Interaction of the B Cell-specific Transcriptional Coactivator OCA-B and Galectin-1 and a Possible Role in Regulating BCR-mediated B Cell Proliferation*

OCA-B is a B cell-specific transcriptional coactivator for OCT factors during the activation of immunoglobulin genes. In addition, OCA-B is crucial for B cell activation and germinal center formation. However, the molecular mechanisms for OCA-B function in these processes are not clear. Our previous studies documented two OCA-B isoforms and suggested a novel mechanism for the function of the myristoylated, membrane-bound form of OCA-B/p35 as a signaling molecule. Here, we report the identification of galectin-1, and related galectins, as a novel OCA-B-interacting protein. The interaction of OCA-B and galectin-1 can be detected both in vivo and in vitro. The galectin-1 binding domain in OCA-B has been localized to the N terminus of OCA-B. In B cells lacking OCA-B expression, increased galectin-1 expression, secretion, and cell surface association are observed. Consistent with these observations, and a reported inhibitory interaction of galectin-1 with CD45, the phosphatase activity of CD45 is reduced modestly, but significantly, in OCA-B-deficient B cells. Finally, galectin-1 is shown to negatively regulate B cell proliferation and tyrosine phosphorylation upon BCR stimulation. Together, these results raise the possibility that OCA-B may regulate BCR signaling through an association with galectin-1.

OCA-B is a B cell-specific transcriptional coactivator that regulates immunoglobulin gene transcription in coordination with OCT factors (1)(2)(3)(4)(5)(6). Studies of OCA-B-deficient mice revealed that OCA-B has a broad role in regulating B cell development, acting in both antigenindependent and antigen-dependent processes (7)(8)(9)(10)(11)(12)(13)(14)(15)(16). Nevertheless, the most dramatic defect in mice lacking OCA-B expression is an inability of these mice to form germinal centers (GCs) 3 or to express normal levels of isotype-switched secondary immunoglobulins (7)(8)(9). The latter defect is due, at least in part, to reduced levels of transcription from normally switched Ig heavy chain loci rather than to defects in isotype switching per se, whereas OCA-B functions relevant to GC formation are still unclear. Although several new OCA-B target genes have been identified (17)(18)(19), none of these are fully responsible for the impaired GC formation in Oca-b Ϫ/Ϫ mice and as yet unidentified OCA-B target genes might be critical for this process (16). Moreover, other mechanisms might be also involved in OCA-B-mediated T cell-dependent immune responses. In fact, OCA-B has been shown to be essential for optimal BCR signaling, as B cell proliferation upon BCR stimulation is dramatically reduced in B cells lacking OCA-B (7). Consistent with these observations, reduced calcium influx and weaker and delayed substrate tyrosine phosphorylation after IgM stimulation are also observed in Oca-b Ϫ/Ϫ B cells (13,14). Therefore, OCA-B plays a positive role in BCR signaling and B cell activation. Insufficient activation of Oca-b Ϫ/Ϫ B cells, along with other defects such as reduced expression of BLR1 and a consequent poorer homing ability (18), could result in impaired GC formation. The molecular mechanism for OCA-B-mediated BCR signaling is unknown. It is possible that OCA-B might modulate BCR signal transduction indirectly through its target genes. Additionally, an effect of OCA-B outside of the nucleus could be also involved in B cell responses to environmental stimuli and, thus, GC formation.
Related to the last possibility is our recent description of a p35 OCA-B isoform that is generated from the single OCA-B mRNA via an alternative translation start site (20). Interestingly, unlike the p34 isoform of OCA-B that is exclusively nuclear and functions as a transcriptional coactivator, p35 can be found both in the cell nucleus and in the cytosol and/or on the cell membrane and may represent a novel type of signaling molecule in B lymphocytes. This distinct property of p35 appears to result, at least in part, from the myristoylation of its N-terminal glycine residue, which represents a surprising and unprecedented modification of a transcription factor (20). Therefore, it is probable that a novel mechanism mediated by OCA-B/p35 is involved in OCA-B-regulated B cell proliferation and differentiation.
Recently, members of the galectin family of lectins have been found to play regulatory roles in the development of immune cells and to influence various immune response events. Among the 14 members, a 14-kDa polypeptide called galectin-1 has been shown to be expressed both in B lymphocytes and in other cell types (21). Galectin-1, like other family members, contains a conserved carbohydrate recognition domain (CRD) and has affinity for ␤-galactosides. Lacking typical transmembrane segments and secretion signal peptides, galectin-1, like other members in this family, is secreted by a nonconventional pathway and binds to various cell surface receptors to mediate cell-cell and cell-matrix adhesion, to control cell growth or to induce apoptosis (22,23). The intracellular galectin-1 is found mainly in the cytosol, although nuclear localization of galectin-1 can also be observed (24). Changes of galectin-1 localization within subcellular compartments have been shown to correlate with the developmental and differentiation status of cells (25).
Galectin-1 has been demonstrated to regulate T cell homeostasis and function (26,27). However, a possible galectin-1 function in the B cell compartment is largely unexplored. As in other cell types, B cell galectin-1 may have both autocrine and paracrine functions. In support of this idea, it has been shown that galectin-1 can be secreted by activated B cells and that it can regulate T cell function (28). Stromal cell surface galectin-1 is crucial for pre-BCR signaling and acts through interaction with the surrogate light chain of the pre-BCR (29). Another B cell surface receptor for galectin-1 is the glycoprotein CD45, a tyrosine phosphatase that is important for positive regulation of BCR signaling cascade and B cell proliferation and maturation (30). By binding to CD45, galectin-1 can modulate its phosphatase activity and, therefore, influence the downstream Src-kinase activity (Lyn in pre-B cells) and B cell proliferation (31). Recombinant galectin-1 has also been reported to induce apoptosis of transformed B cells but not normal B cells (32,33).
