SHIP-2 inositol phosphatase is inducibly expressed in human monocytes and serves to regulate Fcgamma receptor-mediated signaling.

SHIP-2, a recently identified inositol 5'-phosphatase, shares high level homology with SHIP-1. Although the role of SHIP-1 has been extensively studied, the role of SHIP-2 in myeloid cell functions is not known. Here, we have analyzed the expression patterns, molecular mechanism of activation, and function of SHIP-2 in human myeloid cell Fcgamma receptor (FcgammaR) signaling. We report that SHIP-2 is expressed in transformed myeloid cells and in primary macrophages, but not in peripheral blood monocytes. Treatment of peripheral blood monocytes with bacterial lipopolysaccharide induced expression of SHIP-2 in a dose-dependent manner. FcgammaRIIa clustering in THP-1 cells induced SHIP-2 tyrosine phosphorylation, suggesting a role for SHIP-2 in modulating FcgammaR-mediated function. Consistent with this notion, overexpression of wild-type SHIP-2 (but not catalytically deficient SHIP-2) in THP-1 cells almost completely abrogated NFkappaB-mediated gene transcription in response to FcgammaRIIa clustering. Furthermore, FcgammaRIIa-induced Akt activation was blocked by wild-type SHIP-2, but not by a catalytically deficient mutant of SHIP-2. Additional experiments analyzing the molecular mechanism of SHIP-2 induction by FcgammaRIIa revealed that SHIP-2 associated with the phosphorylated FcgammaRIIa immunoreceptor tyrosine-based activation motif via the SHIP-2 SH2 domain. Thus, an SH2 domain mutant of SHIP-2 failed to associate with FcgammaRIIa or to become tyrosine-phosphorylated upon FcgammaRIIa clustering. Finally, we also demonstrate that SHIP-2 phosphorylation was induced by FcgammaRI clustering in THP-1 cells. These findings unravel a novel level of regulation of FcgammaR-mediated activation of human myeloid cells by the expression and function of the inositol phosphatase SHIP-2.

Clustering of Fc␥R 1 on monocytes and macrophages by immune complexes initiates a series of signaling events that result in phagocytosis of the immune complex (1). The process of phagocytosis is accompanied by the generation of reactive oxygen and nitrogen radicals, as well as the production of inflammatory cytokines, all of which cause tissue damage. Thus, the phagocytic process must be subject to tight regulation.
Human monocytes and macrophages express three classes of Fc␥R (2). Fc␥RI, Fc␥RIIa, and Fc␥RIII are activating receptors that are associated with an immunoreceptor tyrosine-based activation motif (ITAM). In contrast, Fc␥RIIb is an inhibitory receptor that bears in its cytoplasmic tail a tyrosine-based inhibitory motif that predominantly recruits negative regulatory phosphatases and causes down-regulation of activation events. In addition to the regulatory role of Fc␥RIIb, recent studies have revealed that the Fc␥R ITAMs are often capable of simultaneously activating positive and negative regulatory proteins such that the final biologic response is tempered. We (3) and others (4) have demonstrated that Fc␥R ITAMs are capable of recruiting and activating the inositol phosphatase SHIP-1. Other studies have shown that the protein-tyrosine phosphatase SHP-1 (5) and the dual phosphatase PTEN (6) also serve to regulate Fc␥R-mediated macrophage functions. Thus, it is clear that Fc␥R-mediated activation of macrophages is subject to multiple levels of regulation, which are not fully understood. In this report, we present molecular details of a novel regulatory influence exerted by the recently identified inositol phosphatase SHIP-2.
Fc␥R clustering in monocytes and macrophages initiates a biochemical cascade that begins with the activation of the Src kinases that phosphorylate the ITAMs of Fc␥R (7). Once phosphorylated, the ITAMs serve as docking sites for several signaling enzymes and enzyme-adapter complexes, including the Syk tyrosine kinase (8); the p85 subunit of phosphatidylinositol 3-kinase (9); and the Ras adapter molecule Shc (3,10), which serves to recruit the Grb2-Sos complex, leading to activation of the Ras pathway. Association of phosphatidylinositol 3-kinase with the ITAM places the enzyme in proximity with its lipid substrates, resulting in the generation of the important lipid second messenger PtdIns-3,4,5-P 3 . PtdIns-3,4,5-P 3 is critical for the activation of PH domain-containing enzymes such as Btk (11); the Tec family tyrosine kinase involved in intracellu-lar calcium mobilization, Vav (12); the guanine nucleotide exchange factor for Rac; and Akt, the serine/threonine kinase that is involved in cell survival (13,14) and in the activation of NFB (15). The hydrolysis of PtdIns-3,4,5-P 3 by SHIP-1 has been demonstrated to down-regulate the activity of the above PH domain-containing enzymes and the downstream functional outcomes (11,13,14). SHIP-2 is an inositol 5Ј-phosphatase that has high level homology to SHIP-1 in its catalytic region (16 -18). The molecules are largely divergent in the C-terminal region, consisting of a proline-rich domain that associates with unique SH3 domain-containing proteins (19). In addition, whereas SHIP-1 has two tyrosine residues in the C terminus that conform to an NPXY motif shown to bind phosphotyrosine-binding domains upon phosphorylation, SHIP-2 has only one NPXY motif (20). Thus, while enzymatically similar, these two enzymes likely differ in functions that are related to their protein interactions via the C-terminal region. These two molecules also differ in their expression patterns: SHIP-1 is expressed predominantly in hematopoietic cells, whereas SHIP-2 is much more ubiquitously expressed (20). Recent studies have revealed a role for SHIP-2 in regulating insulin receptor signaling (21)(22)(23). Other studies have demonstrated a role for SHIP-2 in mediating the inhibitory effect of Fc␥RIIb in B cells (21,24,25). However, the expression and function of SHIP-2 in monocytes and macrophages are not known.
