HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Pawlitzky, I. Articles by Brodeur, P. H. Search for Related Content PubMed PubMed Citation Articles by Pawlitzky, I. Articles by Brodeur, P. H.

Identification of a Candidate Regulatory Element within the 5' Flanking Region of the Mouse Igh Locus Defined by Pro-B Cell-Specific Hypersensitivity Associated with Binding of PU.1, Pax5, and E2A¹

Inka Pawlitzky^*, Christina V. Angeles, Andrea M. Siegel, Michelle L. Stanton^*, Roy Riblet and Peter H. Brodeur^2,*,

^* Immunology Program, Sackler School of Graduate Biomedical Sciences, Tufts University School of Medicine, Boston, MA 02111; Department of Pathology, Tufts University School of Medicine, Boston, MA 02111; and Torrey Pines Institute for Molecular Studies, San Diego, CA 92121

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
The Igh locus is controlled by cis-acting elements, including Eµ and the 3' IgH regulatory region which flank the C region genes within the well-studied 3' part of the locus. Although the presence of additional control elements has been postulated to regulate rearrangements of the V_H gene array that extends to the 5' end of the locus, the 5' border of Igh and its flanking region have not been characterized. To facilitate the analysis of this unexplored region and to identify potential novel control elements, we physically mapped the most D-distal V_H segments and scanned 46 kb of the immediate 5' flanking region for DNase I hypersensitive sites. Our studies revealed a cluster of hypersensitive sites 30 kb upstream of the most 5' V_H gene. Detection of one site, HS1, is restricted to pro-B cell lines and HS1 is accessible to restriction enzyme digestion exclusively in normal pro-B cells, the stage defined by actively rearranging Igh-V loci. Sequence motifs within HS1 for PU.1, Pax5, and E2A bind these proteins in vitro and these factors are recruited to HS1 sequence only in pro-B cells. Transient transfection assays indicate that the Pax5 binding site is required for the repression of transcriptional activity of HS1-containing constructs. Thus, our characterization of the region 5' of the V_H gene cluster demonstrated the presence of a single cluster of DNase I hypersensitive sites within the 5' flanking region, and identified a candidate Igh regulatory region defined by pro-B cell-specific hypersensitivity and interaction with factors implicated in regulating V(D)J recombination.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
Bcell development progresses in multiple steps associated with defined changes in gene expression and the rearrangement of Ig genes. The Ig H chain locus (Igh) is organized as gene segments that are assembled to form Ag receptor coding sequences by the process of V(D)J recombination which is essential for the generation of a vast primary Ag receptor repertoire. The precise targeting of the recombination machinery to regions within the locus depends on an accessible chromatin structure at particular gene segments. Although the molecular mechanisms of accessibility remain to be fully elucidated, the temporal order and complexity of regulation for the initiation, maintenance, and termination of recombinational accessibility suggest the interaction of multiple cis-regulatory elements within or near Ag receptor loci (1, 2, 3).

V(D)J recombination within the H chain locus is tightly controlled to ensure lineage and stage specificity with D to J_H rearrangements on both alleles preceding V_H to DJ_H recombination. In progenitor cells, Igh is anchored via part of its 5' sequence to the nuclear periphery and present in an extended chromatin configuration (4, 5). In pro-B cells, the developmental stage defined by H chain rearrangement, the entire locus moves away from the periphery, a repressive nuclear subcompartment, to a more central position within the nucleus (4, 6). At this stage, possibly as a consequence of DJ_H joining, the D-proximal (3') V_H genes are activated for recombination and attain hallmarks of accessible chromatin, including increased histone acetylation (7) and germline transcription (8), and undergo V_H to DJ_H rearrangements. The D-distal (5') V_H gene segments also become histone acetylated, a process that is linked to IL-7R signaling (7, 9), and two Pax5-dependent processes take place: the removal of histone H3 lysine 9 (H3-K9) methylation (10), a marker of heterochromatin, and the large scale contraction of the locus (11). It is thought that the contraction of the Igh locus is accompanied by the looping of Igh-V domains, a process that moves the D-distal (5') V_H gene cluster into proximity to the DJ_H region and presumably facilitates the rearrangement of the 5' V_H segments (1, 12, 13). After the successful expression of a rearranged (VDJ) H chain, pre-BCR signaling leads to loss of RAG protein expression and induction of locus decontraction on both alleles and thus prevention of further Igh rearrangements (13). Allelic exclusion at the pre-B cell stage also involves the recruitment of the unexpressed allele to regions of heterochromatin which is thought to modulate locus accessibility (6).

The independent regulation of recombination of Igh D-distal (5') vs D-proximal (3') V_H gene segments has been of considerable interest since the seminal observation that, in early B cell ontogeny, a bias toward recombining D-proximal V_H genes exists (14, 15). Further evidence for the existence of independently regulated Igh-V domains was provided by the observation that germline transcripts of D-distal (5') V_H genes, primarily members of the V_HJ558 gene family, are detectable only during early B cell development whereas germline transcription of unrearranged D-proximal V_H segments persists as late as the mature B cell stage (16). A role for IL-7R signaling in D-distal V_H gene accessibility was originally suggested by the observation that IL-7R-deficient mice have a selective defect in recombining V_HJ558 segments (17).

Subsequently, it was demonstrated that IL-7R signaling can induce accessibility of V_HJ558 genes, as measured by nuclease sensitivity, histone acetylation, and germline transcription (9, 18, 19). Furthermore, in pro-B cells lacking Ezh2 or Pax5, the rearrangement of specifically the D-distal V_H segments is drastically reduced compared with D-proximal V_H genes, although the level of accessibility, as measured by germline transcription and histone acetylation of all V_H segments, is unchanged (20, 21). Ezh2 is thought to function in V_H gene recombination by regulating histone H3 methylation and possibly by directing chromatin remodeling complexes through the regulation of nuclear actin polymerization (21, 22). Pax5, in contrast, is required for the contraction of the locus observed during V_H to DJ_H recombination, a process that appears necessary for the efficient recombination of D-distal V_H gene segments (11). However, the elements involved in regulating these processes have not been identified.

Attempts to identify cis-regulatory elements responsible for targeting of gene segments for recombination within the H chain locus, specifically within the large V_H gene array, have mostly focused on transcriptional promoters that flank each V_H gene segment and enhancer elements (2). Whereas transcriptional enhancers exert their control over large genomic regions within Ag receptor loci, germline promoters are required for local chromatin changes associated with individual gene segments (23, 24). The known enhancer elements of Igh, the 3' IgH regulatory region, the intronic enhancer (Eµ), and the DQ52 promoter (PD_Q52) are located within or bordering the D-J_H-C gene clusters that occupy 300 kb of the 3' region of the locus, all of which colocalize with tissue-specific DNase I hypersensitive sites. Although the 3' IgH regulatory region influences class switch recombination and might serve in part as a boundary element for the 3' end of Igh (25, 26), studies have failed to implicate it in V(D)J recombination. Deletion of PD_Q52, an element shown in reporter assays to exhibit promoter and enhancer activity in pre-B cell models (27), has no effect on germline transcription of the D and J_H gene clusters and only minimal impact on Igh rearrangements (28, 29). Only Eµ, positioned between J_H4 and the Sµ region, affects H chain recombination in developing B cells, because mice lacking Eµ have reduced D to J_H rearrangements and substantially impaired V_H to DJ_H rearrangements (30, 31). It has been shown that Eµ influences gene segment accessibility by directly affecting germline transcription across the D-J_H region, whereas germline transcription within the Igh-V genes is unaltered in Eµ-deficient mice (31). Recent studies suggest that the defect in V_H to DJ_H recombination observed in the absence of Eµ can be attributed to a profound reduction in D to J_H joining at the pro-B cell stage of development (28). However, it remains unknown whether Eµ influences additional aspects of locus accessibility within the large Igh-V gene cluster.

