GABPβ2 Is Dispensible for Normal Lymphocyte Development but Moderately Affects B Cell Responses*

GA-binding protein (GABP) is the only Ets family transcription factor that functions as a heterodimer. The GABPα subunit binds to DNA, and the GABPβ subunit possesses the ability to transactivate target genes. Inactivation of GABPα caused embryonic lethality and defective lymphocyte development and immune responses. There are 3 isoforms of the GABPβ subunit, but whether they have distinct functions has not been addressed. In this study, we selectively ablated the expression of GABPβ2 using a gene trap strategy. GABPβ2-deficient mice were viable and had normal T and B cell development, suggesting that loss of GABPβ2 is compensated for by other GABPβ isoforms during these processes. GABPβ2-deficient T cells can be activated and proliferate similarly to wild-type controls. In contrast, B cells lacking GABPβ2 showed 2–3-fold increases in proliferation in response to B cell receptor stimulation. In addition, GABPβ2-deficient mice exhibited moderately increased antibody production and germinal center responses when challenged with T-dependent antigens. These results indicate that albeit GABPβ isoforms are redundant in lymphocyte development, GABPβ2 has a distinct role in restraining B cell expansion and humoral responses.

GA-binding protein (GABP) 2 is a member of the Ets family transcription factors and is comprised of two subunits, GABP␣ and GABP␤ (1). GABP␣ contains a DNA binding Ets domain, which is conserved among all the Ets factors and is of ϳ85 amino acids in length. The Ets domain assumes a winged helixloop-helix configuration and binds preferentially to a purinerich consensus DNA sequence containing GGA(A/T). On the other hand, GABP␤ cannot bind DNA but contains 4 ankyrin repeats in its N terminus, which mediate the protein-protein interaction with GABP␣. GABP␤ also contains a nuclear localization signal, which targets the GABP␣/␤ dimer into the nucleus. Transactivation activity of the GABP␣/␤ complex is considered to reside in the C terminus of the GABP␤ subunit, but the exact location has not been unequivocally mapped.
The GABP␣/␤ complex has versatile roles in regulating basic cellular functions and tissue-specific functions (1)(2)(3). Genetargeting studies of the DNA-binding GABP␣ subunit have revealed its critical roles during embryogenesis (4), reentry into the cell cycle (5)(6)(7)(8), and in synaptic function at the neuromuscular junction (9,10). In lymphocytes, GABP activates an interleukin (IL)-2 enhancer (11), IL-16, and Fas promoters (12,13), and transcription of IL-7R␣ (14,15). Recently, we showed that GABP is a key component of a gene regulatory network programming B lineage commitment and differentiation by directly regulating Pax5 gene expression (16). In contrast to extensive studies on GABP␣, the roles of the GABP␤ subunit in vivo have not been investigated. The most studied GABP␤ isoforms are those encoded by the Gabpb1 allele, which generates two protein products, GABP␤1L (originally named GABP␤1 or GABP␤1-1) and GABP␤1S (originally named GABP␤2 or GABP␤1-2), via alternative splicing (17,18). The N-terminal 332 amino acids of both GABP␤1L and GABP␤1S isoforms are identical, but their C termini differ in length and sequence. GABP␤1L has a longer C-terminal tail (50 amino acids), which adopts a leucine zipper-like structure, forming homodimers and even an ␣ 2 ␤ 2 GABP tetramer when two Ets motifs are adjacent or brought into proximity (17,19,20). In contrast, the C terminus of the shorter isoform, GABP␤1S, is 15 amino acids long, lacks the leucine zipper-like structure, and thus cannot form homodimers or tetramers. Nevertheless, both GABP␤1L and GABP␤1S heterodimerize with GABP␣ with similar affinity (21). It has been disputed over whether the leucine zipperlike structure in GABP␤1L has a unique role in transactivation of GABP target genes. In a recent report, we specifically targeted GABP␤1L by deleting exon 9 of the Gabpb1 gene that encodes the entire leucine zipper-like structure without eliminating the GABP␤1S isoform (22). In contrast to a pre-implantation lethality in a GABP␣-null mutant, GABP␤1L Ϫ/Ϫ mice were viable. On the other hand, targeting both GABP␤1L and GABP␤1S caused embryonic lethality prior to embryonic day 12.5 (22). These findings indicate that the Gabpb1 gene products are required for normal embryogenesis, whereas loss of GABP␤1L expression can be compensated for by other GABP␤ isoforms including GABP␤1S.
