CD40 Stimulation Induces Pax5/BSAP and EBF Activation through a APE/Ref-1-dependent Redox Mechanism*

CD40 is a member of the growing tumor necrosis factor receptor family that has been shown to play important roles in T cell-mediated B lymphocyte activation. Ligation of B cell CD40 by CD154, mainly expressed on activated T cells, stimulates B cell proliferation, differentiation, isotype switching, up-regulation of surface molecules contributing to antigen presentation, development of the germinal center, and the humoral memory response. In this study we demonstrate that the redox factor APE/Ref-1 acts as a key signaling intermediate in response to CD40-mediated B cell activation. The transcription factors Pax5a or BSAP (B cell lineage-specific activator protein) and EBF (early B cell factor) are constitutively expressed in spleen B cells and CD40 cross-linking induces increases in Pax5a and EBF binding activity compared with nonstimulated B cells. We show that upon CD40 antibody-mediated cross-linking, APE/Ref-1 translocates from the cytoplasm to the nucleus of activated B cells, where it modulates the DNA binding activity of both Pax5a and EBF. Moreover, we show that the repression of APE/Ref-1 protein production is able to block CD40-mediated Pax5a activation. We also provide evidence that APE/Ref-1 can modulate the cooperative activation of the blk promoter operated by Pax5a and EBF and that APE/Ref-1 might directly regulate EBF functional activity. Finally, we show that the interaction between Pax5a and EBF enhances EBF binding activity to its consensus sequence, suggesting that Pax5a can physically interact with EBF and modulate its DNA binding activity.

CD40 is a surface receptor expressed on B cells and certain accessory cells that belong to the pleiotropic growing tumor necrosis factor receptor superfamily. The interaction between CD40 and its ligand (CD154), which is mainly expressed on activated CD4ϩ T cells, is critical in the regulation of immune response. Engagement of CD40 on B lymphocytes promotes proliferation, cytokine production, up-regulation of various surface molecules involved in antigen presentation, formation of germinal center, memory B cell, and antibody isotype switching (1,2). CD40 activates the c-Jun NH 2 -terminal kinase-stress-activated protein kinase, NF-B (3,4), p38 kinase (5), and extracellular signal-regulated kinase-mitogen-activated protein kinase (MAPK) pathways (6). In addition to these serinethreonine kinases, a link between CD40 signaling and phosphatidylinositol 3-kinase has been suggested (7)(8)(9). In each of the above pathways, one of the major final outcomes is an alteration in the activity of one or more transcription factors.
Pax5a or B cell lineage-specific activator protein (BSAP) 1 and EBF (early B cell factor) are two transcription factors essential for B cell development. Both are expressed at all stages of B cell development except in the plasma cell and are involved in the transcription control of several B cell restricted genes.
Pax5a, the product of the Pax5 gene, is a member of a multigene family of transcription factors that share the paired box DNA binding domain and are important regulators of early development (10). The Pax5 gene is alternatively spliced during B cell development: the predominant form Pax5a (full-length Pax5), Pax5b, Pax5d, and Pax5e (11). Only Pax5a and Pax5d possess an intact DNA-binding domain, enabling them to interact with and compete for Pax5-binding sites on DNA. Pax5d and Pax5e do not have transactivation, repression, or partial homeodomain homology regions at the C terminus but present a 42-amino acid novel sequence with unknown function (11). Pax5a is essential for the development of B cells; Pax5 gene knockout mice have no mature B cells and serum immunoglobulins, and B cell development in such mice is arrested at the pro-B stage (12)(13)(14). Pax5a binding sites have been identified in the regulatory sequences of a number of genes, including mb-1, CD19 (15), the k-light chain (16 -18), VpreB1,5 (19), the J chain (20), blk (21,22), the mouse engrailed gene (23), the human X-box binding protein-1 (24) and p53 (25), although the functional significance of many of these sites in vivo is unknown. In mature B cells, Pax5a is involved in cell activation, proliferation, and Ig class switching (16, 26 -32).
We have previously shown that Pax5a activity is regulated through a redox mechanism. In particular we demonstrated that an oxidized form of Pax5a is unable to interact with DNA, whereas the reduced form binds strongly, and that an intramolecular disulfide bond within the paired domain of Pax5a causes interference with specific DNA binding (37). Moreover, exposure of B cells to H 2 O 2 results in rapid transfer of the cytoplasmic redox factor APE/Ref-1 into the nucleus, an event correlated with an increase in Pax5a binding activity (33). APE/Ref-1 (also designated HAP-1) is a trifunctional protein involved in apurinic/apyrimidinic endonuclease DNA base repair activity, in proofreading exonuclease activity (34), and in modulating DNA binding activity of several transcription factors including NF-B, Egr-1, p53, and members of Pax family (33,(35)(36)(37). The redox and repair activities of APE/Ref-1 are localized to distinct, nonoverlapping domains of the protein that function independently. The N-terminal domain is essential for redox activity and contains the nuclear localization sequence (residues 1-36) (38), whereas the endonuclease activity resides in the C-terminal region (39). APE/Ref-1 expression is ubiquitous; however, it exhibits a complex and heterogeneous expression pattern that differs among tissue types (40).