In addition to galectin-1, the galectin family member galectin-3 was recently found to be expressed by activated murine B cells (34). Galectin-3 function is critical for signaling during IL-4-induced survival and differentiation of B cells (34). In normal human B cells, galectin-3 expression varies among specific B-cell developmental stages, with the lowest level in GC B cells and higher levels in naïve and memory B cells (35). Galectin-3 and galectin-1 share similar chemical properties. However, unlike galectin-1, which is a proapoptotic factor for B lymphocytes, intracellular galectin-3 exhibits an anti-apoptotic activity that likely is mediated by interactions with Bcl-2 and other proteins implicated in apoptosis (i.e. Fas and Syntexin) (36).
In efforts to understand the role of OCA-B in BCR-mediated signal transduction, an unbiased yeast two-hybrid-based screen for cytoplasmic partners of OCA-B identified an OCA-B association with galectin-8. An extension of this study revealed interactions of OCA-B with the related CRD domain-containing polypeptides galectin-1 and galectin-3 as novel OCA-B interacting proteins in B lymphocytes. An examination of galectin-1 expression and secretion levels in Oca-b Ϫ/Ϫ B cells revealed increased levels of total and surface galectin-1 in comparison with wild-type B cells. Along with the demonstration of a modest, but nonetheless significant, reduction of CD45 phosphatase activity in Ocab Ϫ/Ϫ B lymphocytes and the ability of recombinant galectin-1 to reduce BCR stimulation-mediated growth of B lymphocytes, these observations raise the possibility of BCR regulation by OCA-B through galectin-1 and its receptors, including CD45, on the B cell surface.

MATERIALS AND METHODS
Animals-10 -14-Week-old Oca-b Ϫ/Ϫ mice (7) and C57BL/6J mice (The Jackson Laboratory) were used for all experiments. All mice were maintained under specific pathogen-free conditions at the Laboratory Animal Research Center within The Rockefeller University.
Yeast Two-hybrid Screening and Mammalian Two-hybrid Analysis-Yeast two-hybrid screening was performed as previously described (37). The expression plasmid encoding the Cyto-trap bait was generated by inserting the cDNA sequences for OCA-B (1-140 amino acids) into pSos. Saccharomyces cerevisiae strain Cdc25H was transformed sequentially with pSos-OCA-B and a murine spleen cDNA library (Stratagene).
The mammalian expression vectors encoding hybrid proteins were produced by inserting OCA-B/p35 sequences into the pCMV-GAL4 vector or by inserting galectin-1 cDNA sequences into the pVP-FLAG7 vector. A mammalian two-hybrid assay was conducted in 293T cells as described previously (37).
Protein-Protein Interaction Assay-Recombinant GST fusion or HIStagged proteins were expressed in bacterial cells and immobilized on glutathione or nickel-nitrilotriacetic acid beads. Then 6 l of proteinconjugated beads (containing ϳ6 g of protein) was incubated for 3 h at 4°C with 5 or 10 l of reticulocyte lysate containing 35 S-labeled proteins or with 100 ng of purified HIS-galectin-1 in a final volume of 200 l of BC150 (150 mM KCl, 20% glycerol, 0.2 mM EDTA, pH 8.0, 20 mM Tris-HCl, pH 7.9) containing 2.5 mM dithiothreitol, 0.1% Nonidet P-40, 0.5 mg/ml bovine serum albumin, and protease inhibitors. In some cases, recombinant proteins were coupled to lactosyl-agarose beads. Equal amounts of protein-conjugated beads (containing ϳ20 g of proteins) were incubated for 4 h at 4°C with 150 g of primary B cell lysate in a final volume of 500 l of BC150 containing 0.1% Nonidet P-40, 4 mM ␤-mercaptoethanol, and protease inhibitors. The beads were washed five times with incubation buffer, boiled in 10 l of the 2ϫ SDS loading buffer, and analyzed by autoradiography.
In Vivo Cross-linking and Co-immunoprecipitation-Namalwa cells were treated with 2.5 mM dithiobis(succinimidyl propionate) (DSP) for 25 min at room temperature and a total cell lysate in RIPA buffer (10 mM Hepes, pH 7.4, 1 mM EDTA, pH 7.4, 300 mM NaCl, 1% Nonidet P-40, 0.1% SDS, 0.1% sodium deoxycholate, 2 mM dithiothreitol, and protease inhibitors) was prepared. An equivalent amount of lysate in RIPA buffer from uncross-linked cells was also included as a control. Immunoprecipitation was performed with anti-OCA-B antibody.
For transfection and co-immunoprecipitation assays, 293T cells were co-transfected with expression vectors encoding FLAG-galectin-1 and OCA-B/p35 (either wild-type or mutant with a glycine to alanine change at the N-terminal myristoylation site) (20). In some cases, the parental expression vector was used as a control. Cell lysates were prepared 48 h post-transfection in BC300 (300 mM KCl, 20% glycerol, 0.2 mM EDTA, pH 8.0, 20 mM Tris-HCl, pH 7.9) containing 0.1% Nonidet P-40, 4 mM ␤-mercaptoethanol, and protease inhibitors. The lysates then were subjected to immunoprecipitation with M2-agarose. Proteins that specifically bind to M2-agarose beads were eluted with FLAG peptides and analyzed by immunoblotting with the indicated antibodies.
For tyrosine phosphorylation analysis, cells were treated as indicated and washed with PBS, 1 mM vanadate. Cell lysates were prepared in RIPA buffer containing 1 mM vanadate and 1 mM NaF. Tyrosine phosphorylation was evaluated with anti-phosphotyrosine (4G10, Upsate Biotechnology). Blots were then stripped and probed with ERK-specific antibody (Santa Cruz).
