C. & Tonegawa, S. Diversity and joining segments of mouse immunoglobulin heavy chain genes are closely linked and in the same orientation: implications for the joining mechanism. Proc. Natl Acad. Sci. USA 80, 3030–3034 (1983).

    ADS  CAS  PubMed  Google Scholar 

  11. 11.

    Yancopoulos, G. D. et al. Preferential utilization of the most JH-proximal VH gene segments in pre-B-cell lines. Nature 311, 727–733 (1984).

    ADS  CAS  PubMed  Google Scholar 

  12. 12.

    Hu, J. et al. Chromosomal loop domains direct the recombination of antigen receptor genes. Cell 163, 947–959 (2015).

    CAS  PubMed  PubMed Central  Google Scholar 

  13. 13.

    Zhao, L. et al. Orientation-specific RAG activity in chromosomal loop domains contributes to Tcrd V(D)J recombination during T cell development. J. Exp. Med. 213, 1921–1936 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

  14. 14.

    Lin, S. G., Ba, Z., Alt, F. W. & Zhang, Y. RAG chromatin scanning during V(D)J recombination and chromatin loop extrusion are related processes. Adv. Immunol. 139, 93–135 (2018).

    PubMed  Google Scholar 

  15. 15.

    Sanborn, A. L. et al. Chromatin extrusion explains key features of loop and domain formation in wild-type and engineered genomes. Proc. Natl Acad. Sci. USA 112, E6456–E6465 (2015).

    CAS  PubMed  Google Scholar 

  16. 16.

    Fudenberg, G. et al. Formation of chromosomal domains by loop extrusion. Cell Rep. 15, 2038–2049 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

  17. 17.

    Vian, L. et al. The energetics and physiological impact of cohesin extrusion. Cell 173, 1165–1178 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  18. 18.

    Rowley, M. J. & Corces, V. G. Organizational principles of 3D genome architecture. Nat. Rev. Genet. 19, 789–800 (2018).

    CAS  PubMed  Google Scholar 

  19. 19.

    Kim, M. S., Lapkouski, M., Yang, W. & Gellert, M. Crystal structure of the V(D)J recombinase RAG1–RAG2. Nature 518, 507–511 (2015).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  20. 20.

    Ru, H. et al. Molecular mechanism of V(D)J recombination from synaptic RAG1–RAG2 complex structures. Cell 163, 1138–1152 (2015).

    CAS  PubMed  PubMed Central  Google Scholar 

  21. 21.

    Gapud, E. J., Lee, B. S., Mahowald, G. K., Bassing, C. H. & Sleckman, B. P. Repair of chromosomal RAG-mediated DNA breaks by mutant RAG proteins lacking phosphatidylinositol 3-like kinase consensus phosphorylation sites. J. Immunol. 187, 1826–1834 (2011).

    CAS  PubMed  PubMed Central  Google Scholar 

  22. 22.

    Bredemeyer, A. L. et al. ATM stabilizes DNA double-strand-break complexes during V(D)J recombination. Nature 442, 466–470 (2006).

    ADS  CAS  PubMed  Google Scholar 

  23. 23.

    Choi, N. M. et al. Deep sequencing of the murine IgH repertoire reveals complex regulation of nonrandom V gene rearrangement frequencies. J. Immunol. 191, 2393–2402 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  24. 24.

    Gerstein, R. M. & Lieber, M. R. Coding end sequence can markedly affect the initiation of V(D)J recombination. Genes Dev. 7 (7B), 1459–1469 (1993).

    CAS  PubMed  Google Scholar 

  25. 25.

    Benner, C., Isoda, T. & Murre, C. New roles for DNA cytosine modification, eRNA, anchors, and superanchors in developing B cell progenitors. Proc. Natl Acad. Sci. USA 112, 12776–12781 (2015).

    ADS  CAS  PubMed  Google Scholar 

  26. 26.

    Lutzker, S., Rothman, P., Pollock, R., Coffman, R. & Alt, F. W. Mitogen- and IL-4-regulated expression of germ-line Igγ2b transcripts: evidence for directed heavy chain class switching. Cell 53, 177–184 (1988).

    CAS  PubMed  Google Scholar 

  27. 27.

