Enhancers within the immunoglobulin V-gene region orchestrate chromatin topology and regulate V gene rearrangement frequency to shape the B cell receptor repertoire specificities

Abstract

Effective antibody-mediated responses depend on a highly diverse antibody repertoire with the ability to bind a wide range of epitopes in disease-causing agents. The generation of this repertoire depends on the somatic recombination of the variable (V), diversity (D), and joining (J) genes in the immunoglobulin loci of developing B-cells. It has been known for some time that individual V, D and J gene segments rearrange at different frequencies, but the mechanisms behind this unequal V gene usage have not been well understood. However, recent work has revealed that newly described enhancers scattered throughout the V-gene containing portion of the immunoglobulin loci regulate the V gene recombination frequency in a regional manner. Deletion of three of these enhancers revealed that these elements exert many layers of control during V(D)J recombination, including long range chromatin interactions, epigenetic milieu, chromatin accessibility and compartmentalization.

Introduction

Antibody repertoire development is a tightly regulated and complex process. The B-cell receptor is the membrane-bound version of the antibody that the cell will ultimately secrete. At the early stages of B-cell development, the V, D and J gene segments in the immunoglobulin loci of the B lymphocytes undergo a tightly orchestrated series of complex rearrangements to create functional B cell receptor (1). An antibody molecule is composed of four polypeptide chains, with two identical heavy chains (Igh) and two identical light chains (Igκ or Igλ) linked together by disulfide bonds (2). The rearrangement of V, D and J genes at these loci is mediated by the Recombination Activating Genes (Rag1 and Rag2) (3). The RAG complex binds and cleaves the recombination signal sequences (RSSs) flanking each V, D, and J gene segment. Then, through DNA repair mechanisms, the gene segments fuse to ultimately create a V(D)J coding gene (Figures 1A and 1B). This process occurs sequentially. First, in the Igh locus at the pro-B cell stage, a D_H-J_H rearrangement takes place, followed by V_H-D_HJ_H rearrangement. Afterward, the light chain V-J rearrangement occurs in pre-B cells. Finally, after the rearranged V(D)J genes from the Igh and Igκ or Igλ loci are successfully transcribed and translated, the newly formed B cell receptor is displayed in the cell surface.

Figure 1.

(A and B) Schematic diagrams of the Igh and Igκ loci. (A) Assembly of a V(D)J coding exon in the Igh locus occurs in two main steps. First, D_H to J_H rearrangement, then, D_H-J_H to V_H rearrangement. The orientation and length (12 or 23 bp) of the spacer of the RSS is shown. The Igh is located in the chromosome 12 (5’ J_H-D_H-V_H 3’), while the Igκ locus is located in chromosome 6 (5’ Vκ-Jκ 3’). Genomic location of each loci is shown. (B) Assembly of a Vκ-Jκ coding exon in the Igκ locus occurs in one primary step. However, rearrangement can occur by inversion or deletion, and can continue until the Vκ or Jκ substrates are exhausted. (C) Enhancers are regulatory elements bound by TFs that recruit coactivators. Both TFs and coactivators can have IDRs. These regions mediate the formation of a phase-separated condensate. Enhancer (Enh). (D) Schematic diagrams showing the relative V gene rearrangement frequency (dashed line) after deletion of the E88 enhancer (top), E34 enhancer (middle), and E_VH1 enhancer (bottom). Above (blue) and below (red) 1 indicate an increase or decrease in V gene rearrangement relative to wild-type cells, respectively. (E) The E34 enhancer promotes the deposition of H3K4me1 and H3K27ac across the E34 loop domain (red circles). This epigenetic reprogramming eventually triggers the formation of favorable interactions that brings the Vκ regions dispersed across the E34 loop domain close to the Jκ genes. Likewise, the E88 enhancer instructs a similar process in other loop domains and, concomitantly, the deposition of H3K4me1 and H3K27ac. Beige circles represent the deposition of repressive epigenetic marks. (F) Upon deletion of the E34 enhancer, nearby Vκ elements fail to deposit H3K4me1 and H3K27Ac across the region. Rather repressive epigenetic marks dominate the E34 loop domain landscape. Hence, without E34 enhancer activity, Vκ regions associated with the E34 loop domain fail to move into proximity to the Jκ elements. Instead, the E34 subTAD is associated with genomic regions enriched for the repressive epigenetic mark H3K9me3 positioned outside the Igκ locus. Similarly, the assembly of nuclear condensates instructed by other enhancers across the Vκ region facilitates Vκ-Jκ rearrangement that involves other subsets of Vκ regions altering the ratio of Vκ gene usage in the antibody repertoire. It should be noted that E88 is active in the pro-B cell stage, and is likely one of the the first Vκ enhancers to be active.

