The inositol phosphatase SHIP inhibits Akt/PKB activation in B cells.

The serine-threonine kinase Akt/PKB is activated downstream of phosphatidylinositol 3-kinase in response to several growth factor stimuli and has been implicated in the promotion of cell survival. Although both phosphatidylinositol 3,4,5-trisphosphate (PIP3) and phosphatidylinositol 3,4-bisphosphate (PI 3,4-P2) have been implicated in the regulation of Akt activity in vitro, the relative roles of these two phospholipids in vivo are not well understood. Co-ligation of the B cell receptor (BCR) and the inhibitory FcgammaRIIB1 on B cells results in the recruitment of the 5'-inositol phosphatase SHIP to the signaling complex. Since SHIP is known to cleave PIP3 to generate PI 3,4-P2 both in vivo and in vitro, and Akt activity has been reported to be regulated by either PIP3 or PI 3,4-P2, we hypothesized that recruitment of SHIP through FcgammaRIIB1 co-cross-linking to the BCR in B cells might regulate Akt activity. The nature of this regulation, positive or negative, might also reveal the relative contribution of PIP3 and PI 3,4-P2 to Akt activation in vivo. Here we report that Akt is activated by stimulation through the BCR in a phosphatidylinositol 3-kinase-dependent manner and that this activation is inhibited by co-cross-linking of the BCR to FcgammaRIIB1. Using mutants of FcgammaRIIB1 and SHIP-deficient B cells, we demonstrate that inhibition of Akt activity is mediated by the immune cell tyrosine-based inhibitory motif within FcgammaRIIB1 as well as SHIP. The SHIP-dependent inhibition of Akt activation also suggests that PIP3 plays a greater role in Akt activation than PI 3,4-P2 in vivo.

Recent studies have established the importance of both 3Јphosphorylated inositol phosphates (PI 3,4-P 2 and PIP 3 ) and Akt phosphorylation in activation of Akt. PI 3,4-P 2 has been shown to directly activate Akt in vitro via interaction with the Akt PH domain, while PIP 3 was inhibitory in this experimental system (11)(12)(13). Additionally, Akt enzymatic activity was shown to be dependent on phosphorylation of Akt on a specific serine (Ser 473 ) and a specific threonine (Thr 308 ) (14,15). Akt is phosphorylated by at least two serine-threonine kinases only in the presence of 3Ј-phosphorylated phospholipids (16,17). One of the kinases that phosphorylates Akt, PDK1, has recently been cloned (18 -20). The studies mentioned above demonstrating direct activation of Akt by PI 3,4-P 2 and its inhibition by PIP 3 were performed with immunoprecipitated Akt in the absence of PDK1 (11)(12)(13). In contrast, when purified PDK1 was present, activation of Akt was dependent on the presence of 3Ј-phosphorylated inositol lipids, with PIP 3 being at least 2-3fold more effective than PI 3,4-P 2 in allowing Akt activation (18,19,21,22). Studies utilizing PH mutants of Akt and PDK1 revealed that the regulatory action of PIP 3 in the PDK-mediated activation of Akt is primarily directed toward Akt rather than PDK1 (17,18,20,22). Based on these data, the current model of Akt activation is that Akt is recruited to the membrane by its PH domain binding to 3-OH-phosphorylated phosphatidylinositol phosphates, which causes a conformational change in Akt, which allows phosphorylation and activation by PDK1 and at least one other kinase. Although there is some indication that PIP 3 may be more permissive in allowing PDK1-mediated activation of Akt than PI 3,4-P 2 in vitro (18,19), the relative roles of PIP 3 and PI 3,4-P 2 in Akt activation in vivo remain to be determined.