Early B-cell factor was identified as a B lymphocyte-specific protein that recognizes a site in the mb-1 promoter (41,42). This protein was also identified as olfactory factor 1 in olfactory neurons (43). DNA binding studies showed that EBF binds as an homodimer and recognizes specific nucleotide sequences representing variations of an inverted repeat of a 5Ј-GGGA(A/ T)T half-site separated by 2 bp spacer (42,44). DNA binding activity is mediated by a novel zinc coordination motif in the amino-terminal half of EBF where there are also sequences that mediate DNA-dependent dimerization and transactivation (45). The COOH-terminal region of EBF contains a second, serine/threonine-rich activation domain with homology to the second helix of basic-helix-loop-helix proteins (42). The deletion of these ␣-helical repeats was found to markedly reduce DNA binding and dimerization in solution. EBF was shown to interact with functional regions in the promoters of mb-1 (41), 5 (42, 44, 46 -48), VpreB (46), B29 (49), and blk genes (50). This suggests that EBF has a role in the regulation of several genes encoding proteins of the pre-B and B cell receptor. Pax5a is required also during later stages of B cell development for the maintenance of mature B cell characteristics and function (51), but the role of EBF in mature B cells is currently unknown.
The aim of this work was to determine whether CD40-mediated stimulation of B lymphocytes promotes activation of EBF and Pax5a and to investigate the role of APE/Ref-1 in this process. Our data demonstrate that upon CD40 antibody-mediated cross-linking, APE/Ref-1 translocates from the cytoplasm to the nucleus of activated B cells, where it is able to control the DNA binding activity of Pax5a. In addition, we show that APE/Ref-1 modulates EBF activity and the Pax5a and EBF cooperative activation of the blk promoter. Moreover, we observed that the interaction between Pax5a and EBF enhances EBF binding activity to its consensus sequence, suggesting that Pax5a can physically interact with EBF and modulate its DNA binding activity. Together, these data suggest a novel role of APE/Ref-1 in B cells CD40-mediated activation.

EXPERIMENTAL PROCEDURES
Cell Preparation and Culture Conditions-Purified splenic B cells were obtained from 6 -12-week-old female Balb/c mice as follows: splenocyte cell suspension were depleted of red blood cells by hypotonic lysis with ACK lysing buffer (BioWhittaker) and of T cells by complement-mediated cytotoxic lysis using anti-Thy 1.2 mAb (a gift from K. Hathcock, Experimental Immunology Branch, National Cancer Institute/National Institutes of Health, Bethesda, MD) in conjunction with rabbit complement (Low-Tox M; Cedar Lane).
The resulting cells were checked by cytofluorimetric analysis (BD Biosciences FACScan) and then were incubated with purified antimouse CD40 mAb HM40-3 at 1 g/ml (BD Pharmingen) for 72 h. All experiments reported in this paper were performed at this time point, which represent the maximum value of proliferation, assayed by measuring [ 3 H]thymidine incorporation.
Preparation of Cell Extracts-10 7 cells were washed twice with PBS and then resuspended in 50 l of ice-cold buffer A (10 mM Hepes, pH 7.9, 10 mM KCl, 1.5 mM MgCl 2 , 0.1 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, 0.5 mM DTT). The samples were centrifuged at 800 ϫ g for 10 min at 4°C and the supernatants were collected as cytoplasmic extracts. Then, the pellets were resuspended in 50 l of buffer B (20 mM Hepes, pH 7.9, 400 mM NaCl, 1.5 mM MgCl 2 , 0.1 mM EDTA, 5% glycerol, 0.5 mM phenylmethylsulfonyl fluoride, 0.5 mM DTT) to extract nuclear proteins. After incubation for 20 min on ice, the samples were centrifuged at 10,000 ϫ g for 30 min at 4°C and the supernatants were collected as nuclear extracts.
For whole protein NS-1 extracts, 10 7 cells were re-suspended with 100 l of lysis buffer (10 mM Tris-HCl, pH 7.5, 50 mM NaF, 150 mM NaCl, 1 mM DTT, 0.5% Nonidet P-40, 0.5 mM phenylmethylsulfonyl fluoride) and, after incubation for 20 min at 4°C, the extracts were centrifuged at 22,000 ϫ g for 20 min to remove all cell debris. Protein concentrations were determined using the Bradford method (52) and the samples used immediately for Western blot or electrophoretic mobility shift assay (EMSA) analysis or kept at Ϫ80°C.