The very large size and complex regulation of Igh during B cell development, the lack of identified long-range V_H gene-specific regulatory elements, and the observation that subregions of Ag receptor loci are commonly flanked by control elements (2), suggested the possibility that additional cis-regulating elements associated specifically with the 5' part of the locus remain to be discovered. A 5' flanking position seems advantageous for potential interactions with other Igh chromosomal elements located in different gene clusters within or flanking the locus. In particular, because sequence immediately upstream of the V_H gene cluster is protected from deletion during normal V(D)J recombination, this region fulfills an essential requirement for DNA elements that control Ag receptor loci. In this study, we set out to map the exact 5' end of the coding sequence of the Igh locus and to analyze the immediate 5' flanking sequence for candidate regulatory regions. Using DNase I hypersensitivity assays, we scanned 46 kb of the 5' Igh flanking region and identified a cluster of hypersensitive sites 30 kb upstream of the most 5' V_H gene segment. Significantly, chromatin alterations at one site, HS1, are restricted to pro-B cells, the B cell developmental stage associated with Igh rearrangements. In addition, HS1 contains functional binding sites for PU.1, Pax5, and E2A, factors directly implicated in V(D)J recombination and locus accessibility. These proteins are also recruited to HS1 sequence in a pro-B cell-specific manner and reporter assays indicate a requirement for a functional Pax5 binding site within HS1 sequence.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
Cell lines

The following cell lines were propagated as previously described (16): FL5.12 is an IL-3-dependent pre-pro-B cell (32) and has unrearranged Igh loci (data not shown). 22D6 (Igh^a) is a fetal liver-derived Abelson murine leukemia virus (Ab-MLV)³ transformed pro-B cell line. B14 (Igh^b), F102, F9.9, M4, M9, and M21 (all Igh^a/Igh^b) are bone marrow-derived Ab-MLV transformed pro-B cell lines. CH1, WEHI-231 are lymphomas representing immature B cells. BAL17 is a lymphoma cell line representing mature B cells. BW5147 (Igh^d), a lymphoma cell line, and EL4, a thymoma cell line, represent T cells. MOPC-21 and S107 are plasmacytoma cell lines. KBALB is a murine fibroblast cell line. 25A is an Ab-MLV transformed fibroblast. 6312 is an Ab-MLV transformed pro-B cell derived from RAG2^–/– fetal liver by G. Rathbun and F. Alt.

Bacterial artificial chromosomes (BACs) and cloning

BACs were identified by PCR or hybridization assays from Igh^a 129 strain libraries (Invitrogen Life Technologies). CT7-233D4 was isolated from the CITB-CJ7 (CT7) library. All other BACs used were from the RPCI-22 (RP22) library. BAC restriction fragments (pBR1–8) were subcloned into pGEM3Z vector (Promega). V_HJ558.55 cDNA cloning was performed as previously described (33). Sequencing was done by the Tufts University Core Facility.

DNase I hypersensitivity assays

Nuclei were purified following cell lysis in buffer containing 0.1% Nonidet P-40 (34). Aliquots of nuclei were digested with DNase I (0.25–6.0 µg/ml) at 22°C for 3 min. Digests were stopped by addition of lysis buffer (300 mM sodium acetate, 5 mM EDTA, 0.5% SDS) containing protease and incubation at 55°C overnight. DNA was purified and 15 µg were digested to completion with restriction endonucleases, fractionated on an agarose gel, and transferred to nitrocellulose. Blots were hybridized with fragment probes labeled with [-³²P]dCTP as previously described (35).

Restriction enzyme accessibility assay

Bone marrow cells from 8- to 12-wk-old BALB/c mice were isolated, depleted of RBCs, and sorted based on surface expression of B220, CD43, CD19 (BD Pharmingen), and IgM (Southern Biotechnology Associates). These cells were sorted (MoFlo cell sorter) for B220⁺CD43⁺CD19⁺IgM^– cells (pro-B), B220⁺CD43^lowCD19⁺IgM^– cells (pre-B), and B220⁺CD43^–CD19⁺IgM⁺ cells (immature B). For in vivo digestion, nuclei were washed in NEbuffer 2 (NEB) and 2.5 x 10⁵ nuclei/digest were incubated with increasing units of SphI (NEB) for 8 min at 37°C followed by overnight incubation in lysis buffer containing protease at 37°C. RNase-treated DNA was purified and a one-seventh volume of each preparation was used for ligation-mediated PCR (LM-PCR). The LM-PCR was performed as described (18, 36). Briefly, first strand synthesis was performed using PFU DNA polymerase (Stratagene) followed by the ligation of a linker oligonucleotide using T4 DNA ligase (Invitrogen Life Technologies) at 14°C overnight. PCR amplifications were performed using Vent-exo DNA polymerase (NEB) with a linker-specific primer and a primer/nested primer specific for HS1 sequence. The integrity of each nuclei preparation was tested using a control LM-PCR specific for a SphI restriction site 3.7 kb upstream of HS1. DNA input was determined by PCR using a primer set designed to amplify a sequence 3' of the SphI site in HS1. PCR products were separated on agarose gels and transferred to nitrocellulose. Blots were hybridized with [-³²P]dATP-labeled oligonucleotides. Primer and probe sequences: primer 1, CCATCTTAATCCTGGACCAAGTGC; primer 2, CCTGGACCAAGTGCATTTGGTTAGAAC; primer 3, CATGCCTGTATAAACTCACAG; probe 4, TTGGACACTGACGTCAGCATTACTG; primer 5, GTGAGTGCAGATTGGTACATGAGC; primer 6, AGATTGGTACATGAGCATGACTGTCAC; probe 7, CTGAAATAAGCATAAGGCTACTAGACAATAGC; primer L, CCCGGAGATCTGAATTC.

Nuclear extracts and gel shift assay

Nuclear extracts were prepared by incubating 22D6 cells on ice with lysis buffer (10 mM HEPES, 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM DTT, 1.0 mM PMSF, 0.05% Nonidet P-40). Nuclei were pelleted by centrifugation and nuclear proteins were salt extracted by incubation in extraction buffer (20 mM HEPES, 420 mM NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 0.5 mM DTT, 1.0 mM PMSF, 25% glycerol). After centrifugation, the supernatant was mixed with an equal volume of storage buffer (20 mM HEPES, 0.2 mM EDTA, 0.5 mM DTT, 1.0 mM PMSF, 20% glycerol) and stored at –80°C. All buffers contained an EDTA-free protease inhibitor mixture (Roche). For gel shift assays, double-stranded oligonucleotides were end labeled with [-³²P]dCTP using Klenow fragment (Pharmacia). Approximately 1.0 fmol of probe was added to the binding reaction in a total volume of 20 µl containing 1.0 µg of poly(dIdC) (Pharmacia), 12 mM HEPES, 1.0 mM EDTA, 1.0 mM DTT, 12% glycerol, 300 µg/ml BSA, 4.0 mM Tris-HCl, 10 mM KCl, and 3 µg of 22D6 nuclear proteins or 1.0 µl of in vitro-translated PU.1 protein. For competition analysis, a 50-fold molar excess of unlabeled oligonucleotides was added before the labeled probe. For supershift analysis, 2.0 µg of PU.1-specific, Pax5-specific, or E2A-specific Ab (Santa Cruz Biotechnology) was added to the binding reactions. Complex formation proceeded for 25 min at room temperature. Reactions were loaded onto a prerun nondenaturing 5% polyacrylamide gel and separated at 140 V at 4°C in 0.5% TBE. Dried gels were exposed to films. Oligonucleotide sequences (sense strand, altered nucleotides are underlined): 5' PU.1, GGGTCAGCCAAGAGGAAGTGAGAGT; 5' PU.1mt, GGGTCAGCCAAGAGCTAGTGAGAGT; ETS, GGGCAAAAGTCATTTCCTCTTAGTTCAT; E2A, GGGCTTATTCAGGTGTTCTGGACTGGA; E2Amt, GGGCTTATTCTGGAGTTCTGGACTGGA; Pax5, GGGCTGAATTGTTCCGCAAGGGCCTTTTTATATTT; Pax5mt, GGGCTGAATTGTTCCACGAGGGCCTTTTTATATTT; CD19, GGGGAATGGGGCACTGAGGCGTGACCACCGC.