Most of the V(D)J recombination principles apply equally for the heavy and light chain loci (1-3). However, there are significant differences between the rearrangement in the Igh vs the Igκ loci. First, all the V genes in the Igh locus are located in the same transcriptional orientation as the D-J and C genes segments, and thus all RSSs are in convergent orientation with respect to the D-J rearrangement. In contrast, V genes in the Igκ locus are located in both orientations with respect to the Jκ genes. This specific feature poses a challenge for the currently favored models as to how the RAG complexes find their substrates (V, D, and J genes segments) for recombination (4). Another feature that is different between the Igh vs the Igκ loci is that the V gene rearrangement for the Igh is fixed once it occurs, since all unrearranged D segments are deleted during the first rearrangement. In contrast, the Igκ locus can undergo a series of sequential rearrangements (Figure 1B). This characteristic allows pre-B cells to test and change the specificity of the B-cell receptor if the newly created light chain gene is nonproductive or results in an auto-reactive B cell receptor in a process called receptor editing (5, 6).

The quality of the RSS was considered to be the primary factor behind why certain V genes underwent more frequent rearrangement than others for a number of years (3, 7). RSS sequences have a consensus sequence but very few RSS have a complete consensus sequence. In vitro studies with recombination substrates have shown that variations from the consensus sequence can diminish recombination by vastly different frequencies (7-12). However, the advent of next-generation sequencing technologies such as ChIP-seq, RNA-seq, and VDJ-seq, coupled with bioinformatic analyses, have elucidated several potential reasons why some V genes rearrange more frequently than others, although these are predominantly correlative (13-16). Rearrangement frequency correlates with several factors, including RSS quality, specific transcription factor (TF) binding near a V gene, local epigenetic milieu, nucleosomal accessibility, and three-dimensional (3D) chromatin conformation (13, 15-26). In addition, B cells generated by V(D)J recombination in different physiological locations (bone marrow vs. fetal liver vs. lamina propria) are known to have different repertoires (27, 28). Although many of these features correlate with the levels of rearrangement, they do not fully account for the differences in V gene rearrangement frequency. Here, we will present and summarize the data describing how newly described enhancers located within the large V-gene portion of Igκ and Igh loci work as a higher-order mechanism for the regulation of the antibody repertoire formation by regulating chromatin interactions and epigenetic milieu, which in turn alter the 3D chromatin topology of the loci.

Enhancers in gene transcription and beyond

Enhancers are DNA regulatory elements that play a crucial role in gene expression (29). Enhancers can contain a diverse combination of consensus binding sequences for universal, lineage- and developmental stage-specific TFs (30, 31). Once bound, TFs can recruit additional proteins, such as co-activators, to the enhancer or the target gene's promoter region to augment transcriptional activity (29, 32). Recruitment of these TFs facilitates the upregulation or activation of specific genes or clusters of genes. Although enhancers are often adjacent to the genes they regulate, they can also be located far from the regulated genes, sometimes even hundreds of thousands of base pairs away. Enhancers also function regardless of their orientation relative to the target gene or regulatory circuit. Overall, enhancers play a critical role in controlling gene expression in a precise and context-specific manner, allowing cells to maintain their identity and respond to environmental cues.