A third GABP␤ isoform, GABP␤2 (originally named GABP␤2-1), is encoded by Gabpb2 (18) and is 414-amino acids long with its N-terminal ankyrin repeats (amino acids 1-130) sharing 87% identity with GABP␤1 isoforms. GABP␤2 also has a long C terminus, with its 317-366 amino acid residues sharing 70% identity with that of GABP␤1L and also adopting a leucine zipper-like structure. The 367-414 amino acid residues in GABP␤2 are unique. The C terminus of GABP␤2 cannot only mediate the formation of GABP␤2 homodimers but also mediates the heterodimerization of GABP␤2 and GABP␤1L (18). Since its initial cloning, GABP␤2 has not been studied, and its function is completely unknown. The structural similarity raised the possibility that GABP␤ isoforms may have distinct and overlapping roles in regulating lymphocyte development and functional responsiveness. As one of the first steps aiming to dissect the exact roles of each GABP␤ isoform, we have inactivated the Gabpb2 gene using a gene trap strategy.

EXPERIMENTAL PROCEDURES
5Ј-RACE and RT-PCR-Splenic B cells were negatively selected using EasySep B cell enrichment kits (StemCell Technology), and total RNA was extracted from purified B cells or total thymocytes using the RNeasy Mini kit (Qiagen), and 5Ј-RACE was done with the GeneRacer kit (Invitrogen). Racer and nest racer primers were supplied in the kit, and antisense primers complementary to mouse GABP␤2 cDNA (5Ј-GCAGCTTCTAGCAGCCTCTTC and 5Ј-TTC-CCCAAGTCCACCAGAGAC) were used in RT-PCRs (Fig.  1A). Nested PCR was performed to increase the specificity of amplification. PCR products were subcloned into the pCR4-TOPO vector with a TOPO TA cloning kit (Invitrogen) and then sequenced.
Generation of GABP␤2-targeted Mice-An ES clone (RRJ488) was obtained from the International Gene Trap Consortium through the Mutant Mouse Regional Resource Center of the University of California at Davis. Mapping of the insertion site of the reporter gene is done using the Expand Long Template PCR System (Roche Applied Science). The ES cells were microinjected into C57BL/6 blastocysts, and male chimeras were identified to achieve germline transmission. GABP␤2 ϩ/tp mice were interbred to obtain GABP␤2 tp/tp and littermate controls. All experiments with mice followed protocols approved by the Institutional Animal Care and Use Committee, University of Iowa.
Generation of GABP␤2 Antisera and Western Blotting-To generate antisera that are specific to GABP␤2 with no crossreaction with GABP␤1, we used a recombinant GST fusion protein expressing GABP␤2 C-terminal amino acids 275-414 as an antigen, because the N termini of all GABP␤ isoforms contain highly conserved ankyrin repeat domains. To this end, we PCR-amplified cDNA corresponding to the GABP␤2 C-terminal fragment in a pGEX-4T-1 vector and expressed the fusion protein in BL21 Star competent cells (Stratagene). The fusion protein was expressed at high abundance with expected molecular weight and purified with magnetic resin-based Mag-neGST particles (Promega). The expression of GABP␤2 was confirmed by MALDI-TOF mass spectrometry, and the purified protein was used to immunize rabbits for antibody production (Bio-synthesis). To test the specificity of the GABP␤2 antiserum, we cloned GABP␤1L and GABP␤2 cDNA in pCruz-HA vector (Santa Cruz Biotechnology) so that an HA tag is fused to the N terminus of each expressed protein. We then expressed these proteins in 293 HEK cell lines, and subjected the cell lysates to immunoblot with anti-HA antibody, which detected both GABP␤1L and GABP␤2 of expected molecular weight (Fig. 2B). In contrast, the GABP␤2 antiserum detected GABP␤2 at a 1:2000 dilution but did not cross-react with GABP␤1 (Fig. 2C). For detection of GABP␤2 expression, whole cell extracts were prepared from thymocytes, splenocytes, and splenic B cells as described previously (23). Lysate protein (30 g) was separated on SDS/PAGE gels, transferred to nitrocellulose, and immunoblotted with the GABP␤2 antiserum (14). An anti-GABP␣ antibody (H180, Santa Cruz Biotechnology) was used to detect GABP␣ expression, and an antiserum raised against the N terminus of GABP␤1 (14) was used to detect both GABP␤1L and GABP␤1S.