Western Blot Analysis-Equal amounts of protein, obtained as described above, were electrophoresed on a 12% SDS-PAGE minigel and transferred to nitrocellulose membranes (Schleicher & Schuell) using a semidry apparatus (Multiphor II, Amersham Biosciences). The filters were blocked for 1 h at room temperature in blocking solution: 5% nonfat milk powder in PBS, 0.05% Tween 20 (PBST). After washing with PBST, the membranes were incubated with primary Ab (diluted in blocking solution) overnight at 4°C or 1 h at room temperature. APE/ Ref-1 was detected by a rabbit polyclonal specific antibody (Santa Cruz Biotechnology), actin and nucleoporin by the mouse polyclonal specific antibody (Sigma and BD Biosciences, respectively), and Pax5a by the mouse polyclonal specific antibody D2A8 generated in our laboratory (33). The filters were then washed with PBST and incubated with the horseradish peroxidase-conjugated secondary antibodies (Sigma) in blocking solution for 1 h at room temperature. After washing the immunoreactive bands were visualized using a chemiluminescence substrate (Pierce) and Bio-Max light films (Kodak) according to the manufacturer's protocol.
EMSA-The sequence of the top strand of double-stranded oligonucleotide probes used in EMSAs were: H2A2.2, 5Ј-TCTGACGCAGCGGT-GGGTGACGACT; mb-1, 5Ј-GAGAGAGACTCAAGGGAATTGTGG; CREB, 5Ј-AGAGATTGCCTGACGTCAGAGAGCTAG. Oligonucleotides were end labeled with [␥-32 P]dATP by incubation with T4 polynucleotide kinase (Fermentas), annealed, and purified on column spinX Costar 8160 (Corning) filled with Sephadex G-50 fine (Amersham Biosciences). Nuclear extracts (8 g) were incubated with 0.5-1 ng of labeled probe (10,000 -20,000 cpm) for 30 min at room temperature in binding buffer (100 mM Tris-HCl, pH 7.4, 150 mM KCl, 10 mM EDTA, 10 mM DTT) with 1 g of calf thymus and bovine serum albumin. In the EMSA experiment with the mb-1 probe, the binding buffer contained 0.2 mM ZnCl 2 . Unlabeled oligonucleotides as DNA competitors were added 10 min before the addition of DNA probe at molar excesses indicated in the respective figures. The samples were separated on a 5% polyacrylamide Tris borate-EDTA gel, which was dried and then exposed to a Hyperfilm-MP (Amersham Biosciences) film at Ϫ80°C.
For in vitro experiments with recombinant proteins, EMSA were performed with the mb-1 probe incubated with 100 ng of GST-EBF oxidized protein, obtained after prolonged air exposure, in the presence or absence of 400 ng of APE/Ref-1 recombinant protein or with 10 g of crude, boiled, or APE/Ref-1-depleted NS-1 cellular extracts. The immunoprecipitation of APE/Ref-1 was performed as described before (33). In these experiments binding buffer did not contain EDTA and DTT. To test the role of Pax5a on EBF DNA binding activity, EMSA were performed with the mb-1 probe, equal amounts of oxidized GST-EBF and recombinant APE/Ref-1, and 100 ng of recombinant Pax5a. Moreover, 3 g of stable NS-1 transfected with Pax5a (NS-1/Pax5a) nuclear extracts were incubated with or without 100 ng of GST-EBF.
Oligonucleotide Treatment of B Cells-APE/Ref-1 antisense oligonucleotide and the complementary sense oligonucleotide were synthesized and HPSF purified by MWG-Biotech AG. The oligonucleotides were phosphorothioated at 3Ј and 5Ј ends (at the position marked by *) to confer nuclease resistance (53). The sequence of the APE/Ref-1 antisense probe was 5Ј-T*T*CCCCGCTTTGGCATC*G*C*-3Ј, and sense was 5Ј-G*C*GATGCCAAAGCGGGG*A*A*-3Ј (nucleotides Ϫ3 to 16). The probe did not show homology to other known genes according to the GenBank TM data base.
Spleen B cells were washed once with serum-free media (Opti-MEM, Invitrogen) and then 1.6 ml of serum-free media was added for 4 ϫ 10 6 cells in a 60-mm tissue culture plate. 25 l of Lipofectin (Invitrogen) was diluted in 200 l of serum-free media, incubated for 30 min at room temperature, and then mixed with the DNA (10 g) diluted in 200 l of serum-free media. The mixture was incubated for 15 min and the combined volume of 400 l was added to the cells. The cells were then maintained in 5% CO 2 incubator at 37°C for 12 h, after which 2 ml of growth medium supplemented with 20% fetal calf serum was added to each transfection well. Then transfected cells (1 ϫ 10 6 /ml) were incubated with the appropriate stimulus. At the end of that incubation, culture growth was determined by counting, and viability was confirmed with trypan blue exclusion.
Plasmids and Construction of the Bidirectional Expression Vector-The pBI vector (Clontech) allows to simultaneously express two genes under control of a bidirectional tetracycline-responsive promoter (Pbi-1) (54). Pbi-1 promoter contains the Tet-responsive element, which is repressed in the presence of tetracycline (Tc). pBI-GL (Clontech) allows the simultaneous regulation of both luciferase and ␤-galactosidase genes by one central Tet-responsive element.
The 965-bp BamHI/EcoRI fragment containing hRef-1 cDNA was taken from pGEX-3X-hRef-1 (see below) and filled in with Klenow fragment. The resulting blunt fragment was cloned into the NheI Klenow filled in site of pBI to make pBI-APE/Ref-1.
