Elements between the IgH variable (V) and diversity (D) clusters influence antisense transcription and lineage-specific V(D)J recombination

Abstract

Ig and T-cell receptor (TCR) variable-region gene exons are assembled from component variable (V), diversity (D) and joining (J) gene segments during early B and T cell development. The RAG1/2 endonuclease initiates V(D)J recombination by introducing DNA double-strand breaks at borders of the germ-line segments. In mice, the Ig heavy-chain (IgH) locus contains, from 5′ to 3′, several hundred V_H gene segments, 13 D segments, and 4 J_H segments within a several megabase region. In developing B cells, IgH variable-region exon assembly is ordered with D to J_H rearrangement occurring on both alleles before appendage of a V_H segment. Also, IgH V_H to DJ_H rearrangement does not occur in T cells, even though DJ_H rearrangements occur at low levels. In these contexts, V(D)J recombination is controlled by modulating substrate gene segment accessibility to RAG1/2 activity. To elucidate control elements, we deleted the 100-kb intergenic region that separates the V_H and D clusters (generating ΔV_H-D alleles). In both B and T cells, ΔV_H-D alleles initiated high-level antisense and, at lower levels, sense transcription from within the downstream D cluster, with antisense transcripts extending into proximal V_H segments. In developing T lymphocytes, activated germ-line antisense transcription was accompanied by markedly increased IgH D-to-J_H rearrangement and substantial V_H to DJ_H rearrangement of proximal IgH V_H segments. Thus, the V_H-D intergenic region, and likely elements within it, can influence silencing of sense and antisense germ-line transcription from the IgH D cluster and thereby influence targeting of V(D)J recombination.

Ig and T-cell receptor (TCR) variable-region exons are assembled from germ-line variable (V), diversity (D) and joining (J) gene segments during early stages of B- and T-lymphocyte development, respectively (1). Each germ-line V, D, and J gene segment is flanked by a short recombination signal (RS) sequence that consists of a palindromic heptamer and a nonamer separated by 12- or 23-bp spacer. Joining occurs only between segments flanked by RSs with 12-bp and 23-bp spacers (referred to as 12RSs and 23RSs), respectively. For example, the Ig heavy-chain (IgH) V(D)J variable region is encoded by V_H, D, and J_H segments with V_H segments and J_H segments flanked by 23RSs and D segments flanked on both sides by 12RSs. Thus, complete V(D)J assembly must be mediated via the D segment; because direct V_H to J_H joining is prohibited by the 12/23 rule (2). V(D)J recombination is initiated by introduction of a DNA double-strand break (DSB) between a V, D or J segment and the flanking RS. Early work indicted that a single “V(D)J recombinase” assembles all antigen receptor genes (3). The nature of the common V(D)J recombinase was clarified by discovery of RAG1 and 2, which together form the RAG endonuclease that introduces the DSBs and provides the specificity of the V(D)J recombination reaction (4). The RAG endonuclease cuts paired V, D, and J segments in the context of the 12/23 rule (5, 6). DSB joining components of the V(D)J recombinase are provided by the classical nonhomologous DNA end-joining DSB repair pathway (7).

The activity of the common V(D)J recombinase is regulated in several contexts (8, 9). First, V(D)J recombination is regulated in a lineage-specific manner. Thus, Ig variable-region exons are completely assembled in B cells and not in T cells, whereas TCR variable-region exons are assembled in T but not B cells (6). Second, within a given lineage, V(D)J recombination is highly ordered. The IgH locus contains a large cluster of V_H segments, a cluster of 13 D segments, and a cluster of 4 J_H segments in a several-megabase region (1). In progenitor (pro) B lymphocytes, IgH variable-region exons are assembled via a process in which D to J_H rearrangements occur first and on both alleles, followed by appendage of a V_H segment to the DJ_H complex (10). Direct joining of a V_H to a D is not observed, even though permitted by the 12/23 rule (10). Although not strictly ordered, the most D-proximal V_H families and, in particular, the most D-proximal V_H81X gene segment, rearrange more frequently than the most distal V_H gene families, such as the V_HJ558 family (11). After assembly of a productive IgH variable-region exon and expression of a μ heavy chain, pro-B cells advance to the precursor (pre) stage at which Ig light-chain (IgL) variable-region assembly occurs. Similarly ordered V(D)J recombination occurs during assembly of TCR variable-region exons in developing T lineage cells (12). Finally, V(D)J recombination is regulated in the context of allelic exclusion. For example, a productive V_H to DJ_H rearrangement that generates a μ heavy-chain signals cessation of V_H to DJ_H rearrangement at the other IgH allele if it is in the DJ_H configuration and retargets RAG to IgL loci (13).