For secretion assays, isolated B cells were cultured in Iscove's modified Dulbecco's medium containing 50 mM ␤-mercaptoethanol, 2 mM L-glutamine, and ITS supplement (Sigma). Cell viability (Ͼ90%) was evaluated by trypan blue analysis. Cultured supernatants were collected after 12 h and concentrated 30-fold by centrifugation. Equal amounts of total protein from each supernatant were used for immunoblot analysis. Values were quantified with ImageQuant software.
To evaluate surface-bound galectin-1, cells were washed with PBS three times and then incubated in PBS, 200 mM lactose on ice for 1 h. Galectin-1 levels were analyzed with anti-galectin-1 (R&D Systems).
CD45 Phosphatase Assay-Purified splenic B cells were subjected to CD45 phosphatase assay as described with minor modifications (38). In brief, T-depleted B cells were lysed with buffer containing 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Brij 97 (Sigma), 0.5 mM dithiothreitol and protease inhibitors. Lysates were subjected to immunoprecipitation with anti-CD45 antibody (RA3-6B2, BD Biosciences) and immune complexes were incubated at room temperature for 12 h in 500 l of reaction mixture (100 mM Hepes, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, and 10 mM p-nitrophenyl phosphate). Reactions were terminated by addition of 500 l of 1 N NaOH. Products were quantified by measuring absorbance at 405 nm.
Cell Growth Analysis-T-depleted B cells from wild-type mice were further purified by discontinuous Percoll gradient as described (7) and seeded in 96-well plates at 3 ϫ 10 5 cells/well. Cells were pre-treated with the indicated amounts of galectin-1 (Research Diagnostics Inc., Flanders, NJ) for 4 h at 37°C. The indicated inducers were then introduced into culture medium and cells were subjected to 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assays after a 40-h incubation at 37°C according to manufacturer's protocol (Sigma).

RESULTS
Our previous results showed that OCA-B/p35 is located in both cytoplamic/membrane and nuclear compartments and suggested that it might function as a novel signaling molecule in B lymphocytes (20). To understand the role of OCA-B/p35 in BCR-mediated signal transduction, we sought to identify proteins that bind to cytoplasmic OCA-B in B cells. We employed a yeast two-hybrid (SRS) screening method using the mouse OCA-B/p35 N-terminal domain (140 amino acids) as a bait. After several rounds of screening, we isolated seven positive clones from approximately 10 million transformants of a cDNA library derived from mouse splenocytes. Binding of proteins encoded by these clones to the full-length OCA-B was checked by back transformation of each into the reporter yeast strain containing full-length OCA-B/p35 as a bait. A cDNA clone encoding the C-terminal region (143 amino acids) of mouse galectin-8 showed strong and specific interaction with OCA-B/ p35 in yeast.
Because there have been no reports of galectin-8 expression in B lymphocytes and because our reverse transcriptase-PCR analyses failed to detect galectin-8 mRNA either in primary murine B cells or in most B cell lines (data not shown), these results may reflect an association of OCA-B with other galectins that are expressed in B lymphocytes. In fact, the related galectin-1 is expressed by activated B cells (28) and by B cells at various developmental stages (39), suggesting a function in B lymphopoiesis. To examine an intracellular interaction of galectin-1 and OCA-B, we used a mammalian two-hybrid assay with vectors expressing murine galectin-1 fused to the VP16 transcription activation domain and OCA-B fused to the GAL4 DNA-binding domain. 293T cells were transfected with these vectors, an internal control plasmid, and a reporter containing five GAL4 binding sites fused to the luciferase gene. As shown in Fig. 1A, whereas the coexpression of GAL4 with VP16 or VP16-galectin-1 fusion proteins resulted in basal reporter activity, an 8-fold activation of reporter gene was observed when VP16 and GAL4-OCA-B (p35) were coexpressed. This effect is largely due to the transactivation activity of OCA-B when fused to the GAL4 DNA-binding domain (40). Nevertheless, a more dramatic induction of reporter gene activity (ϳ40-fold) was achieved with the coexpression of GAL4-OCA-B and VP16-galectin-1 in 293T cells, indicating that OCA-B interacts with galectin-1 in mammalian cells. To confirm this interaction, we performed GST pulldown assays by incubating in vitro translated [ 35 S]galectin-1 with immobilized GST-OCA-B fusion protein or GST alone. In this assay, [ 35 S]galectin-1 bound to GST-OCA-B but not to GST (Fig. 1B), consistent with a direct interaction between OCA-B and galectin-1. Similarly, a purified HIS-tagged galectin-1 (expressed in bacteria) bound to GST-OCA-B but not to GST alone (Fig. 1C), strongly suggesting a direct protein-protein association between OCA-B and galectin-1.
To document a possible interaction between OCA-B and galectin-1 in B lymphocytes, whole cell extracts from Namalwa B cells were subjected to immunoprecipitation with anti-galectin-1 or control antibody. As shown by the immunoblot analysis in Fig. 1D, OCA-B, but not the unrelated protein RPC39, was precipitated from Namalwa whole cell extracts by anti-galectin-1, but not control antibody, confirming an association of these two polypeptides in B cells. In a parallel analysis involving co-immunoprecipitation with anti-OCA-B antibody, cells were incubated with the bifunctional cross-linker DSP prior to extract preparation and co-immunoprecipitation. This analysis showed galectin-1 (predominantly monomer) immunoprecipitation with the OCA-B-specific antiserum, but not with preimmune serum or with OCA-Bspecific antiserum in the absence of DSP treatment, and there was no co-immunoprecipitation of unrelated polypeptide (RPC39) under any conditions (Fig. 1E). The observation that DSP cross-linking is required for visulization of an intracellular OCA-B-galectin-1 interaction with anti-OCA-B co-immunoprecipitation, but not with anti-galectin-1 coimmunoprecipitation, may simply reflect the relative efficacies of the two antibodies in recognizing their epitopes in the OCA-B-galectin-1 complex and/or a difference in stability following antibody conjugation. Similarly, we observed co-immunoprecipitation of OCA-B and galectin-1 from BJAB whole cell extracts by anti-galectin-1 (data not shown). To gain further evidence for an intracellular OCA-B/galectin-1 interaction, immunofluorescent staining analyses were conducted with the Namalwa B cell line and with primary murine B cells. Staining of endogenous OCA-B and galectin-1 showed co-localization of these two polypeptides in the cytoplasmic compartment, suggesting that galectin-1 interacts primarily with cytoplasmic OCA-B in B lymphocytes ( Fig. 1F and data not shown). Altogether, these data clearly show that galectin-1 is a novel OCA-B-interacting protein in B cells and that these two proteins interact directly in vivo.