In this study, we first demonstrate that SHIP-2 was expressed in human alveolar macrophages, but was almost undetectable in peripheral blood monocytes derived from the same donors. Interestingly, expression of SHIP-2 in PBMs was induced by bacterial LPS in a dose-dependent manner. Second, Fc␥RIIa clustering in the human myeloid cell line THP-1 induced tyrosine phosphorylation of SHIP-2, suggesting that SHIP-2 may play a role in Fc␥R-mediated function. Analyzing the functional consequence of SHIP-2 in Fc␥RIIa signaling, we report that overexpression of wild-type SHIP-2 (but not catalytically deficient SHIP-2) completely abrogated NFB-dependent gene transcription in response to Fc␥RIIa clustering in THP-1 cells. Transient cotransfection experiments in COS-7 cells demonstrated that wild-type SHIP-2 down-regulated Fc␥RIIa-induced Akt phosphorylation. Additional experiments analyzing the molecular mechanism of SHIP-2 activation by Fc␥RIIa demonstrated that the SH2 domain of SHIP-2 was necessary for optimal association of SHIP-2 with Fc␥RIIa and for optimal tyrosine phosphorylation of SHIP-2. Finally, we also demonstrate that Fc␥RI clustering resulted in phosphorylation of SHIP-2, suggesting that SHIP-2 may regulate both Fc␥RIIa-and Fc␥RI-mediated myeloid cell function.

EXPERIMENTAL PROCEDURES
Cells, Antibodies, and Reagents-THP-1, U937, and COS-7 cells were obtained from American Type Culture Collection. COS-7 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. All other cells were maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum. Anti-Fc␥RI antibody 197 and anti-Fc␥RIIa antibody IV.3 were obtained from Medarex (Annandale, NJ). Rabbit polyclonal anti-SHIP-2 antibody was a generous gift from Dr. Bayard Clarkson (Memorial Sloan Kettering Cancer Center, New York, NY) (19). Goat polyclonal anti-SHIP-2 antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-phosphotyrosine, anti-phospho-Akt, and anti-Akt antibodies were purchased from Cell Signaling Technology (Beverly, MA). Rabbit polyclonal anti-Syk antibody was purchased from Upstate Biotechnology, Inc. (Charlottesville, VA). Anti-Xpress antibody was purchased from Invitrogen.
Isolation of PBMs-CD14-positive PBMs were isolated as described previously (26). Briefly, peripheral blood mononuclear cells were first isolated by density gradient centrifugation over Histopaque (Sigma). Monocytes were then purified from the peripheral blood mononuclear cells by negative selection using the MACs monocyte isolation kit (Miltenyi Biotech). For this, peripheral blood mononuclear cells were first treated with Fc␥R blocking reagent (human IgG), followed by a hapten/antibody mixture (mixture of hapten-conjugated monoclonal anti-CD3, anti-CD7, anti-CD19, anti-CD45RA, anti-CD56, and anti-IgE antibodies). The labeled cells were further treated with MACS antihapten magnetic microbeads conjugated to monoclonal anti-hapten antibody. The cells were then passed over a MACS column, and the effluent was collected as the negative fraction representing enriched monocytes.
Preparation of Human Alveolar Macrophages-Macrophages were obtained from healthy donors by bronchoalveolar lavage. Cells were washed twice with phosphate-buffered saline, counted, and analyzed by Diff-Quick staining for purity. Cell preparations were Ͼ95% positive for macrophages.