Primers EBF-NheI-REV (5Ј-GCAGCTAGCCACTCTGGGACTCATG-3Ј) and EBF-NheI-FOR (5Ј-GTAGCTAGCTGGGCAGCGGCATGAA-3Ј) containing the NheI restriction site were used to amplify the region encoding EBF cDNA, using CMV-EBF as a template. The amplified sequence was cloned into the NheI site of pBI to make pBI-EBF. DNA sequencing confirmed that all PCR products were free of any undesired mutations. The 1.2-kb and 750-bp NotI fragments containing Pax5a and Pax5d cDNA, respectively, were taken from pcDNA3.1-Pax5a and pcDNA3.1-Pax5d (11) and cloned into the NotI site of pBI to make pBI-Pax5a and pBI-Pax5d or cloned into the NotI site of pBI-EBF to make pBI-Pax5a-EBF and pBI-Pax5d-EBF, respectively, or in NotI site of PBI-Ref-1 to make pBI-Pax5a-APE/Ref-1 and pBI-Pax5d-APE/Ref-1, respectively.
Generation of Stable Cells Expressing Various Transactivators and Quantification of Transactivation Activity-HeLa cells were grown in 35-mm dishes to 50 -60% confluence and transfected via the calcium phosphate procedure with 5 g of tTA2/3/4 (Clontech) plasmids. These are three vectors that express at different levels the tetracycline-sensitive Pbi-1 transactivator and cause expression of pBI cloned genes at different levels (55).
After 24 h, cells were transferred to 10-cm dishes and maintained in medium containing 500 g/ml G418 (Invitrogen). After 4 weeks, at least 30 resistant clones were isolated using cloning cylinders (Sigma) expanded separately and frozen as soon as possible. The presence of integrated tTA in neomycin-resistant cell lines were confirmed by PCR on chromosomal DNA using primers ptTA forward (5Ј-GTCTGGATC-CTTACTTAGTTACCC-3Ј) and ptTA reverse (5Ј-ATAGAAGACACCGG-GACCGATC-3Ј) that amplify a portion of the appropriate tTA plasmid. Positive clones were screened by transient transfections with PBI-GL reporter vector to identify G418-resistant clones that meet the criteria for good Tet-Off cell lines: low background and high induction of luciferase and ␤-galactosidase in response to Tc (2 g/ml). A clone from HeLa-tTA2 cells (HeLa cells that constitutively synthesize tTA2) was selected for the highest -fold induction (e.g. showing highest expression with lowest background) for propagation and further testing.
HeLa-tTA2 cells were seeded at a density of 1 ϫ 10 5 cell/35-mm dish and grown in the presence or absence of Tc (2 g/ml). The next day, 2 h prior to transfection, the culture medium was renewed and the cells were incubated at 37°C and 5% CO 2 . The calcium phosphate/DNA precipitate contains 2.5 g of plasmid DNA (consisting of 0.6 g of pBI, 0.8 g of pGLblk, 0.2 g of lacZ expression vector (CMV-␤-galactosidase), included for normalization of transfection efficiency, and pUC18 as nonspecific carrier DNA for equalization the DNA content in each transfection). The precipitate (55 l/dish) was added to HeLa-tTA2 cells that were further incubated at 37°C and 5% CO 2 for 24 h (56). The next day, the cells were washed twice in PBS and grown in the presence or absence of Tc (2 g/ml). After 48 h the cells were scraped using TEN buffer (40 mM Tris-HCl, pH 7.4, 1 mM EDTA, 150 mM NaCl), centrifuged at 300 ϫ g for 5 min at 4°C, and the pellets were resuspended in 150 l of lysis buffer provided by the ␤-galactosidase kit (Roche Diagnostics) plus 1 mM DTT. After incubation for 20 min at room temperature, the samples were subjected to three cycles of freeze-thaw, and after centrifugation at 18,000 ϫ g for 10 min at 4°C, the supernatants were collected and used immediately for assays or kept at Ϫ80°C.
Luc activity was determined using a luciferase assay. In brief, assays were conducted in a luminometer tube with 50 l of the obtained extracts, 350 l of luciferase assay buffer (25 mM glycylglycine, pH 7.8, 2 mM ATP, 10 mM MgSO 4 ), and 100 l of injection solution (0.2 mM luciferin, 25 mM glycylglycine, pH 7.8). The data were normalized for transfection efficiency against the activity of 0.2 g of co-transfected CMV controlled-␤-galactosidase reporter gene. The ␤-galactosidase assays were conducted with 25 l as described in the instruction manual of the kit (Roche Diagnostics).
Results of luc assays are shown as -fold induction of luc conversion Ϯ S.D. Relative -fold induction values for each transfection were calculated by dividing each normalized luc activity by the activity of the reporter construct alone.
