Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Jun 9.
Published in final edited form as: Cell Rep. 2016 May 26;15(10):2159–2169. doi: 10.1016/j.celrep.2016.04.087

Active and inactive enhancers co-operate to exert localized and long-range control of gene regulation

Charlotte Proudhon 1,*, Valentina Snetkova 1,*, Ramya Raviram 1,2, Camille Lobry 1,3, Sana Badri 1,2,4, Tingting Jiang 5, Bingtao Hao 1, Thomas Trimarchi 1, Yuval Kluger 5, Iannis Aifantis 1, Richard Bonneau 2,6,7, Jane A Skok 1,8
PMCID: PMC4899175  NIHMSID: NIHMS788397  PMID: 27239026

Abstract

V(D)J recombination relies on the presence of proximal enhancers that activate the antigen receptor (AgR) loci in a lineage and stage specific manner. Unexpectedly we find that both active and inactive AgR enhancers co-operate to disseminate their effects in a localized and long-range manner. Here we demonstrate the importance of short-range contacts between active enhancers that constitute an Igk super-enhancer in B cells. Deletion of one element reduces the interaction frequency between other enhancers in the hub, which compromises the transcriptional output of each component. We further establish that in T cells long-range contact and co-operation between the inactive Igk enhancer, MiEκ and the active Tcrb enhancer, Eβ, alters enrichment of CBFβ binding in a manner that impacts Tcrb recombination. These findings underline the complexities of enhancer regulation and point to a role for localized and long-range enhancer-sharing between active and inactive elements in lineage and stage specific control.

Keywords: Enhancer-sharing, super-enhancer, nuclear architecture, localized and long-range contacts, Igk, Tcrb, transcriptional output, transcription factor binding, gene regulation

Introduction

B and T lymphocyte development is driven by V(D)J recombination, a process through which V (variable), D (diversity) and J (joining) coding gene segments within each of the seven antigen receptor (AgR) loci are rearranged to create a vast repertoire of receptors (Alt et al., 2013). This process is important because lymphocytes require a set of receptors that can recognize and respond to a wide variety of foreign antigens as part of the adaptive immune response. V(D)J rearrangement is mediated by the recombination-activating gene (RAG) complex (Schatz and Swanson, 2011). RAG targets the recombination signal sequences (RSSs) that flank each V, D and J gene segment and creates a synapse between two segments that are then ligated together via non-homologous end joining.

Recombination occurs in a lineage specific manner so that T cell receptor (Tcr) and immunoglobulin (Ig) loci are only fully rearranged in T and B cells, respectively. In addition rearrangement is ordered by stage within a given lineage. In B cells the Ig heavy chain (Igh) is rearranged at the pro-B cell stage of development prior to Ig light chain (kappa or lambda) rearrangement in pre-B cells, while in T cells the T cell receptor beta locus (Tcrb) is recombined in CD4−CD8− double negative DN2/3 cells prior to T cell receptor alpha (Tcra) recombination in double positive (DP) cells. As the RAG proteins and their RSS targets are the same for each AgR locus, lymphocytes restrict rearrangement by controlling the accessibility of the individual loci (Yancopoulos and Alt, 1985). Opening up of the loci occurs at multiple levels including DNA demethylation, acquisition of active histone marks, initiation of sense and antisense germline transcription and nucleosome repositioning (Johnson et al., 2010).

The AgR loci have served as a rich model system for analyzing the impact of nuclear organization and chromatin architecture on gene regulation (Proudhon et al., 2015). The first evidence that shuttling of loci between repressive and active nuclear compartments (the nuclear lamina or pericentromeric heterochromatin (PCH) and accessible euchromatic regions) has an impact on gene regulation came from tracing the movements of AgR loci during development (Goldmit et al., 2005; Kosak et al., 2002; Roldan et al., 2005). Moreover, studies focusing on chromatin architecture demonstrated that reversible changes in ‘locus contraction’ alter the conformation of each locus, bringing mid and distal V gene segments into contact with proximal DJC domains thereby enabling recombination to occur between widely separated gene segments (Fuxa et al., 2004; Roldan et al., 2005; Sayegh et al., 2005; Skok et al., 2007). These changes are regulated by lineage and stage specific expression of transcription factors that activate enhancers, promoters and other regulatory elements within each locus (Proudhon et al., 2015).

Our previous work demonstrates that AgR enhancers can also act in trans to control accessibility and stage specific regulation of recombination via inter-chromosomal interactions between different loci. Specifically, in developing B cells, the 3’ enhancer of Igk (3’Eκ) mediates transient association of Igh and Igk at the pre-B cell stage after completion of Igh recombination and at the onset of Igk rearrangement (Figure 1A). Igh-Igk pairing repositions the unrearranged Igh allele at PCH and induces its ‘decontraction’. This prevents ongoing Igh rearrangement involving mid and distal VH gene segments (Hewitt et al., 2008). These and other studies indicate that individual enhancers co-operate with other regulatory regions in gene regulation and that control is facilitated by physical contact between participating elements (Collins et al., 2011; Hewitt et al., 2008). Nonetheless, it is not known to what extent enhancer sharing occurs and whether this phenomenon has a widespread impact on gene regulation.

FIGURE 1. Enhancer hubs and their impact on super-enhancer activity.

FIGURE 1

Open in a new tab

(A) Top: Scheme showing the location of the Igk AgR locus with its respective enhancers: MiEκ, 3’Eκ and Edκ on murine chromosome 6. Bottom: Outline of the different stages of B cell development. Stages under investigation are highlighted in orange (pre-B and immature B). (B) Left: Distribution of H3K27Ac signal across the peaks identified by MACS in pre-B and immature B cells with super-enhancers containing an exceptionally high amount of H3K27Ac. Right: H3K27Ac signal at the 3’ end of Igk in pre-B and immature B cells with the region defined as the super-enhancer highlighted. ATAC-seq profiles of the region in wild-type pre-B cells. (C) Detailed scheme showing the location of MiEκ and 3’Eκ 4C baits. (D) 4C signal normalized by DESeq2 in 5kb windows sliding by 0.5kb for ~50kb region neighboring the MiEκ and 3’Eκ baits in WT versus enhancer-deficient cells. Filled circles highlight significant differences in 4C-seq counts identified by DESeq2 analysis of the plotted region. Transcriptional output within the region is represented below each plot by RNA-seq profiles. (E) Model showing the organization of the individual enhancer elements within the Igk super-enhancer in wild-type versus MiEκ−/− and 3’Eκ−/− pre-B cells. See also Figure S1.

Given that chromatin is organized within the nucleus in a manner that promotes contacts between regulatory elements, it is important that we determine the functional significance of these associations. Here we highlight the importance of lineage specific short- and long-range co-operation between enhancer elements, focusing specifically on the impact of the Igk enhancers and their diverse functions in lymphocyte development. The role of the individual Igk enhancers in regulating Igk has been well documented by previous studies as detailed in the results section (Inlay et al., 2002; Inlay et al., 2004). In addition, the Igk MiEκ, 3'Eκ and Edκ enhancer cluster has been identified as being a super-enhancer in mature B cells (Qian et al., 2014). Here we show that according to the criteria defined by Rick Young’s lab, the classification of this cluster as a super-enhancer (Whyte et al., 2013), extends to the pre-B cell compartment where the Igk locus undergoes rearrangement. Although super-enhancers have received a great deal of attention in the scientific press it is not known whether the clustering of enhancer elements is functionally important. To address this question we performed high-resolution circularized chromosome conformation capture coupled with deep sequencing (4C-seq) using different bait sequences within the Igk super-enhancer in B cells. We demonstrate that the three enhancers exhibit strong contacts in wild-type cells leading to the formation of an enhancer hub. Deletion of either the MiEκ or 3'Eκ reduces the interactions in which each enhancer participates and disrupts pairwise interactions between other component enhancers leading to the dissolution of the hub. Importantly we find that the loss of enhancer contacts is linked to a reduction in transcriptional output of all three partner enhancers. These data suggest that synergistic contacts between the individual components of a super-enhancer are important for their activity.