Antigen-mediated activation of B cells through the B cell receptor (BCR) initiates a cascade of intracellular biochemical events including activation of tyrosine kinases, activation of PI3K and PLC␥, subsequent generation of phospholipid and inositol phosphate second messengers, and calcium flux (23). The significance of PI3K activation in BCR signaling has been demonstrated by the ability of the PI3K inhibitor wortmannin to inhibit BCR-induced calcium flux (24) and anti-Ig-induced proliferation of the human B cell line RL (25). BCR-mediated signals are selectively inhibited by co-ligation of the F c ␥RIIB1 receptor to the BCR (26). This F c ␥RIIB1-mediated inhibition is dependent upon a 13-amino acid immune cell tyrosine-based inhibitory motif (ITIM) in the cytoplasmic tail of F c ␥RIIB1 (27,28) and the interaction of the Src homology 2 (SH2) domain of the inositol phosphatase SHIP with the ITIM (24, 29 -31). The enzymatic activity of SHIP has also been shown to be critical for this inhibitory effect (32). SHIP has been shown in vitro (33,34) and in vivo (35) to have 5Ј-phosphatase activity toward PIP 3 , resulting in dephosphorylation of PIP 3 and production of PI 3,4-P 2 . Consistent with the requirement for enzymatically active SHIP for F c ␥RIIB1-mediated inhibition, F c ␥RIIB1 cocross-linking to the BCR diminishes the BCR-induced levels of PIP 3 (35) in vivo. The failure of recruitment by PIP 3 of the kinase Btk (via its PH domain) under F c ␥RIIB1 cross-linking conditions has been shown to be responsible for the F c ␥RIIB1-mediated inhibition of BCR-mediated calcium entry (35)(36)(37). A similar mechanism could be involved in regulation of other downstream effectors whose activity is dependent on phospholipids.
Although BCR stimulation results in PI3K activation and Akt is activated downstream of PI3K in several cell types, BCR regulation of Akt has not been reported. In this report, we show that Akt is activated by BCR cross-linking in a PI3K-dependent manner. Since there is conflicting evidence on the relative importance of PIP 3 and PI 3,4-P 2 in Akt activation (11-13, 16, 18, 19, 21, 22) and F c ␥RIIB1-associated SHIP is known to dephosphorylate PIP 3 to generate PI 3,4-P 2 (33,34), we hypothesized that recruitment of SHIP through F c ␥RIIB1 co-crosslinking to the BCR in B cells would regulate Akt activity. The nature of this regulation, positive or negative, might also reveal the relative contribution of PIP 3 and PI 3,4-P 2 to Akt activation in vivo. We demonstrate here that Akt activity is down-regulated by co-cross-linking F c ␥RIIB1 to the BCR, as compared with BCR cross-linking alone, and that this Akt down-regulation is dependent upon the ITIM motif of F c ␥RIIB1. Finally, we show that F c ␥RIIB1-mediated regulation of Akt is dependent on the 5Ј-inositol phosphatase SHIP.

EXPERIMENTAL PROCEDURES
Cells and Reagents-Murine A20 and IIA1.6 cell lines were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum, penicillin, streptomycin, 2 mM L-glutamine, and 20 M 2-mercaptoethanol, at 37°C and 5% CO 2 . IIA1.6 cells reconstituted with either wild type or CT53 mutant of F c ␥RIIB1 have been previously described (27). The DT40 cell line with deletion of the SHIP gene has been previously described (32). DT40 cells were cultured in the same medium as above additionally supplemented with 1% chicken serum (Sigma) and 50 M 2-mercaptoethanol. Histone-H2B was purchased from Boehringer Mannheim, and wortmannin was obtained from Calbiochem. Polyclonal goat anti-Akt antibody (catalog no. sc-1618) and fluorescein isothiocyanate-conjugated anti-rat IgG were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), and the rabbit anti-Akt antibody was from New England Biolabs (Beverly, MA). Rabbit anti-mouse IgG (F(abЈ) 2 fragment and intact antibody) were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). Murine monoclonal anti-chicken IgM (M4) antibody was obtained from Southern Biotechnology Associates (Birmingham, AL). Rabbit anti mouse anti-IgM (F(abЈ) 2 fragment and intact antibody) were obtained from Zymed Laboratories Inc. (South San Francisco, CA). Anti-F c ␥RIIB1 antibody 2.4G2 was purchased from PharMingen (San Diego, CA). Phosphospecific anti-Akt antibody (recognizing phosphorylated serine 473) was from New England Biolabs. Protein kinase A and protein kinase C inhibitors were obtained from Santa Cruz Biotechnology. Akt kinase assay kit using glycogen synthase kinase-3 peptide as substrate was purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). Sheep polyclonal anti-Akt antibody used for immunoprecipitation and immunoblotting was included in this kit.