NS-1 cells were stably transfected with pCDNA3.1-Pax5a by electroporation using a Bio-Rad gene pulser (Bio-Rad). Aliquots of 1 ϫ 10 7 cells were resuspended in 350 l of RPMI 1640 in a Bio-Rad electroporation cuvette (0.4 cm gap) and pulsed once with 340 V and 960 F. After electroporation, cells were maintained at 4°C for 10 min and then cultured in pre-equilibrated RPMI 1640 supplemented with 10% fetal serum for 24 h. The next day, the cells were washed once in PBS and grown in medium containing 400 g/ml G418 for 1 month. The presence of integrated Pax5a in neomycin-resistant NS-1 cell lines was confirmed by Western blot using D2A8 antibody (33). Transient transfections of HeLa cells were performed by the calcium phosphate method as above described using 2.5 g of total DNA.
Protein Expression and Purification-The APE/Ref-1 cDNA was cloned in the prokaryotic expression vector pGEX-3X (Amersham Biosciences) and used to transform the BL21(DE3)pLysS (Novagen) bacterial strain (Stratagene). Transformed cells were grown at 37°C to A 600 0.6 and then induced by 100 M isopropyl-1-thio-␤-D-galactopyranoside for 3 h. Cells were harvested by centrifugation and resuspended in lysis buffer (200 mM Hepes, 150 mM NaCl, 5% glycerol, 5 mM DTT, pH 7.2). Cells were disrupted by sonication (4 ϫ 30 s at 250 W with 1 min pause) and debris were removed by centrifugation at 23,500 ϫ g for 40 min. The supernatant was loaded onto glutatione-Sepharose 4b resin (Amersham Biosciences) pre-equilibrated with lysis buffer. The resin was washed 3 ϫ 15 min with washing buffer (200 mM Hepes, 400 mM NaCl, 5 mM DTT, pH 7.2) and then equilibrated with 50 mM Tris-HCl, 100 mM NaCl, 1 mM CaCl 2 , pH 8. 10 Units/ml of factor Xa were added to the resin and incubated overnight under rotation at room temperature to obtain the fusion protein APE/Ref-1. The eluant was added to benzamidine resin (Amersham Biosciences) to eliminate factor Xa; the concentration of NaCl was increased to 300 mM and 1 mM DTT was added.
The EBF cDNA was cloned in the prokaryotic expression vector pGEX-4T1 (Amersham Biosciences) and used to transform the BL21(DE3)pLysS bacterial strain. Transformed cells were grown at 25°C to A 600 0.8 and then induced by 100 M isopropyl-1-thio-␤-Dgalactopyranoside for 3 h. Cells were harvested by centrifugation and resuspended in lysis buffer (200 mM Hepes, 150 mM NaCl, 5% glycerol, 5 mM DTT, 2 mM ATP, 10 mM MgSO 4 , pH 7.2, and protease inhibitor EDTA-free) in a volume of 3 ml/g of bacterial pellet. After sonication and centrifugation as described above, the supernatant was loaded with peristaltic pump at 4°C onto a 1-ml pre-equilibrated GSTrap FF column (Amersham Biosciences). The column was washed with wash buffer (50 mM Hepes, 5 mM DTT, pH 7.2) and the protein was eluted with elution buffer (50 mM Hepes, 10 mM reduced glutathione, 5 mM DTT, pH 7.2). The collected fractions were concentrated with Centricon 30MW (Millipore).
The Pax5a cDNA was cloned in the prokaryotic expression vector pQE-31 (Qiagen) and used to transform the BL21(DE3)RIL bacterial strain (Stratagene), previously transformed with pREP4 plasmid (Qiagen). Transformed cells were grown at 37°C to A 600 0.6 and then induced by 1 mM isopropyl-1-thio-␤-D-galactopyranoside for 3 h. Cells were harvested by centrifugation and resuspended in 10 ml of lysis buffer 5 ϫ TE (1ϫ TE, 10 mM Tris-HCl, 1 mM EDTA, pH 7.5) for each gram of bacterial pellet. The protein was purified as described by Tell et al. (57) and then concentrated with Centricon 30MW (Millipore).
The concentration of the three recombinant proteins was determined by the Bradford assay (52). The proteins were then stored at Ϫ20°C in 50% glycerol.
Pull-down Assay-GST and GST-EBF proteins, prepared as above described, were incubated 2 h at 4°C with glutathione-Sepharose beads (Amersham Biosciences) in lysis buffer (see above). After 3 washes in the same buffer, the beads were incubated 2 h with nuclear extracts of HeLa cells transiently transfected with Pax5a, in binding buffer (25 mM Hepes, pH 7.9, 50 mM NaCl, 1 mM DTT, 0.01% Nonidet P-40). For detection of bound protein, beads were washed 3 times with binding buffer, boiled in SDS sample buffer, and bound material was detected by SDS-PAGE and Western blotting.

RESULTS
CD40 Cross-linking Enhances DNA Binding Activity of Pax5a and EBF in Primary Mouse B Cells-The nuclear proteins EBF and Pax5a are important regulators of B lymphocyte development. Their co-expression in B cells at the earliest stages of differentiation allows for collaboration on regulatory modules of early B cell-specific genes (58). Thus, EBF and Pax5a work separately and in concert to activate genes required for B cell differentiation. For instance, Pax5a is a transcriptional regulator of B cell-specific genes and is involved in activation and proliferation of B lymphocytes and Ig class switching. EBF interacts with k promoters and may modify Ig light chain expression (46,59).
