Locus folding mechanisms determine modes of antigen receptor gene assembly

Abstract

The dynamic folding of genomes regulates numerous biological processes, including antigen receptor (AgR) gene assembly. We show that, unlike other AgR loci, homotypic chromatin interactions and bidirectional chromosome looping both contribute to structuring Tcrb for efficient long-range V(D)J recombination. Inactivation of the CTCF binding element (CBE) or promoter at the most 5'Vβ segment (Trbv1) impaired loop extrusion originating locally and extending to DβJβ CBEs at the opposite end of Tcrb. Promoter or CBE mutation nearly eliminated Trbv1 contacts and decreased RAG endonuclease-mediated Trbv1 recombination. Importantly, Trbv1 rearrangement can proceed independent of substrate orientation, ruling out scanning by DβJβ-bound RAG as the sole mechanism of Vβ recombination, distinguishing it from Igh. Our data indicate that CBE-dependent generation of loops cooperates with promoter-mediated activation of chromatin to juxtapose Vβ and DβJβ segments for recombination through diffusion-based synapsis. Thus, the mechanisms that fold a genomic region can influence molecular processes occurring in that space, which may include recombination, repair, and transcriptional programming.

Figure 1.

Homotypic chromatin interactions and chromosome looping fold Tcrb loci in thymocytes. (A) Schematic view of mouse chromosome 6 region spanning Tcrb with tracks presenting nucleotide position, non-Tcrb genes, Tcrb gene segments (functional or non-functional Vβ segments, light blue or black, respectively; Dβ segments, lavender; Jβ segments, purple; Cβ exons, black; Eβ, cyan) and CBEs (red for sense orientation or orange for antisense orientation). (B) HiC heat map of WT Tcrb loci in Rag1^−/− mouse thymocytes. Data are combined from six independent experiments, each performed on cells pooled from at least five mice of either sex and presented at 3 kb resolution. The lower panel indicates loops (black squares) identified by MUSTACHE performed at 3, 5, or 10 kb resolution, and segmented stripes (lines). (C) Genome browser views depicting two replicates each of RNA-Seq (sense and antisense transcripts) or a representative of two replicates of ChIP-Seq for H3K27ac, H3K9me2, Rad21 (GSM2973690), or CTCF for WT Tcrb of Rag1^−/− thymocytes. Below are MUSTACHE loops from panel B now filtered and separated by interactome. The top track shows loops with at least one anchor within the RC (between 5′PC and CBE3). The middle track shows loops with one anchor at Trbv1. The bottom track shows loops with both anchors in the main Vβ cluster. RNA-Seq and ChIP-Seq data are shown as reads per million, except for H3K9me2, which is shown as the log₂ ratio of bound H3K9me2 over input control (enriched for = dark brown above axis; depletion of = light brown below axis).

Figure 2.

The Trbv1 CBE and promoter are determinants of Tcrb recombination and repertoire. (A) Schematic representation of the Trbv1 genomic region of the WT, CBE-inactivated (V1C^Scr), or promoter-deleted (V1P^KO) alleles. WT and scrambled CBE sequences are displayed. (B) Representative flow cytometry plots of thymic αβ T cells expressing TRBV1⁺ TCRβ proteins on their surface in WT, V1C^Scr/Scr, or V1P^KO/KO mice. Each plot displays the TRBV1⁺TCRβ⁺ gate and the percentage of total TCRβ⁺ cells in the gate. (C) Quantification of TRBV1⁺TCRβ⁺ thymic αβ T cells in WT (n = 8), V1C^Scr/Scr (n = 4), or V1P^KO/KO (n = 7) mice. One-way ANOVA followed by Tukey’s post-tests for multiple comparisons. ***P < 0.001 and ****P < 0.0001. (D) Quantification of the percentage of total unique Tcrb genes involving each Vβ gene segment from Adaptive Immunosequencing performed on DNA isolated from DN thymocytes of WT, V1C^Scr/Scr, or V1P^KO/KO mice. n = 2, multiple unpaired t tests. ****, P < 0.0001. (E) Quantification of the percentage of Trbv1 rearrangements to each of the two DβJβ clusters, calculated from data shown in panel D.

Figure 3.

The Trbv1 CBE stimulates recombination by anchoring and terminating loop extrusion. (A) HiC heat maps of Tcrb loci in thymocytes from V1C^Scr/ScrRag1^−/− (V1C^Scr/Scr) or Rag1^−/− (WT) mice. Data are combined from two independent experiments, each performed on cells pooled from at least five V1C^Scr/Scr or WT mice of either sex and presented at 3 kb resolution. Each panel indicates loops (black squares) identified by MUSTACHE performed at 3, 5, or 10 kb resolution. (B) Virtual 4C tracks of the Trbv1 CBE or 5′PC viewpoint from V1C^Scr/Scr or WT thymocytes generated from HiC data of panel A. Black arrowheads below indicate the 4C viewpoint. The bottom track for each viewpoint shows V1C^Scr/Scr data subtracted by WT data so that signals above the axis (red) or below the axis (gray) indicate contacts that are greater in V1C^Scr/Scr or WT cells, respectively. Differentially interacting regions are boxed. (C) Genome browser views depicting two replicates each of RNA-Seq (sense and antisense transcripts) or a representative of two replicates of ChIP-Seq for H3K27ac, H3K9me2, or CTCF of Tcrb loci from V1C^Scr/Scr thymocytes, or MUSTACHE loops on V1C^Scr/Scr (top) or WT (bottom) thymocytes shown in panel A. RNA-Seq and ChIP-Seq data are shown as reads per million, except for H3K9me2, which is shown as the log₂ ratio of bound H3K9me2 over input control (enriched for = brown above axis; depletion of = light brown below axis). (D) Genome browser views depicting representative RNA-Seq (sense and antisense transcripts) or ChIP-Seq for H3K27ac or CTCF spanning the Trbv1 region in WT or V1C^Scr/Scr thymocytes. RNA-Seq and ChIP-Seq data are shown as reads per million.