Chromatin immunoprecipitation (ChIP)

Assays were performed using a ChIP Assay kit (Upstate Biotechnology). Briefly, formaldehyde cross-linked chromatin was sonicated to generate a mean fragment size of 0.5–1.5 kb and diluted in ChIP dilution buffer. After preclearing, the chromatin was incubated overnight with 2.0 µg of one of the following Abs: normal IgG, anti-PU.1, anti-Pax5, or anti-E2A (Santa Cruz Biotechnology). Samples were immunoprecipitated by adding salmon sperm-DNA/protein A-agarose 50% slurry. Beads were collected and washed once in low salt buffer, once in high salt buffer, and twice in TE buffer. Chromatin was eluted using 500 µl of elution buffer and the cross-links were reversed by incubation at 68°C overnight. The purified DNA was resuspended and analyzed by semiquantitative PCR using primers designed to amplify 250- to 360-bp products for HS1, HS1E2A, IL-7R2 (IL-7R- promoter), CD19 (CD19 promoter), µE5 (IgH intronic enhancer), or hypoxanthine guanine phosphoribosyl transferase (HPRT) promoter. The input sample is defined as the total chromatin recovered from the supernatant of the no-Ab sample. Primer sequences (F, forward; R, reverse): HS1F, GAAGGTCAACTGCTCCTGAAAG; HS1R, GCTCTGCTACCTAATGTCATAC; HSE2AF, GGTAGCAGAGCATTACTGTCAGAG; HSE2AR, GTTGCACCTCACGGCTGCTAATG; IL-7RpF, CCAACTGAGAGACATCGTAAGTGTGG; IL-7RpR, CCACAGACAGGGAACTATGAACATC; CD19pF, GTGCGCGCAGTAGTGAAGATG; CD19pR, GCGTGGCTGCGCAGAGGATGC; EµF, GGTCTTGTTTGTGTAGA ACTGAC; EµR, CCACTTCTTCAAACCACAGCTAC; HPRTF, GCGGAGTGATTATCTGGGAATCC; HPRTR, GAAAGCAGTGAGGTAAGCCCAAC.

Reporter constructs

The pJ558 construct consists of a 230-bp fragment containing the functional V_HA1.2 gene promoter inserted upstream of the luciferase gene in pGL-3 Basic (Promega). The pJ558-HS1 construct was made by inserting a 906-bp BamHI-HindIII fragment containing HS1 upstream of the V_HA1.2 gene promoter in pJ558. The pJ558-Emu construct was made by cloning a 585-bp XbaI-HinfI Eµ-containing fragment, shown to retain 97% of the enhancer activity (Ref. 37 ; GenBank no. M12827) upstream of the V_HA1.2 promoter sequence in pJ558. The pJ558-HS1,2,3 construct was made by cloning a 3.0-kb BamHI and ApaI fragment, containing a HS2 and HS3 sequence, upstream of HS1 sequence in pJ558-HS1. The pJ558-HS1 constructs containing the mutated binding sites for either PU.1 (pJ558-HS1mPU.1), Pax5 (pJ558-HS1mPax5), or both factors (pJ558-HS1dm) were made using the QuikChange Site-Directed Mutagenesis kit (Stratagene) following the manufacturer’s protocol. Mutations were confirmed by sequencing.

Transient transfection and luciferase assays

22D6 cells were transfected by electroporation with 20 µg of the luciferase reporter constructs. Each sample was also transfected with 0.4 µg of the Renilla luciferase plasmid, pRL-TK (Promega), to control for transfection efficiency. Cells were harvested after 20 h of incubation and lysed. The luciferase and Renilla luciferase activities were measured using the Dual Luciferase Assay kit (Promega). Shown are representative experiments from at least three independent assays. Each experiment was performed in triplicates and error bars denote SDs. The Student t test was used for statistical analysis.

Computational analysis and GenBank accession numbers

Sequences were analyzed using MatInspector (www.genomatix.de; Ref. 38 or www.essex.ac.uk/bs/molonc) for CTCF binding motifs. AY196298, Igh 5' flanking sequence containing hypersensitive sites; AY197503, Ig H chain V region gene V_HJ558.55; AF305910, Ig H chain V region gene, V_HF102; AF303853, Ig H chain V gene, V_HJ558.22; AF303879, Ig H chain V gene, V_HJ558.48.

	Results

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
Cloning of the 5' end of Igh

Examining the 5' flank of Igh for candidate regulatory sequences first required the precise identification and characterization of the 5' end of the locus. Our strategy relied on previous V_H gene family deletional mapping studies which demonstrated that the large V_HJ558 gene family is the most D-distal (5') V_H gene family of Igh^a and Igh^b (39). We used a mapping cell line, F102 (Igh^a/Igh^b), which has stably rearranged V_H genes on both alleles such that only four unrearranged V_H genes are retained, all members of the V_HJ558 family (39). Because these four V_HJ558 genes must reside at the 5' end of the locus, we used them to identify the 5' boundary of Igh. To accomplish this, we cloned and sequenced a 4.3-kb EcoRI fragment containing one of the remaining, unrearranged V_H segments of F102 (V_HF102). Gene-specific oligonucleotide hybridization showed that V_HF102 is derived from the Igh^a (BALB/c) allele of F102 cells (data not shown). Based on unique sequence flanking the V_HF102 gene segment, a PCR-based assay was designed to screen a BAC library constructed from the Igh^a haplotype strain, 129. We identified one BAC, CT7-233D4 (150 kb), and determined that it contains four V_HJ558 gene segments (Fig. 1A). This BAC also contains the microsatellite marker D12Mit150 which has been mapped to the 5' end of the V_H gene array (40). Comparison of the sequence of each V_H gene on CT7-233D4 to published germline BALB/c V_HJ558 sequences (33) revealed that three of the V_H genes on CT7-233D4 are each identical to different functional germline V_H genes, namely J558.22, J558.25, and J558.48. The sequence of J558.25 matches the coding region of V_HF102 sequence. We verified that the one novel V_H segment on CT7-233D4, designated V_HJ558.55, is functional for rearrangement and expression by isolating a BALB/c cDNA clone that uses this novel V_H segment in a productive rearrangement (data not shown). We conclude that all four V_H genes present on BAC CT7-233D4 (J558.22, J558.25, J558.48, and J558.55; Fig. 1A) are functional based on sequence analysis and their presence in productively rearranged and transcribed Igh alleles.