To determine whether CD40 signaling influences the functional activity of these two B cell-specific transcription factors mouse spleen B cells were activated by CD40 cross-linking with purified anti-mouse CD40 mAb HM40 -3. Because maximal cell proliferation was observed after 72 h (data not shown) we choose this time to perform all further experiments. We performed EMSA with H2A2.2-and mb-1-labeled probes specific for Pax5a and EBF, respectively, using nuclear extracts of mouse B cells before and after CD40-mediated activation. As shown in Fig. 1 we observed that: Pax5a (panel A) and EBF (panel B) are constitutively active in non-stimulated resting B cells and CD40 cross-linking differentially increases Pax5a and EBF DNA binding activity, as compared with the activity observed in non-stimulated B cells. As control, we used a probe specific for the constitutively expressed transcription factor CREB (panel C).
We investigated if the increase in Pax5a DNA binding activity was because of an increase in the protein levels by Western blot. As shown in Fig. 2B, the protein levels of Pax5a remain constant in mouse spleen B cells before and after CD40 cross-linking. Given that the Pax5a protein levels remain constant upon CD40 stimulation, we investigated if post-transcriptional mechanisms such as redox regulation could account for the increase of Pax5a activity. Nuclear extracts (8 g) were incubated with 32 P-labeled probe encompassing the promoter of the sea urchin histone H2A-2.2. B, EMSA detects EBF functional activity in mouse B cells stimulated and not stimulated with anti-CD40 mAb. Nuclear extracts (8 g) were incubated with 32 Plabeled probe encompassing the promoter of mb-1. C, EMSA detects cAMP-response element-binding protein (CREB) functional activity in the same extracts of panels A and B, as control. In all the experiments nuclear extracts from Raji cell were used as positive control, the protein-DNA complexes were separated on a 5% polyacrylamide gel, and competitors were added in molar excess as indicated. F indicates the migration of free probe. The arrows indicate the position of protein-DNA complexes.  (53). As control, a sense oligonucleotide was used. The cells were pretreated for 12 h with oligonucleotide plus Lipofectin as described under "Experimental Procedures" and then treated with anti-CD40 mAb for 72 h. Fig. 3A shows the ability of antisense oligonucleotide to efficiently downregulate APE/Ref-1 protein production. Next, we examined the effect of such down-regulation on Pax5a activity by EMSA. As shown in Fig. 3B, preincubation of anti-CD40 mAb-treated spleen B cells with antisense APE/Ref-1 oligonucleotide significantly reduced Pax5a DNA binding activity in nuclear extracts (Fig. 3B, left panel), suggesting that nuclear translocation of APE/Ref-1 plays a role in the redox modulation of Pax5a DNA binding activity under CD40 stimulation. This decrease in Pax5a DNA binding activity was not because of a decrease in the protein levels, as shown by Western blot for Pax5a protein (Fig. 3A, lower  panel). As control, we used a probe specific for the constitutively expressed transcription factor CREB (Fig. 3B, right panel).

Nuclear Translocation of APE/Ref-1 in Stimulated B Cells-APE/Ref-1 is a ubiquitous protein that has the ability to influence DNA binding of several transcription factors through a redox mechanism (60). We have recently shown that reactive oxygen species (ROS) elicit a nuclear translocation of APE/
Similar experiments were performed with mb-1 probe specific for EBF, but this approach did not lead to conclusive evidence for APE/Ref-1-mediated redox modulation of EBF. This prompted us to design different experimental approaches to investigate the involvement of APE/Ref-1 in EBF activity.

APE/Ref-1 Is Involved in EBF DNA
Binding Activity-To investigate a possible direct role of the redox factor APE/Ref-1 in the regulation of EBF DNA binding activity, we carried out a series of in vitro experiments with recombinant proteins. APE/Ref-1, EBF, and Pax5a were produced from prokaryotic expression systems, as described under "Experimental Procedures," and tested in EMSA assays. First EBF was purified as a GST fusion protein and tested in EMSA using the specific mb-1 promoter probe. Recombinant GST-EBF (Fig. 4A, lane 3) produced a specific band shift of the mb-1 probe, as assessed by the comparison with the shift produced with a nuclear extract of Raji cells (expressing endogenous EBF, Fig. 4A, lane 1). We observed that upon oxidation, recombinant GST-EBF loses its DNA binding activity (Fig. 4B, lane 1). Oxidized GST-EBF protein reacquires its DNA binding activity in the presence of recombinant APE/Ref-1 (shown in Fig. 4B, lane 2) but not in the presence of other reducing agents (e.g. DTT or ␤-mercaptoethanol) (data not shown).
