Analysis of Mice Lacking DNaseI Hypersensitive Sites at the 5′ End of the IgH Locus

The 5′ end of the IgH locus contains a cluster of DNaseI hypersensitive sites, one of which (HS1) was shown to be pro-B cell specific and to contain binding sites for the transcription factors PU.1, E2A, and Pax5. These data as well as the location of the hypersensitive sites at the 5′ border of the IgH locus suggested a possible regulatory function for these elements with respect to the IgH locus. To test this notion, we generated mice carrying targeted deletions of either the pro-B cell specific site HS1 or the whole cluster of DNaseI hypersensitive sites. Lymphocytes carrying these deletions appear to undergo normal development, and mutant B cells do not exhibit any obvious defects in V(D)J recombination, allelic exclusion, or class switch recombination. We conclude that deletion of these DNaseI hypersensitive sites does not have an obvious impact on the IgH locus or B cell development.


Introduction
The variable region of an immunoglobulin heavy chain (IgH) is assembled from V (variable), D (diversity), and J (joining) gene segments that lie upstream of several IgH constant (C) region exons in a process called V(D)J recombination [1]. The mouse IgH locus contains large numbers of V H segments and multiple D and J H segments but an individual IgH V(D)J exon is assembled from only one V H , one D, and one J H segment. V(D)J recombination of the IgH locus takes place in pro-B cells in an ordered way such that D to J H recombination precedes V H to DJ H recombination [2]. In this regard, activation of the IgH locus is thought to progress in a stepwise manner [3]. D to J H rearrangement efficiently occurs on both alleles, however, allelic exclusion ensures that V H to DJ H recombination results in expression of a functional heavy chain (HC) from only one of the two alleles [4].
Mature B-cells can undergo further alterations of their HCs. IgH class switch recombination (CSR) causes expression of different immunoglobulin isotypes which confer different effector functions. During this recombination process one of several sets of downstream C H exons replaces the Cm exons and the intervening sequence is deleted from the chromosome, which results in expression of a new C region without changing the specificity of the IgH variable region [5].
A large effort has been made to elucidate mechanisms of IgH locus regulation and a number of cis-regulatory elements have been described and characterized. The IgH intronic enhancer (Em) resides in the J H -C H intron and was shown to be necessary for efficient V(D)J recombination by promoting both D to J H and V H to DJ H recombination [6,7]. Downstream of the C H genes at the very 39 end of the IgH locus a cluster of DNaseI hypersensitive sites was described, termed 39 IgH regulatory region (39IgH RR). So far two main functions have been assigned to this regulatory region: the 39IgH RR plays an important role in promoting CSR to most IgH isotypes, and the 39IgH RR was shown to be necessary for high level expression of the functionally assembled HC gene from the promoter 59 of the V H DJ H exon [8].
An additional potential regulatory region was identified at the 59 end of the IgH locus, consisting of four DNaseI hypersensitive sites [9]. One of these sites, HS1, was shown to be pro-B cell specific, the stage during which IgH V(D)J recombination takes place, and was suggested to include binding sites for the transcription factors PU.1, Pax5 and E2A [9]. These observations led to the suggestion that this region might represent a new regulatory region for IgH rearrangements. In this regard, the 59 end of the IgH locus is an attractive location for a regulatory element because it would not be deleted during the course of V(D)J recombination, and it might explain control of several unresolved phenomena in the IgH locus. Among these is the regulation of V H germline transcripts as so far no cis-regulatory element has been identified that controls activity of the bulk of unrearranged V H promoters. Furthermore, it is not known how it is achieved that proximal and distal V H segments are activated independently or why usage of distal versus proximal V H gene families varies significantly.
Here we report the targeted deletion of the pro-B cell specific 59IgH HS1 as well as combined deletion of HS1, HS2, HS3a,b in mice. We analyzed potential implications on B cell development, V(D)J recombination, and IgH CSR.

Methods
Targeted deletion of 59IgH DNaseI hypersensitive sites in ES cells and generation of mutant mice All mouse were handled in strict accordance with good animal practice as defined by the relevant national and/or local animal welfare bodies, and all animal work was approved by Animal Research of Children's Hospital Boston (Protocol # 08 11 1253R). The RHS1 targeting vector was assembled in pLNTK [10]. As a 59 homology arm a 2.2 kb PCR product was generated with primers 59 GTCGACGGATTTAGGAGGATACACAAC 39 and 59 GTCGACCTTGGATAACACAGAACTCTG 39 containing a SalI site at their 59 ends, which facilitate cloning of the PCR product into the SalI site of pLNTK. As a 39 homology arm a 7.3 kb AatII -ApaI fragment was blunt end cloned into the XhoI site of pLNTK. The R39HSs targeting vector was generated by blunt end cloning a 4.4 kb EcoRI fragment into the SalI site of pLNTK as a 59 homology arm, and a 7.0 kb KpnI fragment into the XhoI site as the 39 homology arm. Correct targeting events and cre -loxP deletion events were confirmed by Southern blotting (Fig 1). Probe 1 is a 830 bp PCR product amplified with primers 59 GCTCATGTACCAATCTGCACTCAC 39 and 59 CACTGT-GACCTCCATCTTATGTCTG 39. Probe 2 is a 1.2 kb PstI -EcoRI fragment 59 of HS2. Probe 3 is a 0.8 kb PstI -XbaI fragment about 11 kb 39 of HS3b. To confirm single integration of the targeting vectors a 525 bp Neo R probe was used, amplified with primers 59 GCAGCCATATGGGATCGGC 39 and 59 GTTCGGCTGGCGCGAGCCCC 39. EF1 heterozygous IgH a/b embryonic stem (ES) cells, generated in the Alt laboratory, were transfected with PvuI linearized RHS1 targeting vector to obtain RHS1/+ ES cells. To obtain DHS1/+ ES cells, the PGK-Neo R cassette was deleted by applying a Creexpressing adenovirus vector. DHS1/+ ES cells were transfected with PvuI linearized R39HSs targeting vector to obtain R39HSs/+ ES cells. R39HSs/+ ES cells were selected for homozygocity of the targeted allele through increasing concentration of G418 to obtain R39HSs/R39HSs ES cells. Cre -loxP mediated deletion of the PGK-Neo R cassette resulted in DHSs/DHSs ES cells. Targeted ES cells were injected into Rag2-/blastocysts to obtain RDBC chimeras [11] or into C57BL/6 blastocysts to obtain chimeras that could be crossed to 129Sv mice to achieve germline transmission of the targeted allele.

B cell hybridomas
CD43splenocytes were isolated by MACS, stimulated with LPS (20 mg/ml), and fused to NS-1 plasmacytoma cells (TIB-18, ATCC) as described previously [12]. IgH V(D)J rearrangement status was analyzed by Southern blotting of EcoRI digested genomic DNA of clonal hybridomas with three different probes, a 1.6 kb HindIII -EcoRI fragment 39 of J H 4, a 0.38 kb SacI -ApaI fragment 39 of D H Q52, and a 0.75 kb PCR product 59 of D H FL16.1 generated with oligonucleotides 59 GAACAG-CAACCCTTGACTGACTCTG 39 and 59 GATTGGTTCT-TATGGAATGGGTGG 39.

PCR assay for V(D)J rearrangements
Pro-B cells (IgM -B220 + CD43 hi ), pre B-cells (IgM -B220 + CD43 lo ), and double positive T-cells (B220 -CD4 + CD8 + ) were isolated by FACS on a FACSAria (BD Biosciences) and genomic DNA was extracted. 50 ng DNA or 5-fold dilutions were analyzed by PCR for D H -J H , V H -DJ H , V k -J k , and V l -J l rearrangements with primers listed in Table S1. Input DNA amounts were normalized upon PCR amplification within DLG5. PCR was performed at 95uC for 49, 30 cycles of 95uC for 300, 60uC for 900, and 72uC for 29, followed by 72uC for 59. PCR products were transferred from ethidium bromide gels to nylon membranes and visualized with end labeled oligonucleotide probes (Table S1). CDR 3 lengths were generated from IgH VDH rearrangements from mature B cells using oligonucleotides for V558 and JH4 rearrangements. PCR fragments were amplified using iProof (Bio-Rad) polymerase and cloned into Zero Blunt Topo vectors (Invitrogen), and sequenced.
IgH class switch recombination assay CD43splenocytes were isolated by MACS, cultured with LPS or IL4/aCD40, and analyzed by flow cytometry as described previously [13].

RT-PCR analysis
RNA was extracted using TriPure Isolation Reagent (Roche). 200 ng-1 mg of total RNA was reverse transcribed for one hour at 50uC using random hexamers (Roche) and Superscript III (Invitrogen) reverse transcriptase. PCR was performed at 94uC for 49, 30-39 cycles of 94uC for 300, annealing temperature (Table  S1) for 300, 72uC for 300, followed by 72uC for 59. cDNA input amount was normalized upon PCR amplification of b-actin cDNA. PCR products were visualized on ethidium bromide gels and/or subsequently transferred to nylon membranes and visualized with end labeled oligonucleotide probes (Table S1).

Generation of mice with targeted deletion of 59IgH DNaseI hypersensitive sites
To determine the in vivo function of the cluster of DNaseI hypersensitive sites described at the 59 end of the IgH locus [9] we first replaced a ,340 bp BccI -AatII fragment, harboring HS1, with a loxP flanked PGK-Neo R cassette. All targeting experiments were performed in heterozygous IgH a/b EF1 ES cells which have the advantage that IgH a (129 strain) and IgH b (C57BL/6 strain) alleles can be distinguished by antibodies against the different allotypes or by detection of restriction fragment length polymorphisms (RFLP). Targeting vector homology arms were cloned from 129 strain genomic DNA, resulting in correct targeting events only on the IgH a allele. In heterozygous targeted ES cells, the IgH b allele always remained in the untargeted wildtype configuration.
Targetings were performed with the RHS1 targeting vector (Fig. 1A) to obtain the RHS1 allele and, upon cre/loxP deletion,  Figure S1) were confirmed by Southern blotting. Subsequently, targeted ES cells were injected into Rag2-/blastocysts to obtain Rag-deficient blastocyst complementation (RDBC) chimeras, and into wildtype blastocyts to generate mice that carry the RHS1 or DHS1 allele in their germline. In order to delete all four hypersensitivity sites (HS1, HS2, HS3a, and HS3b), ES cells containing the DHS1 allele were targeted with the R39HSs targeting vector to obtain the R39HSs allele (Fig. 1B). Cre/loxP recombination between the loxP site originating from the DHS1 allele and the loxP site 39 of the PGK-Neo R cassette results in the replacement of a 8.9 kb region, harboring all described 59IgH DNaseI hypersensitive sites, with a single loxP site, referred to as the DHSs allele. Germline transmission could not be achieved for either of the R39HSs or DHSs heterozygous ES cell lines. Therefore, we placed ES cells containing the R39HSs allele under increasing concentrations of G418 to select for homozygous mutant ES cells. The homozygous mutant ES cells were subsequently subjected to cre/loxP recombination to delete the Neo r gene and generate ES cells homozygous for the DHSs allele. The homozygous mutant DHSs ES cells were injected into Rag2 -/blastocysts, and chimeras generated by RDBC and lymphocytes were analyzed.

Development of homozygous RHS1, DHS1, and DHSs lymphocytes
Lymphocytes of different developmental stages can be identified by FACS analysis of cells from lymphoid tissues such as bone marrow, thymus, or spleen. We analyzed 8 week old wildtype mice, homozygous RHS1, and homozygous DHS1 mice that carry the mutant alleles in their germline, as well as lymphocytes from RDBC chimeras generated from homozygous DHSs ES cells (Fig. 2). In wildtype bone marrow, pro-B cells can be identified as IgM -B220 + CD43 hi and pre-B cells as IgM -B220 + CD43 lo cells, respectively. Defects in B-cell development can be revealed by the increase or decrease of certain lymphocyte populations. In this regard, impaired IgH V(D)J recombination leads to an accumulation of pro-B cells and to reduced numbers of pre-B cells [7]. We performed FACS analyses of bone marrow from three mice of each genotype to measure the percentage of pro-and pre-B cells in the lymphocyte gate. These analyses revealed the average percentage (6 standard deviation) of pro-B and pre-B cells, respectively of B220 + /CD43 + events in the total lymphocyte gate were 1462 and 50620 for wildtype, 963 and 5666 for DHS1, and 963 and 42611 for RHS1 mice ( Fig. 2A). Thus, there were no obvious differences in early B-cell development in wildtype and mutant mice. However we cannot exclude minor developmental defects not readily detectable by such analyses. Homozygous mutant DHSs bone marrow cells were analyzed in a similar fashion, but only Ly9.1 + cells were included in the analysis. Ly9.1 is exclusively expressed on cells derived from the DHSs ES cells but not on cells derived from the Rag2-/blastocyst. The presence of a large compartment of blastocyst derived Rag-deficient pro-B cells in the bone marrow can interfere with development of ES cell derived B-lymphocytes. However, FACS analysis of DHSs bone marrow B cells indicated the presence of both pro-and pre-B cells and did not suggest a block in B-cell development ( Fig. 2A).
Next we analyzed spleens for IgM + B220 + AA4.1 + transitional B-cells and IgM + B220 + AA4.1mature B-cells (Fig. 2B). In homozygous RHS1, and homozygous DHS1 mice transitional (19.7%-27.5%) and mature (65.3%-72.7%) B-cell compartments similar to wildtype were identified; whereas, in spleens from RDBC chimeras generated from homozygous DHSs ES cells strongly reduced numbers of transitional B-cells were observed (6.44%). This reduction in the transitional B-cell compartment compared to the mature B-cell compartment (75%) might be due to overall reduced numbers of developing B cells in the obtained RDBC chimeras and to the accumulation of mature B-cells in the periphery of these mice and not to a defect in B cell development. Finally, we observed normal development of T-lymphocytes in the thymi of wildtype, homozygous RHS1, and homozygous DHS1 mice as well as RDBC chimeras generated from homozygous DHSs ES cells (Fig. 2C).
The DHS1, RHS1, and DHSs alleles show no significant defect in V(D)J recombination The data indicating that HS1 is pro-B cell specific and contains binding sites for the transcription factors PU.1, Pax5, and E2A led to the suggestion that HS1 could be involved in regulation of V(D)J recombination at the IgH locus [9]. We utilized a PCR based assay to assess V(D)J recombination efficiencies in developing lymphocytes from mice with homozygous deletion of HS1. FACS-sorted pro-B cells (IgM -B220 + CD43 hi ) and pre-B cells (IgM -B220 + CD43 lo ) from bone marrow and double positive (DP) T-cells (B220 -CD4 + CD8 + ) from thymus were analyzed for D to J H , V H to DJ H , V k to J k and V l to J l rearrangements. Intensities of PCR bands for D H Q52 to J H (Fig. 3A) and DSP to J H rearrangements (Fig. 3B) were comparable in pro-B cells, pre-B cells, and DP T-cells from wildtype, homozygous RHS1, and homozygous DHS1 mice indicating that deletion of the pro-B cell specific HS1 site does not detectably affect the D to J H recombination step. DNA input amounts were normalized to the presence of a genomic sequence within the murine DLG5 gene (Fig. 3H).
It was speculated that HS1 might regulate the differential activation of distal versus proximal V H families [9]; therefore, we analyzed the rearrangement efficiencies of the proximal V H 7183 family (Fig. 3C), the distal V H J558 familiy (Fig. 3D), and the distal most V H segment V H J558.55 (Fig. 3E). We found that pro-B cells and pre-B cells from wildtype, homozygous RHS1, and homozygous DHS1 mice rearrange the proximal V H 7183 family at similar levels (Fig. 3C). Also, the distal family V H J558 (Fig. 3D) as well as the distal most V H segment V H J558.55 (Fig. 3E) rearranged at comparable efficiencies in pro-B cells and pre-B cells from the three different genotypes. V H to DJ H recombination was absent in DP T-cells from wildtype, homozygous RHS1, and homozygous DHS1 mice as the V H to DJ H recombination step is restricted to the B-lineage (Fig 3C, D, E). These data show that HS1 is not Figure 2. Development of homozygous DHS1, RHS1, and DHSs lymphocytes. (A) Bone marrow from wildtype (wt), homozygous DHS1, homozygous RHS1 mice, and RDBC chimeras generated from homozygous DHSs ES cells was subjected to FACS analysis. Gates were set on the lymphocyte population, Ly9.1 positive population, and on IgM negative population (upper three blots, left to right) to analyze pro-B cell (IgM -B220 + CD43 hi ) and pre-B cell (IgM -B220 + CD43 lo ) populations (lower blots). (B) FACS analysis of splenocytes from wildtype (wt), homozygous DHS1, homozygous RHS1 mice, and RDBC chimeras generated from homozygous DHSs ES cells. Gates were set on the lymphocyte population and on the IgM positive population (upper two blots, left to right) to analyze transitional B-cell (IgM + B220 + AA4.1 + ) and mature B-cell (IgM + B220 + AA4.1 -) populations (lower blots). (C) FACS analysis of thymocytes gated on the lymphocyte population (upper blot) from wildtype (wt), homozygous DHS1, homozygous RHS1 mice, and RDBC chimeras generated from homozygous DHSs ES cells (lower blots). doi:10.1371/journal.pone.0013992.g002 necessary for rendering the distal part of the V H cluster accessible and, therefore, suggest that HS1 does not play a major role in regulation of usage or accessibility of distal versus proximal V H families.
Recently, it has been shown that IgH and Igk loci can colocalize during B-cell development, mainly at the pre-B cell stage, and it was suggested that this colocalization induces decontraction of the IgH locus [14]. We therefore performed an assay to evaluate Igk (Fig. 3F) and Igl (Fig. 3G) V(D)J recombination efficiencies. Both Igk and Igl loci show similar V(D)J recombination levels in the analyzed developing B cells from wildtype, homozygous RHS1, and homozygous DHS1 mice, while light chain rearrangements were absent in DP T-cells from the three different genotypes. Therefore, we conclude that deletion of HS1 does not markedly affect Ig light chain gene rearrangements.
As an independent method to evaluate D to J H and V H to DJ H recombination efficiencies, we generated clonal hybridoma lines from splenic B-cells of IgH a/b heterozygous RHS1, DHS1 mice carrying the mutant allele in their germline and of RDBC chimeras generated from heterozygous DHSs ES cells (Table 1). In each case the IgH a allele was the mutant allele while the IgH b allele was the wildtype allele. In splenic B-cells, one allele exists as a functional V H DJ H rearrangement, while the second allele can either be in germline configuration, or it exists as a DJ H or an nonproductive V H DJ H rearrangement. The rearrangement status of the second IgH allele was assessed by Southern blot analysis. Consequently, hybridomas expressing the mutant IgH a allele can be analyzed for rearrangement efficiency of the wildtype IgH b allele, and vice versa, in hybridomas expressing the wildtype IgH b allele, the rearrangement status of the mutant IgH a allele can be assessed.
Wildtype B cells undergo D to J H rearrangements on both alleles; but still, consistent with earlier studies, about 5% of hybridomas harbor an IgH allele in germline configuration which presumably originates from tripartite fusions involving non B-cells [12] (not shown). The number of mutant alleles in germline configuration was not increased compared to wildtype indicating that RHS1, DHS1, and DHSs alleles can undergo efficient D to J H recombination (not shown). In 50-60% of wildtype B-cells the nonproductive allele is in DJ H configuration; whereas in 40-50% the nonproductive allele is in V H DJ H configuration [15]. An increased percentage of DJ H alleles could indicate less efficient V H to DJ H recombination: in contrast, an increased percentage of V H DJ H alleles might indicate a break in allelic exclusion. IgM a expressing hybridomas generated from B-cells heterozygous for RHS1, DHS1, and DHSs were analyzed for their rearrangement status of the wildtype IgM b allele and show ratios of DJ H (56%-61%), and V H DJ H alleles (39%-44%) in the expected range (Table 1). IgM b expressing hybridomas were analyzed for the  rearrangement status of their mutant IgM a allele. RHS1, DHS1, and DHSs alleles do not show significantly increased or decreased (Fisher's exact test) rearrangement ratios compared to wt alleles, as 52%-69% of mutant alleles were in DJ H configuration while 31%-48% were in V H DJ H configuration. FACS analysis was performed on B-cells from spleens (Fig. 4A) and bone marrow (Fig. 4B) of RDBC chimeras generated from heterozygous RHS1, DHS1, and DHSs ES cells. IgM a expressing populations, representing the targeted allele, and IgM b expressing populations, representing the wildtype allele, were of similar size both in bone marrow and in spleen from RHS1, DHS1, and DHSs chimeras, suggesting that the RHS1, DHS1, and DHSs alleles can undergo V(D)J recombination at the IgH locus at similar efficiencies as wildtype alleles.
The DHS1, RHS1, and DHSs alleles do not affect allelic exclusion FACS analysis of wt B cells from spleen (Fig. 4A) and bone marrow (Fig. 4B) shows distinct populations of similar size for B cells that are single positive for either IgH a or IgH b , but intact allelic exclusion prevents the appearance of an obvious IgH a , IgH b double producing population. Similarly, RDBC chimeras generated from heterozygous RHS1, DHS1, and DHSs ES cells exhibited IgH a or IgH b single positive B-cell populations of similar size in spleen (Fig. 4A) and bone marrow (Fig. 4B) but no IgH a , IgH b double producing population. These data indicate that the deleted sequences of the targeted alleles do not contain a regulatory element that is necessary for implementation of allelic exclusion. Furthermore, data from hybridoma analysis (Tab. 1) support this notion as in the case of a break in allelic exclusion increased numbers of hybridomas with V H to DJ H rearrangements on both alleles would be expected. Such an increase compared to wildtype alleles could not be observed (Tab. 1), which indicates intact allelic exclusion of RHS1, DHS1, and DHSs alleles.
The 59IgH DNaseI hypersensitive sites are not required for efficient class switch recombination To assess a potential effect of the 59IgH DNaseI hypersensitive sites on CSR, B-cells were stimulated to undergo CSR and analyzed by FACS (Fig. 5). Stimulation with LPS induces IgH isotype switching to c3, while stimulation with IL4+ aCD40 promotes switching to c1. B-cells from AID-/-mice served as negative controls, while wildtype B-cells represented a positive control and therefore switched to the appropriate isotypes under LPS or IL4+ aCD40 stimulation. CSR in homozygous DHSs Bcells occurs at similar levels as in wildtype B-cells implying that the cluster of 59IgH DNaseI hypersensitive sites is not required for efficient CSR to c1 (Fig. 5A) and c3 (Fig. 5B).

Complex phenotypes without an obvious relation to the IgH locus in DHS1 mice
We performed targeted deletion experiments of the 59IgH DNaseI hypersensitive sites to test their suggested function in IgH locus regulation. So far no major IgH related phenotype was identified. However, about 20% of homozygous DHS1 mice develop a complex neurological phenotype and die at 3-5 weeks of Figure 5. Ig class switch recombination in absence of the 59IgH DNaseI hypersensitive sites. MACS purified splenic B-cells were stimulated in culture with LPS or IL4+ aCD40 as indicated. FACS analysis shows B-cells that underwent CSR as B220 + IgG1 + or B220 + IgG3 + cells, respectively. AID -/-B-cells served as negative controls, wildtype (wt) 129 B-cells as positive controls. Homozygous DHSs B-cells were isolated from RDBC chimeras. doi:10.1371/journal.pone.0013992.g005 age, likely due to a lack of food intake. These mice exhibit an abnormal limp grasping phenotype, i.e. mice clasp their front and hind feet almost immediately upon being lifted by their tail (Fig. 6A, B). Furthermore these mice develop a hydrocephalus, which is already visible at about one week of age and is enlarged over the following weeks (Fig. 6C, D). Histological analysis confirmed the presence of a hydrocephalus, revealed abnormal hindbrain development, and revealed retinal abnormalities (Fig. 6E, F, G). The wildtype retina is organized in a delicate layer system (Fig. 6E): stratum opticum and ganglionic layer (1), inner plexiform layer (2), inner nuclear layer (3), outer plexiform layer (4), outer nuclear layer (5), layer of rods and cones (6), pigment layer (7). In the DHS1 mutant mice, the organization of retinal layers is impaired in such a way that nuclei from the outer nuclear layer are aberrantly located in the layer of rods and cones (Fig. 6F). In some more severe cases rosette formation in the outer nuclear layer is evident (Fig. 6G). Currently, we do not know what causes these phenotypes, but we exclude that this phenotype is caused by a second integration of the targeting vector at an undefined site in the genome ( Figure S1). The deletion in the DHS1 allele deletes 340 bp within intron 1 of Zfp386. Therefore, misregulation of that poorly described gene might cause the described phenotypes although other possibilities are conceivable.

Discussion
This study aimed for elucidating the potential regulatory functions of a cluster of recently described DNaseI hypersensitive sites at the 59 end of the IgH locus [9]. We performed targeted deletion of either the pro-B cell specific site HS1 (DHS1) or deletion of the entire cluster of hypersensitive sites (DHSs) in mice or in their lymphocytes, respectively. A potential regulatory element at the 59end of the IgH locus was speculated to regulate processes such as IgH allelic exclusion, V H germline transcription, differential accessibility or usage of distal versus proximal V H gene families. Furthermore, it was suggested that the 59end of the IgH locus might play a role in positioning the IgH locus in distinct subnuclear compartments [16,17,18], and it was suggested to harbor insulator or boundary capacity [19].
B-and T-lymphocytes homozygous for the DHS1, RHS1, and DHSs alleles appear to proceed through lymphocyte development in an unimpaired way. Data from RDBC chimeras generated from Figure 6. Complex phenotypes of homozygous DHS1 mice. Homozygous DHS1 mice exhibit an abnormal limp grasping phenotype (B) whereas wildtype (wt) mice do not (A). DHS1 mice can develop severe hydrocephalus as indicated by arrows in (C) and (D). A wildtype mouse without hydrocephalus is shown in (C). The wildtype retina is organized in distinct layers (E): Stratum opticum and ganglionic layer (1), inner plexiform layer (2), inner nuclear layer (3), outer plexiform layer (4), outer nuclear layer (5), layer of rods and cones (6), pigment layer (7). The retina of homozygous DHS1 mice shows external nuclei from the outer nuclear layer (5) in the layer of rods and cones (6) indicated by arrows in (F), or rosette formation of the outer nuclear layer (5) indicated by arrows in (G). doi:10.1371/journal.pone.0013992.g006 heterozygous DHS1, RHS1, and DHSs ES cells indicated that allelic exclusion is not affected in mutant B-cells and that mutant IgH alleles can undergo efficient V(D)J recombination of their IgH locus. Furthermore, data from PCR assays to analyze V(D)J recombination efficiency in mice with HS1 deleted on both alleles supports the notion that HS1 is not necessary for neither the D to J H nor the V H to DJ H recombination step. Both proximal and distal V H families as well as the distal most V H segment V H J558.55 rearrange as efficiently as on wildtype alleles. Similarly, IgL loci in HS1 deleted B-cells rearrange at the same efficiency as wildtype IgL loci. Analysis of IgH V(D)J rearrangement status in hybridomas generated from heterozygous DHS1, RHS1, and DHSs B-cells also strengthens the idea that the deleted DNAseI hypersensitive sites would not regulate IgH V(D)J recombination. We tested for potential alterations associated with DNA end processing during V(D)J recombination by examining the CDR3 sequence obtained from homozygous DHS1 B cells and found a distribution in length that was similar to wildtype B cells [20] ( Figure S2).
We tested a potential effect of the cluster of DNaseI hypersensitive site on the process of IgH CSR. Assaying class switching upon different in vitro stimulations in wildtype and homozygous DHSs B-cells let us conclude that the cluster of 59IgH DNaseI hypersensitive sites does not play a crucial role in CSR.
The only observed phenotypes so far occurred in homozygous DHS1 mice and seem to be independent of the IgH locus. DHS1 mice show abnormal limp grasping indicating a neurological abnormality, DHS1 mice can develop severe hydrocephalus and exhibit retinal impairments. A possible explanation for these phenotypes is a potential defect in regulation of the zinc finger protein Zfp386. DHS1 deletes a 340 bp region from intron 1 of Zfp386 which might result in different splice forms, impaired expression levels, or expression patterns of this gene.
Overall, our analysis of the deletion of the pro-B cell specific site HS1 or the whole cluster of 59IgH DNaseI hypersensitive sites did not support the existence of a cis-regulatory function of these elements regarding the IgH locus. Figure S1 Single integration of the RHS1 targeting vector. The targeting vector (targeting vector RHS1), the targeted locus (RHS1), and the wildtype (wt) IgH locus with its 59 flanking region are shown. VH, DH, JH indicate representative IgH V, D, and J segments. Exons 1, 2, and 3 of Zfp386 are shown as grey rectangles, DNaseI hypersensitive sites HS1, HS2, HS3a, and HS3b are shown as black ovals, the NeoR specific Southern probe as a black rectangle. X -XbaI. Southern analysis of XbaI digested genomic DNA from the targeted RHS1 clones 5 (lane 1) and 23 (lane 2) utilizing the NeoR specific probe shows a single 16