A recent study has reported expression of galectin-3 in activated B cells and a role in mediating IL-4-induced B cell survival and differentiation (34). Therefore, we also investigated a possible interaction between OCA-B and galectin-3 in B cells. Galectin-3 expression was barely detectable by immunoblot in the B cell lines tested (supplemental Fig. S1A), whereas a low level galectin-3 expression was observed in LPS-activated B cells (supplemental Fig. S1B). Nevertheless, recombinant galectin-3 was found to bind OCA-B as strongly as recombinant galectin-1 (supplemental Fig. S1C, compare lanes 2 and 3). Control experiments failed to detect an OCA-B-galectin-3 association in lysates from OCA-B-null splenic B cells or binding of OCA-B to lactosyl-agarose beads (supplemental Fig. S1C, lanes 1 and 4). Hence, it is possible that OCA-B also interacts with galectin-3 in activated B cells.
Previous results indicated that galectin-1, like several other OCA-Binteracting proteins, bind to N-terminal sequences of OCA-B. To further dissect the interaction between galectin-1 and OCA-B, we constructed a series of vectors expressing N-and C-terminal-truncated mutants of OCA-B fused (at the C terminus) to GFP (Fig. 2A). The recombinant HIS-galectin-1 fusion proteins were expressed in bacteria, purified by affinity chromatography, and immobilized on beads. Interaction assays with in vitro translated [ 35 S]OCA-B-GFP fusion proteins and HIS-galectin-1-conjugated beads showed that the N-terminal 7-43 amino acids of OCA-B are sufficient for galectin-1 binding (Fig. 2B).
Thus, the myristoylation motif that is unique to OCA-B/p35 is not essential for OCA-B interaction with galectin-1 in vitro, although a stimulatory effect is not excluded.
In a similar experiment, and as a control, in vitro translated [ 35 S]OCA-B-(7-140) or [ 35 S]OCA-B-(143-263) were co-incubated with HIS-galectin-1-conjugated beads or beads alone. Consistent with previous results, the N-terminal sequences of OCA-B interacted strongly with galectin-1 but not with control beads. In contrast, the C-terminal sequences of OCA-B showed no detectable binding to galectin-1 in this assay (Fig. 2C).
In an effort to test the role of the OCA-B/p35 myristoylation site in its association with galectin-1 in vivo, we co-expressed FLAG-galectin-1 and OCA-B/p35 (wild-type or G8A mutant, which carries a mutation at the myristoylation site (20)) in 293T cells. Remarkably, the co-immunoprecipitaion results showed that wild-type OCA-B/p35, but not OCA-B/p35G8A bearing mutation at the myristoylation site, specifically associates with galectin-1 in cells (Fig. 2D). The myristoylation requirement for an OCA-B/p35 interaction with galectin-1 in transfected cells may reflect the fact (i) that FLAG-galectin-1 is mainly cytoplasmic (data not shown); (ii) that the binding of OCA-B and galectin-1 occurs predominantly in the cytoplasm (Fig. 1F); and (iii) that myristoylation of OCA-B/p35 is important for its cytoplasmic localization (20).  Nonetheless, it is possible that galectin-1 protein expression and secretion are altered in Oca-b Ϫ/Ϫ B cells. Therefore, we examined galectin-1 expression in primary B cells isolated from wild-type and Oca-b Ϫ/Ϫ mice by immunoblot analysis. We also evaluated the levels of galectin-1 in response to various stimuli in these cells. As has been shown by others (28), low level galectin-1 expression can be detected in normal B cells (Fig. 3A, lane 1). Galectin-1 expression was significantly induced (Ͼ2fold) when B cells were activated by anti-, or by anti-, IL-4 and anti-CD40 (Fig. 3, A, lanes 2 and 3, and B). As expected (41), OCA-B expression was also induced in activated B lymphocytes, whereas OCT-1 expression remained constant (Fig. 3A, lanes 1-3 in middle and lower  panels). In contrast, galectin-1 was expressed at a higher (Ͼ3-fold) level in Oca-b Ϫ/Ϫ B cells than in control B cells, and was further induced in these cells by anti-, IL-4, and anti-CD40 but not by anti-alone (Fig. 3,  A, lanes 4 -6 in upper panel, and B). As controls, the OCT-1 level was comparable in wild-type and Oca-b Ϫ/Ϫ B cells and B cell activation did not alter the level of OCT-1 expression in Oca-b Ϫ/Ϫ B cells (Fig. 3A,  lanes 4 -6 in lower panel).