The Igk enhancer, MiEκ has previously been reported to be important for regulating Igk activation in B versus T lineage cells (Pierce et al., 1991). Thus, we focused our next question on the impact of the MiEκ in regulating Igk recombination in T cells. We did not find any defect in the regulation of lineage specific Igk recombination but much to our surprise we uncovered a role for this enhancer in regulating rearrangement of Tcrb, a locus that is located 26Mb away from Igk on the same chromosome, via long-range intra-chromosomal contacts. The MiEκ (which is active in B cells but inactive in T cells) cooperates with the Tcrb enhancer, Eβ in regulating Tcrb recombination in DN cells via long-range intra-chromosomal contacts. Deletion of MiEκ leads to inefficient Tcrb recombination resulting in a delay in T cell development. These alterations are accompanied by a significant reduction in the enrichment of the RUNX1 DNA binding partner, CBFβ, on Eβ. Since RUNX1 is known to be critical for Tcrb activation this finding explains the impact of MiEκ on Tcrb regulation (Majumder et al., 2015). Thus, despite the large distance separating the Tcrb and Igk loci, Eβ and MiEκ co-operate to regulate Tcrb recombination via a long-range contact that relies on the presence of both enhancers.

In sum, these studies indicate that the Igk enhancers play diverse roles in lymphocyte development as a result of lineage and stage specific chromatin interactions. This data adds to our previous finding that a 3’Eκ-mediated inter-chromosomal interaction with Igh alters locus conformation and prevents ongoing Igh rearrangement after productive rearrangement on one allele (Hewitt et al., 2008), These studies underline the complexity of enhancer-mediated regulation and the importance of enhancer co-operation in optimizing gene function through localized and long-range contacts. Unexpectedly we find that enhancer sharing can involve active as well as inactive regulatory elements, which may be relevant for gene regulation in other settings.

Results

Enhancer hubs and their impact on super-enhancer activity

Super-enhancers are defined as a group of enhancers in close linear genomic proximity that are highly enriched for binding of transcription factors and histone modifications that are characteristic of enhancer elements (Whyte et al., 2013). Despite the definition there has been some controversy concerning the functional significance of enhancer clustering (Pott and Lieb, 2015). Moreover, there is as yet no evidence that contacts between enhancers that make up these super-regulatory elements are functionally important. To investigate this, we focused on the Igk enhancer cluster that is made up of the individual enhancer elements, MiEκ, 3'Eκ and a third enhancer, Edκ (Figure 1A), each of which binds different transcription factors. Three enhancers have overlapping and distinct functions that contribute to the regulation of Igk. MiEκ and 3’Eκ are both important for rearrangement and deletion of either one leads to a reduction in the frequency of kappa expressing B cells, while the double mutant is sufficient to abrogate Igk recombination altogether (Inlay et al., 2002; Inlay et al., 2006). Similarly, individual 3’Eκ and Edκ deletions lead to a reduction in germline and transcription of rearranged Igk in developing pre-B and mature splenic B cells, while in the double knockout mice expression is almost negligible (Zhou et al., 2010).

The Igk enhancer cluster has been identified as a super-enhancer in mature B cells (Qian et al., 2014) but its status in developing B cells is not known. We used iChIP (indexingfirst chromatin immunoprecipitation, (Lara-Astiaso et al., 2014) to confirm that the region is also enriched for high levels of H3K27Ac in pre-B and immature B cells where the Igk locus undergoes recombination (Figure 1B). Next, we asked whether the Igk enhancers that constitute the super enhancer are in physical contact in 3D space in pre-B cells. To investigate this we performed high-resolution circularized chromosome conformation capture coupled with deep sequencing (4C-seq) using baits located adjacent to MiEκ and 3’Eκ (Figure 1C). The bait location was selected to allow us to examine interactions in wild-type as well mutant cells where the whole enhancer region is deleted. In these experiments, for the initial digestion step we used the 4-base pair (bp) cutter restriction enzyme, NlaIII that cuts approximately every 200 nucleotides (van de Werken et al., 2012). Datasets for each of the viewpoints were generated from wild-type as well as MiEκ and 3'Eκ enhancer deficient pre-B cells (Table S1).

In wild-type cells we detected a high frequency of contact between MiEκ, 3'Eκ and Edκ. We found that deletion of either the MiEκ or the 3’Eκ interferes with interactions between each enhancer and the others. Their absence also disrupts interactions between the two other partner enhancers in the hub: in MiEκ−/− cells the contact frequency between 3’Eκ and Edκ was significantly reduced, which is especially apparent from the 4C signal of the 3’Ek bait. Likewise, the contact frequency between MiEκ and Edκ is significantly reduced in 3’Eκ−/− cells (Figure 1D). To determine whether the changes in contact have a functional effect we performed RNA-seq to examine the impact of enhancer deletions on transcription. We identified non-coding transcripts that span or originate at Igk enhancers by assembling a new transcriptome based on our RNA-seq data. We then designed primers to detect the transcripts flanking each enhancer to quantify their expression by qPCR. For the 3’Eκ and Edκ we selected primers that amplify short bidirectional enhancer RNAs (eRNAs) originating at each enhancer, while for the MiEκ the primers amplify transcripts that extend through the enhancer. With this approach we found that a loss of either the MiEκ or the 3’Eκ leads to a reduction in transcriptional output of the other two enhancers (Figure 1D and Figure S1A). We also examined the impact of the enhancers on Igk transcriptional output and found a dramatic decrease in transcription across the locus in mutant cells, with the MiEκ deletion causing the greatest impact (Figure S1B, C). These findings indicate that the presence of MiEκ and 3’Eκ boosts the transcriptional output of the other enhancers in a manner that is linked to their physical contact. They also add new mechanistic insight into super-enhancers and highlight the interdependent nature of the individual enhancer components and the connections between transcriptional output and the contacts between them.

The MiEκ regulates T cell development and Tcrb rearrangement

Previous studies with plasmid constructs carrying Igk enhancers and Igk recombination substrates indicate that the MiEκ is important for regulating Igk activation in B versus T lineage cells (Pierce et al., 1991). However, the effect of the MiEκ on the endogenous locus has not been examined in MiEκ deficient developing T cells. To investigate this we performed immuno-FISH to examine the impact of the MiEκ on Igk rearrangement in developing T cells (Figure 2A). Although we found no evidence for the occurrence of Igk rearrangement in T cells we unexpectedly discovered that an absence of MiEκ disrupts recombination of the Tcrb locus. Immuno-Fluorescent-In-Situ-Hybridization (Immuno-FISH) analysis showed a significant reduction of γH2AX foci associated specifically with Tcrb in MiEκ−/− DN cells, indicating a decreased frequency of breaks in this locus in recombining cells (Figure 2B). Additionally, a germline retention assay, which amplifies a region of Tcrb that is deleted upon rearrangement, revealed a significant reduction in rearrangement in MiEκ−/− versus wild-type DN cells (Figure 2C).

FIGURE 2. The MiEκ regulates T cell development and Tcrb rearrangement.