Enhancers also shape the chromatin topology via enhancer-enhancer and enhancer-promoter interactions (33-35). Enhancers induce promoter activity by moving closer together in space to establish a physical interaction. Current models for enhancer-enhancer and enhancer-promoter interactions propose that local phase separation, mediated by weak multivalent interactions among enhancer-associated molecules, drives spatial proximity between these regulatory elements with analogous states of activity and/or biophysical properties (29, 36-42). TFs and coactivators bound to the enhancers containing intrinsically-disorder regions (IDRs) drive the formation of these phase-separated condensates through a process called liquid-liquid phase separation (LLPS) (Figure 1C) (37-41, 43-46). The formation of biomolecular condensates, also known as membrane-less condensates or droplets, by LLPS is emerging as a fundamental principle in regulating cellular processes, including the initiation and maintenance of gene transcriptional programs (36, 38, 43, 47).

Enhancers in the V gene-containing region of antigen receptor loci

Previously identified and well characterized enhancers in the antigen receptor loci are all located near the J and C region genes. These enhancers regulate chromatin accessibility, histone post-translational modifications, recruitment of RAG complexes, non-coding transcription, rearrangement, and transcription of the rearranged gene (1). For the Igκ locus, there are three enhancers near the Jκ and Cκ genes: iEκ, 3’Eκ, and Ed. Deletion of iEκ or 3’Eκ cause a moderate to severe decrease in total Igκ rearrangement, while the simultaneous deletion of both eliminates all rearrangement (21, 48, 49). A reduction in B cell numbers often accompanies these deletions. In contrast, deletion of the Ed does not cause any rearrangement defect, but it affects the transcription of the rearranged Vκ-Jκ gene segment (50).

Much more recently, enhancers have been identified in the large V gene-containing regions of the T and B cell receptor loci (26, 51, 52). These enhancers have been identified by traditional enhancer characteristics: enrichment of H3K4me1 and H3K27ac, binding of key TFs and often Brg1 and p300. These characteristics are commonalities between these enhancers and the enhancers near the J, D, and C genes. However, unlike enhancers near the J and C genes, deleting these V-region located enhancers did not affect overall levels of rearrangement or total B cell numbers. Instead, their deletion severely affects the rearrangement of nearby V genes, and the levels of V gene rearrangement in a 0.8-1.5 Mb region surrounding the enhancer are also modestly reduced. We will describe the findings about three of these enhancers that have been well characterized recently.

E88.

An enhancer in the Igκ locus, E88, was the first of these V gene-region regulatory elements to be functionally characterized in detail (26). We named these by their proximity to the closest functional Vκ gene, in this case Igkv1-88. E88 deletion in mice shows a profound defect in the rearrangement of the Vκ genes nearby, with a milder effect in the rearrangement of other genes in a wider area (~1.5 Mb) surrounding E88 (Figure 1D). Non-coding (germline) transcription of V genes nearby E88, but not at a distance, was also decreased. High-throughput chromosome-conformation capture (HiC) analysis revealed that the Igκ locus is located inside a discrete topologically associating domain (TAD) and also demonstrated the presence of subTADs in Vκ-gene region. TADs and compartment domains are the primary modes of a high order of chromatin folding (53-56). TADs display a high frequency of interactions between regions within their chromatin domain and a lower frequency between regions in adjacent chromatin domains and they can be observed in HiC studies as dominant triangles off the diagonal (54, 55). SubTADs are smaller domains inside TADs with similar interaction patterns that tend to form in cell-type specific manner (57). The HiC data reveal that the E88 enhancer likely regulates Vκ rearrangement in a subTAD-specific manner, with the 1.5 Mb region that showed reduced rearrangement levels correlating with two subTADs (26). Moreover, we demonstrated that E88 was a hub of local and long-range chromatin interactions across the Igκ locus. These interactions included interactions with several enhancers and promoters, including the traditional enhancers near the Jκ and Cκ genes (iEκ,3’Eκ, and Ed). Deletion of E88 also showed that interactions across the Igκ locus were drastically affected. Furthermore, changes in the interaction patterns from iEκ (~1.8 to 2.3 kb away from the Jκ genes) with the rest of the V-gene region of the Igκ locus in E88^−/− mice correlated with increased or reduced rearrangement frequency. Moreover, this data showed that interactions in wild-type mice correlated with the deposition of H3K4me1, and loss of interactions correlated with the reduction of H3K4me1 (26, 52). Together, this data suggests that the interaction frequency between regulatory elements with enhancer characteristics facilitates Vκ-Jκ rearrangement. Overall, this work revealed that this novel enhancer in the V-gene region shapes the chromatin topology through long range interactions, and that it directly affects the pattern of rearrangement of Vκ genes over a 1.5 MB region, with the most profound affect for Vκ genes near E88.