Flow Cytometry and CFSE Staining-Single cell suspensions were prepared from thymuses, spleens, and bone marrow, and stained with fluorochrome-conjugated antibodies, as described (16). All fluorochrome-conjugated antibodies were from BD PharMingen or eBiosciences. Negatively selected splenic B cells were labeled with CFSE as described (16), stimulated with 10 g/ml of anti-IgM, and dilution of CFSE was determined on different days by flow cytometry.
[ 3 H]Thymidine Uptake-Splenic T and B cells were purified by negative selection using EasySep T and B cell enrichment kits (StemCell Technology), respectively, and the purity of isolated cells was ϳ95%. T cells (1 ϫ 10 5 /well) were stimulated with plate-bound anti-CD3 (0.25 or 0.5 g/ml, clone 145-2C11, BD Biosciences) in the absence or presence of anti-CD28 (1 g/ml, clone 37.51, BD Biosciences). B cells (0.6 ϫ 10 5 /well) were stimulated with 2.5 g/ml anti-IgM chain (Jackson ImmunoResearch Laboratories), 5 g/ml anti-CD40 (BD PharMingen), 5 g/ml LPS (Sigma), or 1 g/ml CpG oligonucleotides (ODN1862, Invivogen). The cells were cultured in 96-well plates in triplicate for each condition for 48 -60 h and pulsed with 1 Ci of [ 3 H]thymidine (PerkinElmer) for the last 12 h of culture. Radioactivity incorporated into cells was collected on a UniFilter-96 (GF/B, PerkinElmer) with a cell harvester and was counted using a TopCount.NXT microplate scintillation and a luminescence counter (Packard).
Immunization and Enzyme-linked Immunosorbent Assay (ELISA)-To determine the immune response to a T-independent antigen, mice were immunized intraperitoneally with 100 g of TNP(52)-AECM-Ficoll (Biosearch Technologies), and the sera were collected on day 8 after immunization. For immune responses to a T-dependent antigen, mice were intraperitoneally injected with 100 g of Imject Ovalbumin (Pierce) mixed with Imject Alum (Pierce, volume 1:1), boosted 1 week later with the same regimen, and sera collected after another week. The sera were diluted in 1:2 series to a total of 12 points, and TNP-or ovalbumin-specific immunoglobulins (Igs) were measured by ELISA. In brief, high binding plates (Immulux, Dynex) were coated with either 20 g/ml of TNP(38)-BSA (Bioresearch Technologies) or 5 g/ml of Imject Ovalbumin to absorb antigen-specific Igs, which were then detected with biotin-conjugated rat anti-mouse antibodies specific for murine IgM, IgG1, and IgG3. Visualization of the antigen-antibody complexes was revealed with the avidin-horseradish peroxidase and TMB substrate set (BD PharMingen), and absorbance at 450 nm was read on an ELX800 microplate reader (Bio-TEK). Linear absorbance readings were observed within one particular range of serum dilutions, and these absorbance readings were used to compare antigen-specific Ig titers in WT and GABP␤2-deficient mice.
Sheep Red Blood Cell (SRBC) Immunization and Immunohistochemistry-Mice were i.p. injected with 0.2 ml of 10% v/v SRBC suspension (Colorado Serum Company) in phosphatebuffered saline, and spleens were harvested on days 8, 12, and 18 postimmunization. Splenocyte suspensions from half of the spleen were stained with anti-B220 mAb along with fluorescein isothiocyanate-conjugated peanut agglutinin (PNA, Vector Laboratories), and analyzed for GC B cell responses by flow cytometry. The other half of the spleen was used to determine GC structures by immunohistochemistry as described (24).