FIGURE 2

Open in a new tab

(A) Outline of the different stages of T cell development. Stages under investigation are highlighted in blue (DN2/3 and DP). (B) DNA-FISH combined with immunofluorescence to detect γH2AX-containing repair foci at the Igk and Tcrb loci. Bar graphs represents three independent experiments. Error bars reflect SEM. Examples of confocal sections are shown in the panels below. (C) Germline retention assay, performed on genomic DNA, to assess the proportion of Tcrb loci in germline versus rearranged conformation. Scheme showing the primer design. Plots showing the percentage of retention in wild-type and MiEκ−/− DN2/3 cells normalized by analysis of RAG1 deficient control cells (a combination of two independent repeats of two wild-type versus two MiEκ−/− samples). (D) Left: FACS analysis of DN cell development in thymi isolated from wild-type and MiEκ deficient mice. Right: Bar plots show quantification of differences in the absolute numbers of wild-type versus MiEκ−/− DN cells per thymus from a combination of three independent experiments. Cells were gated on Thy1.2+TCRbloCD4−CD8− and the DN population defined by expression of CD44 and CD25 as followed: DN1 CD44+CD25−, DN2 CD44+CD25+, DN3 CD44−CD25+ and DN4 CD44−CD25−. Error bars represent SEM. See also Figure S2.

Since AgR rearrangement is intimately connected to developmental progression we next asked whether the absence of MiEκ could affect T cell development. Detailed FACS analysis of the different cell stages revealed a significant accumulation of DN3 cells in MiEκ−/− mice, which is the stage at which Tcrb rearrangement occurs (Figure 2D). Given this finding we checked to see whether γδ T cell development was affected, but found no difference in this compartment. Since αβ T cell development was perturbed we next asked whether the developmental identity of the mutant cells was altered. For this we performed an unsupervised clustering analysis comparing expression levels of key B, T and common lymphoid specific genes in wild-type, MiEκ and 3’Eκ mutant samples. We supplemented this analysis with cells deficient for Eβ, which is a sole documented enhancer of Tcrb. Besides DN and pre-B cells we also examined subsequent developmental stages (DP and immature B cells) from each lineage: to have a fuller picture of early T and B lymphocyte development (Figure S2A). We also checked this at a genome-wide level by performing a Principal Component Analysis based on expression values normalized by DESeq2 (Figure S2B). In both instances the results show clustering of wild-type cells with their mutant counterparts demonstrating that developmental stage and lineage identity is maintained in enhancer deficient cells.

Tcrb has always been thought to be under the regulation of a single enhancer, Eβ. Our findings indicate that the MiEκ also contributes to the regulation of this locus. Indeed, the MiEκ deficient phenotype, although much less pronounced, resembles the phenotype of Eβ deficiency in which a complete absence of Tcrb rearrangement is associated with a block in development at the DN3 stage of T cell development (Bories et al., 1996; Bouvier et al., 1996).

The presence of Jκ-Cκ transcripts in Igk is linked to a long-range contact with Tcrb

The finding that MiEκ has an impact on Tcrb recombination and T cell developmental progression is surprising since Igk enhancers are only known to be active in B cells and inactive enhancers have never been shown to contribute to gene regulation in any way. Furthermore, the Igk locus and its associated enhancers are located 26Mb away from Tcrb on mouse chromosome 6. To investigate the mechanism underlying MiEκ-mediated control of Tcrb recombination we asked if MiEκ is active in DN cells. ATAC-seq analysis (Buenrostro et al., 2013) enabled us to determine that there was no signal associated with the MiEκ (Figure 3A, red box a). Furthermore, there was no enrichment of H3K27Ac, a mark that is used to characterize enhancer activity. However, RNA-seq showed transcription in the region around the enhancer. Assembly of a new transcriptome based on our RNA-seq data identified Jκ-Cκ transcripts (Figure 3A, red box a) that originate at each Jκ segment except Jκ3 and exclude the intervening MiEκ enhancer (Figure S3A). The Jκ-Cκ transcripts are significantly downregulated in MiEκ as well as in Eβ knockout DN cells cells (with an average fold change of 1.41, p-value 0.0107; 2.33, p-value 0.000175, between four Jκ-Cκ transcripts in MiEκ and Eβ knockouts, respectively).

FIGURE 3. The presence of Jκ-Cκ transcripts in Igk is linked to a long-range contact with Tcrb.

FIGURE 3

Open in a new tab

(A) RNA-seq, ATAC-seq and iChIP for H3K27Ac reveal the activity of the 3’ Igk interacting domain. Nc-RNA track displays transcripts identified by transcriptome assembly using cufflinks V2.2.1 (see methods for more details). The bottom track shows the 3’ Igk interacting domain called by 4C-ker from the Eβ viewpoint. (B) Detailed scheme showing the location of the Eβ 4C bait relative to the enhancer. (C) Interaction profile of Eβ for the 6Mb region encompassing Igk in wild-type DN, DP, pre-B and immature B cells using 4C counts in 200kb windows sliding by 20kb. (D) Interaction profile of Eβ with the 3’ Igk interacting domain in WT versus Eβ and MiEκ deficient DN cells using 50kb windows sliding by 5kb. A purple bar delineates a 200kb window, which chromosome-wide DESeq2 analysis identified as having a significantly different 4C signal in Eβ−/− versus wild-type DN cells. A t-test was performed on 4C counts to assess the same 200kb window for differences between wild-type versus MiEκ−/− (gold bar). See also Figures S3 and S4.

As expected there is a major difference between transcriptional output and accessibility in the region in B versus T cells (Figure S3B).

Our finding that the MiEκ has an extra long-range regulatory impact on Tcrb raises the question of whether co-operation between Tcrb and Igk enhancers involves a long-range contact. To examine this we performed 4C-seq using Eβ rather than MiEκ as our bait, because active baits give a more specific contact profile compared to inactive baits (Figure S3C). The bait used in these experiments was located adjacent to the Eβ enhancer as shown in Figure 3B. Although Tcrb and Igk are separated by a considerable linear genomic distance, Eβ appears to establish strong contacts with both 5’ and 3’ ends of the Igk locus in T cells as shown by 4C (Figure S3C). A local analysis of the interaction profile from the Eβ bait across the 6Mb region encompassing and surrounding Igk (Igk occupies 3Mb of DNA) in wild-type DN, DP, pre-B and immature B cells is shown in Figure 3C. This analysis and the 4C profile shown in Figure S3C demonstrate that Eβ contacts with 5’ and 3’ ends of Igk are specific to the T cell lineage and are not detected in B cells. We used 4C-ker, an in-house pipeline that relies on a Hidden-Markov Model (HMM) based analysis to identify regions of chromosome 6 that interact significantly with the bait region (Raviram et al., 2016) (Figure S3C). Using this method of analysis we identified a ‘3’ Igk interacting domain’ that encompasses Jκ and Cκ segments, the three enhancers of Igk, MiEκ, 3’Eκ and Edκ as well as a region that extends 200kb downstream of the 3' end of this locus (Figure 3C, grey box). Analysis of wild-type versus enhancer mutant cells revealed that in DN T cells, the Eβ contact with the 3’ Igk interacting domain is dependent upon the two enhancers Eβ and MiEκ. This contact is abrogated upon deletion of Eβ and reduced in the absence of the MiEκ, in a statistically significant manner as determined by t-test on 4C counts of the 3’ Igk interacting domain (Figure 3C, grey box and Figure 3D, purple and gold bars).

The peak of the 4C signal at the 3’ Igk interacting domain coincides with the actively transcribed coding genes, Rpia and Eif2ak3. Additionally, we identified bidirectional monoexonic transcripts (that are typically found at enhancers) within the Rpia-Eif2ak3 Intergenic Region that were also associated with an ATAC-seq and H3K27Ac peaks (Figure 3A, red box b). The latter highlights the presence of a potential new regulatory element in the area that we named REIR.