View larger version (46K): [in this window] [in a new window]   FIGURE 1. Identification of a cluster of DNase I hypersensitive sites within the 5' flanking region of the Igh locus. A, Genomic organization of the 3.0 Mb mouse Igh locus and its 5' flanking region located within the telomeric region of chromosome 12. The map shows VH, D, JH gene clusters, Cµ, and the order of the four most 5' VH genes. The 3' IgH regulatory region (3'RR) and Eµ, both located within the 3' part of the locus, are also shown. The genomic marker D12mit150, Zfp386 (exons 1 and 2 are shown), and Vipr2 map within the region bordering the 5' end of Igh. The transcriptional orientation of Zfp386 and Vipr2 is opposite to that of VH genes. The identified hypersensitive sites are shown as arrows with their distances from the VHJ558.55 gene segment indicated. Two of the BACs, CT7-233D4 and RP22-207I23, used in our studies, are aligned with the Igh and its 5' flanking sequence. Subclones (pBR1–8) and probes (A-F) used in the DNase I hypersensitive assays are depicted. The 17.8-kb PstI and 14.2-kb BglII restriction fragments and the position of probe A are shown with the identified DNase I hypersensitive sites. B, PstI-digested DNA of DNase I-treated nuclei from 22D6 and KBALB cells were analyzed by hybridization with probe A. C, BglII-digested DNA of DNase I-treated nuclei from 22D6 and KBALB cells was analyzed by hybridization with probe A. DNase I hypersensitive sites appear as subfragments within the 14.2-kb BglII fragment. The approximate size of each fragment and the position of each HS are marked.

Identification of hypersensitivity sites within the 5' flanking region of Igh

Transcription factor interactions at regulatory DNA sequences often result in tissue-specific nuclease hypersensitive sites due to alterations in local chromatin structure (42). A number of regulatory elements associated with Ag receptor loci were initially identified by DNase I hypersensitivity assays (43, 44). To facilitate a detailed analysis of the region immediately 5' of the most D-distal V_H gene segment of Igh, 49 kb of flanking sequence was restriction mapped and subcloned as a series of overlapping restriction fragments. Appropriate unique sequence hybridization probes were prepared to perform assays for DNase I hypersensitivity (Fig. 1A). Our strategy was to analyze the immediate 5' Igh flanking sequence for DNase I hypersensitive sites using 22D6, an Ab-MLV transformed BALB/c fetal liver-derived cell line representing an early B cell with actively rearranging Igh loci and detectable V_H germline transcription (data not shown). For our purposes, we consider Ab-MLV transformed B cell lines, like 22D6, all having either DJ_H or V_HDJ_H rearrangements on both alleles, to be pro-B cells (45, 46). A BALB/c fibroblast cell line, KBALB, served as a control for an inaccessible Igh locus. Approximately 46 kb of sequence immediately upstream (5') of the most D-distal V_H gene segment (J558.55) were scanned for DNase I hypersensitive sites using both cell lines. Within this region, a 17.8-kb PstI restriction fragment positioned 34 kb upstream of V_HJ558.55 and a 14.2-kb BglII restriction fragment located 22 kb 5' of V_HJ558.55 and were analyzed using probe A (Fig. 1). In both cell lines, assays using probe A failed to detect DNase I hypersensitivity within the 17.8-kb PstI restriction fragment, as demonstrated by the lack of subfragments generated with increasing amounts of DNase I (Fig. 1B). Analysis of the 14.2-kb BglII restriction fragment revealed three DNase I hypersensitive sites in the pro-B cell line (Fig. 1C). The three fragments generated by DNase I digestion, referred to as HS1, HS2, and HS3, are 3.5, 8.2, and 10.2 kb in length, indicating their distance from the 5' BglII restriction site. HS2 and HS3 are also detected in KBALB fibroblasts. In contrast, HS1 was detectable only in the pro-B cell line but was not observed in any of multiple experiments using DNase I-treated fibroblast nuclei (Fig. 1C).

Tissue specificity of the identified DNase I hypersensitive sites

Analysis of the 14.2-kb BglII restriction fragment was extended to a panel of cell lines representing both non-B cells and B cells of different developmental stages. HS2 and HS3 were detected in all cell lines examined, including pre-pro-B cells (FL5.12), pro-B cells (22D6, B14, M4, M9, M21, F9.9), immature B cells (WEHI-231, CH1), mature B cells (BAL17), plasmacytomas (MOPC-21, S107), and T cells (BW5147, EL4) (Fig. 2 and Table I). In contrast, HS1 was detected only in the pro-B cell lines. These results indicate that the identified set of 5' Igh hypersensitive sites consists of two constitutive hypersensitive sites, HS2 and HS3, and one pro-B cell-specific hypersensitive site, HS1 (Fig. 2 and Table I). Because Ab-MLV-transformed pro-B cells represent early B cells associated with V_H germline transcription, a feature indicating an accessible Igh-V locus (47), these results suggested that HS-1 detection strictly correlates with the B cell developmental stage characterized by H chain locus rearrangement.

View larger version (47K): [in this window] [in a new window]   FIGURE 2. Tissue and developmental stage specificity of the identified 5' Igh cluster of DNase I hypersensitive sites. BglII-digested DNA from DNase I-treated nuclei of FL5.12 (pre-pro-B cell), M4 (pro-B cell), MOPC-21 (plasmacytoma), and 25A (Ab-MLV-transformed fibroblast) cells were analyzed by hybridization with probe A. DNase I hypersensitive sites appear as subfragments within the 14.2-kb BglII restriction fragment. The identified hypersensitive sites are marked.

Fine mapping of HS1, 2, and 3

We subcloned and sequenced nearly 20 kb of the region containing the identified 5' Igh hypersensitivity sites. Restriction fragment analysis determined that HS1, HS2, and HS3 are located 32.6, 27.9, and 26.2 kb 5' of the V_HJ558.55 gene (Figs. 1A and 3). Hypersensitivity assays using the 14.2-kb BglII fragment map HS1 within a 906-bp BamHI–HindIII fragment, centering around a SphI restriction site, 3.5 kb from the 5' end of the BglII fragment (Fig. 3A). Fine mapping of HS2 and HS3 was achieved by the analysis of a 4.4-kb EcoRI fragment within the 14.2-kb BglII fragment using probe F (Fig. 3C). HS2 maps 1.1 kb from the 5' end of the EcoRI fragment and is contained within a 628-bp BamHI–PstI fragment (Fig. 3B). The higher resolution afforded by smaller fragments generated in these hypersensitivity assays revealed that HS3 consists of two clearly distinguishable sites separated by 250 bp, a weaker 5' site (HS3a) and a stronger 3' site (HS3b). Both HS3a and HS3b map within a 1.6-kb SmaI–ApaI fragment (Fig. 3B). The sequence encompassing the identified hypersensitive sites contains an open reading frame coding for the first two exons of Zfp386, a Kruppel-like zinc finger protein initially identified by expressed sequence tag sequencing. The transcriptional orientation of the detected cDNA is opposite that of V_H gene transcriptional orientation. Sequence alignments show that HS1, 2, and 3a are contained within the intron separating Zfp386 exons 1 and 2, whereas HS3b maps just upstream of exon 1 based on Zfp386 transcriptional orientation (Fig. 3C). A similar placement of control elements, within an intron of a neighboring gene, has been observed for the human -globin locus control regions (LCR) and the Th2 cytokine gene cluster LCR (48, 49). RNAs from spleen, bone marrow, and thymus, as well as RNA isolated from FL5.12, 22D6, MOPC-21, and 25A cells were analyzed by Northern blotting using a full-length Zfp386 cDNA probe (IMAGE 4481593). No Zfp386 gene expression was detected in these samples (data not shown). Publicly available expression data indicate a wide expression of Zfp386, but transcripts were not seen in bone marrow, spleen, and lymph node, and are only modestly expressed in thymus, consistent with our results.