To confirm the result, we incubated GST-EBF with a total extract of the NS-1 cellular line (Fig. 4B, lane 3). This cellular line does not express endogenous EBF and Pax5a but expresses APE/Ref-1. As control, the presence of aspecific interactions was tested with cellular extracts alone before (Fig. 4B, lane 7) and after boiling (Fig. 4B, lane 8). To establish a direct role for APE/Ref-1, we immunodepleted NS-1 cellular extracts with anti-APE/Ref-1 Ab and observed that these extracts were unable to rescue the oxidized EBF DNA binding activity (Fig. 4B,  lane 4). Furthermore, we performed similar experiments using total extracts of HeLa cell line. In the presence of this cellular extract, which does not express EBF but expresses endogenous APE/Ref-1, we observed that recombinant GST-EBF produced a specific band shift of the mb-1 probe (Fig. 4B, lane 9). Taken together the data presented in Fig. 4B indicate that APE/Ref-1 is able to directly reduce the oxidized form of GST-EBF and rescue its DNA binding activity suggesting that APE/Ref-1 might directly regulate EBF functional activity.

APE/Ref-1 Affects the Functional Cooperation between Pax5a and EBF on blk Promoter Activation in a Non-lymphoid Cell
Line-Based on earlier in vitro data indicating that EBF plays an important role in the regulation of blk promoter in early B cell development and that Pax5a and EBF are capable to act in cooperation to induce this target gene (50), we sought to investigate in vivo whether such cooperation is sensitive to APE/ Ref-1 redox regulation. For this set of experiments, we used a reporter construct containing a portion of the blk promoter with responsive elements for both Pax5 and EBF and the HeLa cell line, which offers a neutral functional background as it does not express endogenous EBF and Pax proteins. Earlier studies on the effects of EBF and Pax5a function used cotransfections of two expression vectors, one for EBF and one for Pax5a. With the aim to overcome the problem of disparate levels of expression, we used a tetracycline-regulated bidirectional expression vector that allowed transcription initiation from the same regulatory element. HeLa cells were stably transfected with an expression vector producing a chimerical protein (tTA2) consisting of the DNA binding domain of the Tc repressor and the transactivation domain of the herpes simplex virus 1 VP16. In the absence of Tc the chimerical tTA is able to bind the regulatory element and drive overexpression from two minimal promoters. This system exhibits tight on/off regula- tion and absence of pleiotropic effects. All experiments in this study were done in the absence of tetracycline.
Bidirectional expression vectors were transiently cotransfected with the pGL3-blk luciferase reporter in HeLa-tTA2 cells and transactivation levels were measured. Reporter gene expression from the blk promoter is only detected at background levels when pGL3-blk luciferase is transfected into HeLa-tTA2 cells. As shown in Fig. 5A, the co-transfection of expression vectors containing Pax5a or EBF results in significant transactivation: transfection of pBI-Pax5a and pBI-EBF resulted in a 7-and 5-fold induction, respectively, whereas the reporter gene was expressed at a high level (ϫ29) in the presence of both the proteins suggesting that Pax5a and EBF cooperate on the activation of the blk promoter. This combined effect was de- pendent on a whole functional Pax5a protein, because transfection of pBI-Pax5d-EBF resulted in only a 8-fold induction.
Next, it was tested whether APE/Ref-1 could affect the functional cooperation between Pax5a and EBF. Reporter construct blk-luc was transiently co-transfected into HeLa-tTA2 cells using the same amounts of pBI-APE/Ref-1 and pBI-Pax5a or pBI-EBF (Fig. 5A).
As expected, the construct that induced co-expression of EBF and Pax5a (pBI-Pax5a-EBF) showed higher activity than EBF or Pax5a alone.

Pax5a Physically Interacts with EBF and Modulates Its DNA
Binding Activity-The evidence of a functional cooperation between EBF and Pax5a on the blk promoter (50), prompted us to verify if Pax5a can directly regulate EBF DNA binding activity in vitro. We decided to perform an EMSA assay with the two recombinant proteins incubated with the EBF-specific probe from the mb-1 promoter. As shown before, the oxidized form of GST-EBF did not bind to the DNA (Fig. 6A, lane 1), and reacquires its DNA binding activity in the presence of recombinant APE/Ref-1 (Fig. 6A, lane 2). Surprisingly, we observed that oxidized GST-EBF reacquired DNA binding activity when the protein was incubated with recombinant protein Pax5a (rPax5a) (Fig. 6A, lane 3). Adding rAPE/Ref-1 to this complex further increased its DNA binding activity (Fig. 6A, lane 4). To rule out that this effect could be artificial to the use of recombinant Pax5a, nuclear extracts obtained from stably transfected Pax5a NS-1 cells (NS-1/Pax5a) were used and, as shown in lane 5, a strong increase in EBF DNA binding activity was observed. As control, rPax5a (lane 6), NS-1/Pax5a cellular extracts (Fig. 6A, lanes 8 and 9) were used confirming that rPax5a did not bind the mb-1 probe and there are not specific interactions with the NS-1/Pax5a cellular extracts (Fig. 6A,  lanes 8 and 9). These results demonstrate that the observed interaction is specific for GST-EBF and suggest that Pax5a can establish a molecular complex with EBF and enhance its DNA binding activity.