To confirm the interaction between Eβ and the ‘3’ Igk interacting domain’ in T cells from the reverse viewpoint we designed an additional bait in the REIR region, where the interaction is the highest and where we observe the greatest changes in contact in enhancer deficient cells (Figure S4A). As expected, the REIR bait also identifies a specific contact with Eβ in T cells that is reduced in Eβ and MiEκ knockout DN T cells (Figure S4B, C). Together these data implicate both active and inactive enhancers, Eβ and MiEκ respectively, in the activation of Jκ-Cκ transcripts as well as in the establishment of the contact between Eβ and the 3’ Igk interacting domain in T cells.

The MiEκ regulates CBFβ binding at Eβ in DN cells

To further investigate the impact of MiEκ on Tcrb recombination we focused on transcription factors that are known to contribute to this process. In particular, RUNX1 binding has been shown to be essential for Eβ activation and subsequent Tcrb recombination: mutations of the RUNX binding sites on Eβ leads to a similar phenotype as deletion of Eβ with respect to Tcrb transcription and recombination (Majumder et al., 2015). As we could not detect any ATAC-seq peak at MiEκ in DN cells we assumed that no transcription factors were bound to this element (Figure 3A). Thus we asked whether other sites in the ‘3’ Igk interacting domain’ bound CBFβ (the requisite heterodimeric binding partner of all RUNX proteins, which stabilizes their interaction with DNA). ChIP-qPCR revealed that REIR is bound by CBFβ and the enrichment is not significantly affected by deletion of either the Eβ or the MiEκ (Figure 4A). In contrast, ChIP analyses showed that binding of CBFβ on Eβ was significantly reduced in MiEκ deficient DN cells and the levels were intermediate between wild-type and T cells with RUNX mutated Eβ binding sites (Figure 4B).

FIGURE 4. The MiEκ regulates CBFβ binding at Eβ in DN cells.

FIGURE 4

Open in a new tab

(A) CBFβ ChIP-qPCR for REIR plotted as fold enrichment over negative control. (B) CBFβ ChIP-qPCR for Eβ plotted as fold enrichment over negative control. T-test to determine significance of fold enrichment change at Eβ in EβRUNX (RUNX binding sites at Eβ mutated, p-value = 0.004398) and in MiEκ−/− (p-value= 0.005621) DN cells. Error bars represent SEM of three experiments. (C) Model implicating the CBFβ-bound REIR and MiEκ in co-operatively regulating transcription factor binding at Eβ. In wild-type T and B cells Eβ and MiEκ respectively regulate rearrangement of Tcrb and Igk. Three enhancers of Igk (MiEκ, 3’Eκ and Edκ) comprise the super-enhancer in developing B cells with individual enhancers interacting and co-operatively regulating transcriptional output. In wild-type DN cells Jκ-Cκ transcripts are expressed at the 3’ end of Igk. The contact between Eβ and the 3’ Igk interacting domain encompassing CBFβ-bound REIR promotes Tcrb recombination in an Eβ and MiEκ dependent manner. An absence of MiEκ leads to a reduction in the contact and transcription at the 3’ end of Igk, which in turn leads to a reduction in the localized concentration of CBFβ and RUNX1 in the proximity of Eβ, thereby impairing its function.

A reduction in CBFβ binding at Eβ provides a mechanism that explains the reduction in efficiency of Tcrb rearrangement and the delay in T cell development that result from MiEκ deletion. However, since MiEκ does not bind CBFβ it is not clear how an absence of this enhancer could affect binding of the latter to Eβ. To explain this regulation we put forward a model implicating the MiEκ mediated Jκ-Cκ transcripts and the CBFβ bound element REIR, in co-operatively promoting transcription factor binding at Eβ, as results of a long-range contact (Figure 4C). In the absence of MiEκ, reduced interaction between Eβ and the ‘3’ Igk interacting domain’ encompassing both MiEκ and REIR, leads to a reduction in binding of CBFβ and RUNX1 to Eβ, thereby impairing its function. These findings are consistent with the idea that clustering of co-regulated regulatory elements promotes enhancer sharing and leads to a high local concentration of regulatory factors which influences the levels of their enrichment in a manner that could have functional consequences (Cheutin and Cavalli, 2012). Loss of the MiEκ, which reduces expression of Jκ-Cκ transcripts and impairs contact between REIR and Eβ could dilute out the concentration of CBFβ in the region, altering the regulation of Tcrb.

Discussion

Enhancers are known to be important for activating their associated loci in a lineage and stage specific manner. Here we have examined the impact of the Igk enhancers in B cells where they are active and T cells where they are inactive. Surprisingly we find evidence of enhancer-sharing in both contexts. We chose to center our initial analyses on the Igk superenhancer to address the question of whether clustering of enhancers has functional significance (Pott and Lieb, 2015). Using distinct 4C-seq baits located on different enhancers within the Igk super-enhancer we analyzed the impact of individual enhancer deletions on the contact and transcriptional profile of the region. Here we show that in wild-type cells the individual elements within a super-enhancer are indeed in close contact in 3D space. Importantly, we find that deletion of an individual element reduces pairwise associations between the other components. Thus in terms of chromatin contacts the enhancers act synergistically to promote the formation of a hub that is dependent on the presence of all three regulatory elements as loss of a single enhancer leads to its dissolution. In contrast, deletion of a single enhancer does not totally abolish transcription of the other elements, but leads to a reduction in transcriptional output on the two remaining enhancers. This is consistent with the impact single enhancer deletions have on Vκ transcriptional output. Furthermore, the finding that deletion of either the MiEκ or the 3’Eκ leads to a reduction in Igk recombination and the double MiEκ and 3’Eκ knockout abrogates rearrangement altogether (Inlay et al., 2006), suggests that in terms of their function, these enhancers act additively. Similarly, single and double 3’Eκ and Edκ deletions act additively with regards to their impact on Igk transcriptional output (Zhou et al., 2010). Given that all three enhancers are in contact in pre-B cells and their interaction boosts the transcriptional output of all the partner elements, it is likely that the third enhancer (Edκ or MiEκ, respectively) in each case contributes to the functional outcome of rearrangement and transcription. These findings add new mechanistic insight into the interdependent nature of the individual enhancer components of super-enhancers and the importance of the contacts between them. Further analyses of other super-enhancers will need to be performed to determine whether this is true in other cases.

Using the Igk enhancers as a model system further led to the discovery that the inactive enhancer MiEκ can modestly but significantly influence Tcrb recombination in DN T cells. This is unexpected as inactive enhancers have not previously been documented as having any functional role in gene regulation. It is of note that although we find no evidence for transcriptional activity, enrichment of H3K27Ac or binding of transcriptional regulators at MiEκ in DN T cells, we could show that MiEκ- and Eβ-mediated control of Jκ-Cκ transcripts on the Igk locus is connected to an MiEκ and Eβ-mediated interaction between Eβ and the 3’ Igk interacting domain. Importantly, deletion of MiEκ in DN cells leads to a significant reduction in Tcrb rearrangement and a developmental delay at the DN3 stage that can be explained by a reduction in CBFβ binding to Eβ. We propose that MiEκ and Eβ-mediated contact between Eβ and the 3’ Igk interacting domain brings the CBFβ bound element, REIR into contact with Eβ, thereby increasing the local concentration of CBFβ in the area. To conclusively validate this hypothesis we would need to knockdown Jκ-Cκ transcripts and mutate REIR so that it can no longer bind CBFβ.