Control of V(D)J recombination in a lineage-, stage-, and allele-specific manner is mediated by modulating accessibility of the substrate gene segments to the common RAG V(D)J recombinase (3). Early studies showed that pro-B cells generate sense germ-line V_H transcripts from the 5′ (distal) V_H segments and that these transcripts are down-regulated at later developmental stages, establishing germ-line transcription as a strong correlate of V(D)J recombinational accessibility (14). Over the years, the precise mechanism by which V(D)J recombinational accessibility is generated has remained elusive, despite many additional correlates, including antisense, as well as sense, germ-line transcription of variable-region segments throughout the locus (15). Notably, both B and T lymphocytes have germ-line “μ0” transcripts that initiate from a promoter upstream of the most 3′ D segment (D_HQ52) and run downstream through the J_H segments, consistent with T lymphocytes making low levels of DJ_H rearrangements (16). In addition, pro-B lymphocytes express low levels of antisense transcripts from the D portion of the locus (17, 18). Additional epigenetic modifications associated with V(D)J recombinational accessibility include DNA and/or histone modifications of particular regions and chromosomal positioning of the Ig or TCR loci within the nucleus (15, 19, 20). In particular, locus contraction and looping have been implicated as potential means of bringing distant variable-region gene segments into the proximity of the DJ_H complex to increase their probability of rearrangement (21), and locus decontraction has been suggested to move V_H gene segments away to prevent further rearrangements in the context of allelic exclusion (22, 23).

As outlined above, the V_H to DJ_H step is the regulated event in IgH variable-region gene assembly in developing B cells, both in the context of ordered rearrangement and allelic exclusion. In addition, although developing T cells generate low levels of DJ_H rearrangements, they do not append V_H segments to the DJ_H complexes or generate V_H to D rearrangements, implicating V_H to DJ_H rearrangement as the regulated step with respect to B versus T lineage specificity (14). In these contexts, the ≈100-kb intergenic region between the mouse V_H and D clusters is a likely candidate for harboring sequences that might regulate V_H to DJ_H rearrangement (15, 24, 25). Correspondingly, insertion of a V_H segment just upstream of the most 5′ D segment deregulated its rearrangement both in terms of order and lineage specificity (26). To test for potential V(D)J recombinational regulatory properties of the V_H to D intergenic region, we deleted this region in mice. Our analyses of mice harboring this deletion indicate that it has a silencing function with respect to antisense transcription and that it also can influence lineage-specific IgH V_H to DJ_H rearrangement.

Results

The large intergenic region between the most 3′ functional V_H segment (V_H81X) (11) and the most 5′ D segment (DFL16.1) (27) has a number of properties suggestive of a potential regulatory function, including multiple DNase I hypersensitivity sites, differential histone modifications in V_H and D regions in pro-B versus pro-T cells, and two defined CTCF binding sites located several kilobases upstream of DFL16.1 (25, 28-30). To evaluate potential regulatory properties of the V_H to D intergenic region, we used the Velocigene technology (31) to create an IgH allele termed ΔV_H-D that harbors a 109.8-kb deletion that encompasses the region from 6.4 kb 5′ of V_H81X (V_H7183.a2.3) to 847 bp 3′ of DFL16.1 (Fig. 1A and Fig. S1). We chose to include V_H81X and DFL16.1 in the deletion to ensure that we captured any regulatory elements proximal to or overlapping with gene segments at each end of the intergenic region. We have left intact the downstream D segments, including DSPs, DFL16.2, and DSTs (Fig. 1A). Flow cytometric analyses of mice homozygous for the ΔV_H-D mutation (ΔV_H-D/ΔV_H-D mice) indicated that this deletion had no readily discernable effects on T or B cell development (Fig. S2). More detailed analyses of potential outcomes of the ΔV_H-D mutation on IgH gene rearrangement in B cells are ongoing. In this study, we concentrated on effects of the deletion in T cells, with pro-B cells or pro-B lines being used to analyze certain findings in greater depth.

