E2A^ER/+ and E2A^ER/ER mice were generated by using a knock-in strategy for tamoxifen-inducible E2A function. The tamoxifen-responsive ligand binding domain of the mouse ER 20 was inserted at the carboxyl end of tcfe2a to produce the E2A^ER allele (Figure 1A, B). With this targeting strategy, similar to that used for the E2A^GFP strain previously developed in our lab 21, both alternatively spliced products of the tcfe2a gene, E12 and E47, are translated as ER fusion proteins. Initial characterization of the E2A^ER allele indicated normal expression levels of E2A mRNA in the presence of the ER insertion (Figure 1C). Previous study of E2A knockout mice has demonstrated stunted growth and a high lethality rate of homozygous animals within the first few weeks after birth 1819. In contrast, the E2A^E47bm strain, expressing a dominant negative form of E47, was originally described as indistinguishable from wild-type litter mates in size and survival 22. However, this work was analyzing mice on a mixed genetic background. Upon backcrossing to C57BL/6, the E2A^E47bm/E47bm mice became smaller in size and demonstrated the high lethality rate like that shown with the knockout animals (unpublished data). E2A^ER/ER animals also exhibit stunted growth and reduced survival (see Additional file 1). Fortunately, the lethality rate in our experience has been less severe in litters from the E2A^ER strain than that observed with our E2A knockout and dominant negative strains. However, we do not know if this slight increase in postnatal survival is due to the presence of the E2AER protein or because our E2A^ER strain is currently on a mixed background.

B cells develop from hematopoietic stem cells (HSC) in the bone marrow through a series of developmental stages 23. The pre-proB cell stage is an intermediate stage as lymphoid progenitors develop into committed proB cells. Pre-proB cells can be characterized by the expression of B220 and CD43 and the absence of CD19 expression. As pre-proB cells transition to the proB cell stage, CD19 expression is induced and cells undergo commitment to the B cell lineage. E2A is critical for this transition, as demonstrated by the block in development at the pre-proB cell stage in E2A-deficient animals 181924. Analysis of E2A^GFP mice displays the up-regulation of E2A protein levels from the pre-proB to proB stage (see Additional file 2) 21. This increase in E2A expression is likely critical for E2A's regulation of the B cell lineage gene expression profile given the importance of E2A gene dosage. For example, elimination of one copy of E2A has been shown to greatly reduce the number of proB cells 418.

Investigation of B cell development in E2A^ER/ER mice revealed a block at the pre-proB cell stage, similar to that seen in the E2A-mutant E2A^E47bm/E47bm mice (Figure 2A) 22. Occasionally we have observed a small population of CD19⁺ B cells in the bone marrow of E2A^ER/ER mice, but these incidences of leaky B cell development did not appear to produce a significant population of mature, Ig expressing B cells (Figure 2B). Even though E2A has been suggested to influence proliferation in developing B cells 20252627, no significant difference in expansion at the pre-proB cell stage was observed by BrdU analysis of E2A^ER/ER and wild-type mice (see Additional file 3).

For analysis of E2AER protein expression and DNA binding activity, we chose to derive Abelson transformed preB cells from E2A^ER/+ and E2A^ER/ER bone marrow 28. E2AER protein expression was verified by Western Blot analysis of whole cell lysates from E2A^ER/+ and E2A^ER/ER Abelson cells (Figure 3A). While E2AER protein levels appeared similar to wild-type levels when comparing E2A^+/+ and E2A^ER/ER cells, analysis of E2A^ER/+ cells suggested that E2AER protein expression may be lower than wild-type. Protein analysis of tamoxifen treated cells displayed similar results, indicating that the relative E2AER protein levels are not affected by the presence or absence of tamoxifen. Nuclear extracts from E2A^ER/ER Abelson cells were then used to conduct electrophoretic gel shift analysis of E2AER binding to an E2A binding sequence, μE5. In the absence of tamoxifen treatment, no DNA binding activity was observed from the E2AER protein (Figure 3B). Together, the block in B cell development and lack of E2AER DNA binding activity suggest that the E2A^ER mouse model functions as an E2A-deficient system in the absence of ligand.

Tamoxifen treatment of E2A^ER/ER Abelson cells resulted in rapid E2AER DNA binding activity within 1 hr of treatment (Figure 3B). The specificity of the protein binding to the μE5 probe was verified by using an anti-E2A antibody that effectively super-shifted the protein/DNA complex. The effect of tamoxifen withdrawal was then tested by washing tamoxifen-treated cells and growing them in the absence of tamoxifen for 1 and 6 hr time points. Loss of E2AER DNA binding activity was seen within 6 hrs of tamoxifen withdrawal (Figure 3C), indicating relatively fast reversibility of E2A function. These results demonstrate that the E2A^ER model can be used not only as an inducible model, but this system may also be valuable for providing a tightly regulated window of E2A activity.