Generation of Stably Transfected DT40 Cells-10 7 cells were sus-pended in 0.5 ml of phosphate-buffered saline and transfected with 25 g of linearized pApuro-F c ␥RIIB1 plasmid (directing expression of F c ␥RIIB1 under the actin promoter) at 250 V and 960 microfarads in a Gene Pulser II electroporator (Bio-Rad). Cells were selected in 0.5 g/ml puromycin (Sigma). Surface expression of F c ␥RIIB1 on transfected cells was determined by fluorescence-activated cell sorting analysis (see below). Akt Kinase Assay-Cells were suspended at 2 ϫ 10 7 cells/ml in plain RPMI 1640 medium, and 0.5-ml aliquots were incubated for 10 min at 37°C prior to stimulation. A20 and IIA1.6 cells were stimulated with 2.5 g/ml F(abЈ) 2 anti-mouse IgG (to cross-link the BCR) or 5 g/ml intact anti-mouse IgG (to co-cross-link the BCR to F c ␥RIIB1) for the indicated times at 37°C. DT40 cells were stimulated by first incubating with secondary antibody (either 1.5 g/ml F(abЈ) 2 anti-mouse IgM or 3 g/ml intact anti-mouse IgM) for 2 min at 37°C. Primary antibody (1 g/ml murine anti-chicken IgM (murine IgM class antibody)) was then added, and cells were incubated at 37°C for the indicated times. For experiments involving wortmannin, cells were incubated for 15 min in the presence of 50 nM wortmannin prior to adding antibodies for stimulation as described above. After stimulation, cells were lysed in lysis buffer containing 50 mM Tris, pH 7.6; 150 mM NaCl; 1% Nonidet P-40; 10 mM sodium pyrophosphate; 10 g/ml each of aprotinin, leupeptin, and pepstatin; 10 mM NaF; 1 mM NaVO4; and 2 mM phenylmethylsulfonyl fluoride. Cellular debris was cleared by centrifugation, and lysates were precleared with protein A-Sepharose for 30 min at 4°C. In kinase assays performed using H2B as substrate, Akt was immunoprecipitated using 2 g of goat anti-Akt antibody and protein A-Sepharose. Beads were washed twice with lysis buffer and twice with a solution containing 0.5 M LiCl, 100 mM Tris, pH 7.5, 1 mM EDTA followed by two washes in the kinase buffer (50 mM Tris pH 7.5, 10 mM MgCl 2 , 1 mM dithiothreitol). After the last wash, 15 l of the reaction buffer (2.5 g of histone-H2B, 50 M cold ATP, 5 mM protein kinase A and protein kinase C inhibitors, 3 Ci of [␥-32 P]ATP, 50 mM Tris, pH 7.5, 10 mM MgCl 2 , 1 mM dithiothreitol) was added, and the kinase reaction was performed for 30 min at room temperature. The reaction was stopped by adding 15 l of 2ϫ SDS sample buffer, and proteins were analyzed on an 8 -15% SDS-polyacrylamide gel electrophoresis gradient gel. After electrophoresis, the upper portion of the gel was transferred to a nitrocellulose membrane and immunoblotted with rabbit anti-Akt antibody. The lower portion of the gel was stained with Coomassie Blue to visualize H2B, dried, and subjected to autoradiography. Akt kinase activity was also assayed using a peptide derived from glycogen synthase kinase-3 (GSK-3) as a substrate (after stimulating and lysing cells as described above) using the Akt kinase assay kit from Upstate Biotechnology following the manufacturer's protocol. Akt immunoblots for the experiments using GSK-3 peptide as substrate were performed using sheep anti-Akt antibody.
Fluorescence-activated Cell Sorting Analysis-Cells were incubated at 5 ϫ 10 6 cells/ml in phosphate-buffered saline with 1 g/ml 2.4G2 (anti-F c ␥RIIB1) antibody for 20 min at 4°C. Cells were washed twice in phosphate-buffered saline and stained with fluorescein isothiocyanateconjugated anti-rat IgG at 4 g/ml for 20 min at 4°C. The cells were then washed and analyzed on a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA). As a negative control, cells were stained with fluorescein isothiocyanate-conjugated anti-rat IgG with no primary antibody.