Similar experiments were conducted to investigate levels of galectin-1 expression in murine primary B cells after they were subjected to combined stimulation (anti-, IL-4, and anti-CD40) for various periods of time. Consistent with our previous results (Fig. 3, A and B We also analyzed galectin-3 expression in primary B lymphocytes under the same conditions. As expected, and consistent with published results (34), galectin-3 expression was barely detectable in resting B cells, whereas a dramatic increase in galectin-3 expression was observed when primary B cells were subjected to LPS stimulation. However, the level of galectin-3 expression did not show any significant increase in B cells following activation by antior combined stimuli (anti-, IL-4, and anti-CD40). Moreover, in contrast to what was observed for galectin-1, the levels of galectin-3 expression in resting and activated Ocab Ϫ/Ϫ B cells were comparable with their levels in wild-type B cells (supplemental Fig. S1B, panel b). As controls, levels of OCA-B, actin, and galectin-1 on the same blot were also shown (supplemental Fig. S2B,  panels a, c, and d, respectively).
Thus far, galectin-1 functions in regulating immune cell activity have been shown to rely on the ability of these cells to secrete galectin-1 and/or to bind to galectin-1 with cell surface receptors. Therefore, we employed a serum-free culture system to measure levels of galectin-1 secretion from primary B lymphocytes. After 12 h of culture in vitro, cell-free supernatants were concentrated and galectin-1 levels were determined by immunoblot. As shown in Fig. 4, A and B (densitometric profiles of Fig. 4A), secreted galectin-1 was found in supernatants of both wild-type and Oca-b Ϫ/Ϫ B  1 and 4). Galectin-1 secretion was increased when these cells were subjected to stimulation (lanes 2, 3, 5, and 6). In comparison to wildtype B cells, Oca-b Ϫ/Ϫ B cells showed higher levels of secreted galectin-1 both prior to and following B cell activation (Fig. 4B). The observed increase is comparable in magnitude to the increases in the levels of galectin-1 expression in these cells (Fig. 3). In contrast, no galectin-3 secretion was detected under the same condition (supplemental Fig. S2, lanes 3-5 and  8 -10). Interestingly, however, secreted galectin-3 was observed in supernatants of LPS-activated B cells, and Oca-b Ϫ/Ϫ B cells showed a higher level of galectin-3 secretion than did wild-type B cells upon LPS stimulation (supplemental Fig. S2, lanes 2 and 7), although the intracellular level of galectin-3 is comparable in wild-type and Oca-b Ϫ/Ϫ B cells (supplemental Fig. S1B, lanes 2 and 6).

cells (lanes
To study the cell surface expression of galectin-1 by FACS analysis, splenocytes were isolated from wild-type and Oca-b Ϫ/Ϫ mice. T cells were removed by binding to Thy1.2-magnetic beads. Cells were then labeled with anti-galectin-1 and anti-B220 antibodies. As shown in Fig.  4C, a slight but clear increase in cell surface galectin-1 expression was observed in B220 ϩ B cells from Oca-b Ϫ/Ϫ mice relative to B220 ϩ B cells from wild-type mice. The percentages of galectin-1 ϩ IgM ϩ B cells in wild-type versus Oca-b Ϫ/Ϫ mice were also determined by FACS analysis, and the results revealed a higher (ϳ2-fold) level of galectin-1 ϩ IgM ϩ B cells in mice lacking OCA-B expression (Fig. 4D).
We performed FACS analysis to further study the distribution of galetin-1 ϩ B cells within various B cell compartments. Splenic B cells can be further divided into transitional 1 (T1, IgM high IgD Ϫ CD21 Ϫ ), transitional 2 (T2, IgM high IgD ϩ CD21 bright ), marginal zone (MZ, IgM high IgD Ϫ CD21 bright ), and follicular mature (MT, IgM ϩ IgD ϩ CD21 ϩ ) B cell subsets based on expression of cell surface markers IgM, IgD, and CD21 (42). By co-staining for cell surface galectin-1 and the above mentioned markers, we found that most galectin-1 ϩ cells were follicular mature B cells in both wild-type and Oca-b Ϫ/Ϫ mice. In contrast, few MZ B cells exhibited surface galectin-1 expression (Fig. 5, A and B). Relative to the corresponding wild-type B cell populations, Oca-b Ϫ/Ϫ T1, T2, and MZ compartments showed a slight increase of surface galectin-1 expression (Fig. 5C). Significantly, a subpopulation of follicular mature B cells bearing high levels of cell surface galectin-1 was found from Oca-b Ϫ/Ϫ mice (Fig. 5C, indicated by  arrow in panel MT). Therefore, our results demonstrate higher surface galectin-1 expression on a population of mature B cells lacking OCA-B expression.
Because it has been reported that galectin-1 can modulate B cell signal transduction by binding to CD45 (Ref. 31 and supplemental Fig. S3), and because Oca-b Ϫ/Ϫ B cells express an increased level of galectin-1 (Figs. [3][4][5] while showing reduced BCR signaling (7), we tested galectin-1 effects on the B cell response to BCR-induced proliferation in an in vitro culture system. Resting B cells were isolated from wild-type mice by Percoll gradient separation and cultured in vitro in the presence (squares) or absence (circles) of anti-. In some cultures, the indicated amounts of recombinant galectin-1 also were included. B cell growth was evaluated by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay 40 h later. Consistent with published results (28,32), galectin-1 did not induce B cell apoptosis during this culture period (Fig.  6A, circles). Significantly, galectin-1 showed a negative effect on anti-IgM induced wild-type B cell proliferation (Fig. 6A, squares). As a control, the growth of Oca-b Ϫ/Ϫ B cells was also evaluated in the same conditions. As expected (7), these cells responded poorly to anti-IgM stimulation, and addition of galectin-1 to the culture system had no significant effect on anti-IgM-induced Oca-b Ϫ/Ϫ B cell proliferation in this assay (Fig. 6A).