E34.

Building upon these findings, we tested the hypothesis that enhancers in the Vκ-gene region control the rearrangement of specific Vκ genes with relevant physiological functions. Specifically, we identified and characterized an enhancer, E34, near a Vκ gene, Igkv7-33, that encodes the light chain of the natural antibody EO6/T15 (58, 59). This antibody recognizes the molecule phosphorylcholine (PC) present in the cell wall of many pathogenic bacteria, including Streptococcus pneumoniae (60, 61). EO6/T15 also binds PC in oxidized low-density lipoproteins, which can prevent the progression of atherosclerosis (62, 63). EO6/T15 antibodies are present in naïve mice without any previous foreign antigen exposure, presumably due to natural exposure to oxidized low-density lipoprotein (61). Basal levels of EO6/T15 are sufficient to protect mice after Streptococcus pneumoniae infection (61). EO6/T15⁺ B cells are the dominant cell producing antibodies after exposure to PC-containing antigens. By deleting E34 in mice, we demonstrated that the E34 enhancer controls Igkv7-33 gene rearrangement frequency (59). Furthermore, due to this low level of Igkv7-33 rearrangement, EO6/T15 antibody and EO6/T15⁺ B-cell steady-state levels were both reduced. As expected, EO6/T15⁺ B-cell responses were negatively impacted after stimulation with a relevant antigen due to a low level of EO6/T15⁺ B-cells in the E34^−/− mice. The repertoire changes caused by E34 deletion increased the susceptibility to Streptococcus pneumoniae infection but not to other bacterial or viral pathogens examined.

HiC studies showed that the deletion of E34 changed the chromatin interaction landscape in the Igκ locus (59). Chromatin interactions involving the areas around E34 were substantially depleted in E34^−/− mice. As with E88, interaction pattern alterations from the iEκ to the rest of the Vκ-gene region correlated with rearrangement frequency. The loss of specific chromatin interactions was also correlated with loss or reduction of CTCF and H3K4me1 binding at specific sites. Furthermore, comprehensive analysis of this HiC data reveal that the E34 deletion affected subTAD boundaries. Combining the HiC and the Vκ rearrangement data, we showed that Vκ gene reduced rearrangement was contained in a specific subTAD, similar to the results with E88 mice. A study by Karki et al., arrives to a similar conclusion (64). They showed that non-coding transcription from Vκ-gene promoters is activated in a subTAD specific manner. Using microscopy, they also showed that the actively transcribing subTAD in a specific allele had been translocated to a region with high transcription and enriched RNA polymerase close to the highly transcribed Jκ genes (64). This repositioning facilitated the rearrangement of the Vκ genes in a specific subTAD.

One of the more significant findings in the E34 study was that the deletion of E34 changed the chromatin compartmentalization of the E34 subTAD (59). Compartment domains are defined by a high-order association of distinct genomic loci from HiC data using a principal-component analysis. Each compartment is associated with patterns of active, A compartment, and inactive or repressed, B compartment, histone marks (56, 65, 66). Active transcription is characteristic of A compartments, while repressed transcription is characteristic of B compartments. Compartmental domains are smaller than TADs and are present inside, between, or overlapping CTCF loops in mammals (66-68). They are also classified as active A and inactive B domains. We show that the E34-containing subTAD switches from a compartment that is shared by the rest of the Igκ locus to a compartment that is enriched with repressive epigenetic marks, mainly H3K9me3.