    Qi, L. S. et al. Repurposing CRISPR as an RNA-guided platform for sequence-specific control of gene expression. Cell 152, 1173–1183 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  28. 28.

    Bevington, S. & Boyes, J. Transcription-coupled eviction of histones H2A/H2B governs V(D)J recombination. EMBO J. 32, 1381–1392 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  29. 29.

    Bolland, D. J. et al. Antisense intergenic transcription precedes Igh D-to-J recombination and is controlled by the intronic enhancer Emu. Mol. Cell. Biol. 27, 5523–5533 (2007).

    CAS  PubMed  PubMed Central  Google Scholar 

  30. 30.

    Cong, L. et al. Multiplex genome engineering using CRISPR/Cas systems. Science 339, 819–823 (2013).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  31. 31.

    Yang, L. et al. Optimization of scarless human stem cell genome editing. Nucleic Acids Res. 41, 9049–9061 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

  32. 32.

    Hu, J. et al. Detecting DNA double-stranded breaks in mammalian genomes by linear amplification-mediated high-throughput genome-wide translocation sequencing. Nat. Protoc. 11, 853–871 (2016).

    CAS  PubMed  Google Scholar 

  33. 33.

    Core, L. J., Waterfall, J. J. & Lis, J. T. Nascent RNA sequencing reveals widespread pausing and divergent initiation at human promoters. Science 322, 1845–1848 (2008).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  34. 34.

    Mahat, D. B. et al. Base-pair-resolution genome-wide mapping of active RNA polymerases using precision nuclear run-on (PRO-seq). Nat. Protoc. 11, 1455–1476 (2016).

    PubMed  Google Scholar 

  35. 35.

    Marinov, G. K. ChIP-seq for the identification of functional elements in the human genome. Methods Mol. Biol. 1543, 3–18 (2017).

    CAS  PubMed  Google Scholar 

  36. 36.

    Durand, N. C. et al. Juicer provides a one-click system for analyzing loop-resolution Hi-C experiments. Cell Syst. 3, 95–98 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

  37. 37.

    Durand, N. C. et al. Juicebox provides a visualization system for Hi-C contact maps with unlimited zoom. Cell Syst. 3, 99–101 (2016).

    CAS  PubMed  PubMed Central  Google Scholar 

  38. 38.

    Alt, F. W. & Baltimore, D. Joining of immunoglobulin heavy chain gene segments: implications from a chromosome with evidence of three D-JH fusions. Proc. Natl Acad. Sci. USA 79, 4118–4122 (1982).

    ADS  CAS  PubMed  Google Scholar 

  39. 39.

    Cowell, L. G., Davila, M., Kepler, T. B. & Kelsoe, G. Identification and utilization of arbitrary correlations in models of recombination signal sequences. Genome Biol. 3, research0072 (2002).

    PubMed  PubMed Central  Google Scholar 

  40. 40.

    Cowell, L. G., Davila, M., Yang, K., Kepler, T. B. & Kelsoe, G. Prospective estimation of recombination signal efficiency and identification of functional cryptic signals in the genome by statistical modeling. J. Exp. Med. 197, 207–220 (2003).

    CAS  PubMed  PubMed Central  Google Scholar 

  41. 41.

    Merelli, I. et al. RSSsite: a reference database and prediction tool for the identification of cryptic recombination signal sequences in human and murine genomes. Nucleic Acids Res. 38, W262–W267 (2010).

    CAS  PubMed  PubMed Central  Google Scholar 

  42. 42.

    Lieberman-Aiden, E. et al. Comprehensive mapping of long-range interactions reveals folding principles of the human genome. Science 326, 289–293 (2009).

    ADS  CAS  PubMed  PubMed Central  Google Scholar 

  43. 43.

    Predeus, A. V. et al. Targeted chromatin profiling reveals novel enhancers in Ig H and Ig L chain Loci. J. Immunol. 192, 1064–1070 (2014).

    CAS  PubMed  Google Scholar 

  44. 44.

    Medvedovic, J. et al. Flexible long-range loops in the VH gene region of the Igh locus facilitate the generation of a diverse antibody repertoire. Immunity 39, 229–244 (2013).