Immunoprecipitation and Western Blotting-THP-1 cells and transfected COS-7 cells were activated by clustering Fc␥RI and or Fc␥RIIa with monoclonal antibody 197 or Fab fragments of antibody IV.3 and goat F(abЈ) 2 anti-mouse Ig secondary antibody. Resting and activated cells were lysed in buffer A (50 mM Tris (pH 8.0), 10 mM EDTA, 10 mM Na 4 P 2 O 7 , 10 mM NaF, 1% Triton X-100, 125 mM NaCl, 10 mM Na 3 VO 4 , and 10 g/ml each aprotinin and leupeptin), and post-nuclear lysates were incubated overnight with the antibody of interest and protein G-agarose beads (Invitrogen) or goat F(abЈ) 2 anti-mouse Ig covalently linked to Sepharose, depending on the antibody. Immune complexes bound to beads were washed with buffer A and boiled in SDS sample buffer (60 mM Tris (pH 6.8), 2.3% SDS, 10% glycerol, 0.01% bromphenol blue, and 2% 2-mercaptoethanol) for 5 min. Proteins were separated by SDS-PAGE, transferred to nitrocellulose filters, probed with the antibody of interest, and developed by enhanced chemiluminescence.
Immunoblot Data Quantitation-The ECL signal was quantitated using a scanner and a densitometry program (Scion Image). To quantitate the phosphorylation signals in the activated samples, we first subtracted background, normalized the phosphorylation signal to the amount of total precipitated protein, and plotted the values obtained by expressing them as -fold increase over the values in unstimulated samples.
Transfection-COS-7 cells were transfected as described previously (3). Briefly, cells were grown on culture dishes to 60 -70% confluency. cDNA for Fc␥RIIa in pCEXV3 (kindly provided by Dr. J. Ravetch, Rockefeller University, New York) and cDNAs for wild-type SHIP-2, the inactive SH2 domain mutant of SHIP-2, and the inactive catalytic domain mutant of SHIP-2 in pcDNA3 (generously provided by Dr. Shonna Moodie, Metabolex Inc., Hayward, CA) (27) were mixed in various combinations with LipofectAMINE 2000 reagent (Invitrogen). The DNA mixture was added to cells in serum-free Dulbecco's modified Eagle's medium and incubated for 3 h at 37°C in a CO 2 incubator. The medium was then replaced with Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. The cells were harvested 24 h later and analyzed for expression of the transfected cDNAs by Western blotting. Having ensured that the various transfectants expressed comparable levels of protein, we then examined the ability of the wild-type and mutant SHIP-2 molecules to associate with the Fc␥RIIa ITAM and to become tyrosine-phosphorylated. In other experiments, GST-Akt (a kind gift from Dr. R. B. Pearson, Peter MacCallum Cancer Institute, Melbourne, Australia) (28) was cotransfected with the above molecules (2 g of Fc␥RIIa, 5 g of SHIP-2, and 1 g of GST-Akt) to analyze the influence of SHIP-2 on Akt activation by Fc␥RIIa.
Transfection of THP-1 Cells and Luciferase Assays-For analysis of SHIP-2 influence on NFB transcriptional activity, THP-1 cells were transfected as described previously (3). Briefly, THP-1 cells were electroporated (310 V, 950 microfarads; Bio-Rad Gene Pulser II) with 5 g of wild-type SHIP-2 or catalytically deficient SHIP-2 in pcDNA3, 1 g of NFB-luciferase plasmid, and 0.5 g of pEGFP to normalize for transfection efficiency. Transfectants were harvested 24 h later and activated by clustering Fc␥RIIa by the methods described above for 6 h at 37°C. The cells were lysed in 100 l of cell culture lysis reagent (Promega). Luciferase activity was measured using the Promega luciferase assay reagent. Data are represented as graphs indicating the percent induction of NFB activity in cells activated by clustering Fc␥RIIa over the activity in cells that were not activated. Data points are expressed as the mean Ϯ S.D. of three independent experiments. Statistical analysis was performed by Student's t test.

SHIP-2 Expression Pattern in Myeloid
Cells-Expression of SHIP-2 in myeloid cells was assessed by Western blotting protein-matched lysates obtained from the THP-1 and U937 hu-man monocytic cell lines, the P388D1 mouse myeloid cell line, and primary human PBMs and alveolar macrophages. The results shown in Fig. 1A (upper panels) demonstrate that SHIP-2 was expressed in the transformed myeloid cell lines and in primary human alveolar macrophages. Interestingly, PBMs appeared to express very low levels of SHIP-2 compared with alveolar macrophages from the same donor. The results were consistently observed in five separate donors. A reprobe of the membranes with anti-␤-actin antibody (Fig. 1A, lower panels) showed equivalent loading of protein in the different lanes. Treatment of PBMs with LPS, a reagent often used to activate PBMs, resulted in increased expression of SHIP-2 in a dose-dependent manner (Fig. 1B, upper panel). These data suggest that SHIP-2 is inducibly expressed in human monocytes.