Fig. 1.

Deletion of V_H to D_H intergenic region leads to generation of long V_H- and D-containing transcripts. (A) Schematic diagram of proximal WT murine 129sVe IgH locus before (Upper) and after gene targeting to delete the 109-kb intergenic region (Lower). The order of the D segments is indicated. (B) Northern analyses of total RNA from WT (^+/+), ΔV_H-D/+, and ΔV_H-D/ΔV_H-D thymocytes (Left) as well as total RNA from RAG2^−/−(^+/+) and RAG2^−/−(ΔV_H-D/ΔV_H-D) A-MuLV-transformed pro-B lines after probing with (from top to bottom): a 1-kb fragment encompassing DSP2.9 (see A); a 511-bp fragment derived from the germ-line V_H7183.a4.6 segment (see A); a 631-bp probe derived from the germ line corresponding to the V_HJ558.29 segment; and a GAPDH cDNA. Ribosomal RNA loading is shown at the bottom as additional loading control.

To assess potential effects of the ΔV_H-D on germ-line IgH transcription, we conducted RT-PCR assays for germ-line IgH DSP and V_H7183 transcripts in purified CD4⁻/CD8⁻ (double negative or DN) and CD4⁺CD8⁺ (double positive or DP) T cells, as well as in purified pro-B and pre-B cells from WT and ΔV_H-D /ΔV_H-D mice. We observed low levels of DSP germ-line transcripts in the absence of detectable V_H7183 transcripts in WT DN and DP T cells but found increased levels of both types of transcripts in ΔV_H-D /ΔV_H-D DN and DP T cells (Fig. S3). Both types of transcripts were also detectable in WT pro-B and pre-B cells, but their levels were greatly increased in pro-B and pre-B cells of ΔV_H-D /ΔV_H-D mice (Fig. S3). The size of germ-line transcripts in ΔV_H-D /ΔV_H-D cells was consistent with unprocessed transcripts, suggesting the possibility, confirmed below, that they were antisense (Fig. S3).

To further characterize the apparent germ-line DSP and V_H7183 transcriptional activation by the ΔV_H-D mutation, we performed Northern blotting with total RNA from ΔV_H-D/+ or ΔV_H-D/ΔV_H-D thymocytes. Northern blotting with a DSP2.9 probe revealed high-level expression of very long (>10 kb) steady-state transcripts in both ΔV_H-D/+ and ΔV_H-D/ΔV_H-D thymocytes that were undetectable in WT thymocyte or sperm RNA (Fig. 1B). Likewise, we detected extremely high levels of large DSP2.9-hybridizing transcripts in total RNA from a RAG2^−/− ΔV_H-D/ΔV_H-D A-MuLV-transformed pro-B line but not a RAG2^−/− WT A-MuLV-transformed pro-B line (Fig. 1B); because Rag-2^−/− cells do not undergo V(D)J recombination, these transcripts must come from unrearranged IgH alleles. Assays of the same RNA preparations revealed that the ΔV_H-D allele led to similarly high levels of large V_H7183-specific transcripts that were not detectable in WT thymocytes, WT A-MuLV-transformed pro-B cells, or sperm (Fig. 1B). Additional Northern blot analyses with a V_HJ558 probe revealed expected germ-line V_HJ558 transcripts in RAG-2^−/−A-MuLV-transformed pro-B lines that appeared unchanged in RAG2^−/−, ΔV_H-D/ΔV_H-D A-MuLV-transformed pro-B lines (Fig. 1B). Germ-line V_HJ558 transcripts are not observed in WT or ΔV_H-D/ΔV_H-D thymocytes (Fig. 1B). We conclude that the ΔV_H-D mutation activates the expression in both developing B and T cells of long germ-line transcripts that potentially appear to be antisense transcripts running from the IgH D region to the juxtaposed proximal V_H region.