    CAS  PubMed  PubMed Central  Google Scholar 

Download references


We thank B. P. Sleckman for the v-Abl transformed, RAG2-deficient mouse pro-B cell line and Rag2-expressing shuttle plasmid, and Alt laboratory members for stimulating discussions. This work was supported by NIH R01 AI020047 (to F.W.A). F.W.A. is an investigator of the Howard Hughes Medical Institute. Y.Z. is a special fellow of the Leukemia and Lymphoma Society. Z.B. was a Cancer Research Institute Irvington fellow. A.P.A was supported by a Thrasher Research Fund Early Career Award. E.L.A. was supported by an NSF Physics Frontiers Center Award (PHY1427654), the Welch Foundation (Q-1866), a USDA Agriculture and Food Research Initiative Grant (2017-05741), an NIH 4D Nucleome Grant (U01HL130010), and an NIH Encyclopedia of DNA Elements Mapping Center Award (UM1HG009375).

Author information




Y.Z., X.Z. and F.W.A. designed the study; Y.Z., X.Z., Z.L., H.H., J.L. and E.D. performed experiments, except for Hi-C experiments which were performed by A.P.A. and analysed by A.P.A., M.S.S. and E.L.A. Z.B. provided critical reagents and advice on 3C-HTGTS. N.K. and J.Z. designed some bioinformatics pipelines; Y.Z., X.Z. and F.W.A. analysed and interpreted all data other than Hi-C data. Y.Z., X.Z. and F.W.A. designed figures and Supplementary Video and wrote the manuscript. Z.B., Z.L., H.H., J.L., A.P.A. and E.L.A. helped to polish the manuscript.

Corresponding author

Correspondence to Frederick W. Alt.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Peer review information Nature thanks Andre Nussenzweig and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Extended data figures and tables

Extended Data Fig. 1 Working model for role of loop-extrusion-mediated RAG scanning in driving deletion-biased D-to-JH recombination.

a, Illustration of the Y-structured RAG heterodimer complex. b, Schematic of Igh highlighting the RC and 3′ Igh loop domain bounded by IGCR1 and 3′ CBEs. c, Working model for RAG scanning to Ds upstream of DQ52. Cohesin (red ring) initiates loop extrusion upon being loaded into the upstream portion of the RC within the IGCR1–iEμ/RC subdomain. Proximal downstream active RC chromatin impedes cohesion-mediated extrusion of downstream chromatin and thereby serves as a downstream sub-loop anchor, allowing continued extrusion of upstream chromatin past RC-bound RAG. d, Continued upstream loop extrusion brings DHs upstream of RC-based DQ52 past the open RAG1 subunit active site opposite the JH-bound active site in the other RAG1 subunit. This linear process aligns a downstream D-12RSS with the RAG-bound JH-23RSS for orientation-specific, deletional D-to-JH recombination. e, Upstream Ds are frequently passed without being used and most loop extrusion-mediated RAG scanning continues until reaching the 5′ CBE loop anchor (IGCR1), which strongly impedes (nearly blocks) loop extrusion and RAG scanning. The latter prolonged interaction may contribute to robust DFL16.1 utilization. fh, Owing to the location of the RC, DQ52 can bind to the open RAG active site by diffusion (f), which allows it to bind in both deletional (g) and inversional (h) configurations. In this case, deletion-biased usage of DQ52 is achieved through a much stronger RSS-DN that, in this location, dominates RAG binding and cleavage compared to its weaker RSS-UP. Other schematics in bh are as described in Fig. 1 legend.

Extended Data Fig. 2 HTGTS V(D)J-seq analysis of V(D)J recombination outcomes in DH–JH+/− line and its mutant derivatives.