Because CD45 is a galectin-1 receptor whose function is negatively regulated by galectin-1 interaction (31), CD45 phosphatase activity in wild-type and Oca-b Ϫ/Ϫ B cells was assessed. Thus, equal amounts of murine splenic B cell lysates from wild-type and Oca-b Ϫ/Ϫ mice were subjected to immunoprecipitation with CD45-specific monoclonal antibody. The phosphatase activity of the immune complex was deter-mined using p-nitrophenyl phosphate as substrate. As shown in Fig. 6B, Oca-b Ϫ/Ϫ B cells had a slightly lower, but statistically significant (p Ͻ 0.05), level of CD45 phosphatase activity than did wild-type B cells. This is consistent with the finding that only a subpopulation of Oca-b Ϫ/Ϫ B cells bears a higher level of surface-bound galectin-1 (Fig. 5C). The specificity of the assay was confirmed by showing that addition of sodium vanadate (a tyrosine phosphatase inhibitor) to the reaction mixture completely eliminated the phosphatase activity (data not shown).
To confirm the binding of galectin-1 with murine primary B cell surface CD45, we first carried out in vitro interaction assays. As shown in supplemental Fig. S3A, the galectin-1-conjugated lactosyl-agarose beads can specifically bind CD45 from B cell lysates of both wild-type and Oca-b Ϫ/Ϫ splenocytes. Immunofluorescent staining of B cell surface CD45/B220 and galectin-1 showed that these two polypeptides co-localize to certain areas on the surface of some murine primary B cells (supplemental Fig. S3B). Thus, as seen in assays with a B cell line (31), CD45 on murine primary B cells can function as a galectin-1 receptor. Because CD45 phosphatase activity is required for BCR signaling, our data raise the possibility that galectin-1 may influence the B cell proliferation response to BCR cross-linking, at least in part, by modulating CD45 phosphatase activity. We also evaluated effects of ectopic galectin-1 on B cell responses to BCR cross-linking by comparing tyrosine phosphorylation patterns in galectin-1-treated wild-type and Oca-b Ϫ/Ϫ B cells. Our results show that Oca-b Ϫ/Ϫ B cells respond weakly to anti-IgM stimulation. This difference is more obvious at 15 min after BCR cross-linking (compare lanes 9 and 11 in Fig. 6, C and D). This observation is consistent with results of our studies with pre-B cells 5 and of studies with marginal zone B cells (13). To investigate the possible role of galectin-1 in lowering the Oca-b Ϫ/Ϫ B cell response to anti-IgM stimulation, we treated wild-type primary B cells with ectopic recombinant galectin-1. Immunoblot analysis revealed that galectin-1-treated wildtype B cells, compared with Oca-b Ϫ/Ϫ B cells, bear similar levels of surface-bound galectin-1 (Fig. 6E). A tyrosine phosphorylation analysis showed that galectin-1-treated wild-type B cells have a globally reduced level of phosphorylated tyrosine relative to that of untreated wild-type B cells after BCR cross-linking (Fig. 6, C and D, compare lanes 5 and 6 and lanes 9 and 10). In addition, and importantly, the overall tyrosine phosphorylation pattern of wild-type cells treated with ectopic galectin-1 is very similar to that of Oca-b Ϫ/Ϫ B cells, especially 15 min after anti-IgM stimulation (Fig. 6, C and D, lanes 10 and 11). This indicates that surfacebound galectin-1 affects BCR-mediated signal transduction, and this might be one of the molecular mechanisms accounting for the difference in BCR signaling between wild-type and Oca-b Ϫ/Ϫ B lymphocytes.

DISCUSSION
We previously reported the identification of a cytoplasmic/membrane isoform (p35) of the OCA-B coactivator and proposed that it might be a novel signaling molecule to regulate BCR function in B lymphocytes (20). In further exploration of the role of OCA-B/p35 in BCR signal transduction, a yeast two-hybrid assay that identified galectin-8 as an OCA-B-interacting protein was followed by additional studies that identified galectin-1 (and galectin-3) as a novel OCA-B binding partner. Complementary in vitro interaction, mammalian two-hybrid, co-immunoprecipitation, and immunofluorescence assays confirmed an association of galectin-1 and OCA-B both in vitro and in vivo. Compar-FIGURE 6. Roles of galectin-1 in regulating B cell proliferation. A, resting B lymphocytes were purified from spleens of wild-type mice. 5 ϫ 10 3 cells were seeded into each well and cells were pretreated with the indicated amounts of recombinant galectin-1 or with control PBS buffer at 37°C for 4 h. Then the indicated stimuli were added into cell cultures. B cell proliferation was analyzed after 40 h using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide method (following the manufacturer's protocol). B, whole cell extracts of purified B lymphocytes from wild-type and Oca-b Ϫ/Ϫ spleens were subjected to immunoprecipitation with anti-CD45 antibody. The phosphatase activity in the immune complexes was determined using p-nitrophenyl phosphate as substrate. * indicates p Ͻ 0.05. ** indicates p Ͻ 0.01. C, galectin-1 effects on BCR-stimulated tyrosine phosphorylation in B lymphocytes. C, purified resting B lymphocytes were treated with 10 g/ml galectin-1 for 3 h. Unbound galectin-1 was washed out and cells, together with the corresponding untreated cells, were subjected to anti-IgM treatment as indicated. Western blot analysis was performed to evaluate tyrosine phosphorylation responses under these conditions. The total ERK level in these cells was also determined by immunoblot as control. Data represent one of the two independent experiments. D, quantities of the tyrosine phosphorylation level of C using the total ERK kinase level as control. E, levels of surface-bound galectin-1 determined by Western blot.