The high levels of steady-state transcripts from the DSP and V_H7183 loci of ΔV_HD alleles might result either from increased transcriptional initiation or from stabilization of transcripts that also are generated in WT cells. To distinguish these possibilities, we did run-on transcription assays with nuclei from WT and ΔV_H-D/ΔV_H-D A-MuLV-transformed RAG-2-deficient pro-B lines. For the nuclear run-on experiments we used probes from various regions within the D and V_H portions of the IgH locus (Fig. 2 A and B). These assays demonstrated robust transcription of the V_HJ558 and D_HQ52 regions, but not of tested intervening regions of IgH (e.g., DSP or V_H7183 sequences), in RAG2-deficient pro-B lines. These results are consistent with substantial steady-state levels of germ-line V_HJ558 transcripts and μ0 transcripts but lack readily detectable DSP or V_H7183 transcripts in these cells. In contrast, the run-on transcription assays of ΔV_H-D/ΔV_H-D A-MuLV-transformed RAG-2-deficient pro-B lines revealed nascent transcription throughout the DSP cluster, through the “intergenic probe” region, and through V_H7183 segments as well as through the D_HQ52 and V_HJ558 portions of IgH (Fig. 2 C and D). Although we could not do run-on experiments with RAG2-deficient thymocytes because of lack of sufficient material, we obtained similar run-on results in experiments done with nuclei from total WT and ΔV_H-D/ΔV_H-D thymocytes (Fig. S4).

Fig. 2.

Deletion of V_H to D_H intergenic region leads increased germ-line transcription. Nuclear run-ons were conducted with nuclei from RAG2^−/− (^+/+) and RAG2^−/−(ΔV_H-D/ΔV_H-D) A-MuLV-transformed pro-B lines. ³²P-UTP–labeled nascent transcripts were hybridized to a series of double-stranded restriction endonuclease-generated DNA fragments (named on x axes) that were separated by electrophoresis and immobilized on Southern membranes. (A) The location of nuclear run-on probes (A through S) within the WT (Upper) and ΔV_H-D (Lower) IgH alleles is shown. (B) Schematic diagram of location of probe fragments in gels shown in C and D. Open squares depict backbone vector sequences, and filled squares depict IgH fragments/probes. (C and D) (Upper) Ethidium bromide–stained gel of Southern membrane. (Lower) Same membrane exposed after probing with labeled transcripts from RAG2^−/− (^+/+) (Left) and RAG2^−/−(ΔV_H-D/ΔV_H-D) (Right) A-MuLV-transformed pro-B lines. Results are representative of three independent experiments with different lines of the indicated genotype. See SI Materials and Methods for details.

To further characterize the germ-line transcription from ΔV_H-D alleles, we performed run-on experiments with nuclei from WT and ΔV_H-D/ΔV_H-D A-MuLV-transformed, RAG-2-deficient pro-B lines, using single strand sense and antisense DNA probes for V_HJ558, V_H7183, and DSP sequences, as well as for a positive (N-myc) and negative (albumin) control probes. As expected (32), sense N-myc transcription was found in both cell types. Notably, we found a consistent (n = 3) 10- to 20-fold excess of antisense versus sense transcription across V_HJ558 gene segments in both cell types (Fig. 3). Although we failed to detect transcription through the DSP and V_H7183 sequences in RAG-2-deficient pro-B lines, we found readily detectable nascent antisense, but not sense, transcription of V_H7183 sequences, as well as substantial antisense and, to a lesser extent, sense transcription through DSP sequences in the ΔV_H-D/ΔV_H-D, RAG-2-deficient pro-B lines. Given that there are many more 129sVe V_HJ558 segments than V_H7183 segments, we cannot directly compare their signal strength. Finally, although we cannot exclude a very low level of V_H7183 sense transcription in either WT or ΔV_H-D/ΔV_H-D RAG-2-deficient pro-B lines, we note that RT-PCR failed to show spliced sense germ-line V_H7183 transcripts in these cells (Fig. S3). We conclude that there is a strong bias toward nascent antisense versus sense transcription in both the D and V_H7183 regions of the ΔV_H-D IgH loci.