a, Schematic of the two Igh alleles of the DH–JH+/− v-Abl pro-B line. This C57BL/6, 129/Sv mixed background line was derived by deleting the indicated region from the 129/Sv allele to inactivate it for V(D)J recombination. b, Southern blotting confirmation of allele deletion in the DH–JH+/− line. Done twice with similar results. c, C57BL/6 versus 129/Sv DH usage in parental versus DH–JH+/− lines, as analysed via HTGTS V(D)J-seq (JH1 CE primer). The lack of 129/Sv-specific DHs in DH–JH+/− libraries confirmed the retention of C57BL/6 and absence of 129/Sv allele in this line. d, Bar chart shows utilization frequency of each VH, DH and JH from JH-distal to JH-proximal locales (n = 3 independent libraries). Pie chart shows indicated V(D)J recombination products as percentage of total Igh junctions. Beyond predominant DJH1 junctions, both low-level VHDJH1 joins4,12 and inversional JH(D)JH1 joins38 were detected. There were very low levels of JH1 joins to ‘cryptic RSSs’, or to a different JH-RSS (other) that is likely to occur in extra-chromosomal excision circles13. e, Utilization of each D as a percentage of total DJH1 joins (n = 3 independent libraries). f, Strategy for analysis of D-RSS-DN versus D-RSS-UP utilization. The orientation of D coding sequences relative to the JH1 CE primer is preserved in primary and secondary joins for both D-RSS-DN and D-RSS-UP, allowing calculation of the relative utilization of D-RSS-DN versus D-RSS-UP. g, Utilization frequency of D-RSS-DN versus D-RSS-UP in the DH–JH+/− line. h, Effect of DFL16.1-RSS mutations on utilization of D-RSS-DNs versus D-RSS-UPs. Libraries in d, e, g, h were normalized to 40,000 total junctions. Data represent mean ± s.d. Data for DH–JH+/− line in dg, h are two sets of three repeats each, with the latter done along with DLF16.1 mutants. Source data

Extended Data Fig. 3 Generation and analyses of the DH–JH1+/− line and its mutant derivative lines.

a, Coding and flanking D-RSS-UP and D-RSS-DN sequences and their RSS recombination information content (RIC) scores39,40 generated from a publicly available program ( Predicted ‘functional’ 12RSSs have a RIC of at least –38.81, with increasing RIC scores proposed to reflect increasing RSS strength. b, Illustration of potential DJH1 recombination on excision circle. Joining of JH1 to DHs downstream of DFL16.1, which occurs on excision circles generated by primary joining between distal JHs (JH2–JH4) and distal DHs, is not subject to the same mechanistic constraints as chromosomal D-to-JH recombination13. To obviate such joins, we deleted JH2–JH4 in the DH–JH+/− line to generate the DH–JH1+/− line. c, d, Southern blotting confirmation of DH–JH1+/− (done once after PCR confirmation; c) and intervening DH inversion (done twice with similar results; d) lines. e, Utilization of D-RSS-UP and D-RSS-DN in the DH–JH1+/− line and its mutant derivatives with intervening DH inversion (n = 3 libraries for each genotype). f, Relative utilization of DFL16.1-RSS-DN versus DFL16.1-RSS-UP for normal DFL16.1 (left) or DFL16.1 inversion (right) located in place of DQ52 in DH–JH1+/− cells with endogenous DFL16.1 deleted (n = 3 libraries for each genotype). e, f, Data represent mean ± s.d. from biologically independent samples, normalized to 70,000 total junctions for each library. g, Model of low level inversional RC-distal D joining involving loop-extrusion-based mechanism, which could bring distal upstream D-RSSs into diffusion radius of the RC. See Supplementary Discussion. Source data

Extended Data Fig. 4 Directional RAG scanning from a DQ52-based RC within 3′ Igh CBE-anchored loop domain.

a, HTGTS V(D)J-seq analysis with DQ52-RSS-DN bait in DH–JH+/− line. Major junctional outcomes are deletional DQ52-RSS-DN-to-JH joins (77%) and deletional DQ52-RSS-DN joins to cryptic RSSs near the immediately upstream DH3-2 region (20%), with the latter probably resulting from secondary events on excision circles following primary JH-to-distal DH joins (left; also, see below). b, Southern blot confirmation of JHΔ lines (done once after PCR confirmation). c, Repeats of HTGTS V(D)J-seq experiments shown in Fig. 3a. Each library was normalized to the same number of DQ52-RSS-UP SE junctions. d, Repeats of Fig. 3b HTGTS V(D)J-seq experiments. Each library was normalized to the same number of DQ52-RSS-UP CE junctions captured by the DQ52-RSS-DN bait (see Methods). Note the near abrogation of cryptic deletional joins near DH3-2 in JHΔ lines, which is consistent with their excision circle origin. Also, unlike the DH–JH+/− line with germline JHs, robust RC downstream cryptic scanning activity is readily detected in the JHΔ lines. e, Repeats of GRO-seq experiments shown in Fig. 3d. Each library was normalized to a coverage of 10 million 100-nt reads for display. fi, Model of cohesin loop-extrusion-meditated directional RAG scanning from RC DQ52-RSS-UP to upstream regions until reaching IGCR1 loop anchor. jm, Model of extrusion-meditated directional RAG scanning from RC DQ52-RSS-DN to downstream regions until reaching the 3′ CBEs loop anchor. Transparent yellow rectangles in f, j indicate, respectively, the upstream and downstream RAG scanning regions with DQ52 upstream and downstream RSS joining to cryptic RSSs shown in schematic form. Other schematics are as described in Fig. 1 and Extended Data Fig. 1. The two models are supported by the directional RAG scanning activity in c, d and Fig. 3a, b.