Using multiomics, loss-of-function, and in vivo pathogenesis approaches, we connected functional layers to decipher the mechanisms of the selection of the V gene usage during V(D)J recombination. We merged the chromatin conformation data from HiC with epigenetic profiling and transcriptomic data to reveal the depth of the changes caused by E34 deletion. Enrichment for H3K4me1, H3K27ac, CTCF, and H3K27me3 was reduced in a subTAD that contained the E34 enhancer. Chromatin accessibility was also severely affected. Chromatin accessibility and H3K4me1 enrichment have been associated with rearrangement fitness while H3K4me1, H3K27ac, and CTCF are have been associated with chromatin interactions (13, 15, 16, 69, 70). Furthermore, the translocation of chromatin domains between active (A) and inactive (B) compartments is correlated with changes in epigenetic signatures, transcription, chromatin accessibility, and intra- and inter-chromatin interaction frequency (71-74). All these features are intimately interconnected and play critical roles in cell development, identity, and function. Our findings that E34 controls all of these features illustrate how enhancers in the Igκ locus have co-opted these mechanisms in order to regulate specific Vκ gene rearrangement. To our knowledge, these findings are the first demonstration that proper chromatin architecture, and ensuing recombination within a locus are essential for an effective immune response.

E_VH1.

Four enhancers spread across the V_H gene portion of the Igh locus were recently characterized by one of us (AJF), together with colleagues from the Kenter lab, based upon the same characteristics as we used to identify the Igκ enhancers described above (75). We deleted one of these enhancers (E_VH1) in mice. Similar to E34 and E88 deletion, E_VH1^−/− mice show defects in rearrangement of a large area of V_H genes, especially nearby E_VH1 (Figure 1D). Interestingly, B1a B cells specific for phosphatidylcholine (PtC) were reduced in E_VH1^−/− mice. PtC is exposed in senescent cells and is also present in some pathogenic bacteria, including Brucella melitensis and Pseudomonas aeruginosa (76, 77). B1a B cells specific for PtC frequently use the IGHV11-2, which was reduced ~2-fold in E_VH1^−/− mice, which may explain the reduction in splenic B1a B cells (78, 79).

Virtual 4C studies reveal the presence of an enhancer network coordinated by E_VH1 (75). A viewpoint at E_VH1 shows interactions with E_VH2, E_VH3 and E_VH4 as well as with Eu, and deletion of E_VH1 eliminates interactions with all of these other enhancers. With a viewpoint at Eu, wild type mice show strong interactions at E_VH1. Deletion of E_VH1 eliminates the interaction of Eu with E_VH1 but now a new interaction can be seen at E_VH2, and the weaker interaction at E_VH3 is maintained. These observations were confirmed using FISH analyses. This enhancer network likely instructs the chromatin topology of the Igh locus.

Unlike the data from the E34 and E88 studies, relative chromatin interaction frequency from the Eu enhancer, which is located near the D_H-J_H genes, does not correlate with relative rearrangement frequency (E_VH1^−/− over WT V gene rearrangement) (75). In contrast, this correlation is very clear in the E88 and E34 studies (26, 59). Other studies also show a correlation between TF binding and enrichment of H3K4me1, an enhancer-associated epigenetic modification near the Vκ genes, with their rearrangement frequency (13, 16). There is also a correlation between active histone marks and rearrangement frequency for many V_H genes. However, in contrast, studies in the Igh locus show a correlation between rearrangement frequency and CTCF binding near proximal V_H genes, and deletion of some of those CTCF sites shows their requirement for proximal V_H gene rearrangement (14, 15, 80, 81). In sum, this suggests that the function of the enhancer network on the Igh locus differs in some ways from that in the Igκ locus.

Ig loci enhancers determine the timing of V gene rearrangement.