Fig. 3.

Deletion of V_H to D_H intergenic region leads to antisense transcription of adjacent regions. Nuclear run-ons were conducted with nuclei from RAG2^−/− (^+/+) and RAG2^−/−(ΔV_H-D/ΔV_H-D) A-MuLV-transformed pro-B lines. ³²P-UTP–labeled nascent transcripts were hybridized to indicated double strand (DS), antisense, and sense DNA fragments from the IgH and control loci as indicated (see text and SI Materials and Methods for further details). (Left and Center) Run-on results. (Right) Loading as revealed by ethidium bromide staining. SS, single strand.

We took advantage of the high level of DSP region transcription in ΔV_H-D/ΔV_H-D thymocytes to localize potential start sites via 5′ RACE of antisense transcripts. However, we were unable to detect RACE products from the DSP region in WT thymocyte RNA, consistent with very low steady-state transcript levels from this region. In contrast, PCR amplification, with two independent sets of nested primers, of 5′ RACE products from ΔV_H-D/ΔV_H-D thymocyte RNA revealed a heterogeneous array of products (Fig. S5), potentially representing heterogeneous transcription initiation sites such as those observed in the TCRα locus (33). However, we cannot rule out potential “strong stops” attributable to secondary RNA structures as an origin of some products. In this regard, nuclear run-on analyses suggested transcription initiated in a region centered on DST4, given that a probe −958 bp 5′ of DST4 (5′DST4) had a positive run-on signal, whereas a probe −365 bp 5′ of DST4 (3′DST4) did not (Fig. 2). Together, these data are consistent with the possibility that antisense transcription through the DSP region may initiate heterogeneously but that there may be a downstream boundary near the DST4 region.

The normal B cell development in ΔV_H-D/ΔV_H-D mice (Fig. S2) suggests that the increase in antisense transcription in the D region does not markedly interfere with IgH locus V(D)J recombination. Therefore, we asked whether high levels of antisense transcription coupled with low-level sense transcription on the ΔV_H-D allele might enhance IgH D to J_H recombination in T cells. For this purpose, we generated splenic T cell hybridomas from WT, ΔV_H-D/+, or ΔV_H-D/ΔV_H-D mice. We then analyzed rearrangement status of the IgH locus in StuI-digested DNA from each hybridoma by Southern blotting with a probe that detects J_H rearrangements (Fig. 4A). In these assays, the fusion partner contributes a germ-line J_H band; thus, we score rearrangements distinct non-germ-line–sized bands with two additional non-germ-line bands representing “biallelic” rearrangement. Consistent with earlier reports (34), ∼30% of WT T cells had J_H rearrangements and the majority (>90%) were monoallelic (Fig. 4B). Strikingly, however, >60% of ΔV_H-D/+ T cells had J_H rearrangements with ∼20% having biallelic rearrangements, and nearly 90% of ΔV_H-D/ΔV_H-D T cells had J_H rearrangements, over a third of which were biallelic (Fig. 4B).

Fig. 4.

V_H to D_H intergenic deletion up-regulates D_H to J_H recombination in thymocytes. (A) Two sets of independent splenic T-cell hybridomas were obtained from WT (n = 186), ΔV_H-D/+ (n = 188), and ΔV_H-D/ΔV_H-D (n = 167) mice. DNA was digested with StuI, and Southern blot analysis was performed with an EcoRI/HindIII fragment probe downstream of J_H4. The germ-line J_H locus yields a 6.2-kb band (GL), and D_H to J_H rearrangements are scored as rearranged bands with a size distinct from the germ-line band. DNA from the fusion partner BW5147 and kidney DNA is shown as controls. A representative set of analyses is shown. (B) Percentage of cells containing monoallelic (red) and biallelic (blue) J_H rearrangements as determined by either one or two non-germ-line bands in analyses of A. The average of two independent experiments is shown.