Structure of 5Ј-Untranslated Regions in the Gabpb2
Gene-In our previous studies, we used ES clones that have a Gabpa allele inactivated by a gene trap strategy to generate GABP␣-deficient mice (14,16). In this method, ES cells were retrovirally transduced with a reporter gene, such as ␤geo, which is a fusion of ␤-galactosidase and a neomycin resistance gene. When the reporter gene is integrated into the ES genome and inserted in an intron of a gene, the gene is thus trapped. Because a strong splice acceptor is placed immediately 5Ј of the reporter gene after the trapped gene is transcribed, the splicing will occur between the upstream introns and the reporter gene, resulting in truncation and/or inactivation of the trapped gene. By searching the data base of the International Gene Trap Consortium, we identified an ES clone (named RRJ488) which has an insertion of the ␤geo reporter gene in the Gabpb2 locus. The insertion presumably occurred in an intron that is downstream of a non-coding exon at the 5Ј-end of the locus. The information on the Gabpb2 gene structure suggests existence of three 5Ј non-coding exons and alternative usage of the exons. If this approach is to be used to study the effect of Gabpb2 inactivation in lymphocytes obtained from RRJ488 ES clone-derived mice, it is important to determine whether these exons are utilized in B and T cells. We therefore used rapid amplification of cDNA 5Ј-end assay (5Ј-RACE) to determine the transcription initiation sites and the usage of 5Ј non-coding exons in the Gabpb2 gene. We performed nested RT-PCRs on splenic B cells (Fig.  1A) and observed three major transcripts (Fig. 1B). Analysis of the sequences of each transcript revealed that transcription of the Gabpb2 gene can be initiated from three different sites. If we arbitrarily define a location 12,480 bp upstream of exon 2 as "ϩ1," transcription can be initiated from ϩ959, ϩ1042, and ϩ1118 bp, and that the three 5Ј non-coding exons are utilized differently in each transcript (Fig. 1C). We also characterized Gabpb2 transcription initiation sites in thymocytes and observed two major transcripts that correspond to Transcripts 1 and 3 in B cells (data not shown). These results indicate that although transcription of the Gabpb2 gene can be initiated from different locations and there is different usage of 5Ј noncoding exons, the protein product from these transcripts is not altered.
GABP␤2 Expression Is Completely Abolished in GABP␤2 tp/tp Mice-We then mapped the insertion site of the ␤geo reporter gene in the RRJ488 ES cells using long template PCR. Based on the sequence tags provided for the ES clone, sense primers were designed based on the genomic sequence starting from exon 1a, and an antisense primer was based on the ␤geo coding sequence. Sequence analysis of the PCR product revealed that the insertion occurred between exons 1b and 1c ( Fig. 2A). As shown in Fig. 1C, because at least one part of the first noncoding exon (exon 1a) is used in all transcripts, we hypothesized that the insertion of the ␤geo reporter between exons 1b and 1c will interfere with normal splicing and inactivate the Gabpb2 gene. ES cells were injected into blastocysts to generate GABP␤2 ϩ/tp mice. Germline-transmitted GABP␤2 ϩ/tp mice gave birth to GABP␤2 tp/tp mice at a normal Mendelian ratio and were grossly normal. To determine GABP␤2 protein expression, we raised an antiserum specific for GABP␤2 using its C-terminal portion as an antigen, which can specifically recognize GABP␤2 but does not cross-react with GABP␤1L (Fig.  2, B and C). By Western blotting, we found that GABP␤2 B, Gabpb2 gene is transcribed from three locations. Total RNA was extracted from purified splenic B cells and subjected to 5Ј-RACE assay. First round PCR was performed using Racer primer and antisense primer, and 1 l of the PCR product was amplified using nested primers (nested racer primer and nested antisense primer) to increase the amplification specificity. C, differential usage of the three 5Ј non-coding exons in Gabpb2 transcripts. The Gabpb2 gene structure is based on information available at the University of California Santa Cruz (UCSC) genome browser, and the three non-coding exons are defined as 1a, 1b, and 1c. For the purpose of description clarity, a location that is 12,480 bp upstream of the coding exon, i.e. exon 2, is arbitrarily defined as "ϩ1," and the location of each non-coding exon is indicated based on the UCSC data base. The actual exon 1a equivalents in each transcript and their relative locations are marked separately. expression is easily detected in both thymocytes and splenocytes in WT mice and is not detectable in GABP␤2 tp/tp mice (Fig. 2D). This is in contrast to the situation of GABP␣ tp/tp embryos, which had hypomorphic expression of GABP␣ protein (14,16). The difference is likely ascribed to the location of reporter gene insertion and/or the gene context. We further confirmed that ablation of GABP␤2 did not alter the expression of GABP␤1L, GABP␤1S, and GABP␣ (Fig. 2E).