A defining characteristic of some enhancers is their cell type specific activity (30, 82). Enhancers in the Ig loci can ensure the proper timing for rearrangement during development by recruiting lineage-specific TFs. Pioneer TFs, such as EBF, have an important role because they can bind heterochromatin and drive the recruitment of histone acetyltransferases, such as p300, and methyltransferases that eventually allow recruitment chromatin remodeler complexes such as BRG1 (30, 45, 82, 83). EBF, p300, and Brg1 bind to E34, E88, and E_VH1. E2A is another key TF bound to these three enhancers. In contrast, Pax5 enrichment is much higher in E34 than in E88 or E_VH1, which might explain some developmental differences in activation between these enhancers. E88 and E_VH1 are more active in pro-B cells compared to E34, which is more active in pre-B cells. All of these TFs are associated with B cell-specific enhancer function, recruitment of chromatin remodeling complexes, chromatin interactions, and the establishment and maintenance of the 3D genome structure (82, 84-88). Collectively, this allows the binding of co-activators, such as Mediator. Finally, these processes, coupled with the recruitment of RNA polymerase II, eventually activate transcription and chromatin accessibility.

Enhancers located in the V-gene region of the Ig loci serve a variety of functions. First, they help to determine the proper chromatin compartmentalization and topology of the V-genes in a large region surrounding them, which is likely delimited by loop domains or subTADs. Deletion of any of these three enhancers leads to significant changes in the subTAD structure and a decrease in long-range interactions emanating from the region of the deleted enhancer. Second, they influence the epigenetic landscape of the surrounding area, including the enrichment of activation-associated epigenetic marks and the binding of other factors. Third, they open the chromatin over the V-gene RSS, making it available for RAG-mediated cleavage. Therefore, the deletion of any of these enhancers results in a similar pattern of severe defects in the rearrangement of nearby genes and a more modest decrease in a larger area, often demarcated by subTAD boundaries (Figures 1E and 1F). These multiple layers of control ensure proper V gene rearrangement, guaranteeing the appropriate representation of specific V genes in the B-cell receptor and antibody repertoires.

A tale of two principles

Loop extrusion and compartmentalization are the two major mechanisms of chromosome organization (66, 68, 89-95). TADs and subTADs are formed by cohesin-dependent loop extrusion and are typically, but not always, demarcated by CTCF bound sites at the boundaries (68, 89, 91, 94) (Figure 2A). Cohesin complexes are loaded onto the DNA fiber by NIPBL to initiate loop extrusion (89, 96, 97). NIPBL is usually present at enhancer and promoter regions throughout the genome (32, 98). Once loaded into the chromatin fiber, cohesin complexes extrude DNA until they reach a boundary element, typically CTCF-enriched sites that are usually in a convergent orientation. Compartment and compartmental domains are created by a process called LLPS, which depends on the multivalent interactions of factors associated with chromatin and histone modifications specific to each type of compartment (66, 99, 100) (Figure 2B). Chromatin with repressive epigenetic signatures tends to cluster driven by the HP1 proteins and the Polycomb complex (72, 95, 100-102). On the other hand, the clustering of active chromatin is partially mediated by IDRs in proteins like BRD4, Mediator, RNA polymerase II, and a variety of TFs (37-41, 44-46, 103, 104). Active transcription by RNA Polymerase II and the produced RNA can also contribute to this process (39, 41, 46, 105-108).

Figure 2.

(A) Loop extrusion initiates by uploading cohesin to the chromatin fiber by NIPBL. In an ATP-dependent process, the DNA extrudes through the cohesin ring until it reaches a boundary element usually a set of convergent CTCF binding sites. This process forms chromatin loops, TADs or subTADs. (B) Loop domains tend to cluster depending on their activation or repressive histone marks in a process called compartmentalization. (C) Loop extrusion-mediated RAG scanning helps bring the V_H genes close to the D_H-J_H segments in the recombination center to facilitate rearrangement. Here, only V_H genes with RSS in convergent orientation with the RSS of the D_H-J_H segment are able to rearrange. (D) Illustration showing how enhancers instruct the assembly of nuclear condensates to regulate Vκ-Jκ rearrangement. Across the pre-B cell population, the iEκ-Ed-3’κ super-enhancer instructs the assembly of a nuclear condensate, which is present in most cells. Likewise, enhancers that span the Vκ region orchestrate the assembly of nuclear condensates. Each cell can have a different set of nuclear condensates in the V-gene region depending on the activation of specific enhancers. Due to homologous activation status and epigenetic signatures, the iEκ-Ed-3’κ super-enhancer condensate and the V-gene region enhancer-associated condensates physically merge positioning Vκ and Jκ genes within close spatial proximity enabling Vκ-Jκ rearrangement. (E) Model for how enhancers facilitate the rearrangement of V genes in both orientations. Direct enhancer-enhancer interactions and association of chromatin fragments with homologous epigenetic signatures can mediate this process independently of loop extrusion-mediated RAG scanning . Here, the V gene directionality is irrelevant as the proximity is mediated by multivalent low-affinity interaction between IDRs or active epigenetic signatures such as H3K4me1 and H3K27ac.