Although the effect of the MiEκ deletion on Tcrb regulation is nowhere near as profound as Eβ deletion, given the linear distance separating the 3’ Igk interacting domain from Tcrb (29 Mb), it is surprising that it can exert any effect at all, let alone a significant impact on transcription factor binding and recombination. Certainly it is now well established that co-regulated genes associate in the nucleus and it has been proposed that their interactions may be important for streamlining gene expression (Schoenfelder et al., 2010). It has also been suggested that contact between loci that come together as a consequence of binding of a common regulator, increases the local concentration of the factor reinforcing its downstream effects (Cheutin and Cavalli, 2012). However, it was not previously known that association of regulatory elements could have a functional impact on the enrichment of transcription factor binding on partner loci. These findings demonstrate a new paradigm for extra-long range control and emphasize the complexities of gene regulation and the impact of enhancer sharing.

The way that chromatin is packaged with the nucleus is clearly a major determinant of enhancer sharing. Indeed, it has been proposed that TAD structures are important for limiting the influence of enhancers because by definition sequences within a TAD interact more frequently than those in different TADs. However, sequences from different TADs on the same and different chromosomes can and do interact with each other although these contacts are less well documented than proximal higher-frequency interactions. Nonetheless, our data indicates that these interactions can play a significant role in fine-tuning the regulation of target genes.

In sum these studies reveal that localized and long-range enhancer-sharing between active and inactive elements can impact gene regulation in a lineage and stage specific way. Our findings should be taken into consideration in gene targeting analyses of enhancer deletions as the phenotype they generate may not solely relate to the primary biological function of the target gene, but could include other unrelated effects that may not be easy to predict. In lineages and cell types where individual enhancers are known to be transcriptionally inert, they may still exert a significant influence on gene regulation as we have revealed here.

Experimental Procedures

Mice

Wild-type C57BL/6 mice 5–7 weeks of age were used as controls in these experiments. Eβ−/− mice were kindly provided by Dr. Pierre Ferrier (Bouvier et al., 1996). MiEκ−/− and 3’Eκ−/− mice (Inlay et al., 2002) were kindly shared by Dr. Mark Schlissel’s lab. The whole enhancer region is deleted in all these mutant mice. Animal care was approved by Institutional Animal Care and Use Committee (protocol number 150606-01, NYU School of Medicine).

Cell sorting

Single cell suspensions were prepared from thymus or bone marrow (BM) from mice of various genotypes. CD4 and CD19 microbeads (Miltenyi Biotec) selections were performed on thymocytes and BM cell suspension respectively using a Manual MACS® Cell Separator (Miltenyi Biotec) to reduce cell sorting time. Flow cytometry cell sorting was performed on Beckman Coulter MoFlo and Sony SY3200 machines as follow: from CD4 negative selection, DN2/3 cells were purified as Thy1.2+TCRbloCD4−CD8−CD25+ CD44−/+, from CD4 positive selection DP cells were purified as Thy1.2+TCRbintCD4+CD8+ and from CD19 positive selection, pre-B cells were purified as IgD−B220+IgM−c-kit−CD25+ and immature-B cells as IgD−B220+IgM+. Antibodies used were as follows: Thy1.2 PE-Cy7 (clone 53-2.1, eBioscience, 1:5,000 dilution), TCR-b APC-eFluor780 (clone H57-597, eBioscience, 1:500 dilution), CD4 APC (clone RM4-5, BD Biosciences, 1:500 dilution), CD8a FITC (clone 53-6.7, BD Biosciences, 1:500 dilution), CD25 PE (clone PC61, BD Biosciences, 1:500 dilution), CD44 PerCP-Cy5.5 (clone IM7, BD Biosciences, 1:500 dilution), IgD APC-Cy7 (clone 11-26c.2a, BioLegend, 1:500 dilution), CD45R/B220 PE-Cy7 (clone RA3-6B2, BD Biosciences, 1:500 dilution), IgM FITC (clone II/41, BD Biosciences, 1:500 dilution), CD117/c-kit APC (clone 2B8, eBioscience, 1:500 dilution).

FACS analysis

The same surface markers used for sorting cells were used for FACS analysis of DN cells from whole thymi derived from WT or MiEκ−/− mice. Cells were gated on Thy1.2+TCRbloCD4−CD8− and the DN population defined by expression of CD44 and CD25 as followed: DN1 CD44+CD25−, DN2 CD44+CD25+, DN3 CD44−CD25+ and DN4 CD44−CD25−. TCRγδ cells were also gated on CD4−CD8− population and defined as CD3+TCRγδ+. The analysis was performed using FlowJo.

RNA extraction

Following flow cytometry cell sorting, RNA was extracted using the RNeasy plus kit from Qiagen.

RNA-Seq methodology

Libraries where prepared according to the directional RNAseq dUTP method adapted from http://wasp.einstein.yu.edu/index.php/Protocol:directional_WholeTranscript_seq that preserves information about transcriptional direction. We generated at least 2 biological replicates per cell type. Additional replicates were generated for important samples: 2 more for DN WT and 1 more for DN MiEκ. Our libraries were sequenced with Illumina Hi-Seq 2000 in paired-end 50 cycles mode.

RNA-Seq analysis

Paired-end reads were mapped to the mm10 genome using TopHat2 (parameters: --no-coverage-search --no-discordant --no-mixed --b2-very-sensitive –N 1). Counts for Refseq genes were obtained using htseq-counts (Anders et al., 2015). Segments of antigen receptor loci were added as exons to RefSeq annotation file (Table S2). DESeq2 version 1.4 (Love et al., 2014) was used to normalize expression counts.

ncRNA scan

RNA-seq reads from DN and pre-B samples were mapped to the mm10 genome using TopHat2 (-N 1 –no discordant --no-mixed --library-type fr-firststrand -G (RefSeq GTF file provided). Transcripts from chromosome 6 were assembled using Cufflinks version 2.2.0. Transcript assemblies from the same cell type were merged using Cuffmerge. Merged transcripts were annotated with Cuffcompare using RefSeq GTF file. Following the lncRScan (Sun et al., 2012; Trimarchi et al., 2014) pipeline, we extracted lncRNAs with class code ‘u’,’x’,’i’ or ’s’. Multi-exonic transcripts with more than 200 nucleotides as well as with less than 300 nucleotides putative Open Reading Frame were selected. Mono-exonic transcripts overlapping antigen receptor loci enhancers were manually added back to the final non-coding transcriptome of pre-B cells. Bedtools multicov function was used to count the reads per each transcript, followed by differential expression analysis with DESeq2. Transcripts with an adjusted p-value <0.05 together with an absolute value of log2 fold change greater than 1 were assigned as differentially expressed.

qPCR of non-coding transcripts at Igk enhancers and Igk locus itself

RNA was extracted from 3 different biological replicates that were distinct from the samples used for the RNA-seq libraries. These were used for reverse transcription using the Superscript kit from Life technologies with random hexamers as primers. qPCR reactions were run in duplicates for each replicate using Applied Biosystems instrument. Each reaction was run with 4ng of template and 200nM final concentration of primers. Standard curves were run in parallel for each primer set to correct for primer efficiency. Actin and Tbp (TATA-box binding protein) loci were used as reference genes as they displayed stable expression across the 16 samples in our RNA-seq results. Relative expression to the reference genes was calculated using the 2−ΔCt method. Since no eRNAs originating at the MiEκ enhancer were detected by the ncRNA scan, qPCR primers for quantification of expression were set right on the enhancer in order to detect RNA that is transcribed through it (F: GGGGAAAGGCTGCTCATAAT; R: ACTGTAATCTGGGCCACCTG (188 bp)). For both 3’Eκ and Edκ, the primers were chosen to detect eRNAs originating at the respective enhancer. 3’Eκ antisense transcript, F: GATGGGATCACAGGGGTGTA; R: AACTTGGCCTGATCAAGAGG (198 bp). Edκ antisense transcript, F: GCATTTCCCTGGCTTCTATG; R: CAGCATGACTGGGAGAATCA (184 bp). Edκ sense transcript, F: AGTAAGGGCCTAGGGTTCCA; R: GCTTGTCATTTTCCCACTGC (142 bp). qPCR primers for Igk constant region Cκ F: CTGATGCTGCACCAACTGTA; R: ACGCCATTTTGTCGTTCACTG (151 bp).