Loss of GABP␤2 Expression Does Not Perturb T and B Cell Development-Previously we have demonstrated that GABP␣ is critically required for IL-7R␣ expression in T cells as well as
for normal B cell development (14,16). T cells develop in the thymus, with the most immature thymocytes being double negative (DN) for CD4 and CD8 expression. DN thymocytes then develop into CD4 and CD8 double positive (DP) cells, which further mature to CD4 ϩ or CD8 ϩ single positive cells (25). The thymocyte, splenocytes, splenic T and B cells numbers between GABP␤2 tp/tp and WT control were comparable (Fig. 3A). Staining of thymoyctes with anti-CD4 and anti-CD8 antibodies revealed that the percentages of DN, DP, CD4 ϩ , and CD8 ϩ subset were not altered in GABP␤2 tp/tp mice (Fig. 3B). DN thymocytes can be further divided into four developmental stages based on CD25 and CD44 expression, i.e. DN1 (CD44 ϩ CD25 Ϫ ), DN2 (CD44 ϩ CD25 ϩ ), DN3 (CD44 Ϫ CD25 ϩ ), and DN4 (CD44 Ϫ CD25 Ϫ ). Fractionation of DN thymocytes with CD25 and CD44 staining showed that all four DN subsets appeared at similar frequencies between WT and GABP␤2 tp/tp mice (Fig. 3C). These results suggest that the loss of GABP␤2 expression does not affect T cell development, including early stages. In the periphery, splenic CD4 ϩ and CD8 ϩ T cells in GABP␤2 tp/tp mice appeared at a normal ratio (Fig. 3D), and IL-7R␣ expression on GABP␤2 tp/tp CD4 ϩ and CD8 ϩ T cells was not diminished (Fig. 3E and data not shown for CD8 cells), indicating that GABP␤2 is dispensable for IL-7R␣ expression on peripheral T cells.
We also examined B cell development in GABP␤2 tp/tp mice. Fractionation of bone marrow cells reveals three populations of sequentially developing B cells, i.e. pro-B and pre-B (B220 ϩ IgM Ϫ ), immature (B220 med IgM ϩ ), and recirculating B cells (B220 high IgM ϩ ) (26). All three populations in the bone marrow of GABP␤2 tp/tp mice were detected at similar percentages to those observed in WT controls (Fig. 3F). In the periphery, the frequency of total splenic B cells and marginal zone B cells (CD21 high CD23 dim ) was comparable between WT and GABP␤2 tp/tp mice (Fig. 3H). The CD23 bright B220 ϩ subset is heterogeneous and can be further fractionated to sequentially maturation stages, transitional 1 (CD24 high CD21 low ), transitional 2 (CD24 high CD21 high ), and mature follicular (CD24 dim CD21 dim ) B cells (27), and these maturing B cells showed similar frequency in both WT and GABP␤2 tp/tp mice (Fig. 3I). These data collectively demonstrate that inactivation of GABP␤2 expression did not detectably perturb B cell development in the bone marrow or further maturation in the spleen.