In the Igh locus, proximity between the V_H and D_HJ_H region is thought to be achieved through loop extrusion, likely aided by the enhancer network (4, 18, 75, 80, 81, 109-112). Loop extrusion drives V_H to D_HJ_H rearrangement by a process that involves RAG-mediated chromatin scanning for compatible convergent RSSs between the D_HJ_H segment and the V_H genes (Figure 2C). As discussed by Zhang et al.(4), the loop extrusion mechanism that assists RAG-scanning does not necessarily need to begin at the recombination center over the D_H-J_H genes, but it can initiate in the V_H-gene region. Perhaps, the enhancers in this area, such as E_VH1, serve as cohesin-loading elements, through NIPBL, for loop extrusion initiation favoring rearrangement of the V_H genes nearby. The distinct configuration of the Igh locus, where all the V_H gene RSSs are in convergent orientation to the D_HJ_H RSSs, facilitates RAG-scanning-mediated rearrangement involving deletion of the intervening DNA (18, 80, 81, 109, 112). Furthermore, the inversion of a part or all of the Igh locus was shown to inhibit the rearrangement of the V_H genes present in these inverted DNA segments, indicating that correct RSS directionality is a requirement for a RAG-scanning-mediated rearrangement (111, 112). In contrast, Vκ-Jκ rearrangement involves either deletion or inversion of the intervening DNA sequences due to the fact that Vκ genes are positioned in either forward or reverse orientations with respect to the Jκ genes. This renders loop extrusion-dependent RAG-mediated chromatin scanning insufficient to drive Vκ to Jκ rearrangement, raising the question of whether Igκ locus rearrangement fundamentally differs from that of the Igh locus.

Our current and previous data suggest that enhancers instruct Vκ-Jκ rearrangement by bringing the Vκ genes into close physical proximity to the Jκ genes (26, 52, 59). This data also indicates that the super-enhancer (formed by the iEκ, 3’Eκ, and Ed enhancers) located near the Jκ genes is in direct physical contact with other enhancers, including E34 and E88, and with enhancer-associated epigenetic marks across the Vκ gene region (26, 52, 59, 113, 114). These enhancer-associated epigenetic marks are frequently found on the RSSs, gene body and promoters of the Vκ genes, likely facilitating the linking of the Jκ gene RSS with the Vκ gene RSS. Furthermore, enhancers in the Vκ gene region of the Igκ locus regulate the compartmentalization of the chromatin domain by controlling the epigenetic status of the domain (59). This agrees with the observations that compartmental domains correlate directly with chromatin features (66-68, 115). A number of studies suggest that enhancer-enhancer and enhancer-promoter interactions are mediated by LLPS (29, 36-42). Chromatin domains with similar epigenetic signatures tend to cluster in 3D space (66-68, 115). Moreover, recent studies revealed that nuclear condensates act as mechano-active chromatin to draw paired genomic regions that are separated by vast genomic distances together, while mechanically opposing non-targeted regions located in the vicinity, to merge into the condensate (103). In a similar fashion, we propose that that the Igκ super-enhancer pulls the Jκ region and the E34 or E88 enhancers and loop domains into close physical proximity in a shared nuclear condensate (Figure 2D).