ATAC-seq

We sorted 50,000 cells per replicate for pre-B, immature B and DN wild-type samples. The assay was performed in biological duplicates as described previously (Buenrostro et al., 2013) with several modifications. We amplified our libraries with KAPA HiFi polymerase and sequenced them with Illumina Hi-Seq 2000 using 50 cycles paired-end mode. Reads were aligned to mm10 genome with Bowtie2 (parameters: --no-discordant -p 12 --no-mixed -N 1 -X 2000). Potential PCR duplicates were removed from the reads with Picard-tools. ATAC-seq peaks were called with PeaKDEck (McCarthy and O'Callaghan, 2014) (parameters: -sig 0.0001 -PVAL ON).

iChIP

The genome-wide status of H3K27 acetylation was assessed with indexed chromatin immunoprecipitation (iChIP) as previously described (Lara-Astiaso et al., 2014). Briefly, chromatin from wild-type pre-B, immature B and DN cells was incubated with H3 antibody (Abcam, ab1791) overnight at 4°C. Two replicates were processed with 100K cells per replicate. Chromatin bound to H3 antibody was indexed with NEXTflex™ DNA Barcode adapters (Bio Scientific) and samples were pooled together for precipitation with H3K27Ac antibody (Abcam, ab4729) for four hours at 4°C. DNA was de-crosslinked, purified and amplified for fourteen cycles. Libraries were sequenced with HiSeq2500 paired-end 50 cycles. Sequencing reads were aligned to the mouse genome mm10 using Bowtie2 with parameters --no-discordant --nomixed -N 1. Enriched regions of H3K27Ac were identified using MACS 1.4.2 (Zhang et al., 2008) with parameters -p 1e-5 -g mm -B --single-profile with input libraries as background signal. To identify super-enhancers, we used ROSE (Hnisz et al., 2013), with H3K27Ac MACS peaks as constituent enhancers with the following parameters: -g MM10 -s 20000 -t 2500.

4C-Seq

The 4C template was produced by two successive rounds of digestion-ligation using two different 4bp cutters, NlaIII and DpnII, as previously described (van de Werken et al., 2012). 4C templates were amplified with PCR primers specific for our bait regions, MiEκ (NIaIII primer 5’-ACATTCTTTTCAGTTCCATG-3′, DpnII primer 5′-CTTCTACCCCAAACATCA-3′), 3’Eκ (NIaIII primer 5’-CCATCTGGTGCAGGAGCATG-3′, DpnII primer 5′-TCCCTGACTGTGAACTGAAGG-3′), Eβ (NIaIII primer 5’-TGTGGATTGATTAAGCCATG-3′, DpnII primer 5′-TGAGCATTTCTTTCTCCTAGTGG-3′) and REIR (NIaIII 5’-CACCCAGAGCTCCTTCCATG-3′, DpnII 5′-CAAGTCTTGGAGCCTTCCTG-3′), respectively. The Illumina-specified adapters for Illumina sequencing together with multiplexing indexes were included at the 5’ end of each primer. Our 4C libraries were sequenced on the Illumina Hi-Seq 2000 using single-read 100-cycle runs. The number of reads obtained for each dataset is listed in Table S1.

Sequence reads from 4C-Seq were mapped to a reduced genome derived from mm10 containing unique sequences adjacent to restriction enzyme sites. Reads were mapped using Bowtie2 (Dryden et al., 2014; Langmead and Salzberg, 2012) (−N=0 and −5 was used to trim the barcode and primer sequences). The counts for each fragment were then obtained from the mapped SAM files. To determine significant differences in interactions between three Igk enhancers, DESeq2 analysis of 4C counts from MiEκ and 3’Eκ baits was performed on 5kb windows sliding by 0.5kb for regions (~50–80kb) around the baits in WT versus enhancerdeficient cells (FDR adjusted p-value<0.05).

To find chromosome-wide interacting domains, an in-house Hidden Markov Model based pipeline (4C-ker) was used (Raviram et al., 2016). 4C-ker uses a 3 state HMM to define regions of high-interaction using an input of window counts corrected for the effect of linear distance from the bait. High interacting domains were defined across the entire bait chromosome (4C-ker parameter of k=10) for Eβ bait.

Differential 4C-Seq analysis was performed using DESeq2 on 4C counts in 200kb windows overlapping by 20kb throughout the entire chromosome 6 (FDR adjusted p-value<0.05) comparing wild-type versus Eβ−/− DN cells. T-tests were performed using DESeq2 normalized 4C counts within a 200kb region at the Igk 3’ interacting region to compare Eβ bait 4C signal in wild-type versus MiEκ−/− DN cells.

Immuno-DNA-FISH

We used oligo probes covering the entire Tcrb and Igk loci to label respective genes. Combined detection and co-localization analysis of γH2AX and Tcrb or Igk probes was carried out as previously described (Hewitt et al., 2009). Briefly, cells were adhered to poly-L-lysine coated coverslips, fixed with 2% paraformaldehyde/PBS and permeabilized with 0.4% Triton/PBS. Samples were blocked for 30 min with 2.5% BSA, 10% normal goat serum and 0.1% Tween-20 in PBS at 22°C. Samples were incubated for 1 h at 22°C with an antibody to phosphorylated serine 139 of H2AX (γH2AX; JBW301, Millipore) diluted 1:400 in blocking solution. Cells were rinsed with 0.2% BSA / 0.1% Tween-20/PBS and stained for 1 h with goat anti-mouse IgG Alexa Fluor 488 diluted 1:500 (Life Technologies). Cells were rinsed with 0.1% Tween-20 in PBS. Cells were post-fixed in 1% paraformaldehyde, incubated with RNaseA (0.1 mg/ml in PBS, 30 min at 37°C) and re-permeabilized in 0.7% Triton-X-100 / 0.1 M HCl. Cells were then denatured for 30 min at 22°C with 1.9 M HCl and were rinsed with cold PBS. DNA probes were denatured for 5 min at 95°C and applied to coverslips, which were then sealed onto slides with rubber cement, followed by incubation overnight at 37°C. Cells were then rinsed with 2×SSC at 37°C, 2×SSC at 22°C and 1×SSC at 22°C, all for 30 min each and were mounted in Prolong Gold (Life Technologies) containing 1.5 µg/ml 4,6-diamidino-2-phenylindole (DAPI, Sigma).

Confocal microscopy and analysis

DNA FISH with immunofluorescence was imaged by confocal microscopy on a Leica SP5 AOBS system (Acousto-Optical Beam Splitter). Optical sections separated by 0.3 µm were collected and only cells with signals from both alleles (typically over 95%) were analyzed using Image J software (NIH). For γH2AX colocalization analysis alleles were defined as colocalized with the protein if the DNA probe signals and immunofluorescence foci directly overlapped (at least two pixels of colocalization).

Germline retention assay

Germline retention of Tcrb alleles was determined by real-time PCR using primers amplifying the Dβ1 RSS, which is excised when Tcrb rearranges. A primer pair located downstream that is not deleted during rearrangement was used as a control. We calculated the percentage of germline retention using Rag1 knockout cells as 100% reference since recombination does not occur in these cells. Dβ1 RSS primer pair: Fwd 5’-GTGGTTTCTTCCAGCCCTCA-3’; Rev 5’-GGCCTTGGGACAGACAGAAT-3’; Control primer pair: Fwd 5’-CAAACGCTACCTCACCCCAT-3’; Rev 5’-CAGAGGCAGGAGAGCTCAAC-3’.