To test whether deletion of the intergenic V_H-D region altered V_H to DJ_H rearrangement in developing T cells, we first used a PCR approach to assay rearrangements of proximal and distal V_H families in thymocytes from WT, ΔV_H-D/+, or ΔV_H-D/ΔV_H-D mice, with ΔV_H-D/ΔV_H-D B cells and ES cells serving as positive and negative controls. Consistent with prior findings (12), normal T cells did not generate IgH locus V_H to DJ_H rearrangements of any V_H gene family tested. Strikingly, however, both ΔV_H-D/+ and ΔV_H-D/ΔV_H-D DP thymocytes showed rearrangements of the two most proximal V_H gene families (V_H7183 and V_HQ52) but not of more distal family members (V_HJ558 and V_HGAM3.8) (Fig. 5A). Indeed, cloning and sequencing of V_H7183 rearrangements and V_H7183 germ-line transcripts indicated that, within the V_H7183 gene family, mainly D proximal V_H segments were germ-line transcribed and rearranged (Fig. S6). To quantify V_HDJ_H rearrangements, we prepared large panels of splenic T cell hybridomas from WT and mutant mice and assayed for V_H(D)J_H versus DJ_H rearrangement by a Southern blotting and PCR approach (35, 36). These analyses confirmed lack of V_HDJ_H rearrangements in WT T cells and revealed that >5% of splenic ΔV_H-D/ΔV_H-D T cells contained V_HDJ_H rearrangements (Fig. 5B).

Fig. 5.

Deletion of the V_H to D_H intergenic region leads to V_H to DJ_H recombination in thymocytes. (A) PCR assay for V_H to DJ_H for indicated gene families (Right) in DNA from sorted DP-T cells. Size of rearrangements to various J segments is indicated on Left. Analysis of Igκ rearrangement (VJκ) controlled for potential B cell contamination. (B) The level of V_H to DJ_H rearrangements in the same panel of hybridomas shown in Fig. 4 was determined by a combination of hybridization to a JH probe, a probe between V_H and D_H segments, plus PCR for proximal V_H to DJ_H rearrangements in each hybridoma (see SI Materials and Methods for details).

Discussion

The intergenic region between the most 3′ V_H segment and the most 5′ D segment has been speculated to harbor elements that control V_H to DJ_H rearrangement. Correspondingly, we find that deletion of the V_H-D intergenic region in T cells dramatically increases germ-line transcription and rearrangement of V_H segments just upstream of the deletion as well as that of the D segments downstream of the deletion. In developing T and B cells, absence of this region leads to high-level antisense transcription through both regions with the accumulation of high steady-state levels of germ-line antisense transcripts that contain both V_H and D sequences. Various lines of evidence suggest that these antisense transcripts initiate in the D cluster and proceed upstream through proximal V_H segments. This antisense transcription strongly correlates with increased rearrangement potential of the transcribed germ-line V_H and D segments, at least in developing T lineage cells. The direct correlation between antisense transcription through and V(D)J recombination of particular germ-line IgH V and D segments indicates that, in this situation, antisense transcription has a positive role in mediating V(D)J recombinational accessibility. Our findings implicate a regulatory element or elements in the V_H-D intergenic region that may modulate germ-line transcription, V(D)J recombination, or both.

Currently, we do not know the nature of the regulatory elements in the V_H-D intergenic region or their function in normal development. Although there are several potential scenarios, this region clearly exerts a silencing effect on transcription from downstream antisense and, to a lesser extent, sense promoters within the D cluster. In this context, there may be insulators or other elements within the V_H-D intergenic region that silence downstream D promoters by modulating their interaction with positive regulatory elements in the IgH locus (e.g., the intronic or 3′ IgH enhancers) or elsewhere via looping or related mechanisms (21, 37). One candidate would be the CTCF binding sites within the V_H-D intergenic region that are postulated to serve as boundary elements between the V_H versus D compartments (25, 28). Alternatively, the ΔV_H-D deletion might activate D region promoters by bringing them into proximity to a putative positive transcriptional regulatory element in the V_H region.