Figure 4.

Promoter-mediated homotypic chromatin interactions drive Trbv1 rearrangement. (A) HiC heat maps of Tcrb loci in thymocytes from V1P^KO/KORag1^−/− (V1P^KO/KO) or Rag1^−/− (WT) mice. Data are combined from two independent experiments, each performed on cells pooled from at least five V1P^KO/KO or WT mice of either sex and presented at 3 kb resolution. Each panel indicates loops (black squares) identified by MUSTACHE performed at 3, 5, or 10 kb resolution. (B) Virtual 4C tracks of the Trbv1 CBE or 5′PC viewpoint from V1P^KO/KO or WT thymocytes generated from HiC data of panel A. The black arrowheads below indicate the 4C viewpoint. The bottom track for each viewpoint shows V1P^KO/KO data subtracted by WT data so that signals above the axis (cyan) or below the axis (gray) indicate contacts that are greater in V1P^KO/KO or WT cells, respectively. Differentially interacting regions are boxed. (C) Genome browser views depicting two replicates each of RNA-Seq (sense and antisense transcripts) or a representative of two replicates of ChIP-Seq for H3K27ac, H3K9me2, or CTCF of Tcrb loci from V1P^KO/KO thymocytes, or MUSTACHE loops on V1P^KO/KO (top) or WT (bottom) thymocytes shown in panel A. RNA-Seq and ChIP-Seq data are shown as reads per million, except for H3K9me2, which is shown as the log₂ ratio of bound H3K9me2 over input control (enriched for = brown above axis; depletion of = light brown below axis). (D) Genome browser views depicting representative RNA-Seq (sense and antisense transcripts) or ChIP-Seq for H3K27ac or CTCF spanning the Trbv1 region in WT or V1P^KO/KO thymocytes. RNA-Seq and ChIP-Seq data are shown as reads per million.

Figure 5.

Tcrb employs diffusion-based synapsis as an underlying mechanism of V(D)J recombination. (A) Schematic representation of the Trbv1 and DβJβ genomic regions of the Tcrb allele of mice with the Trbv1 RSS in native configuration (WT) or replacement of the Trbv1 RSS with the 3′Dβ1 RSS in the same (V1^R) or inverted (V1^Ri) genomic orientation. The locations of PCR primers used to analyze Tcrb rearrangement coding joins (green arrows) or signal joins (orange arrows) with relevant probes (purple lines) in each genotype are depicted below the V1^Ri allele. (B) Representative flow cytometry plots (left) and quantification (right) of thymic αβ T cells expressing TRBV1⁺ TCRβ proteins on their surface in WT, V1^R/R, and V1^Ri/Ri mice. n = 6; one-way ANOVA followed by Tukey’s post-tests for multiple comparisons. ***P < 0.001 and ****P < 0.0001. (C) Schematic representation of V1^Ri alleles with rearrangement between the inverted 3′Dβ1 RSS at Trbv1 and the 5′Dβ1 RSS (top) or 5′Dβ2 RSS (bottom). The locations of PCR primers used to analyze Tcrb rearrangement coding joins (green arrows) or signal joins (orange arrows) with relevant probes (purple lines) are indicated. (D) Taqman PCR quantification of indicated coding joins (left) or signal joins (right) generated by inversional rearrangements between the inverted 3′Dβ1 RSS at Trbv1 and the 5′ Dβ1 RSS or 5′Dβ2 RSS. n = 5. (E) Schematic representation of primer and probe placement for cleavage assay. (F) Taqman PCR quantification of cleavage assay where the plotted frequency represents intact alleles normalized to uncut alleles (tail DNA) and CD19. n = 4, 6.

Figure 6.

The mechanisms of Tcrb folding determine modes of long-range Vβ recombination. Left: Loop extrusion emanating from the Trbv1 CBE promotes rearrangement of only Trbv1 by generating chromosome loops with RC CBEs, positioning Trbv1 in spatial proximity with the RC to facilitate diffusion-based synapsis. Homotypic chromatin interactions between active Trbv1 and RC chromatin stabilize these loops to increase the chance for synapsis. Center: Loop extrusion emanating from an RC CBE promotes rearrangement of upstream Vβ segments by forming chromosome loops with any convergent Vβ CBE, juxtaposing adjacent Vβ segments and the RC to facilitate diffusion-based synapsis. Homotypic chromatin interactions between active RC and Vβ chromatin stabilize these loops to increase the chance for synapsis. Right: Loop extrusion anchored by an RC promotes rearrangement of upstream Vβ segments by enabling the open active site of RAG bound at a DβJβ complex RSS to scan Vβ chromatin for synapsis with a Vβ RSS. Impediments to cohesin activity in Vβ chromatin, including homotypic interactions with RC chromatin, enhance the ability of RAG scanning to capture a Vβ RSS. Depicted in all panels are Tcrb gene segments as rectangles (Vβ, blue; Dβ, lavender; Jβ,purple) and RSSs as triangles (Dβ and Jβ RSSs match their gene segment color; Vβ RSSs are colored according to their availability [green] or unavailability [yellow] for synapsis and rearrangement), CBEs as triangles (red = sense strand; orange = antisense strand), cohesin as a ring (cyan), Eβ as an oval (teal), and RAG (dark red).