ChIP assay

Purified DN2/3 cells (Thy1.2+TCRbloCD4−CD8−CD44−/+CD25+) from thymi of wild type, Eb−/− and MiEk−/− mice (4*106 cells) were fixed and immunoprecipitated with 10ug anti-CBFb or normal rabbit IgG as described previously (Setoguchi et al., 2008). Anti-CBFb was a kind gift from Ichiro Taniuchi. Post DNA purification enriched DNA fragments were assessed by qPCR. Three biological replicates were processed. qPCR reactions were run in duplicates for each replicate sample using the Applied Biosystem instrument. Each reaction was run with 250nM final concentration of primers. Standard curves were run in parallel for each primer set to correct for primer efficiency and determine product quantity. Promoter of MageA2 was used as a reference gene (it is known to be silent in T cells) for fold enrichment. The following primers were used for assessment of IgG or CBFb enrichment at the respective sites.

  • Eb: Fwd 5’-AGCTCCATCTCCAGGAGTCA-3’; Rev 5’-CTGCATGAGAAGGGTTTGAAG-3’

  • REIR: Fwd 5’-TCACTGTCCCAGCTCTGCTA-3’; Rev 5’-GTCACCTGACCCTGAAGCTC-3’

  • Arhgef5: Fwd 5’-CACGGAGGAAGAGACACG-3’; Rev 5’-GGGATTGAGCTCTACCCACA-3’

  • Cd8a: Fwd 5’-CCTGGGCTACAGAAAGCAAG-3’; Rev 5’-ATGCCAAAGCAACATGTCAA-3’

  • Bid: Fwd 5’-CCTCCTGAGTGGGAAGTCTG-3’; Rev 5’-CCGGATTCAGGTTCAGAAGA-3’

  • Clec12a: Fwd 5’-GGGGCTGGTTACCACATTAC-3’; Rev 5’-TCAAAGGGATGGAAAATTTGA-3’

  • MageA2: Fwd 5’-AACGTTTTGTGAACGTCCTGAG-3’; Rev 5’-GACGCTCCAGAACAAAATGGC-3’

Supplementary Material

Supplemental figures

Acknowledgments

We would like to thank members of the Skok lab for helpful discussions, especially Pedro Rocha and Lili Blumenberg for input in the revision of the manuscript. C.P. was supported by an NCC fellowship; V.S. is supported by NYSTEM institutional training grant (contract #C026880); J.A.S. is a Leukemia & Lymphoma Society (LLS) scholar. J.A.S. and Y.K. are supported by NIH grant R01 GM086852, J.A.S. and R.B. are supported by NIH grant R01 GM112192. We would also like to thank NYU CHIBI, the NYUMC sorting facility and NYUMC genome technology center for their contributions to this work.

Footnotes

Author Contributions

C.P., V.S. and J.A.S. conceived the project. C.P., V.S. and J.A.S. wrote the manuscript. C.P., V.S., C.L. and B.H. performed experiments. C.P., V.S., C.L., S.B., R.R., and J.A.S. analyzed the data. T.T., and I.A. provided expertise.

Accession Numbers

Raw reads and processed files for our RNA-seq, ATAC-seq, 4C-seq and iChIP-seq datasets are available under the NCBI Gene Expression Omnibus accession number GSE80272.

Conflict of interest

The authors declare that they have no competing financial interests.

