• V(D)J recombination control
• antibody repertoires
• B-cell development

The B-lymphocyte antigen receptor (BCR) is made of Ig heavy (IgH) and light (IgL) chains. T cells recognize antigen through related T-cell receptors (TCRs). The N-terminal variable regions of Ig and TCR chains form the antigen-binding site. Each B cell expresses a single unique antigen receptor. However, there is an immense diversity of different antigen receptors expressed population-wide in different B cells, a phenomenon that relies on the assembly of exons that encode variable regions from germline V, D, and J gene segments (“Vs,” “Ds,” and “Js”) during early B- and T-cell development (1). V(D)J recombination at both Ig and TCR loci is carried out by a common V(D)J “recombinase.” Recombination activating genes 1 and 3 (RAG1 and RAG2) comprise the lymphocyte-specific endonuclease (“RAG”) that initiates V(D)J recombination by generating DNA double-strand breaks at recombination signal sequences (RSSs) adjacent to participating Vs, Ds, and Js (2). Subsequently, RAG-cleaved segments are joined by classical nonhomologous end-joining (3).

In B cells, V(D)J exons are assembled within the IgH locus (Igh) and within either of two Igl loci (Igκ or Igλ). The murine Igh spans 3 Mb on chromosome 12 (Fig. 1). Approximately 150 V_Hs are scattered within a 2.7-Mb upstream portion of Igh, followed by a 100-kb intergenic region that separates the most downstream V_H (V_H81X) from the first of 13 Ds (4). Four J_Hs lie downstream of the Ds and are separated from the first of eight sets of IgH constant region exons by a several-kilobase region that harbors the intronic Igh enhancer (iEμ). A second set of enhancers within the Igh 3′ regulatory region (3′ Igh RR) lies ∼200 kb downstream, just beyond the last set of C_H exons (5). V(D)J recombination at Igh is regulated on multiple levels, in each case by modulating the accessibility of the substrate Vs, Ds, and Js and their flanking RSSs to RAG cleavage (6, 7). In this regard, Igh V(D)J recombination events are ordered and are stage specific; D-to-J_H joining occurs first, usually on both alleles, in preprogenitor B cells, followed by the joining of a V_H to a preassembled DJ_H complex in progenitor B (pro-B) cells (8, 9). In addition, V(D)J recombination is lineage specific; thus, although both developing B and T cells generate Igh DJ_H joins, complete V_HDJ_H joins only occur in B cells (10, 11). Finally, V(D)J recombination is feedback regulated in the context of allelic exclusion, with productive V_HDJ_H rearrangements that lead to IgH chain production inhibiting V_H-to-DJ_H recombination events on the other DJ_H allele in developing B cells (8, 12).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Mutation of single IGCR1 CBEs. Schematic of the murine Igh locus showing the IGCR1 region in WT relative to IGCR1/CBE1-mutated and IGCR1/CBE2-mutated configurations. Symbols are indicated on the figure. CBE orientation is indicated by the direction of the pink arrowheads (see the Introduction for details).

In the context of the regulatory events outlined above, accessibility of particular RSSs to RAG cleavage correlates with various factors, including germline transcription of target gene segments and certain chromatin modifications (9, 13). Normal Igh V(D)J recombination also depends on iEμ, although residual V(D)J recombination in the absence of this element implicates additional cis elements (14, 15). In addition, regulatory factors such as Pax5 and Yin Yang 1 (YY1), which bind to sites throughout Igh, participate in a locus-contraction process that brings distal V_Hs closer to DJ_H joins subsequent to their formation to promote a more diverse set of primary V(D)J rearrangements (16⇓–18). During B-cell development, cells deficient for such factors fail to undergo Igh locus contraction and, as a result, predominantly rearrange D-proximal V_Hs versus distal V_Hs (16, 18). Notably, the most D-proximal V_H, V_H81X, is used preferentially during primary V_H-to-DJ_H rearrangements even in normal pro-B cells (19, 20). However, productive V_H81X rearrangements are counter selected because of autoreactivity and poor pairing with IgL chains, greatly reducing their representation in antibody repertoires (21).