isons of the levels of galectin-1 in wild-type versus Oca-b Ϫ/Ϫ murine splenic B cells by immunoblot and FACS analysis revealed that galectin-1 levels, including intracellular galectin-1 and secreted and surfacebound galectin-1, are significantly increased in B cells lacking OCA-B. Most galectin-1-positive B cells in the spleen are mature B cells, whereas very few marginal zone B cells exhibit surface galectin-1 expression. In Oca-b Ϫ/Ϫ mice, a subpopulation of mature B cells, which is missing in the wild-type mature B cell compartment, has a high level of surface galectin-1. In a more detailed analysis of the effect of changes in galectin-1 levels, recombinant galectin-1 showed a selective negative effect on BCR-induced B cell proliferation. This effect could involve an association of galectin-1 with CD45 and modification of CD45 phosphatase activity because Oca-b Ϫ/Ϫ B cells showed lower overall CD45 activity relative to wild-type B cells. However, this pathway could be important only for a subpopulation of B cells, and additional galectin-1 receptors on B cell surfaces might also be involved in this process. Nonetheless, our results raise the possibility that OCA-B could help regulate BCR signaling by its interaction with galectin-1.
Because galectin-3 has recently been shown to be expressed in activated B cells (34), we also tested the interaction of OCA-B and galectin-3. In vitro binding assays clearly showed that galectin-3 binds OCA-B as strongly as galectin-1. This is not surprising because OCA-B also binds other galectins (such as galectin-8) through their CRD domains. Although we could not detect any significant increases in galectin-3 expression in activated B cells stimulated by anti-or by combined stimuli (anti-, IL-4, and anti-CD40) in our culture system, we did observe an induction of galectin-3 expression following LPS stimulation. However, the levels of galectin-3 expression in B cells were relatively low. The difference between our results and recently published results (34) could result from different experimental systems and/or reagents used in these assays. Nevertheless, it is very possible that OCA-B also interacts with galectin-3 in vivo under certain conditions and that this interaction may be involved in certain signaling pathways mediated by OCA-B.
OCA-B interacts with galectin-1 through the N-terminal 7-43 amino acids (Fig. 2, A and B), suggesting that the myristoylation of OCA-B/p35 is not essential for association of OCA-B with galectin-1 in vitro. However, and interestingly, the co-immunoprecipitation results from transfected 293T cells show that mutation of the OCA-B/ p35 myristoylation site abolish the interaction of galectin-1 in cells (Fig.  2D), indicating that myristoylation of OCA-B/p35 is required for association of OCA-B with galectin-1 in vivo. The effect of myristoylation on the OCA-B/p35-galectin-1 interaction may reflect the important role of this modification in regulating the subcellular localization of OCA-B in cells (20). This is also consistent with our observation that co-localization of OCA-B and galectin-1 occurs predominantly in the cytoplasmic compartment of Namalwa B cells (Fig. 1F). The interaction of OCA-B and galectin-1 appears to be direct because we can detect binding of bacterially expressed galectin-1 and bacterially expressed OCA-B and because the presence of lactose did not affect the co-immunoprecipitation of OCA-B and galectin-1 from B cell lysates (data not shown). This is not surprising because the interaction of galectin-1 with Gemin4, another intracellular partner for galectin-1, also does not involve carbohydrates (43). Nonetheless, there exists the possibility that saccharides might be able to strengthen the interaction between OCA-B and galectin-1. Although the p34 and p35 isoforms of OCA-B both can be coimmunoprecipitated with galectin-1 from B lymphocytes, this may relate to the experimental procedures involved in co-immunoprecipitation, the possibility of dissociation and reassembly of galectin-1 with both isoforms. However, it also may be physiologically relevant because galectin-1, whereas predominantly cytoplasmic, can be detected in B cell nuclei (data not shown).
Galectin-1 can exist in both monomeric and homodimeric forms in cells (44). Our results showed that dimerization is not required for galectin-1 binding to OCA-B, because a purified galectin-1 monomer can bind OCA-B in in vitro pulldown assays (Fig. 1C) and because one CRD domain is sufficient for OCA-B binding in yeast two-hybrid analysis. Nonetheless, further experiments are needed to distinguish which form of galecitn-1, if any, preferentially binds OCA-B in vivo.
The BCR plays a central role in B cell differentiation and maturation. Besides the core components of the BCR complex, additional factors have been found to be important for efficient BCR signaling (7). These factors include both positive regulators (e.g. CD45) and negative regulators (e.g. CD22). Oca-b Ϫ/Ϫ B cells show an impaired response to BCRinduced proliferation. Although the overall tyrosine phosphorylation pattern in Oca-b Ϫ/Ϫ B cells after BCR cross-linking is similar to that of wild-type B cells, the response is weaker and delayed in these cells (13). Moreover, the calcium influx response to anti-IgM stimulation is significantly lower in Oca-b Ϫ/Ϫ B cells (45). These observations suggest that OCA-B plays a positive role in BCR signal transduction. In support of this idea, disruption of CD22 expression in Oca-b Ϫ/Ϫ B cells can partially rescue the proliferation defects and restore the normal calcium influx response of these cells upon BCR stimulation (45), suggesting that the imbalance between positive and negative regulators of BCR signaling in the absence of OCA-B is re-established by reducing CD22 expression. The positive role of OCA-B in BCR signaling could also be due to its ability to regulate transcription of target genes (i.e. Lck) whose function is related to the B cell response to BCR stimulation (46). Our finding that galectin-1 is a partner for OCA-B suggests additional mechanisms for OCA-B function in BCR signaling.
