Our ability to fight the multitude of potential disease-causing agents that we encounter depends on a process called recombination, which can occur in different ways. Recombination manipulates DNA sequences to enable our bodies to generate an enormous diversity of the immune system’s recognition components: antibodies and T-cell receptors (TCRs). Two papers in Nature from the same laboratory, by Zhang et al.1 and Zhang et al.2, reveal an unexpected similarity in how these types of recombination event occur.
Read the paper: The fundamental role of chromatin loop extrusion in physiological V(D)J recombination
In developing immune-system cells, a process called V(D)J recombination rearranges DNA sequences to assemble genes that will encode either an antibody or a TCR, using a large pool of three classes of gene segments, termed V, D and J. These gene segments are flanked by evolutionarily conserved DNA sequences called recombination signal sequences (RSSs), which direct the enzyme RAG to join together one V segment and one J segment, and sometimes also one D segment, in an astonishing variety of combinations. The intervening DNA between these joined segments is usually deleted, although in rare instances it is instead inverted and retained when two gene segments are joined. This recombination process enables antibodies and TCRs to have diverse protein domains called variable domains, which recognize protein fragments called antigens. It is this diversity in antigen-recognition domains that allows the immune system to respond effectively to a variety of disease-causing agents.
The genes encoding antibodies can sometimes undergo further refinements that change single DNA nucleotide bases (generating what are known as somatic point mutations) to boost an antibody’s ability to recognize antigens. The DNA in these genes can also go through a series of alterations, called antibody class-switch recombination (CSR), that do not affect antigen recognition — instead, they endow the antibody with diverse effector functions, such as the ability to bind to mucosal surfaces or to help other immune cells tackle the infection.
V(D)J recombination is initiated by RAG, whereas somatic point mutations and CSR in antibody-encoding sequences are initiated by a DNA-mutating enzyme called AID. The potential power of RAG and AID to cause widespread alterations to the genome is dangerous, so their action needs to be limited to the target sequences at which DNA alterations can be exploited for host defences.
The DNA–protein complex chromatin, which is tightly packaged inside the nucleus of human cells, forms thousands of loops of varying size that are anchored at their base by the ring-like structure of a protein complex called cohesin3. These loops form when a molecular-motor component of cohesin actively extrudes chromatin through the cohesin ring until chromatin hits a ‘roadblock’. This probably forms before or when chromatin enters the ring, and typically if DNA has bound to the protein CTCF. Cohesin-dependent extrusion of large loops partitions chromatin into discrete regions known as topologically associated domains, and smaller loops enable regulatory DNA sequences, such as enhancers and promoters that are located far apart in the linear DNA sequence, to be placed next to each other to drive gene expression. Zhang and colleagues’ work1,2 shows that chromatin-loop extrusion also underlies the control of both V(D)J recombination and CSR (Fig. 1).
During V(D)J recombination, RAG is recruited to modified DNA-binding histone proteins that accumulate at high levels in a small region of the chromosome containing antibody- or TCR-encoding J genes. This generates a VDJ recombination centre4, at which RAG binds the RSS motifs that flank the J gene segments. RAG then scans the rest of the chromosome in a linear fashion to locate the RSS of another, more distant gene segment5. Once compatible RSSs are aligned, RAG introduces DNA breaks to initiate recombination between these two RSSs. RAG is anchored in the VDJ recombination centre, which raises the question of how DNA moves during this scanning process.
Zhang et al.1 realized that chromatin-loop extrusion might explain this DNA movement. In this model, after cohesin has assembled in the VDJ recombination centre with an RSS-bound RAG, cohesin ‘reels’ DNA through its ring, enabling the RSSs in the loop to possibly find a compatible RSS bound to RAG with which to recombine (see Supplementary Video 1 in ref. 1). This model is supported by the authors’ experiments, including their demonstration that blocking DNA movement through the cohesin ring biases recombination events to favour recombination targeting RSSs near the site where DNA movement was impeded. Importantly, the directional DNA-scanning mechanism in this model also explains the overwhelming predominance of deletion rather than inversion events during V(D)J recombination, which has long been an unexplained conundrum. Together with earlier studies6 demonstrating that cohesin-binding elements (DNA motifs next to certain antibody V gene segments) are major determinants of DNA-rearrangement patterns, and the resulting antibody repertoires, a convincing model emerges that chromatin-loop extrusion aids V(D)J recombination.
CSR is a conceptually similar, although enzymatically distinct, process to V(D)J recombination. Zhang et al.2 investigated whether cohesin-driven extrusion of DNA loops also underlies CSR. During CSR, AID introduces multiple point mutations of DNA nucleotide bases in specific ‘switch regions’ in antibody-encoding genes, which eventually leads to DNA breaks7. Unlike V(D)J recombination — in which RAG-mediated cleavage of DNA depends on the assembly of a pair of compatible RSSs — AID causes mutations at individual DNA sites that can lead to DNA breakages before or after the alignment of the switch regions that will subsequently join together7.
The authors propose that, analogous to events that occur in a VDJ recombination centre, a CSR centre forms over one particular switch region (called Sµ) in an antibody-encoding gene. Previous studies7 favoured diffusion as the mechanism leading to the alignment of DNA during class switching, whereas Zhang and colleagues’ work supports the idea that cohesin-based loop extrusion aligns the two switch regions to enable their recombination (Fig. 1). These two studies thus offer compelling evidence for a unified model for V(D)J recombination and CSR. It also links these processes to gene-expression regulation, on the basis of the dynamic modulation of chromatin architecture.
This model offers testable predictions and raises numerous questions. For example, how is cohesin recruited to VDJ and CSR recombination centres? Cohesin depletion from particular cellular lineages causes defective V(D)J recombination8, and cohesin loss eliminates all loops across chromosomes9. However, the effect of such alterations on CSR remains to be determined.
Loop extrusion generates torsional stress in DNA10, and cohesin recruits the enzyme topoisomerase IIB to relieve this stress by transiently breaking DNA10. Therefore, reeling in DNA to regulate gene expression or to enable recombination-based immune diversification might drive a type of chromosomal abnormality known as a chromosomal translocation, which could lead to cancer. Much like the DNA loops themselves, these insights into the role of chromosomal architecture might help to reveal connections between areas that were previously considered to be separate.
Nature 575, 291-293 (2019)