Extended Data Fig. 5 RAG cryptic targeting activity from DQ52-RSS-UP and DQ52-RSS-DN in JHΔ lines.

a, HTGTS V(D)J-seq profile of upstream RAG cryptic scanning activity from DQ52-RSS-UP with indicated peak regions at IGCR1 and DH3-2 locales (grey transparent bars). Top, junctions plotted at 100-bp bin size. Bottom, examples of most robust peak near IGCR1 (I) and DH3-2 (II) plotted at single-base-pair resolution. Letters next to the peaks show DNA duplex sequences of the targeted cryptic heptamers. See text for more details. b, HTGTS V(D)J-seq of downstream RAG cryptic scanning activity from DQ52-RSS-DN with indicated peak regions in Sγ2b and 3′ CBEs locales and lower frequency peaks in iEμ-Sμ, DH3-2 and IGCR1 (grey transparent bars). Top, junctions plotted at 100-bp bin size. Bottom, examples of most robust Sγ2b (III) and 3′ CBEs (IV) locale peaks plotted at single-base-pair resolution.c, Low frequency DQ52-RSS-DN junctions upstream of RC detected by DQ52-RSS-DN bait. Top, expanded views of IGCR1 and DH3-2 locales in b plotted at 20-bp bin size with representative junctions labelled (V–X). Bottom, single-base-pair resolution plot of junctions for V–X. Deletions are mediated by cryptic RSSs in divergent orientation (forward CAC) and inversions are mediated by cryptic RSSs in the same orientation (reverse CAC) as DQ52-RSS-DN. Also illustrated are junctions resulting from joining DQ52 CEs to cryptic CEs12, mediated by DQ52-RSS-UP and cryptic convergent RSSs. A likely explanation for these low level joins is that loop extrusion brings them into proximity with the RC where their location or transcription impedes extrusion, allowing them to access RC-bound RAG by local diffusion12, analogous to diffusion-mediated DQ52-to-JH1 joining.

Extended Data Fig. 6 RAG targeting and transcriptional activity analysis in the DFL16.1JH4-inv lines.

a, Generation of the DFL16.1JH4-inv line. Schematic shows two Igh alleles of DFL16.1JH4 line and DFL16.1JH4-inv line. In the DFL16.1JH4 line, one Igh allele contains a nonproductive VDJ join involving VH1-2P and JH3, and the other allele harbours the DFL16.1JH4 join. The DFL16.1JH4-inv line was derived from the DFL16.1JH4 line by inverting a 1-kb segment encompassing the DJH join using CRISPR–Cas9. b, Illustration of mechanism for RAG cryptic scanning activity from the RC DJH-RSS in DFL16.1JH4 line (top), DFL16.1JH4-inv line (middle) and DFL16.1JH4-inv 3′ CBEs−/− line (bottom). c, Representative HTGTS V(D)J-seq profiles showing RAG cryptic scanning patterns of DFL16.1JH4 line (top; n = 3 technical repeats), DFL16.1JH4-inv line (middle; n = 3 biological replicates) and DFL16.1JH4-inv 3′ CBEs−/− line (bottom; n = 3 biological replicates). Black line indicates bait primer position. Yellow shadows highlight RAG scanning regions. Purple arrows underneath the RAG targeting profiles indicate positions of forward and reverse CBEs. d, Representative GRO-seq profile of three repeats of the DFL16.1JH4-inv Rag2/ line (n = 3 biological replicates).