Stimulation through the B Cell Receptor Induces Akt Kinase
Activity-We used murine A20 cells to test whether BCR stimulation leads to Akt activation. Cells were either left unstimulated or stimulated by cross-linking the BCR with the F(abЈ) 2 fragment of anti-mouse IgG. Cells were then lysed, Akt was immunoprecipitated, and in vitro kinase assays were performed using histone-H2B as a substrate, as described under "Experimental Procedures." As shown in Fig. 1, BCR stimulation led to an increase in Akt kinase activity (lanes 1 and 2). It has been shown that platelet-derived growth factor/insulinmediated activation of Akt occurs downstream of PI3K and is inhibited by the specific PI3K inhibitor wortmannin (1-3). Since BCR stimulation leads to PI3K activation (23,24), we examined whether Akt activation in B cells is also downstream of PI3K. BCR activation of Akt was inhibited by preincubation of the cells with the PI3K inhibitor wortmannin prior to BCR cross-linking (Fig. 1, top panel, lane 3). These results show that BCR cross-linking activates Akt in a PI3K-dependent manner.
BCR-mediated Activation of Akt Is Inhibited by Co-crosslinking F c ␥RIIB1-F c ␥RIIB1 can inhibit several BCR-mediated downstream signals when co-cross-linked to the BCR, including the PI3K dependent calcium entry into cells (23,26). Since Akt activity can be modulated in vitro by 3Ј-phosphorylated phosphoinositides (16) and F c ␥RIIB1 recruits SHIP (26), which converts PIP 3 to PI 3,4-P 2 in vivo (35), we hypothesized that F c ␥RIIB1 recruitment would affect BCR-induced activation of Akt in vivo. However, since Akt has been reported to be activated in vitro by both the substrate (PIP 3 ) and the product (PI 3,4-P 2 ) of SHIP's enzymatic activity depending on the experimental system used (11-13, 16, 18, 19, 21, 22), F c ␥RIIB1 co-cross-linking could either positively or negatively regulate Akt activity. To address this question, A20 B cells were left unstimulated, stimulated by cross-linking the BCR only (with F(abЈ) 2 fragment of anti-mouse IgG), or cross-linking BCR with F c ␥RIIB1 (using intact anti-mouse IgG) in a time course of activation. Akt was immunoprecipitated, and its activity was analyzed by an in vitro kinase assay using histone-H2B as a substrate. As shown in Fig. 2A, F c ␥RIIB1 co-cross-linking to the BCR resulted in a decrease in Akt activity as compared with BCR cross-linking alone (compare lanes 2-5 with lanes 6 -9). The decrease in Akt activation observed upon F c ␥RIIB1 co-cross-linking was not simply due to delayed kinetics of activation, since a time course of activation extended to 60 min ruled out this possibility (data not shown). The kinase activity at these time points was correlated to phosphorylation of Akt on serine 473 as assayed by immunoblotting with phosphospecific anti Akt antibody (Fig. 2B). As controls, Akt and H2B levels are shown in Fig. 2, C and D, respectively. The decrease in Akt activity is not due to a general lack of activation in the BCR plus F c ␥RIIB1 cross-linked cells, since several proteins become tyrosine-phosphorylated in the stimulated cells as compared with the unstimulated cells (Fig. 2E).
To further confirm these results, we also assayed Akt kinase activity following stimulation performed in triplicate using a different anti-Akt antibody for immunoprecipitation (Sheep anti-Akt, Upstate Biotechnology) and using a peptide whose se-quence was derived from GSK-3 (a known substrate of Akt) as substrate. As shown in Fig. 3A, stimulation of A20 cells by co-cross-linking F c ␥RIIB1 to the BCR resulted in a substantial decrease in Akt activity as compared with BCR cross-linking alone. Immunoprecipitation of proteins from BCR-activated A20 lysates with control antibody failed to show any significant kinase activity toward GSK-3-derived peptide (Fig. 3A, column  4), indicating that the kinase activity precipitated with anti-Akt antibodies is due to Akt. Amounts of Akt precipitated were equivalent among the different lanes (Fig. 3B). These results show that recruitment of F c ␥RIIB1 to the BCR signaling complex results in a specific inhibition of BCR-mediated Akt activation.