T and B Cell Proliferation in the Absence of GABP␤2-To determine if mature T and B cells that developed in the absence of GABP␤2 are functional, we isolated splenic T and B cells by negative selection. We stimulated T cells with different doses of plate-bound anti-CD3 in the presence and absence of anti-CD28 mAb and found that both WT and GABP␤2-deficient T cells proliferated similarly (Fig. 4A). In contrast, when the purified B cells were stimulated with anti-IgM, which cross-links the B cell receptors (BCRs), GABP␤2-deficient B cells showed increased proliferation in the presence or absence of IL-4 (Fig.  4B). However, anti-CD40-elicited proliferation was comparable in both WT and GABP␤2 tp/tp B cells (Fig. 4C). In addition to the clonally rearranged BCRs, B cells express nonclonal pattern recognition receptors, notably Toll-like receptors (TLRs) including TLR4 and TLR9 (28 -30). We used lipopolysaccharide (LPS) to stimulated TLR4, and ODN1826 (a synthetic oligonucleotide containing unmethylated CpG) to stimulated TLR9. WT and GABP␤2-deficient B cells proliferated similarly in response to both stimulants (Fig. 4D). These data collectively suggest that B cells lacking GABP␤2 have an enhanced proliferative response specifically to BCR stimulation.
We next tested if the increased B cell proliferation is a result of suppressed apoptosis or increased cell division. We activated B cells with anti-IgM and monitored apoptosis using Annexin V and 7-AAD staining on various days poststimulation. We did not find any consistent differences in percentages of Annexin V ϩ 7-AAD ϩ FIGURE 2. Inactivation of the Gabpb2 gene using a gene-trapped ES clone. A, mapping of the ␤geo insertion in the gene-trapped Gabpb2 gene. By long template PCR, the insertion of the ␤geo reporter gene is mapped between exons 1b and 1c, and the insertion site is at position 6754 based on the location indicated in Fig. 1C. The hatched box illustrates scrambled sequences at the insertion site, and these sequences are derived from the gene trap vector. SA, splice acceptor. B and C, specificity of anti-GABP␤2 antiserum. 293T cells were transfected with expression plasmids expressing HAtagged GABP␤1L or GABP␤2, and the cell lysates were fractionated using SDS-PAGE, transferred to membrane, and immunoblotted with an anti-HA antibody (B) or anti-GABP␤2 antiserum (C). D, expression of GABP␤2 in thymocytes and splenocytes. Thymocytes (Thy) and total splenocytes were isolated from mice of indicated genotypes. Whole cell lysates were prepared, separated by SDS-PAGE, and blotted with antiserum to GABP␤2 or ␤-actin (loading control). KO, Gabpb2 tp/tp . NS, nonspecific bands detected by the antiserum. E, expression of GABP␤1L, GABP␤1S, and GABP␣ in GABP␤2-deficient B cells. Cell lysates from splenic B cells were analyzed for expression of GABP subunits/isoforms using indicated antibodies. apoptotic cells (data not shown). In contrast, when labeled with carboxy-fluorescein diacetate succinimidyl ester (CFSE) before anti-IgM stimulation, GABP␤2 tp/tp B cells showed moderately accelerated CFSE dilution on all days examined (Fig. 4E). These results indicated that BCR-activated B cells can proliferate faster in the absence of GABP␤2.

GABP␤2-deficient B Cells Showed Increased in Vivo Responses to
Immunization-To investigate if the increased B cell division is associated with enhanced humoral immune responses, we immunized the mice with both T-independent (TNP-Ficoll) and T-dependent (ovalbumin) antigens. Although TNP-specific IgM and IgG3 were produced at similar levels in WT and GABP␤2-deficient mice (Fig.  5A), production of ovalbumin-specific IgM and IgG1 in GABP␤2 tp/tp mice was ϳ1.5-fold of those in littermate controls (Fig. 5B). In the immune response to protein antigens, B cells migrate into the follicle in lymphoid organs and form germinal centers (GCs), where antigenspecific B cells further expand and undergo class-switch recombination. To further delineate if loss of GABP␤2 expression can enhance GC responses to protein antigens, we immunized the mice with SRBC and examined GC formation in the spleen. Confocal microscopic analysis showed that both WT and GABP␤2 tp/tp mice can form GCs 8 days after immunization, and no apparent alteration in GC size was observed (data not shown). To quantitatively determine the SRBCelicited GC responses, we used fluorochrome-conjugated PNA, which selectively binds to GC B cells. In most cases, B220 ϩ PNA high cells were detected at an increased frequency in GABP␤2 tp/tp mice (Fig.  5C). On average, we observed a 33 and 40% increase in GC B cell frequency on day 8 and day 12 postimmunization, respectively (Fig. 5D). Similarly, the absolute numbers of splenic GC B cells were also increased by 51% on day 8 and by 31% on day 12 (Fig. 5E), albeit the difference did not reach statistical significance. These data collectively suggest that GABP␤2deficient B cells exhibit moderately increased responses to protein antigens.