Extended Data Fig. 7 dCas9-binding impedes RAG scanning and corresponding loop formation.

a, Illustration of the dCas9-block system. An Sγ1-gRNA that has 16 binding sites (blue lines) within a 4-kb highly repetitive Sγ1 region on the C57BL/6 allele was introduced into the JHΔ-dCas9 line. b, Western blot confirmation of dCas9 expression in JHΔ-dCas9 lines but not the parental JHΔ line (done twice with similar results). c, RT–PCR confirmation of Sγ1-gRNA expression in the JHΔ-dCas9-Sγ1-sgRNA lines but not parental lines (done twice with similar results). d, Additional HTGTS V(D)J-seq repeats (DQ52-RSS-DN bait) for JHΔ-dCas9 lines and JHΔ-dCas9-Sγ1-sgRNA lines shown in Fig. 3e. Each library was normalized to the same number of DQ52-RSS-UP CE junctions captured by the DQ52-RSS-DN bait (see Methods). e, Expanded view of Sγ1 region from HTGTS V(D)J-seq profiles in d, showing accumulation of RAG activity at the dCas9-bound Sγ1 region. f, GRO-seq analysis of JHΔ-dCas9 and JHΔ-dCas9-Sγ1-sgRNA lines. Each library was normalized to a coverage of 10 million 100-nt reads for display. Bar graph compares transcriptional activity of indicated regions (n = 3 libraries for each genotype). Data represent mean ± s.d. from biologically independent samples. P values were calculated via two-tailed paired t-test. NS, P ≥ 0.05. The modest decrease in Sγ2b transcription upon Sγ1–dCas9 binding is potentially due to impaired loop extrusion between Sγ2b and iEμ. Source data

Extended Data Fig. 8 dCas9 binding impedes downstream loop formation in association with cohesin loading and accumulation at the impediment locale.

a, Hi-C analysis of the 3′ Igh domain interaction of the JHΔ-dCas9 line compared with the JHΔ-dCas9-Sγ1-sgRNA line. We compared 1.3 billion contacts in the control line with 1.2 billion contacts in the dCas9-impediment line. Letters annotate the interactions between the two indicated loci, and the numbers next to the letters reflect relative interaction intensity. Black and blue arrows highlight Sγ1 interaction with the RC (B) and 3′ CBEs (F) locales, respectively, in the JHΔ-dCas9-Sγ1-sgRNA line. b, 3C-HTGTS repeats with iEμ bait (green stars) for JHΔ-dCas9 and JHΔ-dCas9-Sγ1-sgRNA lines shown in Fig. 3f. The iEμ bait primer strategy is shown above. Each library was normalized to 192,000 total junctions for analysis. While these lines retain downstream CH sequences on their 129/Sv allele (Extended Data Fig. 2b), the iEμ bait should have very low interactions in trans42. Blue and grey transparent boxes extending through all panels are as described in Fig. 3. In addition, an interaction between the RC and an Iγ2b upstream enhancer named hRE1, an enhancer of unknown activity43,44, was evident (see also Fig. 4d) and was accompanied by accumulation of RAD21 and NIPBL (see below) and a low level of RAG scanning activity (Extended Data Fig. 7d). c, RAD21 ChIP–seq profiles of JHΔ-dCas9 lines versus JHΔ-dCas9-Sγ1-sgRNA lines (n = 2 biological replicates). Each library was normalized to a coverage of 10 million 100-nt reads. d, NIPBL ChIP–seq profiles of JHΔ-dCas9 lines versus JHΔ-dCas9-Sγ1-sgRNA lines (n = 2 biological replicates). Each library was normalized to a coverage of 10 million 100-nt reads.