View larger version (66K): [in this window] [in a new window]   FIGURE 3. Fine mapping of the identified 5' Igh DNase I hypersensitive sites. A, A Southern blot hybridized with probe A is shown. Using probe A and BglII-digested DNA from DNase I-treated 22D6 nuclei reveals HS1, HS2, and HS3 (lane 2). Lane 1 contains BglII-digested BALB/c liver DNA. For fine-mapping BglII-digested BALB/c liver DNA was further digested with either BamHI (lane 3), SphI (lane 4), or AseI (lane 5). The size of each restriction fragment is indicated and the position of HS1, 2, and 3 is shown. B, To fine-map HS2 and HS3 a 4.4-kb EcoRI fragment that is located within the 14.2-kb BglII restriction fragment was analyzed using probe F. EcoRI-digested DNA from DNase I-treated 22D6 nuclei reveals HS2, HS3a, and HS3b (lane 2). Lane 1 contains EcoRI-digested BALB/c liver DNA. To fine-map the sites, EcoRI-digested BALB/c liver DNA was further digested with BamHI (lane 3), PstI (lane 4), SmaI (lane 5; note partial digest), or ApaI (lane 6). The size of each restriction fragment and the position of each HS are indicated. C, A detailed map of the 14.2-kb BglII fragment with relevant restriction sites is shown and the position of each hypersensitive site is marked. The 4.4-kb EcoRI fragment used to fine-map HS2 and HS3 is located within the BglII fragment. The hybridization probes are marked A and F. Zfp386 exons 1 and 2 are depicted as black boxes. P (PstI), B (BamHI), Sp (SphI), H (HindIII), A (AseI), RI (EcoRI), S (SmaI), Ap (ApaI).

DNase I hypersensitivity studies demonstrated that HS1 is detectable only in cell lines representing early B cells having accessible Igh-V regions. To monitor the accessibility state of HS1 during B cell differentiation in vivo, we used a LM-PCR assay. This sensitive assay allowed the assessment of restriction enzyme accessibility (i.e., open chromatin structure) within HS1 sequence in primary B cell populations.

Nuclei from cell lines previously analyzed in DNase I hypersensitivity assays and nuclei isolated from primary sorted bone marrow B cell populations were incubated with increasing amounts of SphI. After DNA isolation the 3' overhangs resulting from SphI digestion were blunt ended and a unidirectional linker was ligated to the modified ends. The extent of SphI cleavage, an indication of accessible chromatin, was then analyzed in each sample by LM-PCR using a linker-specific and HS1 sequence-specific primers (Fig. 4A). As shown in Fig. 4B, PCR product was detectable in 22D6 and 6312 (RAG2^–/– pro-B cells) samples. Consistent with DNase I hypersensitivity assays, HS1 was not accessible to SphI digestion in FL5.12 and BAL17 cells (Fig. 4B). The amount of LM-PCR product detected in assays of 6312 cells, pro-B cells that have unrearranged, but germline transcribe Igh-V regions (9), was less than observed in assays of 22D6 cells (Fig. 4B). These results suggest that the extent of HS1 accessibility to restriction endonucleases can vary and that D to J_H recombination may be required for complete chromatin alterations at HS1.

View larger version (37K): [in this window] [in a new window]   FIGURE 4. In primary B cells, HS1 detection is restricted to pro-B cells. A, A map of the SphI restriction sites used in the accessibility assay with primer sets used in the LM-PCRs is shown. The LM-PCR assay testing accessibility within HS1 was performed using the linker-specific primer (L) and primers 1 and 2 as sequence-specific primer/nested primer. The control LM-PCR assay testing the specificity of the assay for accessible chromatin was performed using primer L and primers 5 and 6 as primer/nested primer specific for sequence flanking a SphI restriction site 3.7-kb upstream of HS1. DNA input for all samples was determined using primers 1 and 3. HS1 and input PCR products were detected by hybridization with oligonucleotide 4. Control LM-PCR products were detected by hybridization with oligonucleotide 7. B, Isolated nuclei from 6312, 22D6, FL5.12, and BAL17 cells were treated with SphI and analyzed for the extent of SphI cleavage. Purified 22D6 gDNA digested with SphI and undigested 22D6 cell gDNA were used as specificity control for the PCR assay. C, Nuclei from sorted fractions of BALB/c bone marrow B cells were analyzed for SphI accessibility using the LM-PCR assays described above. Product is detectable in the in vivo SphI-treated pro-B cell fraction and 22D6 cells, which were used as control for the PCR, using the HS1-specific LM-PCR assay. The integrity of each nuclei preparation was confirmed using the control LM-PCR, because product is only detected in the in vitro-digested 22D6 gDNA sample.

These results demonstrated that HS1 detection is specific to the pro-B cell stage at which the entire Igh-V region is accessible based on DNase I sensitivity (50), V_H germline transcription (16), and histone hyperacetylation (9). In addition, it has been clearly demonstrated that pro-B cells initiate Igh-V rearrangement based on the detection of RAG-dependent double-strand breaks at the V_H gene RSS (51).

In vitro binding studies for sequence motifs within HS1

Detailed sequence analysis was primarily focused on HS1 because of its pro-B cell restriction. HS1 sequence contains two binding motifs for PU.1 located 5' and 3' of the SphI restriction site. The more 5' binding motif is a perfect match (AGAGGAAGT) for the consensus PU.1 binding sequence (52) and is part of a PU.1/IFN regulatory factor (IRF) composite element (Fig. 5D). After demonstrating by Western blot that 22D6 cells express PU.1 (data not shown), we used 22D6 nuclear proteins and an oligonucleotide containing the 5' PU.1 motif in binding reactions that resulted in the formation of four distinct complexes (complex 1–4, Fig. 5A, lane 2). The specificity of three of the complexes (complex 1–3) was confirmed by competition assays using 50-fold molar excess of unlabeled self-, mutated-, and ETS-consensus (53) oligonucleotides (Fig. 5A, lanes 3–5) because the mutant oligonucleotide having two altered nucleotides within the core binding motif (GGAA to GCTA) was unable to compete for complex formation. Consistent with these results, when the mutated PU.1 oligonucleotide was directly used as probe, no complex formed (Fig. 5A, lane 6). Addition of a PU.1-specific Ab, but not control Ab, to the binding reaction resulted in a shift of the fastest moving complex (complex 1) confirming the presence of PU.1 in this complex (Fig. 5A, lanes 7 and 8). Using in vitro-translated PU.1 protein in similar binding reactions resulted in the formation of a single complex that was shifted upon addition of the PU.1-specific Ab (Fig. 5A, lanes 11 and 12). We observed a small difference in the migration between the complex containing the in vitro-translated PU.1 and the PU.1 present in 22D6 nuclear proteins. This may reflect differences in posttranslational modifications or the interaction of PU.1 with other proteins in the nucleus of 22D6 cells. Based on these in vitro binding studies, we conclude that the 5' PU.1 motif is a specific binding site.

View larger version (77K): [in this window] [in a new window]   FIGURE 5. Binding motifs within HS1 are functional in vitro. A, Gel shift assays for the PU.1 motif within HS1 were performed using 22D6 nuclear proteins or PU.1 protein and an oligonucleotide containing the PU.1 motif in binding reactions (lanes 2 and 11). Complexes are indicated 1–4. The specificity of the complexes for PU.1 was confirmed using competition and supershift analyses (lanes 3–5, 7, 12). The supershifted complex is marked S. Lane 8 contains control IgG. Lane 9 contains the PU.1-specific Ab and the oligonucleotide encompassing the PU.1 motif to control for nonspecific interaction of the Ab. Lane 6, The mutated PU.1 oligonucleotide was directly used as probe. Lanes 1 and 10 contain free PU.1 or free mutated PU.1 oligonucleotide probes. B, Gel shift assays for the Pax5 binding motif were performed using 22D6 nuclear proteins. Binding reactions result in the formation of one prominent complex, marked 1 (lane 2). The complex also formed when the CD19 promoter oligonucleotide containing a known Pax5 motif was used as probe (lane 9). The specificity of the complex for Pax5 was confirmed by competition and supershift analyses (lanes 3–5, 6, 10–13). The supershifted complex is marked S. Lanes 7 and 14 contain control IgG. Lane 8, The mutated Pax5 binding motif was used in binding reactions with 22D6 nuclear extracts. Lanes 1 and 15 contain free oligonucleotide probes. C, Gel shift assays were performed using 22D6 nuclear proteins and E2A or the mutated E2A binding motif as oligonucleotide probes (lanes 1 and 7). The binding reactions result in the formation of one prominent complex, labeled 1. Competition experiments were performed using increasing amounts of either self- or mutated E2A oligonucleotides in the binding reactions (lanes 3–6). Lanes 8 and 9 contain either an E2A-specific Ab or control IgG. Lanes 1 and 10 contain free probes only. Lane 11, The E2A-specific Ab was incubated directly with the E2A oligonucleotide to test for nonspecific interaction. Nonspecific binding is marked NS. D, Nucleotide sequences of the identified binding motifs within HS1 for PU.1, Pax5, and E2A are shown. The PU.1 binding motif is part of a PU.1/IRF composite element. mt, Mutated binding motif. *, Opposite strand.