Many aspects of Igh V(D)J recombination regulation, including order, lineage specificity, and feedback control, as well as the balanced use of proximal and distal V_Hs, depend on the integrity of the intragenic control region 1 (IGCR1), which, as a working definition, was ascribed to a 4.1-kb DNA segment within the 100-kb V_H81X-to-D intragenic region (22). Moreover, known IGCR1 functions depend on the integrity of two 20-bp CCCTC-binding factor (CTCF)-binding elements (CBEs), termed “CBE1” (upstream) and “CBE2” (downstream) (Fig. 1), that are separated by about 2 kb within IGCR1 (22, 23). CTCF is a highly conserved zinc-finger protein with diverse roles in gene regulation, including promoter activation/repression, long-range chromosome interactions, and insulating activity (24). Roles for IGCR1 and the two IGCR1 CBEs in regulating Igh V(D)J recombination were discovered by the common phenotypes of mice in which either the 4.1-kb IGCR1 segment was deleted or in which IGCR1 CBEs within this segment were simultaneously replaced with scrambled sequences (to generate IGCR1/CBE1&2^−/− alleles) that do not bind CTCF (22). Both types of mutant mice displayed increased germline transcription and premature rearrangement of V_H81X and, to a lesser extent, several closely linked V_Hs on mutant alleles, leading to breaks in ordered, lineage-specific, and feedback-regulated Igh V(D)J recombination (22). Thus, IGCR1, via a process that requires one or both CBEs, promotes diverse antibody repertoires by preventing premature proximal V_H rearrangement, which can deplete DJ_H substrates for distal V_H joining subsequent to locus contraction.

The Igh has a large number of CBEs with a strikingly suggestive organization (Fig. 1 and Fig. S1) (9, 17, 22). In this regard, CBEs lie just downstream of most proximal V_H RSSs or are dispersed among distal V_H clusters (25, 26). In addition there is a cluster of 10 CBEs termed the “3′ CBEs” just downstream of the Igh 3′ RR (27). The IGCR1 locale interacts with the 3′ CBEs and with some V_H CBEs in a manner dependent on the integrity of its CBEs (22, 28). This body of findings led to proposals that interaction of the IGCR1 CBEs with 3′ CBEs or other downstream elements may sequester the D-to-J_H portion of Igh in a loop that contains iEμ and the 3′ Igh RR and/or other elements to promote D-to-J_H recombination and that such loops exclude transcriptional activation and rearrangement of proximal V_Hs (9, 26, 29). Interaction of IGCR1 CBEs with upstream V_H-associated CBEs also is speculated to play a positive role in promoting V_H-to-DJ_H pairing for V(D)J recombination (17, 30).

To elucidate potential roles for CBE1, CBE2, or both with respect to known IGCR1 functions in V(D)J recombination, we now have generated and analyzed mice in which one or the other IGCR1 CBE was individually replaced with scrambled sequence mutations.

Results

Replacement of CBE1 or CBE2 in Mice with a Scrambled Control Sequence.

To determine whether the two CBEs within IGCR1 have overlapping, distinct, or redundant roles in Igh V(D)J recombination control, we introduced a scrambled sequence that has no CTCF binding activity (22) into 129SV ES cells, either in place of the CBE1 sequence to generate an IGCR1/CBE1⁻ allele or in place of CBE2 to generate an IGCR1/CBE2⁻ allele. We then introduced these two mutations into the 129SV mouse germline to assay potential effects on the control of V(D)J recombination (Fig. 1 and Fig. S2). Based on our gene-targeted replacement strategy, both IGCR1/CBE1⁻ and IGCR1/CBE2⁻ alleles retain an inserted loxP site after cre-mediated deletion of a neomycin-resistance selection cassette. However, our prior studies confirmed that control mice with a WT IGCR1, but which contain the upstream loxP site, were indistinguishable from WT mice with respect to Igh V(D)J recombination control (22).