Galectin-1 bears multiple biological activities and has been reported to be involved in many physiological processes, such as cell growth and apoptosis, cell adhesion, immune responses and inflammation, and mRNA splicing (21)(22)(23). Although galectin-1 function in the induction of apoptosis of activated T cells has been well studied (27), only recently has there an appreciation of the expression and function of galectin-1 in B cells and macrophages. Galectin-1 mRNA can be detected in various developmental stages of B cells (39), suggesting a broad role for this molecule in B lymphopoiesis. Mature B cells have the lowest level of galectin-1 mRNA and the level increases when B cells are activated (39). In fact, recent data have shown that galectin-1 expression and secretion are induced in activated B cells (28), and our current results confirm this observation. Interestingly, the levels of galectin-1 in Oca-b Ϫ/Ϫ B cells are dramatically increased (Fig. 3). Because OCA-B does not affect galectin-1 gene transcription in B cells, 6 OCA-B might modulate galectin-1 expression post-transcriptionally or regulate galectin-1 degradation, such that lack of OCA-B expression results in increased accumulation of galectin-1 in B cells. For instance, Siah1 might also mediate galectin-1 degradation because both proteins interact with OCA-B and because the Siah1-interacting domain in OCA-B is different from the galectin-1-interacting domain (47,48). In addition and more interestingly, the secretion of galectin-1 from Oca-b Ϫ/Ϫ B cells is also higher (Fig. 4, A and B), implying that the interaction of OCA-B with galectin-1 in B lymphocytes may modulate galectin-1 secretion from these cells. In fact, we have observed an interaction of galectin-1 and OCA-B with tubulin in B cells, supporting the idea that OCA-B may affect the trafficking of galectin-1 in these cells (data not shown). Recent data have suggested that the CRD domain of galectin-1, by binding to ␤-galacto-sides on its counter-receptors, mediates galectin-1 secretion in Chinese hamster ovary cells (49). Therefore, it is plausible to speculate that OCA-B may regulate galectin-1 secretion in B cells by interacting with the CRD domain of galectin-1. Although galectin-1 can function as a monomer or homodimer, depending on the cell type and/or the biological process, our analyses employed conditions under which all galectin-1 would be present as monomers such that total galectin-1 levels in cells would be scored.
More importantly, we found that cell surface galectin-1 levels are increased in OCA-B knock-out mice (Figs. 4C, 5, and 6E), consistent with an increased level of galectin-1 secretion in Oca-b Ϫ/Ϫ primary B cells. In addition, Oca-b Ϫ/Ϫ B cells may have a surface receptor profile that is distinct from that of wild-type B cells and that provides more galectin-1 binding sites. In fact, our recent cDNA microarray study has revealed that OCA-B is required for optimal transcription of B4galt1 (␤1,4-galacosyltransferase) and thus may affect carbohydrate modification of certain molecules (such as IgG) (46). One B cell surface receptor for galectin-1 is CD45. By binding to CD45, galectin-1 can modulate its downstream tyrosine kinase activity (31), and thus the BCR signal transduction pathway. Here, we have observed higher surface galectin-1 binding and lower CD45 activity in Oca-b Ϫ/Ϫ B cells. We also showed that recombinant galectin-1 has negative effects on anti-IgM-induced B cell growth and affects tyrosine phosphorylation upon BCR cross-linking (Fig. 6). Altogether, our results lead to the proposal that OCA-B may stimulate positive regulators (such as CD45) of BCR function, and hence BCR signaling and B cell proliferation, via an intracellular association with galectin-1 that effectively sequesters galectin-1 and reduces its secretion and subsequent interaction with CD45. Conversely, OCA-Bdeficient B cells have increased galectin-1 secretion, decreased CD45 activity, and decreased BCR signaling/B cell proliferation. Detailed analysis of BCR-mediated signal transduction in galectin-1 ϩ B cells from wild-type versus Oca-b Ϫ/Ϫ mice will provide more information about the significance of the OCA-B/galectin-1 interaction in BCR signaling. Future studies involving introduction of wild-type OCA-B and mutant OCA-B lacking the galectin-1 interacting domain back into Oca-b Ϫ/Ϫ B cells could provide further information about the physiological importance of the OCA-B/galectin-1 interaction in B cell development and differentiation. Nevertheless, our current combined data on the OCA-B/galectin-1 interaction, the effect of OCA-B deficiency on galectin-1 levels (including intracellular expression, secretion, and B cell surface binding) and CD45 phosphatase activity, and recombinant galectin-1 effects on anti-IgM-stimulated BCR signaling and B cell proliferation lead us to propose that OCA-B, through its association with galectin-1, may regulate BCR signaling, at least in part, by modulating CD45 activity. Although the physiological significance of the marginal difference in CD45 phosphatase activity between wild-type and Oca-b Ϫ/Ϫ B cells remains a key issue, this may reflect the fact that only a subpopulation of mature B cells from Oca-b Ϫ/Ϫ mice have high levels of surface galectin-1 (Fig. 5C). In addition, there are B cell surface molecules other than CD45 that may bind and be modulated by galectin-1 (50). For instance, galectin-1 may bind surface molecules that contain ␤1 integrin (51), and ligand and ␤1 integrin interactions can also lead to signaling transduction to B lymphocytes (52). Hence, an OCA-B/galectin-1 interaction may be functional in modulating BCR signaling through receptors other than CD45.
It also is noteworthy that non-B cells (B220 Ϫ lymphocytes) from Ocab Ϫ/Ϫ mice also bear more galectin-1 on their surfaces (data not shown), suggesting that OCA-B deficiency may have a broad effect for immune response regulation. It has been shown that B cell-secreted galectin-1 can modulate T cell function (28). Because B cell-T cell collaboration is crucial for germinal center formation and efficient T-dependent immune responses, increased galectin-1 secretion from Oca-b Ϫ/Ϫ B cells to B-T synapses could result in early T helper cell death and, consequently, premature termination of T cell help and associated B cell activation. Increased galectin-1 may also affect lymphocyte movement, as seen for other cell types (53)(54)(55). It will be informative to investigate these possibilities in mice lacking both OCA-B and galectin-1.