We first tested for a functional outcome of E2A induction by in vivo tamoxifen treatment of E2A^ER/ER mice followed by analysis of B cell populations in the bone marrow and spleen. All in vivo treatment efforts, including intraperitoneal injection and treatment in drinking water, unfortunately resulted in low efficiency rescue of B cell development. The emergence of B cells in tamoxifen-treated animals was rarely great enough to determine if the resulting B cells were generated in response to the tamoxifen treatment or were simply the incidence of leaky B cell development described above. The most significant recovery of B cells observed from in vivo treatment, which resulted from a 27 day tamoxifen treatment, was still considerably less than the B cell population in control mice (Figure 4A, B). However, the presence of Ig μ heavy chain (IgM) positive B cells in the spleen does suggest that a low level of tamoxifen-dependent B cell development occurred in vivo.

In contrast to in vivo tamoxifen treatment, treatment in an ex vivo B cell culture system effectively rescued the development of CD19⁺ B cells. Sorted E2A^ER/ER pre-proB cells were cultured in hormone-free media on an S17 stromal layer in the presence of IL-7, with or without tamoxifen. Control DMSO treated E2A^ER/ER pre-proB cells failed to develop efficiently into CD19⁺ B cells over the course of 5 days, whereas tamoxifen treated E2A^ER/ER pre-proB cells effectively gave rise to CD19⁺ B cells (Figure 5A). Interestingly, the kinetics of B cell development from tamoxifen treated E2A^ER/ER pre-proB cells appeared delayed compared to that of control cells. In addition to using CD19 expression to validate the rescue of B cell development, we analyzed Pax5 expression throughout the 5 day culture. Pax5, initiated downstream of E2A expression, is a transcription factor critical for B cell lineage commitment 29. Consistent with CD19 expression, induction of Pax5 expression was observed in tamoxifen treated E2A^ER/ER pre-proB cells, also appearing delayed compared to control cultures (Figure 5B).

Another key event during early B cell development, downstream of E2A function, is rearrangement of the Ig heavy chain locus (IgH) 21819. We next used this culture system to determine if induction of E2AER activity could also support IgH V to DJ recombination. To amplify V to DJ rearrangements, we chose primers recognizing members of the V_H1 gene family 30, which represent a large percentage of the total IgH V genes, and primers recognizing the J_H4 gene segment. Rearrangements were analyzed in Day 8 cultured samples because we did not observe consistent V to DJ PCR signals until this point in the time course analysis. Wild-type mice were used as positive controls for detecting recombination events, and mice deficient in the recombination activating gene RAG1, required for V(D)J recombination, were used as negative controls. Analysis of Day 8 cultures demonstrated a clear V to DJ rearrangement product from tamoxifen treated E2A^ER/ER cells (Figure 6). V to DJ rearrangement was further verified by sequencing analysis (see Additional file 4). The faint product detected from DMSO control treated E2A^ER/ER cells was accompanied by a small "leaky" CD19⁺ population generated at this relatively late time point in the culture system (see Additional file 5). However, only one unique rearrangement product was identified out of all of the colonies sequenced for this sample. In addition to providing functional proof of induced E2A activity, the detection of IgH recombination, along with the induction of CD19 and Pax5 expression described above, suggests that restored E2A function can rescue B cell development from the pre-proB cell stage.

E2A^E47bm mice have been described previously 22. Generation of the E2A^ER allele is described below. All research with mice was performed in accordance with relevant guidelines, and protocols were approved by the Duke University Animal Care and Use Committee.

The gene targeting strategy used was a modification of the strategy for generation of the E2A^GFP strain 21. The tamoxifen-responsive region of the mouse estrogen receptor ligand binding domain containing the G525R mutation 44 was PCR amplified from the MigR1-E47R vector 20 using the primers ERfpA: 5'-CGGATCCACGAAATGAAATGGGTGC-3' and ERrpA: 5'-CCGGCCGCTAGAATTCGATCGTGTTGGGGAAGCCCTC-3' to introduce a 5' BamHI site and 3' EcoRI and EagI sites for subsequent cloning steps. The ER fragment was inserted, replacing EGFP, at the BamHI position in frame with E2A. The targeting construct also contained a positive selection marker, PGKNeo cassette, and a negative selection marker, PGK driven thymidine kinase (TK) cassette. Mouse ES cells used were derived from a 129/sv strain obtained from Phillippe Soriano's lab in 1995 and then maintained in our own lab. E2A^ER/+ and E2A^ER/ER mice were maintained on a C57BL6 and 129/sv mixed background. Three primers were used for detection of wild-type and mutant alleles, yz164: 5'-AAGAACGAGGCCTTCCGTGTC-3', yz29: 5'-TCGCAGCGCATCGCCTTCTA-3', and bjE2Ar3: 5'-CAAGAGACTAGGATGCCACTG-3'.

RNA extraction, DNase I treatment and reverse transcription have been described previously 41. Quantitative real-time PCR analysis for Pax5 expression was performed using a Roche LightCycler and Fast-Start DNA master SYBR green kit I (Roche) as per manufacturer's instructions. The following primers were used, E2A: f1 5'-CCAGTCTCAGAGAATGGCAC-3' and r1 5'-CCTTCGCTGTATGTCCGGCTAG-3'; Pax5 and GAPDH primers 41.