HS1 sequence also includes a Pax5 binding motif that matches the 17-bp consensus sequence across 13 bp (54) (Fig. 5D). The expression of Pax5 in 22D6 nuclear extracts was confirmed by Western blots (data not shown). Incubation of the Pax5 motif containing oligonucleotide with 22D6 nuclear proteins resulted in the formation of a single prominent complex (Fig. 5B, lane 2). The sequence specific binding of Pax5 was confirmed by competition analysis. An oligonucleotide containing a known Pax5 binding site within the CD19 promoter (55) and unlabeled self-oligonucleotide competed for complex formation whereas the mutated HS1 Pax5 oligonucleotide, in which two nucleotides were altered (CGCAA to CACGA), failed to compete (Fig. 5B, lanes 3–5). We demonstrated the presence of Pax5 in the observed complex by showing that the Pax5-specific Ab shifted the complex (Fig. 5B, lane 6). To validate the specificity of our binding assays for Pax5, we showed that using the CD19 promoter Pax5 oligonucleotide directly as probe resulted in the formation of an indistinguishable complex that was specific for Pax5 based on competition and supershift analyses (Fig. 5B, lanes 9–14). In repeated binding assays, the Pax5 binding motif within HS1 competed only partially for the CD19 Pax5-containing motif, suggesting that the HS1 motif might be a lower affinity Pax5-specific binding site.

Within the HS1 sequence, we also identified a binding motif for the basic helix-loop-helix proteins E47 and E12, alternative splice variants of the E2A gene, positioned 5' of the SphI restriction site (Fig. 5D) and analyzed the specificity of this binding sequence for E2A using in vitro binding assays. Binding reactions of 22D6 nuclear extracts, that were confirmed to contain E2A protein by Western blotting (data not shown), and the HS1 E2A oligonucleotide resulted in the formation of two prominent complexes (Fig. 5C, lane 2). The faster moving complex is present when either wild-type or mutant probes are used in the binding reactions (Fig. 5C, lane 2 and 7) indicating that the formation of this complex is independent of the core E2A binding motif and therefore a nonspecific complex (complex NS). The formation of the other complex (complex 1) depends on the core E2A motif because, unlike the wild-type oligonucleotide, the mutant oligonucleotide in which the core was altered (CAGGTG to CTGGAG) failed to compete for complex formation (Fig. 5C, lanes 5 and 6) and binding reactions using the mutant E2A oligonucleotide directly as probe revealed only nonspecific complex formation (Fig. 5C, lane 7). Addition of an E2A-specific Ab to the binding reactions did not produce a compelling change in the observed complex formation (Fig. 5C, lane 8), which could indicate that the available epitope on the isolated E2A protein is not recognized in these in vitro binding studies. These results demonstrate that E2A present in 22D6 nuclear extracts displays sequence-specific binding to the E2A motif identified within HS1 in vitro.

In vivo recruitment of PU.1, Pax5, and E2A to HS1 sequence is pro-B cell specific

To determine whether PU.1, Pax5, and E2A are recruited to HS1 in vivo, we performed ChIP assays using 22D6, a pro-B cell line having accessible HS1 sequence, and BAL17, a mature B cell line having HS1 sequence inaccessible. Using in vivo cross-linked chromatin from 22D6 pro-B cells and a PU.1-specific Ab, the recovered chromatin fraction in the anti-PU.1-immunoprecipitated sample was greatly enriched for HS1 sequence relative to the chromatin recovered from the control-Ig and no-Ab samples (Fig. 6A, upper panel), suggesting that PU.1 is bound to HS1 in 22D6 pro-B cells. In contrast, we did not detect a similar specific enrichment of HS1 sequence using BAL17 chromatin recovered after immunoprecipitation with the PU.1-specific Ab (Fig. 6A, upper panel). Comparable PU.1 expression in 22D6 and BAL17 nuclear extracts was demonstrated by Western blotting (data not shown). As control for the specificity of the assay, we examined the IL-7R2 promoter region that encompasses a functional PU.1 binding site (56) and showed specific PU.1 recruitment to the IL-7R2 promoter region in both 22D6 and BAL17 samples immunoprecipitated with anti-PU.1 (Fig. 6A, middle panel). Expression of the IL-7R chain was confirmed by flow cytometry and RT-PCR for both cell lines (data not shown). A region within the HPRT promoter, encompassing only unrelated binding sites, served as negative control in the assay (Fig. 6A, lower panel).

View larger version (100K): [in this window] [in a new window]   FIGURE 6. Recruitment of PU.1, Pax5, and E2A to HS1 sequence in vivo is pro-B cell specific. A, ChIP assays were performed for the PU.1 binding motif using cross-linked chromatin from 22D6 and BAL17 cells. The chromatin from both cell lines was immunoprecipitated using a PU.1-specific Ab or control IgG. PCRs were performed on 4-fold diluted samples of 5% input or immunoprecipitated DNA recovered from the no-Ab, anti-PU.1, and control IgG samples. In the upper panel, the PCR was designed to amplify HS1 sequence containing the functional PU.1 binding motif. PCR for the IL-7R2 promoter (middle panel) and the HPRT gene promoter (lower panel) were used as specificity control for the PU.1 Ab. B, ChIP assays specific for Pax5 were performed using cross-linked chromatin from 22D6 and BAL17 cells. Four-fold dilutions of the recovered chromatin from anti-Pax5 immunoprecipitated, control IgG, and no-Ab samples or 5% input sample were used as template for PCR with primers designed to amplify HS1 sequence encompassing the Pax5 binding motif (upper panel). Sequence of the CD19 promoter and the HPRT gene promoter were amplified as positive and negative control for the specificity of the Pax5 Ab, respectively (middle and lower panel). C, Recruitment of E2A to HS1 sequence was tested using ChIP assays. 22D6 and BAL17 cross-linked chromatin was immunoprecipitated using an E2A-specific Ab or control IgG. Four-fold dilutions of the recovered chromatin from the immunoprecipitated samples, the no-Ab control, and 5% input were analyzed by PCR with primers specific for HS1 sequence encompassing the E2A binding motif (upper panel). Sequence of the µE5 and the HPRT promoter region were amplified as positive and negative control for the E2A Ab, respectively (middle and lower panel).

Using an E2A-specific Ab, we performed ChIP assays to determine whether E2A interacts with HS1 sequence in 22D6 and BAL17 cells. Immunoprecipitated DNA/protein complexes were analyzed for enrichment of HS1 sequence in 22D6 and BAL17 samples. As shown in Fig. 6C, upper panel, the recovered chromatin fraction containing the HS1 sequence is enriched in the E2A-immunoprecipitated 22D6 sample as compared with the control-Ig and no-Ab samples, whereas such specific enrichment was not observed using BAL17 samples. As specificity control for the E2A Ab, a similar sized region within the intronic enhancer Eµ that contains known E2A binding sites (57, 58) was analyzed for both cell lines. In vivo interaction of E2A with this enhancer region was confirmed in 22D6 and BAL17 samples (Fig. 6C, middle panel). Comparable expression of E2A in 22D6 and BAL17 nuclear proteins was demonstrated by Western blotting (data not shown). The HPRT promoter region served as negative control in the assay (Fig. 6C, lower panel).