F c ␥RIIB1-mediated Inhibition of Akt Activation Is Dependent upon the ITIM in the Cytoplasmic Tail of F c ␥RIIB1-Previous work has demonstrated that the inhibitory effects of F c ␥RIIB1 can be recapitulated by a 13-amino acid sequence within the cytoplasmic tail of F c ␥RIIB1, the ITIM (27,28). We therefore tested the role of the ITIM in the F c ␥RIIB1-mediated inhibition of Akt activation. The IIA1.6 cell line, an F c ␥RIIB1-deficient derivative of the A20 B cell line, reconstituted with the wild type F c ␥RIIB1, or a truncation mutant of this receptor lacking the ITIM (CT53), was used to address this question. As expected, IIA1.6 cells stably transfected with the wild type F c ␥RIIB1 exhibited a pattern of Akt activation similar to the parental A20 cells, both in response to BCR cross-linking alone and following BCR plus F c ␥RIIB1 cross-linking (Fig. 4, left  panels). In IIA1.6 cells reconstituted with the CT53 mutant of F c ␥RIIB1 (which lacks the ITIM), co-cross-linking of F c ␥RIIB1 to the BCR failed to show inhibition of BCR-mediated Akt activation (Fig. 4, right panels). These data indicate that F c ␥RIIB1-mediated inhibition of BCR-induced Akt activation is dependent upon the ITIM of F c ␥RIIB1.   (32) demonstrated that F c ␥RIIB1-mediated inhibition of calcium flux and NFAT activation are dependent on the recruitment of enzymatically active SHIP. Since we have shown that F c ␥RIIB1-mediated inhibition of Akt activation is dependent upon the ITIM, we sought to test whether it was also dependent on SHIP. To evaluate this, we utilized parental DT40 cells (SHIP-expressing) and a DT40 line in which the SHIP genes were deleted by homologous recombination (32). SHIP-expressing and SHIP-deficient DT40 cells were stably transfected with plasmids encoding wild type F c ␥RIIB1, and clones expressing F c ␥RIIB1 on the surface were selected. Calcium flux patterns in these clones following BCR and BCR plus F c ␥RIIB1 cross-linking were similar to what has been previously reported (data not shown) (32). Next, we assessed the state of Akt activation in these cells following BCR or BCR plus F c ␥RIIB1 cross-linking. Co-cross-linking of F c ␥RIIB1 to the BCR in SHIP-expressing DT40 cells resulted in inhibition of Akt activation in a pattern similar to A20 cells (Fig. 5A). In contrast, when SHIP-deficient DT40 cells were used, F c ␥RIIB1 co-cross-linking to the BCR did not result in a strong inhibition of Akt activation (Fig. 5B). Akt activation by the BCR could be inhibited by the PI3K inhibitor wortmannin in the SHIP-deficient DT40 cells, indicating that Akt activation in these cells still occurs in a PI3K-dependent manner (Fig.  5C).
We also confirmed these results in DT40 cells by assaying Akt kinase activity following stimulation in three independent replicates toward GSK-3-derived peptide. Co-cross-linking of F c ␥RIIB1 to the BCR in SHIP-expressing DT40 cells resulted in inhibition of Akt activation in a pattern similar to A20 cells (Fig. 6A). In contrast, when SHIP-deficient DT40 cells were used, F c ␥RIIB1 co-cross-linking to the BCR did not result in a significant inhibition of Akt activation (Fig. 6B). Immunoblotting for Akt indicated that the same level of Akt was precipitated in all lanes (Fig. 6, bottom panels). In this immunoblotting we also observed a shift in Akt mobility upon BCR stimulation, which was inhibited by F c ␥RIIB1 cross-linking in SHIP-expressing but not in SHIP-deficient DT40 cells (Fig. 6, bottom panels). In addition, we observed in several experiments higher levels of Akt activation in SHIP-deficient cells as compared with SHIP-expressing cells upon BCR stimulation (Fig. 6). DISCUSSION Co-ligation of F c ␥RIIB1 and the BCR results in a potent inhibitory signal, which leads to a selective attenuation of BCR-mediated signals (26). This phenomenon represents a negative feedback mechanism triggered by immune complex or anti-idiotypic antibodies to suppress excessive B cell immune response (23,26). The negative regulation by F c ␥RIIB1 requires the recruitment and the enzymatic activity of the inositol phosphatase SHIP (32). The mechanism of this F c ␥RIIB1/ SHIP-mediated inhibitory action is not fully understood. Recently, it has been suggested that SHIP-mediated dephosphorylation of PIP 3 and consequently the inhibition of Btk membrane localization and activation plays a central role in inhibition of calcium flux by F c ␥RIIB1 (35)(36)(37).