For sorting, bone marrow was harvested and pooled from 2-3 mice per genotype. Cells positive for lineage markers Mac-1, Gr-1, Ter-119, and CD3 were depleted with Dynal Dynabeads (Invitrogen) according to manufacturer's instructions. Dead cells stained with 7-aminoactinomycin D (7AAD, Molecular Probes) were excluded. Pre-proB cells were further distinguished as B220⁺CD43⁺CD19^-. FACS analysis was done with a FACSCalibur (BD Biosciences) or FACSVantage SE with DiVa option (BD Biosciences) and FlowJo software (Tree Star). FACSVantage SE with DiVa option was used for cell sorting.

Tamoxifen (Sigma) was prepared as a 1 mM stock (1000×) dissolved in cell culture grade dimethyl sulfoxide (DMSO) and stored at -20°C.

The E2A^ER/ER Abelson preB cell line was derived by Abelson Murine Leukemia Virus transformation of bone marrow cells from an E2A^ER/ER mouse. Briefly, whole bone marrow was plated on an S17 stromal layer in the presence of 1 uM tamoxifen and 10 ng/mL IL-7 in 5% FBS RMPI media. This culture was performed prior to transduction to ensure cells were proliferating and were at the optimal target stage for Abelson transformation. Once an expanding B cell population was observed, cells were infected with Abelson virus in the presence of 4 ug/mL polybrene. Abelson transformed cells were then removed from the stromal layer, and tamoxifen and IL-7 were withdrawn. The established E2A^ER/ER Abelson preB cell line was maintained in 10% FBS RPMI media (also containing 100 units/ml penicillin, 100 ug/mL streptomycin and 55 uM 2-mercaptoethanol) prior to experimental analysis. Additional Abelson lines were established as described previously 32.

Abelson preB cell lines were cultured without tamoxifen or with 1 uM tamoxifen for 24 hr prior to analysis. Cells were lysed in RIPA lysis buffer (1% Triton, 0.5% sodium deoxycholic acid, 0.1% SDS, 25 mM Tris-Cl pH 7.6, 150 mM NaCl, 5 mM EDTA) with protease inhibitors. Whole cell lysates were resolved by SDS-PAGE and blotted with anti-E2A (G127-32, BD Biosciences, 554077) and anti-ERK2 (C-14, Santa Cruz Biotechnology, sc-154) antibodies.

E2A^ER/ER Abelson preB cells were cultured with or without 1 uM tamoxifen as indicated. For withdrawal analysis, tamoxifen-treated cells were washed and re-plated in the absence of tamoxifen for the indicated times. Nuclear extracts were incubated with a ³²P-labeled μE5 oligonucleotide probe, with or without Yae anti-E2A monoclonal antibody (Santa Cruz Biotechnology, sc-416), and resolved on a 5% polyacrylamide gel. Gels were dried and exposed to a phosphor screen for phosphorimager analysis (Amersham Biosciences). Oligos used for μE5 probe: 5'-TCGAAGAACACCTGCAGCAGCT-3' and 5'-TAGAGCTGCTGCAGGTGTTCTT-3'.

Mice were treated with tamoxifen in the drinking water for 27 days. A 68 mg/mL tamoxifen in ethanol stock was used to bring the concentration in drinking water to approximately 26 ug/mL, resulting in 0.04% ethanol in water. A fresh bottle of tamoxifen water was given every 5 days.

Sorted pre-proB cells were plated on an S17 stromal layer in 24-well plates at approximately 1.5 × 10⁴ cells per well and cultured with 5% FBS RPMI hormone-free media containing 10 ng/mL IL-7. Hormone free media consisted of phenol-red free RPMI 1640 supplemented with 5% charcoal/dextran treated FBS (Hyclone), 100 units/mL penicillin, 100 ug/mL streptomycin, and 55 uM 2-mercaptoethanol. Treated wells contained 1 uM tamoxifen and untreated controls were given DMSO alone (0.1%). Cells received fresh media, cytokine, and tamoxifen or DMSO every other day. Cells were harvested at time points indicated, and samples were split in half for FACS analysis and RNA or DNA extraction.

Sorted pre-proB cells were ex vivo cultured as described above. DNA was extracted from Day 8 cultures and then analyzed for IgH V to DJ rearrangements using V_H1 gene family and J_H4 specific primers. A nested PCR strategy was used to amplify rearrangements involving V_H1 family gene segments. The following primers were used for Round 1: V_H1 ext 5'-AGRTYCAGCTGCARCAGTCT-3' 30 and J_H4 YZB6 5'-TCCCTCAAATGAGCCTCCAAAGTCC-3' 45 and for Round 2: V_H1 int 5'-GARGATRTCCTGYAAGGCTTC -3' 30 and J_H4 YZB5 5'-CCTGAGGAGACGGTGACTGAGGTTCCTTG-3'46. CD14 primers were used to demonstrate the presence of DNA in all samples 41. V_H1-D_HJ_H4 PCR products were cloned into a pCR4 TOPO vector (Invitrogen) and sequenced (Direct Sequencing from Colonies Service, Eton Bioscience, Inc). Rearrangement product sequences were analyzed by SoDA 47.