The normal role of the antisense transcriptional silencing function of the V_H-D intergenic region remains to be determined. In its absence, the generation of long antisense transcripts that appear to activate proximal V_H gene segments for rearrangement in T cells provides a striking example of its ability to maintain lineage-specific V_H to DJ_H rearrangement. Whether this region serves such a role in normal cells, where it remains intact following D to J_H rearrangement, is not clear. Conceivably, deregulated T-cell rearrangement of proximal V_H segments on the ΔV_H-D allele, reminiscent of the Pax-5 overexpression phenotype (38, 39), reflects a physiological suppressive function of the intergenic region in T lineage cells. In so, this suppressive function must normally be inactivated in pro-B cells. Alternatively, the region might function to regulate downstream D segment transcription, with deregulated V_H to DJ_H rearrangement in ΔV_H-D T cells reflecting up-regulation of downstream antisense transcription combined with artificial juxtaposition of the V_H segment to the transcription origin. In this context, prior studies concluded that low-level antisense transcription in the D_H cluster suppresses D segment rearrangement (18). The apparent discrepancy between those findings and our current findings might have several explanations, including relative transcription rates, steady-state transcript levels, or transcript structure differences. Another possibility is that activation of downstream transcription on the ΔV_H-D allele might reflect a function normally activated by deletion of this region via V_H to DJ_H recombination, in which case the role may be in transcriptional activation of the completely assembled V(D)J exon. Finally, additional studies will be important to test whether the intergenic region, as postulated (26), contains elements important for regulating allelic exclusion or ordered rearrangement of V_H, D, and J_H segments in B cells.

We describe a gain of function mutation that dramatically up-regulates nascent noncoding RNA transcription. This finding suggests potential existence of other silencing regions/sequences within the mammalian genome that may dictate levels of antisense intergenic noncoding RNAs through suppression of transcription. Thus, further characterization of mechanisms by which the V_H-D intergenic region suppresses antisense noncoding RNA transcription could have more broad implications. In this regard, antisense transcription is often considered to be synonymous with transcriptional suppression stemming from paradigms such as imprinting (40–42). However, we now demonstrate that antisense transcription, even in the presence of sense transcription, does not necessarily lead to negative regulation. Finally, we find that antisense transcription of V_H558 gene segments is surprisingly more robust than their sense transcription. However, steady-state antisense transcripts from V_HJ558 segments are not more abundant than sense transcripts (14, 43), suggesting V_HJ558 antisense transcripts are preferentially processed or degraded or that V_HJ558 sense transcripts are stabilized (44).

Materials and Methods

Gene Targeting.

To make the ΔV_H-D mutant IgH locus, Velocigene technology was used (31). Briefly, a targeting vector made from 129sVe DNA (Fig. S1A) was constructed via recombineering. The vector upstream and downstream arms are sequences flanking the intergenic V_H to D_H region. A neomycin resistance gene flanked by loxP sites allows removal of the selection cassette via Cre recombinase. F1 ES cells with IgM^a (129sVe) and IgM^b (C57BL/6) haplotypes were electroporated with the targeting construct. To screen for targeted clones, quantitative PCR was used to measure the copy number of the unmodified gene. Targeting was detected by a copy-number reduction from two to one per cell by a quantitative TaqMan PCR and confirmed by Southern analysis. See Fig. S1A and SI Materials and Methods for further details.

Nuclear Run-On Assays.

Nuclear run-on analyses were performed as described (32, 45). Further details are in SI Materials and Methods and Table S1.

Northern Blot Analysis.

Isolation of RNA and Northern analyses were done as described (14). Particular dsDNA probes used for indicated Northern blots were generated with Klenow and random hexamers (see Table S2 for further details).

V(D)J Rearrangement Assays.

PCR assays for D to J_H or V_H to DJ_H rearrangements were performed as described (46). See SI Materials and Methods for further details and Table S2 for primers.

T-Cell Hybridoma Analyses.

Splenic T cell hybridomas and analysis of IgH and Igκ rearrangements were done as described (36, 47). See the legend to Fig. 5 and SI Materials and Methods for more details.

RT-PCR Analyses.

RT-PCR assays of germ-line transcripts of IgH gene segments were performed as described (16). See Table S2 for primers.