It has been shown that co-ligation of F c ␥RIIB1 and BCR induces apoptosis in mouse splenocytes (38,39). The serine threonine kinase Akt, an enzyme activated downstream of PI3K, is a major signaling protein involved in protection from apoptosis (5)(6)(7)(8)40). Although BCR stimulation activates PI3K, as demonstrated by the ability of wortmannin to inhibit BCR signaling (23, 24), activation of Akt by BCR engagement has not been reported previously. Here, we show the activation of Akt by BCR stimulation in a PI3K-dependent manner. Fur-  1-3) or control antibodies (lane 4), and kinase activity toward glycogen synthase kinase-3 peptide was assayed. A, cpm of incorporated radioactivity were measured in a scintillation counter after binding the phosphorylated peptide to phosphocellulose paper. The nonspecific binding of unincorporated radioactivity was measured and subtracted from each individual assay. In this experiment, the nonspecific binding was 316 cpm. Each value represents the average of three independent replicates, with S.D. indicated by the error bars. B, level of immunoprecipitated Akt from this experiment was determined by immunoblotting using sheep ␣-Akt antibody. This immunoblot represents one set of the three replicates. Akt and IgH are indicated by the arrows. The other two immunoblots yielded similar results. Ipt, immunoprecipitation.
FIG. 4. Inhibition of Akt activation is dependent on the ITIM in the cytoplasmic tail of F c ␥RIIB1. IIA1.6 cells stably transfected with wild type (wt) F c ␥RIIB1 (left panels) or CT53 mutant of F c ␥RIIB1 (lacking ITIM, right panels) were stimulated by BCR cross-linking alone or BCR plus F c ␥RIIB1 co-cross-linking as described under "Experimental Procedures," and Akt kinase assays were performed (top panels). Levels of histone-H2B and immunoprecipitated Akt are shown as controls (middle panels). The expression of transfected F c ␥RIIB1 was analyzed by flow cytometry using anti-F c ␥RIIB1 antibody 2.4G2 (gray line) as compared with control (black line). thermore, our data demonstrate that co-ligation of F c ␥RIIB1 with BCR results in inhibition of Akt kinase activity in murine A20 cells as well as in chicken DT40 cells transfected with murine F c ␥RIIB1. Using an F c ␥RIIB1-deficient subclone of A20 (IIA1.6) reconstituted with either wild type F c ␥RIIB1 or a carboxyl-terminally truncated F c ␥RIIB1 (lacking ITIM), we demonstrate that the inhibition of Akt activity is dependent on the ITIM motif. Additionally, in SHIP-deficient DT40 cells stably transfected with murine F c ␥RIIB1, co-ligation of BCR and F c ␥RIIB1 could not significantly inhibit Akt activation. These data indicate that the inhibition of Akt kinase activity is mediated by the inositol phosphatase SHIP. However, we consistently observed a slight inhibition of Akt by BCR plus F c ␥RIIB1 stimulation in SHIP-deficient cells. This may reflect minor roles played by other molecules such as SHP-1, known to interact with F c ␥RIIB1 (41). Alternatively, there could be some redundancy for SHIP, since a second SH2-containing inositol phosphatase (SHIP2) has been recently described (42,43).