Extended Data Fig. 9 Working model for loop-extrusion-mediated RAG downstream scanning.

ai, Model for cohesin-mediated loop extrusion of chromatin past nascent Igh RC in JHΔ v-Abl lines based on RAG2-deficient background analyses. For all examples, increased interactions of impediment sites with RC targets scanning activity in RAG-sufficient cells. a, Cohesin (red rings) are loaded at multiple sites in the RC–3′ CBEs Igh subdomain. Illustrations show cohesin loading at the region downstream of RC. b, Cohesin-mediated extrusion promotes a linear interaction between the nascent RC and downstream regions. c, Robust transcription (green arrow) across the Iγ2b–Sγ2b impedes loop extrusion. d, In a subset of cells, loop extrusion proceeds past the Iγ2b–Sγ2b impediment to the 3′ CBEs loop anchor. ei, Loop extrusion in JHΔ-dCas9-Sγ1-sgRNA lines is impeded, directly or indirectly, by the dCas9-bound Sγ1. As dCas9 impediment is not a complete block, loop extrusion in a subset of cells proceeds downstream, allowing dynamic formation of sub-loops of RC with Iγ2b-Sγ2b or 3′ CBEs. jl, In RAG-sufficient cells, RC-bound RAG might enhance the dCas9-bound Sγ1 extrusion impediment. mp, Elimination of Iγ2b-promoter-driven transcription permits unimpeded RAG-bound RC extrusion to the 3′ CBEs anchor, increasing RAG scanning activity there. q, r, 3C-HTGTS analysis of RC interactions with DH and flanking regions in the JHΔ-dCas9 line (q) and DH–JH+/− line (r). DpnII (n = 4, biological replicates) and NlaIII (n = 3, biological replicates) digestions are shown for the JHΔ-dCas9 line. The NlaIII digestion more clearly reveals an interaction peak near DH3-2 owing to a paucity of DpnII sites in that region. NlaIII digestion of the DH–JH+/− line shows a similar RC interaction pattern to that of the JHΔ-dCas9 line (r, n = 2 technical repeats). Bar graphs show relative RC interaction of the 25-kb intervening DH region (from DH2-3 to DH2-8) versus that of the same-sized neighbouring regions (n as indicated above). Data represent mean ± s.d. (q) or mean (r). P values calculated via two-tailed paired t-test. Source data

Extended Data Fig. 10 Ectopic transcription of Iγ2b-Sγ2b region impedes downstream loop formation and RAG scanning.

a, GRO-seq repeats for JHΔ-dCas9 lines (Iγ2bWT) and JHΔ-dCas9-Iγ2b-del lines (Iγ2bΔ/Δ) shown in Fig. 4b. Each library was normalized to a coverage of 10 million 100-nt reads. b, HTGTS V(D)J-seq repeats with DQ52-RSS-DN bait for Iγ2bWT versus Iγ2bΔ/Δ lines shown in Fig. 4c. Each library was normalized to the same number of DQ52-RSS-UP CE junctions captured by the DQ52-RSS-DN bait. c, 3C-HTGTS repeats from iEμ bait for Iγ2bWT and Iγ2bΔ/Δ lines for data shown in Fig. 4d. Each library was normalized to 150,000 total junctions for analysis. d, RAD21 ChIP–seq analysis for Iγ2bWT and Iγ2bΔ/Δ lines. Each library was normalized to a coverage of 10 million 100-nt reads for display. Bar graph shows comparison of RAD21 accumulation at the Sγ2b region (Sγ2a region as control) in Iγ2bWT lines versus Iγ2bΔ/Δ lines (n = 3 libraries for each genotype). Data represent mean ± s.d. from biologically independent samples. For bar graph presentation, the junction number recovered from the Sγ2b region of Iγ2bWT control samples was normalized to represent 100%, relative values of Sγ2a region in the control and Sγ2b and Sγ2a regions in the Iγ2bΔ/Δ samples are listed as a percentage of the control Sγ2b values. P values were calculated by a two-tailed paired t-test. NS, P ≥ 0.05. Source data

Supplementary information

Supplementary Information

Supplementary Discussion and References, Uncropepd Gel Images and Supplementary Tables 1-3.

Reporting Summary

Video 1

Loop extrusion-mediated RAG chromatin scanning during Igh D-to-JH recombination. This video provides the animation of our working model for the role of loop extrusion-mediated RAG chromatin scanning in enforcing deletional orientation D-to-JH chromosomal V(D)J recombination. The details of the model shown in the video are described in Extended Data Figure 1 and legend. The video has also incorporated self-explanatory text that allows it to be understood on its own.

Source data

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Zhang, Y., Zhang, X., Ba, Z. et al. The fundamental role of chromatin loop extrusion in physiological V(D)J recombination. Nature 573, 600–604 (2019).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing