Surface Igμ expression in the absence of Igκ rearrangements in Ik^Pax5/+ thymocytes

Sequence analysis of the V_HDJ_H2 PCR fragments revealed that V(D)J recombination in Ik^Pax5/+ pre-T cells involved different members of the V_H7183 and V_HQ52 gene families, although with a strong bias for more proximally located genes within each family (data not shown). Moreover, 13% of the sequenced PCR fragments contained a functional in-frame V_H–DJ_H rearrangement (data not shown). This finding was confirmed by flow cytometric analysis demonstrating that functionally rearranged Igμ chains were expressed not only in the cytoplasm, but also on the cell surface of Thy1.2⁺ Ik^Pax5/+ thymocytes at the expected frequency (Fig. 2A). However, no V_κ–J_κ rearrangements could be detected in Ik^Pax5/+ pre-T cells, demonstrating that Pax5 is unable to induce rearrangements at the Igκ light-chain locus (Fig. 2B). As a consequence, the productively rearranged Igμ protein is likely to be expressed as part of the pre-BCR on the cell surface of Ik^Pax5/+ thymocytes, which is consistent with expression data shown below. Most Ik^Pax5/+ thymocytes expressed TCRβ on the cell surface (Fig. 2A), in agreement with the observation that D_β2–J_β2 and V_β5.1–DJ_β2 rearrangements were present at similar frequency in Ik^Pax5/+ and control Ik^neo/+ pre-T cells (Fig. 2B). These data further demonstrate that the Ik^Pax5/+ thymocytes are of T-lymphoid origin despite ectopic expression of the B-lineage commitment factor Pax5 (Souabni et al. 2002).

Open in a separate window

Figure 2.

Surface Igμ expression in the absence of Igκ rearrangements in Ik^Pax5/+ thymocytes. (A) Flow cytometric analysis. Cytoplasmic (c) and cell surface (s) expression of functionally rearranged Igμ protein and cell surface TCRβ expression were analyzed in wild-type (wt) and Ik^Pax5/+ thymocytes (gated as Thy1.2⁺ cells). The percentage of Igμ⁺ cells is indicated. (B) Normal TCRβ recombination and absence of V_K–J_K rearrangements in Ik^Pax5/+ thymocytes. The same sorted Ik^Pax5/+ and Ik^neo/+ pre-T cells, which were used for the experiment shown in Figure 1, were analyzed by PCR for V_K–J_K rearrangements at the Igκ locus and for D_β2–J_β2 and V_β5.1–DJ_β2 rearrangements at the TCRβ locus. Splenic B220⁺ B cells and stromal ST2 cells served as positive and negative controls, respectively. The numbers next to the B-cell lane indicate rearrangements to the J_K1, J_K2, J_K4, and J_K5 segments. GL denotes the position of the germ-line PCR product.

Conditional Pax5 activation induces V_H–DJ_H rearrangements within the T-lymphoid lineage

The expression of RAG1 and RAG2 is initiated in the earliest lymphocyte progenitor (ELP), resulting in subsequent D_H–J_H recombination (Igarashi et al. 2002). Hence, the V(D)J rearrangement machinery is already active in lymphoid progenitors of the bone marrow prior to B- and T-lineage commitment. On the other hand, expression of the Ik^Pax5 allele is initiated in the hematopoietic stem cell and is maintained in the progenitors and differentiating cells of all major hematopoietic lineages (Souabni et al. 2002). It is therefore conceivable that the Pax5-expressing lymphoid progenitors of Ik^Pax5/+ mice may undergo V_H–DJ_H rearrangements before migration to the thymus and initiation of T-cell development. Alternatively, the V_H–DJ_H rearrangements present in Ik^Pax5/+ thymocytes may originate within the T-lymphoid lineage. To distinguish between these two possibilities, we took advantage of the conditional Ik^neo allele, which contains a floxed neomycin stop cassette upstream of the Pax5 minigene in the Ikaros locus (Souabni et al. 2002). Pax5 expression from this allele is only activated upon Cre recombinase-mediated deletion of the neomycin gene (Souabni et al. 2002). To this end, we crossed the Ik^neo/+ mouse with a transgenic mouse, which expressed the Cre recombinase under the control of the proximal lck promoter (lck-cre) during pro-T-cell development (Lee et al. 2001; Wolfer et al. 2002). The lck-cre transgene was shown to initiate Cre-mediated inactivation of a floxed Notch1 allele in CD44⁺ CD25⁺ (DN2) pro-T cells and to complete gene deletion in CD44^– CD25⁺ (DN3) pro-T cells (Wolfer et al. 2002). However, the same lckcre transgene deleted the floxed neomycin gene of the Ik^neo allele with lower efficiency, as the activated Ik^Pax5 allele was detected in only ~60% of pre-T cells (Fig. 3). Nevertheless, these thymocytes of Ik^neo/+ lck-cre mice were characterized by an increase in D_H–J_H rearrangements and the presence of V_H–DJ_H-rearranged IgH alleles (Fig. 3). The V_H–DJ_H rearrangement frequency was highest for the proximal V_H7183 genes and lowest for the distal V_HJ558 gene family (Fig. 3), thus recapitulating the V(D)J recombination phenotype of Ik^Pax5/+ thymocytes (Fig. 1B). The rearrangement frequency was, however, lower in Ik^neo/+ lck-cre pre-T cells, possibly because of the inefficient activation of the Ik^neo allele during T-cell development. Together these data unequivocally demonstrate that Pax5 is able to induce V_H–DJ_H rearrangements within the T-lymphoid lineage.

Open in a separate window

Figure 3.

V_H–DJ_H rearrangements upon conditional Pax5 activation in thymocytes. Pre-T cells were isolated by FACS sorting as DP thymocytes from Ik^neo/+ mice carrying the lck-cre transgene followed by PCR quantification of D_H–J_H and V_H–DJ_H rearrangements as described in Figure 1B. Ik^neo/+ pre-T cells and stromal ST2 cells served as negative controls and Pax5^+/+ pro-B cells as positive control for the detection of V_H–DJ_H rearrangements. Cre-mediated conversion of the Ik^neo to the Ik^Pax5 allele was determined by PCR (bottom row).

Pax5 activates germ-line V_H transcription and multiple B-lymphoid genes in pro-T cells

To study the extent of B-cell-specific gene activation by Pax5 in thymocytes, we purified pro-T cells from Ik^Pax5/+ mice by depleting Lin⁺ cells (including B220⁺ B cells), followed by sorting for Thy1.2⁺ Lin^– DN thymocytes. Flow cytometric reanalysis demonstrated that the sorted Thy1.2⁺ Lin^– pro-T cells were purified to homogeneity (Fig. 4A). Moreover, RT–PCR analysis failed to detect rearranged Igκ and Igλ1 mRNAs in these sorted cells (Fig. 4B). We conclude, therefore, that the sorted Ik^Pax5/+ pro-T cells were free of contaminating B cells. In addition, we sorted wild-type pro-T and pro-B cells as well as Pax5^–/– pro-B cells. cDNA was prepared from all four cell types and normalized for equal expression of the control HPRT gene prior to semiquantitative RT–PCR analysis of B-cell-specific transcripts (Fig. 4C).

Open in a separate window

Figure 4.

Activation of germ-line V_H transcription and B-cell-specific gene expression in Ik^Pax5/+ pro-T cells. (A) Flow cytometric reanalysis of sorted Ik^Pax5/+ DN thymocytes. DN pro-T cells of 2-week-old mice were sorted as Thy1.2⁺ Lin^– cells after depletion of Lin⁺ cells, which were stained with antibodies recognizing B220, CD4, CD8, DX5, CD11c, Mac-1, Gr-1, and Ter119. (B) Absence of B cells in the sorted Ik^Pax5/+ pro-T cell population. Rearranged Igκ (V_K–JC_K) and Igλ1 (V_λ–JC_λ1) mRNAs could not be detected by RT–PCR in purified Ik^Pax5/+ pro-T cells in contrast to IgM⁺ IgD⁺ B cells isolated from wild-type (wt) spleen. The hypoxanthine phosphoribosyltransferase (HPRT) gene was equally expressed in both cell types. (C) B-cell-specific gene expression in Ik^Pax5/+ pro-T cells. Transcripts of the indicated genes were analyzed by semiquantitative RT–PCR of five-fold serial dilutions of cDNA that was prepared from ex vivo sorted pro-B cells and pro-T cells of the indicated genotypes. The cDNA input was normalized according to the expression of the control HPRT gene. V_HJ558-DJCμ and V_H7183–DJCμ refer to rearranged Igμ mRNAs and GLT to the corresponding germline transcripts. Transgenic human (h) and endogenous mouse (m) Pax5 transcripts were amplified with conserved primers, giving rise to the same PCR fragment (denoted by an asterisk) for both transcripts.

The accessibility of a particular V_H gene in active chromatin can be monitored by expression of its germline transcript (GLT), whereas the abundance of spliced Igμ (V_H–DJCμ) mRNA is a direct measure of the V_H–DJ_H recombination frequency in progenitor cells (Yancopoulos and Alt 1985). The germ-line transcripts of the proximal V_H7183 gene family were expressed at a fivefold lower level in Pax5^–/– pro-B cells and Ik^Pax5/+ pro-T cells compared with wild-type pro-B cells. The rearranged mRNAs of the V_H7183 genes were, however, present at similar abundance in all three cell types (Fig. 4C), consistent with efficient V_H7183–DJ_H recombination in these cells (Fig. 1B; Hesslein et al. 2003). In contrast, the rearranged mRNA of the distal V_HJ558 gene family was reduced ~100-fold in Ik^Pax5/+ pro-T cells and Pax5^–/– pro-B cells (Fig. 4C), which undergo V_HJ558–DJ_H rearrangements with a similarly low efficiency compared with wild-type pro-B cells (Fig. 1B; Nutt et al. 1997). Germ-line V_HJ558 transcripts were, however, present at a similar abundance in Ik^Pax5/+ pro-T cells relative to wild-type and Pax5^–/– pro-B cells (Fig. 4C). Germ-line transcripts of the TCR V_β5 gene could not be detected in pro-B cells, but were expressed in Ik^Pax5/+ pro-T cells, further confirming the T-lymphoid origin of these cells (Fig. 4C). These data indicate, therefore, that ectopic expression of Pax5 establishes an accessible chromatin state at both the proximal and distal V_H genes in Ik^Pax5/+ pro-T cells, although only the proximal V_H genes undergo efficient V_H–DJ_H recombination.

Ectopic Pax5 expression activated the Pax5 target genes CD19 (Nutt et al. 1998), BLNK (Schebesta et al. 2002b), and Igα (mb-1; Nutt et al. 1998) in Ik^Pax5/+ pro-T cells (Fig. 4C). More surprisingly, these pro-T cells also expressed the genes Igβ (B29), λ5, and VpreB (Fig. 4C), which are known to be cooperatively regulated by the transcription factors EBF and E2A (Sigvardsson et al. 1997, 2002; O'Riordan and Grosschedl 1999). Consistent with this finding, Pax5 induced expression of the B-cell-specific EBF gene in Ik^Pax5/+ pro-T cells to a level that is normally seen in wild-type and Pax5^–/– pro-B cells. In contrast, the expression of E2A was similar and independent of Pax5 in pro-T cells as in pro-B cells (Fig. 4C). None of the B-cell-specific transcripts analyzed could be detected in wild-type pro-T cells except for low-level expression of the Igβ gene (Wang et al. 1998). Moreover, the efficient expression of all pre-BCR components (λ5, VpreB, Igα, and Igβ) strongly suggests that the rearranged Igμ protein is transported as part of the pre-BCR to the cell surface of Ik^Pax5/+ thymocytes (Fig. 2A).

The Pax5 gene gives rise to two distinct mRNAs by alternative promoter usage and splicing of the different 5′-exons onto common coding sequences (Busslinger et al. 1996). The presence of the mouse Pax5A and Pax5B transcripts in Ik^Pax5/+ pro-T cells demonstrated that the ectopically expressed human Pax5 protein is able to induce transcription of the endogenous Pax5 gene in thymocytes (Fig. 4C). This Pax5 activation is likely to be an indirect effect of EBF expression, which is known to regulate the Pax5 gene in pro-B cells (O'Riordan and Grosschedl 1999). PCR conditions, which detect human and mouse Pax5 mRNAs with equal efficiency, revealed that the human Pax5 minigene of the Ik^Pax5 allele is expressed at a fivefold higher level in pro-T cells compared with the endogenous Pax5 gene in wild-type pro-B cells (Fig. 4C). In conclusion, moderate overexpression of Pax5 activates its own gene as well as many other B-cell-specific genes in Ik^Pax5/+ pro-T cells.

Relocation of the IgH locus from the nuclear periphery to central positions in Ik^Pax5/+ thymocytes

A recently published FISH analysis indicated that the subnuclear location and chromatin state of the IgH locus may regulate V(D)J recombination in non-B versus pro-B cells (Kosak et al. 2002). In particular, the 3-Mb-long IgH locus with its 150–200 V_H genes was shown to be present in an extended chromatin state at the nuclear periphery of thymocytes (Kosak et al. 2002). To further investigate the discrepancy between distal and proximal V_H gene recombination in Ik^Pax5/+ thymocytes, we next used three-color 3D FISH analysis (Skok et al. 2001) to localize different IgH gene segments in three-dimensionally (3D) preserved nuclei by confocal laser scanning microscopy (Fig. 5). The locations of the different V_H and Cγ1 gene probes used are shown in Figure 5A together with the minimal number of base pairs separating these probes in the IgH locus.

Open in a separate window

Figure 5.

Subnuclear compartmentalization and extended chromatin state of V_H genes in thymocyte nuclei. (A) Schematic diagram indicating the positions of the DNA probes in the IgH locus (not drawn to scale). The minimal number of megabase pairs (Mb) separating the probes was estimated according to the mouse genome database (Ensembl release of July 1, 2003). (B) Three-color 3D DNA-FISH analysis of the IgH locus. Confocal laser scanning microscopy identified the subnuclear locations of different gene segments of the two IgH alleles, which are shown on representative single optical sections through the nucleus of sorted wild-type (wt) and Ik^Pax5/+ pro-T (DN) cells. The different probes with their colors are indicated. The V_HJ558 BAC (526A21) was directly labeled with Cy5-dUTP (blue). The V_H7183 (167C1) BACs as well as V_H15 (243G9) BAC (not shown in panel B) were labeled with digoxygenin-dUTP and detected with rhodamine- and Texas red-coupled anti-digoxygenin antibodies (red). The Cγ1 probe (HE17) was labeled with biotin-dUTP and detected with FITC-avidin and biotinylated FITC-coupled anti-avidin antibodies (green). The IRES-GFP sequences, which were flanked by frt sites in the Ik^Pax5 allele (Souabni et al. 2002), were specifically deleted for the three-color FISH analysis by mating Ik^Pax5/+ mice with the transgenic ACTB:FLPe deleter line (Rodriguez et al. 2000). The contour of the cell is indicated by a broken line. (C) Statistical analysis of the distance between proximal and distal V_H genes in thymocyte nuclei of the indicated genotypes. Signals of the different IgH gene probes, which were separated in the nucleus by 0.3–0.5 μm(separate; dark color) or 0.5–1.5 μm(far apart; light color), were scored as percentage of all signals analyzed. The actual numbers and sample sizes are shown in Supplementary Table 1A. (D) Statistical analysis of the peripheral location of distal V_H genes in thymocytes and pro-B cells of the indicated genotypes.

As illustrated by the confocal images of Figure 5B, the distal V_HJ558 and V_H15 genes were colocalized at the nuclear periphery in 92%–95% of all wild-type pro-T and pre-T cells (Fig. 5D), in agreement with published data (Kosak et al. 2002). The signals of the distal V_H genes were separated in only 2% of all wild-type thymocytes (Fig. 5C). In marked contrast, the distal V_H genes were separated from the proximal V_H7183 and Cγ1 genes by a distance of 0.3–1.5 μmin 89%–98% of the wild-type nuclei (Fig. 5C). In addition, the proximal IgH domain was positioned away from the nuclear periphery toward the center in 96% of wild-type pro-T and pre-T cells (Fig. 5B; data not shown). Hence, the IgH locus is anchored via the distal V_H genes at the nuclear periphery and is oriented in its extended chromatin state toward the center, which may facilitate access of the V(D)J recombinase to the proximal IgH domain, thus accounting for the D_H–J_H rearrangements observed in wild-type thymocytes.

Ectopic Pax5 expression resulted in subnuclear repositioning of the IgH locus in pro-T and pre-T cells, as the majority (62%) of IgH loci were located at central positions in the nuclei of Ik^Pax5/+ thymocytes (Fig. 5B,D). A similar percentage of centrally located IgH alleles was observed in RAG2^–/– and Pax5^–/– RAG2^–/– pro-B cells (Figs. ​(Figs.5D,5D, ​,6A),6A), indicating that the relocation of the IgH locus is as efficient in Ik^Pax5/+ thymocytes as in pro-B cells. Importantly, the IgH locus remained in an extended chromatin state even in its more central location, as the distal V_HJ558 and V_H15 genes were still separated from the proximal V_H7183 and Cγ1 genes in 92%–100% of all Ik^Pax5/+ pro-T and pre-T cells (Fig. 5C). The physical separation of distal V_H genes from the rearranged DJ_H segment is therefore a likely cause for the low V_H–DJ_H recombination efficiency of distal V_H genes in Ik^Pax5/+ thymocytes.

Open in a separate window

Figure 6.

Absence of IgH locus contraction in the nucleus of Pax5^–/– pro-B cells. (A) Three-color 3D DNA-FISH analysis of the IgH locus. In vitro cultured RAG2^–/– and Pax5^–/– RAG2^–/– pro-B cells were analyzed by FISH and confocal microscopy as described in the legend for Figure 5B. (B) Statistical analysis of the distance between proximal and distal V_H genes in pro-B cells of the indicated genotypes. A distance of 0.3–0.5 μm(separate) or 0.5–1.5 μm(far apart) between the signals is indicated by dark or light color, respectively. See Supplementary Table 1B for further information.

The IgH locus is present in an extended chromatin state in Pax5^–/– pro-B cells

As the Pax5^–/– pro-B cells and Ik^Pax5/+ thymocytes exhibit a similar V_H–DJ_H recombination phenotype, we next studied the nuclear location and chromatin state of IgH loci in wild-type and Pax5^–/– pro-B cells. The IgH alleles were centrally located in ~70% of the pro-B cell nuclei regardless of the presence or absence of Pax5 (Figs. ​(Figs.5D,5D, ​,6A).6A). Interestingly, the distal V_HJ558 and V_H15 genes were separated from the proximal V_H7183 and Cγ1 genes in 84%–94% of all Pax5^–/– pro-B cells (Fig. 6A,B), similar to the situation observed in Ik^Pax5/+ pro-T cells (Fig. 5B,C). In contrast, the distal and proximal IgH gene segments were colocalized in 76%–86% of wild-type pro-B cells, whereas they were separated by only a short distance (0.3–0.5 μm) in the remaining 14%–24% of pro-B cells (Fig. 6B). These data therefore point to an essential role for Pax5 in regulating large-scale contraction of the IgH locus. To rule out the possibility that this Pax5 function depends on V(D)J recombination, we analyzed Pax5^+/+ and Pax5^–/– pro-B cells on a RAG2 mutant background, which prevents immunoglobulin gene rearrangements (Shinkai et al. 1992). The IgH locus was still in an extended state in Pax5^–/– RAG2^–/– pro-B cells and in a contracted state in Pax5^+/+ RAG2^–/– pro-B cells (Fig. 6A,B), indicating that Pax5-mediated contraction of the IgH locus can precede V(D)J recombination in pro-B cells.

FACS sorting and analysis

The following phycoerythrin (PE)- or allophycocyanin (APC)-coupled antibodies were used for flow cytometry: anti-B220 (RA3-6B2), CD4 (L3T4), CD8 (53-6.7), CD11c (HL3), CD19 (1D3), DX5 (DX5), Gr-1 (RB6-8C5), c-Kit (2B8), IgM (M41.42), Mac-1 (M1/70), TCRβ (H57-597), Ter119 (TER-119), and Thy1.2 (30-H12) antibodies. Unspecific antibody binding was suppressed by preincubation of cells with CD16/CD32 Fc-block solution (PharMingen). Intracellular staining of cytoplasmic Igμ protein was performed as described (Thévenin et al. 1998). Pro-B cells were sorted as B220⁺ c-Kit⁺ cells after enrichment of c-Kit⁺ bone marrow cells by magnetic cell sorting (MACS). DP thymocytes were sorted as CD4⁺ CD8⁺ cells after depletion of B220⁺ cells. DN thymocytes were sorted as Thy1.2⁺ Lin^– cells following depletion of Lin⁺ cells (stained with PE-anti-B220, CD4, CD8, DX5, CD11c, Mac-1, Gr-1, and Ter119 antibodies) with anti-PE MACS beads (Miltenyi Biotec).

V(D)J recombination analysis

Sorted cells from 2-week-old mice were digested with proteinase K, and DNA was isolated by phenol extraction and ethanol precipitation. PCR analyses of immunoglobulin genes were performed with published primers (Supplementary Table 2) as described (Schlissel et al. 1991; Angelin-Duclos and Calame 1998). TCRβ rearrangements were analyzed as described (Wolfer et al. 2002). PCR cycle numbers were adjusted to be in the linear range, based on the analysis of serially diluted DNA. PCR products were separated on agarose gels, transferred to a porablot NYamp membrane and analyzed by Southern blotting using published oligonucleotide probes (Supplementary Table 2; Schlissel et al. 1991).

RT–PCR analysis

RNA was prepared from sorted cells, using the Trizol Reagent (GIBCO-BRL). Reverse transcription (with randomhexamers) and semiquantitative PCR were performed as described (Horcher et al. 2001), using the primers shown in Supplementary Table 3. PCR products were separated on agarose gels and visualized by ethidiumbromide.

3D DNA-FISH and confocal analysis

Three-color 3D FISH experiments were carried out with sorted pro-T and pre-T cells as well as with in vitro cultured pro-B cells as previously described in detail (Skok et al. 2001). The fixation conditions used were designed to preserve nuclear integrity (Skok et al. 2001). Cells were analyzed by confocal microscopy on a Leica SP2 AOBS (Acousto Optical BeamSplitter) system. Optical Z-sections were collected at 0.3-μmsteps through individual nuclei. Only cells containing signals of both IgH loci were evaluated. DNA probes were prepared from the BACs 526A21 (V_HJ558), 243G9 (V_H10), and 167C1 (V_H7183; Kosak et al. 2002), as well as from the plasmid pγ1/HE17 containing a 17-kb genomic insert of the Cγ1 region (Skok et al. 2001). The distance between the signals of the different IgH gene probes in the nucleus was measured on individual confocal images. A distance of 0.3–0.5 μmwas evaluated as “separate,” whereas “far apart” referred to a distance of 0.5–1.5 μm(Supplementary Table 1; Figs. ​Figs.5,5, ​,6,6, ​,77).

We thank C. Wilson for providing the lck-cre transgenic mouse, D. Hesslein for advice on rearrangement PCR assays, and G. Stengl for FACS sorting. This research was supported by Boehringer Ingelheim, the Austrian GEN-AU initiative (financed by BMBWK), and a Wellcome Trust University Award (J.S.).

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC section 1734 solely to indicate this fact.