There have been conflicting data in the literature on the role of PIP 3 versus PI 3,4-P 2 in the activation of Akt. In the absence of PDK1, one of the kinases that phosphorylates Akt, the addition of PI 3,4-P 2 containing micelles to immunoprecipitated Akt increased its enzymatic activity, whereas PIP 3 either had no effect or was inhibitory (11)(12)(13). In contrast, when purified PDK1 was present, activation of Akt was dependent on the presence of 3Ј-phosphorylated inositol lipids, with PIP 3 being at least 2-3-fold more effective than PI 3,4-P 2 in allowing Akt activation (18,19,21,22). Studies utilizing PH mutants of Akt and PDK1 revealed that the regulatory action of PIP 3 in the PDK-mediated activation of Akt is primarily directed toward Akt rather than PDK1 (17,18,20,22). Thus, it has not been clear whether PIP 3 or PI 3,4-P 2 mediates the activation of Akt in vivo. SHIP is known to dephosphorylate PIP 3 both in vitro (33,34) and in vivo (35) to generate PI 3,4-P 2 . It has been proposed that SHIP functions upstream of Akt and that the generation of PI 3,4-P 2 positively regulates Akt activation (10 -13). Inhibition of BCR-induced signals by F c ␥RIIB1 involves recruitment of SHIP to the BCR signaling complex (26). Recently, Scharenberg et al. (35) have demonstrated in A20 cells that F c ␥RIIB1 co-cross-linking to the BCR decreases cellular PIP 3 levels generated by BCR stimulation. Therefore, this system provides a useful tool to indirectly address the role of PIP 3 versus PI 3,4-P 2 in in vivo activation of Akt. Our data strongly suggest that, in vivo, dephosphorylation of PIP 3 inhibits Akt activation and that SHIP-mediated generation of PI 3,4-P 2 does not positively regulate Akt. Therefore, in this system, PIP 3 appears to be a more potent activator of Akt than PI 3,4-P 2 . Additionally, in several experiments, we observed a greater level of BCR-mediated activation of Akt in SHIP-deficient DT40 cells compared with SHIP-expressing cells, most likely due to the absence of PIP 3 dephosphorylation. This suggests that SHIP also directly regulates the BCR-mediated rise in PIP 3 levels. This is consistent with the reported increase in Btk activation and calcium flux (another PIP 3 -dependent event) in SHIP-deficient cells as compared with SHIP-expressing cells (37). Thus, SHIP, through its effects on PIP 3 , could set the threshold for the magnitude of Akt activation and the downstream consequences.
Others have suggested a biphasic pattern of Akt activation, with PIP 3 being responsible for the early phase and PI 3,4-P 2 mediating the late activation of Akt in a calcium-and calpaindependent manner in platelets (44). Since a rise in calcium levels is an important event in BCR-mediated signaling, we tested a possible role for calcium flux upstream of Akt activation in B cells. However, we did not observe any effect of either EGTA or the calpain inhibitor calpeptin on BCR-induced Akt activation in A20 cells (data not shown). Additionally, in Btkdeficient DT40 cells, which fail to flux calcium upon BCR stimulation (45), we could demonstrate Akt activation in response to BCR cross-linking (data not shown). It therefore appears that a calcium-and calpain-dependent pattern of regulation of Akt is not operational in B cells. F c ␥RIIB1 cross-linking to the BCR results in an increased sensitivity to apoptosis in mouse splenocytes (38,39). Since FIG. 5. Inhibition of Akt activation by F c ␥RIIB1 requires the inositol phosphatase SHIP. Parental DT40 (SHIP-expressing) cells (A) or SHIP-deficient DT40 cells (B) stably transfected with wild type F c ␥RIIB1 were stimulated by BCR cross-linking only or BCR plus F c ␥RIIB1 co-cross-linking as described under "Experimental Procedures," and Akt kinase assays were performed. The level of immunoprecipitated Akt was determined by immunoblotting with goat anti-Akt antibody (middle panels), and the level of H2B was determined by Coomassie staining (bottom panels). The expression of transfected F c ␥RIIB1 was analyzed by flow cytometry using anti-F c ␥RIIB1. C, SHIP-deficient DT40 cells were left unstimulated (lane 1) or activated for 5 min by cross-linking the BCR in the absence (lane 2) or presence (lane 3) of wortmannin. Akt was immunoprecipitated and assayed for its kinase activity toward the substrate H2B. activation of Akt leads to protection from apoptosis in several cell types (40), it is reasonable to hypothesize that F c ␥RIIB1/ SHIP-mediated inhibition of Akt might be responsible for this increased sensitivity to apoptosis. However, Ono et al. (32) reported that in the DT40 cell line, the absence of SHIP correlates with an increased sensitivity to F c ␥RIIB1-mediated apoptosis. This increased sensitivity to apoptosis may reflect differences in cell types and/or experimental conditions. Our results show that SHIP-deficient DT40 cells do not exhibit F c ␥RIIB1-mediated inhibition of Akt, suggesting that the increased sensitivity to apoptosis in these cells cannot be explained by lack of Akt activation. The reason for this inconsistency at present is not clear. The correlation between F c ␥RIIB1-mediated inhibition of Akt activation and its relevance to survival/apoptosis of B cells during an immune response is currently under investigation.