Roles of IGCR1 CBE1 Versus CBE2 in B-Cell Development.

Bone marrow (BM) pro-B cells undergo V_H-to-DJ_H rearrangements, which, if productive (in-frame), generate μ IgH chains that signal proliferation and progression to the precursor (pre) B-cell stage in which assembly of the Igl variable region exon occurs (1). Productive Igl rearrangements produce an IgL chain that can combine with the μ IgH chain to form a surface BCR and can lead to differentiation of B cells that migrate to peripheral lymphoid organs such as the spleen. To probe for B-cell developmental defects associated with the mutant Igh alleles, we bred 129SV strain mice (which express the IgM^a allotype) harboring either the IGCR1/CBE1⁻ or the IGCR1/CBE2⁻ allele with C57BL/6 mice (which express the IgM^b allotype) to generate F1 mice that harbor a WT Igm^b allele and either an IGCR1/CBE1⁻ or an IGCR1/CBE2⁻ Igm^a allele. These F1 mice were assayed for the proportion of BM and splenic IgM⁺ B cells that express surface IgM^a versus IgM^b. F1 mice with WT alleles have roughly equal numbers of IgM^a- and IgM^b-expressing B cells in BM or the spleen, because there is an equal chance of forming a productive rearrangement on both alleles and because of the phenomenon of allelic exclusion (Fig. 2 and Fig. S3 A and B) (1, 9). However, although mice harboring the IGCR1/CBE2⁻ Igh^a allele were indistinguishable from WT mice with respect to IgM^a versus IgM^b expression, mice harboring the IGCR1/CBE1⁻ Igh^a allele had only half as many IgM^a-expressing as IgM^b-expressing B cells (Fig. 2 and Fig. S3 A and B), despite having similar total numbers of splenic B cells (Fig. S3C). Although mutation of IGCR1 CBE1 had a greater impact than mutation of IGCR1 CBE2 on the ability of an Igh allele to support B-cell development, the defect was not nearly as severe as that observed for IgM^a/b F1 mice in which both IGCR1 CBEs were mutated simultaneously on the IgM^a allele (22).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Effect of IGCR1/CBE mutations on B-cell development. IgM^b:IgM^a expression ratios of IgM^a and IgM^b allotypic markers in BM (A) and spleen (B) from WT F1 IgM^a/IgM^b mice [BM: mean (m) = 1.08 ± 0.0502, n = 12; spleen: m = 0.857 ± 0.0196, n = 11], heterozygous mutant IGCR1/CBE1⁻ IgM^a/WT IgM^b mice (BM: m = 2.79 ± 0.108, n = 5; spleen: m = 1.904 ± 0.128, n = 5), and IGCR1/CBE2⁻ IgM^a/ WT IgM^b mice (BM: m = 1.22 ± 0.086, n = 3; spleen: m = 0.94 ± 0.032, n = 3). Horizontal lines and whickers indicate mean and SEM. P values were generated by unpaired Student’s t test. Each data point was collected from different 4- to 6-wk-old mice; at least three mice were used for each genotype.

Roles of IGCR1 CBE1 Versus CBE2 in Generation of the Igh V(D)J Repertoire.

We used a PCR-Southern approach (22) to assay for proximal (V_H7183) and distal (V_HJ558) V_H-to-DJ_H rearrangements in pro-B and pre-B cells of WT and mutant mice. Primers for either the proximal V_H7183 family or the distal V_HJ558 family were paired with a primer downstream of J_H4. PCR products were run on a gel, probed with a ³²P-labeled oligo that bound a region between J_H3 and J_H4, and then were exposed via autoradiography. The mouse Dlg5 gene was used as a loading control. The IGCR1/CBE2^−/− pro-B cells and, even more so, IGCR1/CBE1^−/− pro-B cells displayed increased V_H7183- versus V_HJ558-to-DJ_H rearrangements compared with WT, but not nearly to the extent observed in IGCR1/CBE1&2^−/− pro-B cells, with V_H7183/V_HJ558 ratios averaging ∼1/1, 10/1, 3/1, and 40/1 for WT, IGCR1/CBE1^−/−, IGCR1/CBE2^−/−, and IGCR1/CBE1&2^−/− pro-B cells, respectively (three representative experiments are shown in Fig. 3A). Similar comparative analyses of WT pre-B cells among the WT and various CBE-mutant pre-B cells indicated much more modest differences in V_H7183/V_HJ558 utilization ratios (Fig. S4A), likely because of selective expansion of pre-B cells expressing more distal V_H rearrangements, as discussed previously (22).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