We conclude that PU.1, Pax5, and E2A are recruited to HS1 sequence in vivo in a pro-B cell, but are not associated with HS1 in a mature B cell line. The recruitment of these factors to HS1 sequence in 22D6, but not BAL17, does not correlate with the amount or activity of each protein in BAL17 because we show that these factors are specifically recruited to known binding sites in BAL17. Thus, these results suggest that the binding of PU.1, Pax5, and E2A to HS1 depends on the developmental stage-specific processes that result in altered chromatin structure at HS1. The restriction of HS1 detection to the pro-B cell stage of development together with the pro-B cell-specific recruitment of PU.1, Pax5, and E2A, factors implicated in V(D)J recombination (59, 60, 61), support our hypothesis that HS1 might be part of a novel regulatory element involved in Igh regulation.

In vitro functional analysis of the 5' Igh hypersensitive site

Because locus control regions (LCR) and other cis-acting elements often regulate transcriptional activities, we tested the ability of HS1 alone, or in combination with the other identified hypersensitive sites, to influence transcription from a V_H gene promoter in transient transfection assays. We inserted the HS1 sequence, or sequence containing the entire cluster of hypersensitive sites, upstream of a V_HJ558 gene promoter in luciferase reporter constructs. Transient transfection of 22D6 pro-B cells with HS1-containing constructs showed a consistent 3-fold (65%) repression of transcriptional activation relative to V_HJ558 promoter-only construct (Fig. 7A). Transfection of the construct containing the combined sequences of all hypersensitive sites resulted in a 6-fold (83%) decrease of reporter activity compared with the V_HJ558 promoter only construct (Fig. 7A). Conversely, when a reporter construct containing the Eµ sequence upstream of the V_HJ558 promoter was transfected into 22D6 cells, a consistent 7-fold increase of transcriptional activation was observed (Fig. 7A).

View larger version (35K): [in this window] [in a new window]   FIGURE 7. Analysis of transcriptional activity of the 5' Igh hypersensitive sites. A, 22D6 pro-B cells were transfected with luciferase constructs containing a VHJ558 gene specific promoter (pJ558) and either HS1 (pJ558-HS1) or the combined sequence of the cluster of hypersensitive sites (pJ558-HS1,2,3) upstream of the VH gene promoter. The luciferase activity of each construct was measured. The J558.muE construct contains enhancer sequence of the intronic enhancer upstream of the VH gene promoter and served as a transcriptional activator in these reporter assays. B, Reporter assays testing the pJ558-HS1 construct containing mutant binding motifs for either PU.1 (pJ558-HS1mPU.1), Pax5 (pJ558-HS1mPax5), or both factors (pJ558-HS1dm) in similar transfection assays are shown. Assays were performed in triplicates and the error bars denote SDs. RLU, Relative luciferase units. *, Statistical significance.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
Ig H chain recombination and establishment of allelic exclusion are associated with a complex set of epigenetic changes at the Igh locus. It has been proposed that the regulation of these developmentally programmed changes is likely to involve novel cis-regulatory elements within or flanking the locus. We therefore chose to examine the 5' flanking region of Igh for candidate regulatory elements by scanning this region for chromatin alterations associated with early B cell development and Igh locus rearrangement. A significant obstacle to initiating these studies was that, unlike the well-characterized 3' region of Igh (40), the 5' boundary for the mouse Ig H chain locus was not defined. To identify the 5' end of the locus, we used information from V_H gene deletional mapping to identify a BAC clone containing the four most D-distal V_H gene segments. These four V_HJ558 gene segments are functional and therefore delineate both the physical and functional 5' end of Igh coding region. We mapped and subcloned nearly 49 kb of the immediate 5' flanking region and scanned this previously uncharacterized region for DNase I hypersensitive sites. In this study, we have identified a cluster of four DNase I hypersensitive sites within the region flanking the most D-distal V_H gene segment. One site, HS1, is detectable exclusively in pro-B cells, the developmental stage defined by actively rearranging Igh loci. The other sites in the cluster, HS2, HS3a, and HS3b, are constitutive hypersensitive sites detectable in all cells tested, including non-B lineage cells and B cells representing a range of developmental stages. The location of these hypersensitive sites within the 5' boundary of Igh and the pro-B cell specificity of HS1 suggest that this cluster of hypersensitive sites may define a novel regulatory region for Igh. This possibility is supported by the fact that complex cis-regulatory elements are often initially identified as one or more hypersensitive sites. For example, the 3' IgH regulatory region located within an 28-kb region downstream of the C exon (63) is composed of at least seven DNase I hypersensitive sites that function as enhancer or insulator/boundary elements and regulate the expression of germline promoters at distances reaching 130 kb (26, 64). Other important regulatory elements within the Igh locus, including Eµ and the PD_Q52 promoter region, are associated with single tissue-specific DNase I hypersensitive sites (7). Moreover, novel Ig loci regulatory elements continue to be revealed using nuclease hypersensitivity to detect alterations in chromatin structure. For example, a recently described Igk regulatory element, the plasmacytoma cell-specific transcriptional enhancer Ed (HS9) was identified downstream of Ek3' based on its associated DNaseI hypersensitivity (65).

Although our mapping and sequencing studies have focused on the Igh^a haplotype (BALB/c), we detected an identical cluster of hypersensitive sites in a pro-B cell line derived from Igh^b mice (C57BL/6). Furthermore, 13.5 kb of contiguous sequence from the Igh^a allele, containing all four sites, are 98% identical to sequence retrieved from the C57BL/6 genome database and 100% identical at HS1 and HS3a/b sequence. Significantly, the four hypersensitive sites are similarly located telomeric to the most 5' V_H gene segment in both haplotypes. Our 20 kb of the mouse Igh^a sequence containing the 5' Igh hypersensitive sites is also conserved in the rat genome (82% identity) and aligns near the 5' end of the rat Ig H chain locus on chromosome 6. In addition, HS1 sequence is 90% identical to the homologous rat sequence and a nearly perfect match at the identified binding motifs. HS3a/b sequence is 94% conserved in the rat Igh 5' flanking region. The sequence conservation of the identified cluster of hypersensitive sites between two strains having highly polymorphic Igh-V regions (66) and between rodent species further supports the hypothesis that the set of hypersensitive sites 5' of Igh has an important regulatory function. Comparison of the mouse sequence containing the identified cluster of hypersensitive sites with human Igh 5' flanking region was not possible because the 5' end of the human V_H region is located within a few kilobases of the 14q telomere. It is possible that the subtelomeric position of the human 5' V_H genes requires distinct regulatory mechanisms to escape silencing due to the telomeric position effect (67). In contrast, the mouse Igh-V domain must be isolated from genes encoded beyond its 5' end. Other differences between the human and mouse Igh loci, including the length of the V_H gene array and possibly the role of IL-7R signaling on V_H gene accessibility (68), might also be associated with distinct regulatory elements.