To confirm this hypothesis we performed a pull-down assay using nuclear extracts of HeLa cells transiently transfected with Pax5a (Fig. 6B, lane 1) using GST-EBF (Fig. 6B, lane 2) or GST alone (Fig. 6B, lane 3) proteins as baits. As shown in Fig.  6B Pax5a was found to be specifically associated with GST-EBF, thus confirming that the two proteins can physically interact. factors (33,(35)(36)(37). Stress signals such as ROS and other exogenous agents can modify the activity of APE/Ref-1 (33,35,36). For instance, APE/Ref-1 acts as a pivotal signaling factor in the induction of early stress response genes, such as c-Fos and c-Jun. At least one mechanism regulating c-Fos/c-Jun DNA binding is mediated by a conserved cysteine (Cys) located in the basic DNA-binding domain of both proteins (62). In vitro these regulatory cysteines are not permissive for DNA binding under oxidized conditions, whereas reduction to a sulfhydryl state promotes DNA binding (35,62). As such, these critical cysteines act as a redox-sensitive "sulfhydryl switch" that reversibly modulates DNA binding (36). In the absence of reducing agents, the redox factor-1 (APE/Ref-1) protein regulates c-Fos/ c-Jun DNA binding via the same conserved cysteine. APE/Ref-1 is found exclusively in the cytoplasm of some cells, in the nucleus in others, and in some cases it is localized both in the cytoplasm and nucleus (40). The biological relevance of APE/Ref-1 compartmentalization is not understood, but the complexity of staining patterns suggests that localization is regulated. The redox role of APE/Ref-1 may be relevant to its cytoplasmic localization, where it may be required to maintain newly synthesized transcription factors in a reduced state while they are being transported to the nucleus (40). We have previously demonstrated that the DNA binding activity of the Prd domain of Pax proteins is regulated through the oxidationreduction of conserved cysteine pairs and that the Pax5a activity is regulated through a redox mechanism involving APE/ Ref-1 (57). Furthermore, we have shown that exposure of B cells to H 2 O 2 results in rapid transfer of the cytoplasmic redox factor APE/Ref-1 into the nucleus and this provides evidence with an increase in Pax5a binding activity (33).
CD40 is involved in different important aspects of B cells activation such as B cell proliferation, differentiation, isotype switching, up-regulation of surface molecules, development of the germinal center, and the humoral memory response. CD40 engagement results in the production of reactive oxygen intermediates, which serve as second messengers in cell signaling (63). In this work we investigate if CD40-mediated stimulation of B lymphocytes could promote activation of EBF and Pax5a. We show that both of these transcription factors are constitutively expressed in spleen B cells. Moreover, we provide evidence for the first time that APE/Ref-1 acts as a key signaling intermediate in response to CD40-mediated B cell activation. We observe that APE/Ref-1 translocates from the cytoplasm to the nucleus of activated B cells. We demonstrate that CD40mediated activation of spleen B cells enhances the DNA binding activity of Pax5 and EBF. APE/Ref-1 appears to be required for CD40-mediated Pax5 activation, as the repression of APE/ Ref-1 protein production is able to block CD40-induced Pax5 DNA binding activity.
Based on earlier in vitro data indicating that Pax5a and EBF are capable to act in cooperation to induce blk gene expression (50)  In in vivo experiments we provided evidence that the cooperative activation of the blk promoter operated by Pax5a and EBF (50) is dependent on a whole functional Pax5a protein and that APE/Ref-1 can modulate this functional cooperation. In fact, EBF does not show cooperative activation of the blk promoter when cotransfected with Pax5d (an isoform of Pax5 that possesses an intact DNA-binding domain but does not have a transactivation domain and partial homeodomain homology region). This suggests that the transactivation domain, the partial homeodomain homology region of Pax5, or both play a role in the functional cooperation with EBF. Moreover, for the first time we provide evidence that APE/Ref-1 is able to enhance the cooperative activation of Pax5a and EBF on the blk promoter, because cotransfection of APE/Ref-1 with Pax5a and EBF has more than additive effects on reporter expression. We also show that APE/Ref-1 increases EBF-and Pax5a-mediated blk promoter activation, confirming our in vitro data. Taken together, these data show that both Pax5a and EBF are sensitive to APE/Ref-1 regulation and also that their functional cooperation on the blk promoter can be modulated by APE/Ref- 1. In addition, in vitro EMSA experiments show a gain of EBF DNA binding activity on its consensus sequence when Pax5a is present and in particular that Pax5a can influence the EBF functional activity establishing a molecular complex. This is confirmed by the GST-EBF pull-down assay that shows that there is a physical interaction between the two proteins and this interaction appears to be independent from the presence of the DNA. APE/Ref-1 might contribute to B cell activation linking extracellular generated signals with the regulation of key target genes, as it translocates from the cytoplasm to the nucleus and regulates B cell-specific transcription factors following the triggering of CD40 or the exposure to ROS. Given the ability of APE/Ref-1 to control the DNA binding activity of Pax5a and Pax5a-EBF complexes, we propose that APE/Ref-1 could act as a key signaling intermediate in B cell activation and thus in the regulation of immune responses.