IGCR1/CBE1 and IGCR1/CBE2 mutations alter proximal V_H use and ordered rearrangement. (A) PCR analyses of the V_H7183, V_HJ558 V_H family V(D)J rearrangements in pro-B cells from indicated mice. Data from three independent experiments, each using a different mouse, are shown. The Dlg5 lanes represent a loading control. The same starting DNA was assayed for each sample, along with two fivefold serial dilutions. Bands from the V_HDJ_H products contain J_H1, J_H2, and J_H3 rearrangements, respectively, from largest to smallest products. We calculated by inspection the approximate use of V_H7183 relative to V_HJ558 for each sample in a given set based on signals obtained with the 1×, 5×, and 25× dilutions and then expressed the relative values for each sample from a given type of mutant pro-B cells relative to that of the WT sample which was set as 1 (shown at the bottom of the panels for each experimental set). This approach is analogous to one we have described previously (19, 20). (B) We used PCR to assay for direct V_H81X-to-D_HQ52 rearrangements using the same three sets of DNA samples (and dilutions) as described for A; thus, the DLG5 loading controls in A and B are identical for each set of repeats.

Roles of IGCR1 CBE1 Versus CBE2 in Ordered Rearrangement.

Our previous studies on the IGCR1^−/− and IGCR1/CBE1&2^−/− mice revealed a break in ordered V(D)J recombination, with V_H-to-D rearrangements often preceding D-to-J_H rearrangements (22). Thus, we sought to determine whether one or both of the IGCR1 CBEs are required for maintaining ordered rearrangement. For this purpose, we assayed for direct V_H-to-D joins by using a forward V_H81X primer with a reverse primer located between DQ52 and J_H1. By this method, we clearly detected direct V_H–D joins in both pro-B and pre-B cells in the IGCR1/CBE1^−/− mice, albeit to a substantially lesser extent than observed in the IGCR1/CBE1&2^−/− mutants (Fig. 3B and Fig. S4B). We also observed a variable increase in direct V_H81X-to-DQ52 joins in IGCR1/CBE2^−/− mutants (Fig. 3B and Fig. S4B). Thus, both CBE1 and CBE2 appear to be required for fully maintaining ordered Igh V(D)J recombination, with disordered rearrangement occurring more abundantly when both are mutated.

Roles of CBE1 Versus CBE2 in Maintaining Lineage-Specific V_H-to-DJ_H Rearrangement.

We used the PCR-based assay described above to assay for V_H-to-DJ_H rearrangements in developing thymic T cells. A PCR assay for Igl Vκ–Jκ joins confirmed low to no detectable B-cell contamination in the CD4⁺/CD8⁺ thymocyte samples; WT splenic B-cell DNA served as a positive control (three representative experiments are shown in Fig. 4). These analyses clearly revealed complete V_H7183-to-DJ_H but not V_HJ558-to-DJ_H joins in both IGCR1/CBE1^−/− and IGCR1/CBE2^−/− thymocytes, albeit in either case not nearly to the same extent as when both of the CBEs were mutated simultaneously (Fig. 4). We also assayed for direct V_H81X-to-D_Q52 joins in CD4⁺/CD8⁺ T cells from WT, IGCR1/CBE1^−/−, IGCR1/CBE2^−/−, and IGCR1/CBE1&2^−/− mutants and detected direct joins in the IGCR1/CBE1^−/− and, to a lesser extent, in IGCR1/CBE2^−/− thymocytes, albeit again in both cases at substantially lower levels than observed in IGCR1/CBE1&2^−/− mutant thymocytes (Fig. S5). Thus, CBE1 and CBE2 also appear to function cooperatively to maintain lineage-specific proximal V_H rearrangements.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