We have shown that PU.1 binds HS1 in vitro and is recruited to HS1 sequence in pro-B cells. PU.1 binding at HS1 is particularly intriguing because this factor was originally identified through its binding to Eµ and was subsequently shown to bind sites within other Ig enhancers, including the E3' enhancer (69) and the Ig 2–4 enhancers (70). PU.1 has also been shown to increase chromatin accessibility proximal to its binding sites in Eµ, leading to the suggestion that this factor may be capable of opening local chromatin and facilitating the assembly of tissue- and gene-specific protein complexes (59, 71). In contrast, a recent analysis of PU.1^–/–, Spi-B^–/– pro-B cells suggests that PU.1 may actually inhibit Igh transcription and V(D)J recombination (72), thereby highlighting that the physiological functions of PU.1 are context dependent and their precise in vivo role in Ig locus transcription and recombination are unknown.

The PU.1 binding site identified in HS1 is part of a PU.1/IRF composite element, similar to the PU.1 sites within Eµ, E3', and the Ig2–4 enhancers (57, 69, 73). These sites bind PU.1 which in turn recruits either IRF-4 (Pip) or IRF-8 (ICSBP) to form complexes implicated in regulating Ig locus transcription. Although we have not determined whether IRF-4 or IRF-8 bind to the composite element in HS1, the formation of multiple PU.1-containing complexes in our in vitro binding assays suggests that PU.1 does interact with coregulatory proteins in pro-B cells. A recent study, however, showed IRF-4 or IRF-8 to be required for L chain rearrangements but found no similar requirement for H chain rearrangements (74). Nevertheless, because these studies did not examine Igh rearrangements or V_H gene repertoire in detail, it remains to be determined whether IRF-4 or IRF-8 have a role in accessibility of V_H gene segments or some other aspect of Ig H chain rearrangement.

Pax5 is critical for the development of committed B lineage cells and Pax5-deficient bone marrow cells are blocked at the pro-B cell stage. Pax5 is also one of the few transcription factors known to regulate V(D)J recombination and is directly required for the rearrangement of D-distal V_H gene segments (75). Pax5 function in H chain recombination involves its role in mediating Igh locus contraction (11) and presumably its function in removing inhibitory H3-K9 methylation associated with V_H genes (10). In terms of the known Ig locus cis-regulating elements, Pax5 binds both E3' and 3' IgH HS-1, 2 and represses these enhancers by directly targeting the transcriptional function of PU.1 in a highly context-dependent manner (62, 76). In light of the central and diverse roles of Pax5 in B cell biology and V(D)J recombination, our finding that Pax5 recruitment to HS1 correlates with H chain rearrangement during B cell development supports the possibility that HS1 functions as a novel Igh locus regulatory element.

Transient transfection assays did not detect enhancer activity for HS1 alone, or in combination with HS2 and HS3a/b, in the context of transcriptional activation of a V_HJ558 gene promoter in pro-B cells. In fact, these experiments revealed a modest repressive function for these tested sequences. Mutation of PU.1 and Pax5 binding sites demonstrated that the Pax5 motif is responsible for the observed repression. These results indicate that, in the context of these reporter constructs, Pax5 may have a dominant role in the repression mediated by HS1. Although an in vivo function cannot be attributed to HS1 based on our reporter assays, they suggest that Pax5 may mediate topological changes at HS1 and therefore support our hypothesis that HS1 might be part of a functional regulatory element. Obviously, deletion of these hypersensitive sites from the endogenous locus will be required to determine their role, if any, in V(D)J recombination and Igh expression.

HS1 sequence has an E2A binding motif which is identical to the E2A binding site within Eµ (µE5). E2A-deficient mice have greatly reduced levels of germline (µ0) transcripts and a complete absence of Igh rearrangements (77). E2A binds the HS1 E-box in vitro and, based on ChIP assays, is recruited to HS1 in pro-B cells. The presence of E2A at HS1 is especially interesting in light of recent studies that demonstrated the association of E47 at the site of physical interaction between widely spaced enhancers of rearranged Ig loci (65). Furthermore, it has been recently proposed that long-range physical interactions between Igh-V regions and the DJ_H cluster in pro-B cells facilitates H chain rearrangements and that these interactions are significantly reduced in E2A-deficient hemopoietic progenitor cells (12). Taken together, these results are consistent with a "looping model" in which widely spaced cis-regulatory elements interact to form an "active chromatin hub" capable of recruiting transcription factors and chromatin remodeling complexes (78).

In summary, our studies have identified a cluster of four hypersensitive sites within the 5' flanking sequence of Igh, one of which (HS1) binds multiple transcription factors important in B cell development and Ig locus recombination. The constitutive hypersensitive sites within the cluster, HS2 and HS3a/b, have no enhancer activity in our reporter assays potentially because they might constitute boundary or insulator elements within the 5' flank of Igh similar to elements that have been identified as part of the 3' IgH regulatory region and the -globin locus (26). Although not tested, HS2 sequence contains a binding motif for CTCF, a conserved NF involved in mediating chromatin insulator function as part of DNA elements that function to separate control elements between subdomains (79). As part of a potential novel 5' Igh regulatory region, HS2 could thus protect neighboring sequences from recombination events or accessibility changes associated with these events. Sequence analysis of HS3a revealed binding motifs for cut-like protein x/CCAAT-displacement proteins (Cux/CDP) and special AT-rich sequence binding protein (SATB1), both factors frequently colocalized with matrix attachment regions (80). An intriguing possibility is that HS3a has a role in defining chromatin borders necessary for V(D)J recombination and Igh expression.

Analyses of both cell lines and normal B-lineage populations revealed a striking restriction of HS1 detection to an early stage of B cell development, namely the pro-B cell stage. Thus, chromatin alterations involving HS1 sequence are tightly associated with the window of B cell development during which the Igh-V locus is accessible and V_H to DJ_H rearrangements take place. Specifically, the Igh loci of pro-B cells are marked by germline transcription of D-distal V_H genes (16), Igh-V region genic and intergenic antisense transcription on both alleles (8), covalent histone modifications (81), the repositioning of the locus toward the center of the nucleus (4, 5), and the contraction of the entire locus on both alleles (11). In addition, it has recently been observed that, before recombination in pro-B cells, individual Igh-V domains localize into DJ_H proximity through a looping mechanism which appears to be monoallelic (12). The complexity of this emerging picture of pro-B cell specific Igh locus alterations, together with an ever-increasing appreciation of the pivotal role of chromatin dynamics in regulating gene expression (78, 82) suggest that, in combination with Eµ and the 3' IgH regulatory region, additional long-range cis-regulating elements are likely to be involved in rearrangement, expression, and allelic exclusion of Igh. Because HS1 defines a region that undergoes chromatin changes and factor binding at the B cell developmental stage marked by numerous Igh locus alterations, it is tempting to speculate that the region defined by the 5' Igh hypersensitive site cluster may function in one or more aspects of V(D)J recombination.

	Acknowledgments

 
We thank Dr. Edouard Vannier for providing the in vitro-translated PU.1 protein. We thank Dr. Ananda Roy for critical review of this manuscript.

	Disclosures

Top Abstract Introduction Materials and Methods Results Discussion Disclosures References

 
The authors have no financial conflict of interest.

	Footnotes

 
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

¹ This work was supported by National Institutes of Health Grant GM36064 (to P.H.B.), AI23548 (to R.R.), and National Institutes of Health Training Grant 5T32AI07077 (to I.P. and M.L.S.).

² Address correspondence and reprint requests to Dr. Peter H. Brodeur, Department of Pathology, Tufts University School of Medicine, 150 Harrison Avenue, Boston, MA 02111. E-mail address: peter.brodeur{at}tufts.edu

³ Abbreviations used in this paper: Ab-MLV, Abelson murine leukemia virus; BAC, bacterial artificial chromosome; LM-PCR, ligation-mediated PCR; ChIP, chromatin immunoprecipitation; HPRT, hypoxanthine guanine phosphoribosyl transferase; LCR, locus control region; IRF, IFN regulatory factor.

Received for publication December 21, 2005. Accepted for publication March 22, 2006.

	References

Top </