Furthermore, recent studies in cell lines by the Alt lab (112), or in mice by the Busslinger lab (116), support the model that the Igh and Igκ loci use different chromatin folding principles to achieve V gene recombination. Their findings demonstrate that the structure of the Igκ locus and Vκ-Jκ recombination patterns remain mostly unaffected by changes Wapl levels. Wapl is responsible for unloading the cohesin complex from the chromatin, which in turn impacts its resident time. Thus, changes in Wapl dynamics directly affect loop extrusion. In contrast with the lack of effect of Wapl levels on Igk rearrangement, the Igh locus chromatin folding and V_H to D_H-J_H recombination were highly susceptible to changes in Wapl expression (111, 112).

Studies using acute protein depletion and high-resolution chromatin conformation capture approaches reveal that most short- or mid-range enhancer-promoter interactions are largely cohesin-independent, with the caveat that cohesin-dependent loops created before the acute depletion may be maintained after the short-term depletion. (117-119). However, some enhancer-promoter interactions that rely on cohesin were also found in these studies, primarily long-range contacts (120, 121). Furthermore, depletion of cohesin or NIPBL caused super-enhancers to co-localize, generating many de novo interactions (68, 89). This observation suggests that the disruption of loop extrusion may facilitate enhancer-enhancer interactions located in different TAD, subTADs, or CTCF-loops. The lack of significant changes in Vκ-Jκ rearrangement in cell lines with low levels of Wapl and pre-B cells with high levels of Wapl suggest that enhancers interactions and the enhancer network in the Igκ locus are predominantly cohesin-independent, reinforcing the idea that the Igκ and the Igh locus use different chromatin folding principles to bring the V genes to the recombination centers to facilitate V(D)J recombination (111, 112).

While the loop extrusion-mediated RAG scanning mechanism elegantly explains the rearrangement process in the Igh locus, the orientation of the Vκ genes in the Igκ locus makes this mechanism insufficient to facilitate Vκ to Jκ rearrangement (4, 112, 116). Mechanistic principles of enhancer-mediated interactions and chromatin compartmentalization could explain how Vκ gene segments' inversional rearrangement is facilitated. Unlike loop extrusion-mediated V_H to D_HJ_H proximity, the juxtaposition between the Vκ and Jκ genes segments created by the enhancers' physicochemical properties is not bound by directionality. This would allow the chromatin fiber in the Vκ-gene region to engage the RAG complexes over the Jκ genes in both orientations (Figure 2E). This orientation-independent interaction would facilitate the rearrangement of the Vκ genes in forward or reverse orientations.

Conclusions

Enhancers in the V-gene region of the immunoglobulin loci add a critical layer of control to the formation of the B cell receptor repertoire by V(D)J recombination, and likely the T-cell receptor repertoire as well. To date, germline deletion of these elements, two in the Igκ locus and one in the Igh locus, show a remarkably similar phenotype in that V genes in the vicinity of the deleted enhancer are significantly impaired in rearrangement. These deletions also revealed that these enhancers are involved in dominant long-range chromatin interactions with a few other enhancers in the V gene containing part of the locus, along with the traditionally defined enhancers at the C region end of the loci. In this way, they shape the 3D structure of their cognate Ig locus. It remains to be determined whether other enhancers within the V gene-containing part of the loci regulate the 3D chromatin structure and/or other specific V-gene rearrangement to generate proper and specific antibody responses to additional evolutionary conserved microbial epitopes and autoantigens.

A diverse repertoire is critical for fighting harmful pathogens and eliminating cancerous cells, among others. However, in the arms race between the host and pathogens, the immune system must be prepared for an overwhelming and forceful attack from pathogens in order for the host to survive those critical early stages of infection. Effectively including V gene segments in antibodies that target specific conserved epitopes in pathogens confers an advantage to the host. There are several examples of V gene germline-encoded specificities for pathogen or virulent factors-associated epitopes (122, 123). However, without a functional level of these V gene segments in the repertoire, this protective characteristic would be ineffective. Enhancers provide appropriate functionality to ensure that these V genes are represented in the repertoire at optimal levels that enable early responses to pathogens.

Abbreviations:

TF

transcription factor

LLPS

liquid-liquid phase separation

IDR

intrinsically disordered regions

TAD

topologically associating domains

PC

phosphorylcholine

PtC

phosphatidylcholine