IGCR1/CBE1 and IGCR1/CBE2 mutations deregulate IgH V(D)J recombination in thymocytes. PCR analyses of Igh V_H7183 and V_HJ558 family V(D)J rearrangements in CD4⁺CD8⁺ thymocytes from indicated mice. The VκJκ panels are controls for potential B-cell contamination, and the Dlg5 lanes represent a loading control. The 1x, 5x, and 25x serial dilutions were assayed for each sample as described in Fig. 3. Bands from the V_HDJ_H products are as described in Fig. 3. Each type of PCR for all three panels of samples was done at the same time and run on the same gel together with the WT splenic B-cell control, which is shown only in the top panel. Three independent experiments, each from a different mouse, are shown. We also used PCR analyses for direct V_H81X-to-DQ52 rearrangement in CD4⁺CD8⁺ thymocytes from this same set of samples (shown in Fig. S5).

Roles of IGCR1, CBE1, and CBE2 in Feedback-Regulated V_H-to-DJ_H Recombination.

Surface staining of IgM⁺ splenic B cells from F1 mice containing a WT Igm^b allele and a IGCR1/CBE1⁻ or a IGCR1/CBE2⁻ Igm^a allele did not reveal IgM^aIgM^b double-expressing cells, indicating that neither mutation leads to a detectable break in allelic exclusion (Fig. S3B). This finding is consistent with the prior observation that the IGCR1/CBE1&2⁻ IgM^a allele in this F1 background did not break allelic exclusion (22). On the other hand, prior studies showed that IGCR1/CBE1&2⁻ IgM^a V_H-to-DJ_H rearrangements were not suppressed by a productive V_HDJ_H knockin (VB1-8 KI) on the other Igh allele (22). Thus, we similarly assayed for the ability of a VB1-8 KI allele to suppress V_H-to-DJ_H rearrangements on IGCR1/CBE1⁻ and IGCR1/CBE2⁻ Igh alleles. As expected (22), only very low levels of proximal V_H7183DJ_H rearrangements were detected in splenic B cells from VB1-8 KI mice in which the second allele was WT (Fig. 5). In contrast, we observed increased V_H7183DJ_H rearrangements in splenic B cells from VB1-8 KI mice in which the second Igh allele harbored the IGCR1/CBE2⁻ mutation and an even greater increase in those in which the second allele harbored the IGCR1/CBE1⁻ mutation (Fig. 5). Nucleotide sequencing showed that V_HDJ_H rearrangements on the IGCR1/CBE1⁻ allele were predominantly nonproductive V_H81X rearrangements (Fig. S6), consistent with patterns observed in IGCR1/CBE1&2⁻ mutants in this experimental setting (22). Rearrangements from the IGCR1/CBE2⁻ alleles also were nonproductive; but in the limited dataset analyzed appeared to involve a wider range of V_H7183 segments (Fig. S6). Thus, CBE1 also appears to play a more major role in enforcing feedback regulation of proximal V_Hs than CBE2, but both together contribute to achieving WT levels. Finally, consistent with previous observations in IGCR1/CBE1&2^−/− mice (22), distal V_HJ558 segments on IGCR1/CBE1⁻ and IGCR1/CBE2⁻ alleles were still feedback regulated by the VB1-8 KI allele (Fig. 5).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

IGCR1/CBE1 and CBE2 are both required for feedback regulation of proximal V_H-to-DJ_H recombination. V_HDJ_H rearrangements in pre-B and splenic B cells from VB1-8 knockin mice in which the second Igh allele was WT, IGCR1/CBE1⁻ (CBE1⁻), or IGCR1/CBE2⁻ (CBE2⁻). The 1×, 5×, and 25× serial dilutions were performed for each sample. Other details are as described for Fig. 3.