DISCUSSION
A functional GABP␣/␤ complex requires two subunits, GABP␣ and GABP␤. The DNA binding Ets domain of GABP␣ directs the GABP complex to its target genes and determines binding specificity. Activation or repression of GABP target genes is mediated through the C terminus of GABP␤. As revealed in crystal structure studies of GABP, the interaction between GABP␣ and GABP␤ augments and stabilizes the DNA binding (31). In contrast to embryonic lethality caused by ablation of GABP␣ (4,14), we show here that mice deficient for GABP␤2 were viable. Previously we have demonstrated that GABP␣ is required for IL-7R␣ expression in peripheral T cells (14) and normal B cell development (16). In GABP␤2-deficient mice, both T and B cells developed normally and the expression of IL-7R␣ was not affected. These results indicate that GABP␤2 is dispensable for normal embryogenesis and lymphocyte development. These observations are quite similar to those in GABP␤1L Ϫ/Ϫ mice as we recently reported (22). Given the structural similarity between GABP␤1L and GABP␤2, loss of one GABP␤ isoform is likely compensated for by the other. Indeed, we have shown that both GABP␣/GABP␤1L and GABP␣/GABP␤2 heterodimers can bind to the Ets motif in the IL-7R␣ promoter region (22).
In contrast to the possible redundancy between GABP␤1L and GABP␤2 in lymphocyte development and T cell activation, GABP␤2 appears to have a distinct role in moderately restraining B cell proliferation and humoral responses to protein antigens. GABP␤2-deficient B cells manifested enhanced proliferation specifically by BCR stimulation but not by CD40 ligation or TLR stimulation. This observation suggests that GABP␤2 may have a specific role in negatively regulating BCR signaling, rather than signals that are derived from CD40-CD40 ligand interaction, TLR4 (activated by LPS), or TLR9 (activated by CpG nucleotides). Ligation of BCRs by antigens leads to rapid phosphorylation and activation of Syk, which further activate downstream signal components, including the phosphoinositide 3-kinase (PI3K)/Akt, extracellular signal-regulated kinase (Erk), and nuclear factor-B (NF-B) pathways, all three of which are known to be crucial for the survival and proliferation of B cells (32). We probed BCR signaling pathways by stimulating purified B cells with anti-IgM and measuring phosphorylation of tyrosines 519/520 in Syk, serine 473 in Akt, threonine 202 and tyrosine 204 in p44/42 Erk, and serine 32 in IB␣. However, no apparent changes were observed with the potency and duration of these signal pathways, and the protein levels of these signaling molecules were similar between WT and GABP␤2-deficient B cells (data not shown). The precise molecular mechanism by which GABP␤2 negatively regulates BCR signaling awaits further investigation.
Consistent with increased BCR-stimulated proliferation of GABP␤2-deficient B cells in vitro, GABP␤2-deficient mice displayed increased IgM and IgG1 antibody levels after ovalbumin immunization and a heightened GC response after challenge with SRBC. The increase is somewhat moderate, and a possible explanation is that only BCR-elicited responses were enhanced in the absence of GABP␤2, which may be blunted by similar responses derived from CD40-CD40L interaction. It is noteworthy that mice lacking GABP␤2 did not show apparent difference in antibody levels after a challenge with TNP-Ficoll. This can be explained by the somewhat modest proliferation of B cells observed in vivo after immunization with a T-independent polysaccharide antigen, and the restriction of expansion to the first few days after exposure (33). In contrast, antigen-selected B cells undergo marked proliferation for extended periods in GCs after T-dependent challenge (34), allowing for a longer period of time during which the absence of GABP␤2 can generate a higher number of antigen-specific clones. In sum-mary, our data revealed that GABP␤2 has both redundant and distinct roles in the immune system. Lack of GABP␤2 did not affect B cell development but showed increased B cell proliferation and humoral responses to protein antigens. Thus, manipulation of GABP␤2 expression may be a useful approach to modulate B cell responses without interfering with normal B cell development.