References

  1. Alt FW, Zhang Y, Meng FL, Guo C, Schwer B. Mechanisms of programmed DNA lesions and genomic instability in the immune system. Cell. 2013;152:417–429. doi: 10.1016/j.cell.2013.01.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Anders S, Pyl PT, Huber W. HTSeq--a Python framework to work with high-throughput sequencing data. Bioinformatics. 2015;31:166–169. doi: 10.1093/bioinformatics/btu638. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bories JC, Demengeot J, Davidson L, Alt FW. Gene-targeted deletion and replacement mutations of the T-cell receptor beta-chain enhancer: the role of enhancer elements in controlling V(D)J recombination accessibility. Proc Natl Acad Sci U S A. 1996;93:7871–7876. doi: 10.1073/pnas.93.15.7871. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bouvier G, Watrin F, Naspetti M, Verthuy C, Naquet P, Ferrier P. Deletion of the mouse T-cell receptor beta gene enhancer blocks alphabeta T-cell development. Proc Natl Acad Sci U S A. 1996;93:7877–7881. doi: 10.1073/pnas.93.15.7877. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Buenrostro JD, Giresi PG, Zaba LC, Chang HY, Greenleaf WJ. Transposition of native chromatin for fast and sensitive epigenomic profiling of open chromatin, DNA-binding proteins and nucleosome position. Nat Methods. 2013;10:1213–1218. doi: 10.1038/nmeth.2688. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Cheutin T, Cavalli G. Progressive polycomb assembly on H3K27me3 compartments generates polycomb bodies with developmentally regulated motion. PLoS genetics. 2012;8:e1002465. doi: 10.1371/journal.pgen.1002465. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Collins A, Hewitt SL, Chaumeil J, Sellars M, Micsinai M, Allinne J, Parisi F, Nora EP, Bolland DJ, Corcoran AE, et al. RUNX Transcription Factor-Mediated Association of Cd4 and Cd8 Enables Coordinate Gene Regulation. Immunity. 2011;34:303–314. doi: 10.1016/j.immuni.2011.03.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Dryden NH, Broome LR, Dudbridge F, Johnson N, Orr N, Schoenfelder S, Nagano T, Andrews S, Wingett S, Kozarewa I, et al. Unbiased analysis of potential targets of breast cancer susceptibility loci by Capture Hi-C. Genome Res. 2014;24:1854–1868. doi: 10.1101/gr.175034.114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Fuxa M, Skok J, Souabni A, Salvagiotto G, Roldan E, Busslinger M. Pax5 induces V-to-DJ rearrangements and locus contraction of the immunoglobulin heavy-chain gene. Genes Dev. 2004;18:411–422. doi: 10.1101/gad.291504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Goldmit M, Ji Y, Skok J, Roldan E, Jung S, Cedar H, Bergman Y. Epigenetic ontogeny of the Igk locus during B cell development. Nat Immunol. 2005;6:198–203. doi: 10.1038/ni1154. [DOI] [PubMed] [Google Scholar]
  11. Hewitt SL, Farmer D, Marszalek K, Cadera E, Liang HE, Xu Y, Schlissel MS, Skok JA. Association between the Igk and Igh immunoglobulin loci mediated by the 3' Igk enhancer induces 'decontraction' of the Igh locus in pre-B cells. Nat Immunol. 2008;9:396–404. doi: 10.1038/ni1567. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Hewitt SL, Yin B, Ji Y, Chaumeil J, Marszalek K, Tenthorey J, Salvagiotto G, Steinel N, Ramsey LB, Ghysdael J, et al. RAG-1 and ATM coordinate monoallelic recombination and nuclear positioning of immunoglobulin loci. Nat Immunol. 2009;10:655–664. doi: 10.1038/ni.1735. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Hnisz D, Abraham BJ, Lee TI, Lau A, Saint-Andre V, Sigova AA, Hoke HA, Young RA. Super-enhancers in the control of cell identity and disease. Cell. 2013;155:934–947. doi: 10.1016/j.cell.2013.09.053. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Inlay M, Alt FW, Baltimore D, Xu Y. Essential roles of the kappa light chain intronic enhancer and 3' enhancer in kappa rearrangement and demethylation. Nat Immunol. 2002;3:463–468. doi: 10.1038/ni790. [DOI] [PubMed] [Google Scholar]
  15. Inlay MA, Lin T, Gao HH, Xu Y. Critical roles of the immunoglobulin intronic enhancers in maintaining the sequential rearrangement of IgH and Igk loci. J Exp Med. 2006;203:1721–1732. doi: 10.1084/jem.20052310. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Inlay MA, Tian H, Lin T, Xu Y. Important roles for E protein binding sites within the immunoglobulin kappa chain intronic enhancer in activating Vkappa Jkappa rearrangement. J Exp Med. 2004;200:1205–1211. doi: 10.1084/jem.20041135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Johnson K, Chaumeil J, Skok JA. Epigenetic regulation of V(D)J recombination. Essays Biochem. 2010;48:221–243. doi: 10.1042/bse0480221. [DOI] [PubMed] [Google Scholar]
  18. Kosak ST, Skok JA, Medina KL, Riblet R, Le Beau MM, Fisher AG, Singh H. Subnuclear compartmentalization of immunoglobulin loci during lymphocyte development. Science. 2002;296:158–162. doi: 10.1126/science.1068768. [DOI] [PubMed] [Google Scholar]
  19. Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nat Methods. 2012;9:357–359. doi: 10.1038/nmeth.1923. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Lara-Astiaso D, Weiner A, Lorenzo-Vivas E, Zaretsky I, Jaitin DA, David E, Keren-Shaul H, Mildner A, Winter D, Jung S, et al. Immunogenetics. Chromatin state dynamics during blood formation. Science. 2014;345:943–949. doi: 10.1126/science.1256271. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15:550. doi: 10.1186/s13059-014-0550-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Majumder K, Koues OI, Chan EA, Kyle KE, Horowitz JE, Yang-Iott K, Bassing CH, Taniuchi I, Krangel MS, Oltz EM. Lineage-specific compaction of Tcrb requires a chromatin barrier to protect the function of a long-range tethering element. J Exp Med. 2015;212:107–120. doi: 10.1084/jem.20141479. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. McCarthy MT, O'Callaghan CA. PeaKDEck: a kernel density estimator-based peak calling program for DNaseI-seq data. Bioinformatics. 2014;30:1302–1304. doi: 10.1093/bioinformatics/btt774. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Pierce JW, Gifford AM, Baltimore D. Silencing of the expression of the immunoglobulin kappa gene in non-B cells. Molecular and Cellular Biology. 1991;11:1431–1437. doi: 10.1128/mcb.11.3.1431. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Pott S, Lieb JD. What are super-enhancers? Nat Genet. 2015;47:8–12. doi: 10.1038/ng.3167. [DOI] [PubMed] [Google Scholar]
  26. Proudhon C, Hao B, Raviram R, Chaumeil J, Skok JA. Long-Range Regulation of V(D)J Recombination. Adv Immunol. 2015;128:123–182. doi: 10.1016/bs.ai.2015.07.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Qian J, Wang Q, Dose M, Pruett N, Kieffer-Kwon KR, Resch W, Liang G, Tang Z, Mathe E, Benner C, et al. B Cell Super-Enhancers and Regulatory Clusters Recruit AID Tumorigenic Activity. Cell. 2014 doi: 10.1016/j.cell.2014.11.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Raviram R, Rocha PP, Müller CL, Miraldi ER, Badri S, Fu Y, Swanzey E, Proudhon C, Snetkova V, Bonneau R, et al. 4C-ker: A Method to Reproducibly Identify Genome-Wide Interactions Captured by 4C-Seq Experiments. PLoS computational biology. 2016;12:e1004780. doi: 10.1371/journal.pcbi.1004780. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Roldan E, Fuxa M, Chong W, Martinez D, Novatchkova M, Busslinger M, Skok JA. Locus 'decontraction' and centromeric recruitment contribute to allelic exclusion of the immunoglobulin heavy-chain gene. Nat Immunol. 2005;6:31–41. doi: 10.1038/ni1150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Sayegh C, Jhunjhunwala S, Riblet R, Murre C. Visualization of looping involving the immunoglobulin heavy-chain locus in developing B cells. Genes Dev. 2005;19:322–327. doi: 10.1101/gad.1254305. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Schatz DG, Swanson PC. V(D)J recombination: mechanisms of initiation. Annu Rev Genet. 2011;45:167–202. doi: 10.1146/annurev-genet-110410-132552. [DOI] [PubMed] [Google Scholar]
  32. Schoenfelder S, Sexton T, Chakalova L, Cope NF, Horton A, Andrews S, Kurukuti S, Mitchell JA, Umlauf D, Dimitrova DS, et al. Preferential associations between co-regulated genes reveal a transcriptional interactome in erythroid cells. Nat Genet. 2010;42:53–61. doi: 10.1038/ng.496. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Setoguchi R, Tachibana M, Naoe Y, Muroi S, Akiyama K, Tezuka C, Okuda T, Taniuchi I. Repression of the transcription factor Th-POK by Runx complexes in cytotoxic T cell development. Science. 2008;319:822–825. doi: 10.1126/science.1151844. [DOI] [PubMed] [Google Scholar]
  34. Skok JA, Gisler R, Novatchkova M, Farmer D, de Laat W, Busslinger M. Reversible contraction by looping of the Tcra and Tcrb loci in rearranging thymocytes. Nature Immunology. 2007;8:378–387. doi: 10.1038/ni1448. [DOI] [PubMed] [Google Scholar]
  35. Sun L, Zhang Z, Bailey TL, Perkins AC, Tallack MR, Xu Z, Liu H. Prediction of novel long non-coding RNAs based on RNA-Seq data of mouse Klf1 knockout study. BMC Bioinformatics. 2012;13:331. doi: 10.1186/1471-2105-13-331. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Trimarchi T, Bilal E, Ntziachristos P, Fabbri G, Dalla-Favera R, Tsirigos A, Aifantis I. Genome-wide mapping and characterization of Notch-regulated long noncoding RNAs in acute leukemia. Cell. 2014;158:593–606. doi: 10.1016/j.cell.2014.05.049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. van de Werken HJ, Landan G, Holwerda SJ, Hoichman M, Klous P, Chachik R, Splinter E, Valdes-Quezada C, Oz Y, Bouwman BA, et al. Robust 4C-seq data analysis to screen for regulatory DNA interactions. Nat Methods. 2012;9:969–972. doi: 10.1038/nmeth.2173. [DOI] [PubMed] [Google Scholar]
  38. Whyte WA, Orlando DA, Hnisz D, Abraham BJ, Lin CY, Kagey MH, Rahl PB, Lee TI, Young RA. Master transcription factors and mediator establish super-enhancers at key cell identity genes. Cell. 2013;153:307–319. doi: 10.1016/j.cell.2013.03.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Yancopoulos GD, Alt FW. Developmentally controlled and tissue-specific expression of unrearranged VH gene segments. Cell. 1985;40:271–281. [PubMed] [Google Scholar]
  40. Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE, Nusbaum C, Myers RM, Brown M, Li W, et al. Model-based analysis of ChIP-Seq (MACS) Genome Biol. 2008;9:R137. doi: 10.1186/gb-2008-9-9-r137. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Zhou X, Xiang Y, Garrard WT. The Igkappa gene enhancers, E3' and Ed, are essential for triggering transcription. J Immunol. 2010;185:7544–7552. doi: 10.4049/jimmunol.1002665. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental figures
NIHMS788397-supplement-Supplemental_figures.pdf (351.4KB, pdf)

RESOURCES