SHIP-2 Is Tyrosine-phosphorylated and Serves to Down-regulate Fc␥RIIa-mediated Function-To test whether SHIP-2 is involved in Fc␥R-mediated myeloid cell activation, SHIP-2 phosphorylation was assessed in THP-1 cells stimulated by clustering Fc␥RIIa (Fig. 2A). For this, Fc␥RIIa receptors were clustered using Fab fragments of monoclonal anti-Fc␥RIIa antibody IV.3, followed by F(abЈ) 2 fragments of goat anti-mouse IgG. SHIP-2 was immunoprecipitated from resting and activated cells with goat polyclonal anti-SHIP-2 antibody and analyzed by Western blotting with anti-phosphotyrosine antibody. The results indicate that Fc␥RIIa clustering induced SHIP-2 phosphorylation within 1 min and that the phosphorylation signals peaked at ϳ7 min and subsided by 30 min post-stimulation ( Fig. 2A, upper panel). A reprobe of the same membrane with anti-SHIP-2 antibody revealed equal loading of SHIP-2 in all lanes. The seventh lane is a negative control immunoprecipitation with normal goat IgG ( Fig. 2A, lower  panel). In a second set of experiments, human PBMs cultured overnight with 100 ng/ml LPS were stimulated by clustering Fc␥RIIa, and SHIP-2 phosphorylation was assessed by Western blotting (Fig. 2B). As shown in Fig. 2B (upper panel), SHIP-2 phosphorylation was induced by Fc␥RIIa clustering in PBMs. Fig. 2B (lower panel) demonstrates equivalent loading of SHIP-2 in the two lanes.
Having ensured that SHIP-2 is phosphorylated during Fc␥RIIa signaling, we next assessed whether SHIP-2 influences Fc␥RIIa-mediated functional outcomes. We (3) and others (10) have previously reported that Fc␥R clustering induces NFB-mediated gene transcription, which is down-regulated by the inositol phosphatase SHIP-1. Because SHIP-1 and SHIP-2 are both capable of hydrolyzing PtdIns-3,4,5-P 3 and influence downstream events, we analyzed whether SHIP-2 plays a role in regulating NFB-mediated gene transcription.
In these experiments, we analyzed NFB-dependent transcription from a reporter plasmid encoding the luciferase gene in the presence of excess wild-type SHIP-2 or the catalytically deficient SHIP-2 mutant D608A. Thus, THP-1 cells were transiently cotransfected with the NFB-luciferase plasmid and with plasmids encoding wild-type SHIP-2 or catalytically deficient SHIP-2. The transfected cells were activated by clustering Fc␥RIIa for 5 h at 37°C. Transcription of the luciferase gene was measured by a luciferase enzyme assay as described above. The results are expressed as percent increase in luciferase activity in cells activated by clustering Fc␥RIIa over the activ- ity in resting cells (Fig. 2C). The data indicate that NFB-dependent transcription of the luciferase gene occurred upon Fc␥RIIa clustering. However, overexpression of wild-type SHIP-2 blocked the induction of gene transcription (p ϭ 0.027, presence of exogenous wild-type SHIP-2 versus pcDNA3 empty vector). Importantly, gene transcription was significantly enhanced in the presence of dominant-negative, catalytically deficient SHIP-2 (p ϭ 0.006, presence of exogenous catalytically deficient SHIP-2 versus pcDNA3 empty vector), suggesting that SHIP-2 serves to down-regulate NFB-dependent gene transcription by Fc␥RIIa.
Recent studies have demonstrated an upstream role for Akt in NFB activation (15). Because Akt is a PH domain-containing enzyme whose activity is regulated by PtdIns-3,4,5-P 3 , a substrate of SHIP-2, we next analyzed Akt activation in the presence of overexpressed wild-type SHIP-2 or catalytically deficient SHIP-2. For these experiments, COS-7 fibroblasts were used to achieve high levels of transfection. Thus, COS-7 cells were transiently transfected to express epitope (GST)tagged Akt along with Fc␥RIIa (COS-7 cells do not express any endogenous Fc␥R) and either wild-type SHIP-2 or the catalytically deficient SHIP-2 mutant D608A. The transfected cells were harvested 24 h later and activated by clustering Fc␥RIIa by the methods described above. Whole cell lysates from resting and activated cells were separated by SDS-PAGE and analyzed by Western blotting with anti-phosphoserine Akt antibody (Fig. 3A, upper panel). GST-Akt migrated at ϳ95 kDa in comparison with endogenous Akt, which migrated at 65 kDa. The slower migration of GST-Akt allowed for analysis of the effect of overexpression of the SHIP-2 proteins on cotransfected GST-Akt. The membrane was reprobed with anti-Akt antibody to detect the total amount of GST-Akt present in each lane (Fig.  3A, middle panel). Fig. 3A demonstrates that GST-Akt was serine-phosphorylated upon Fc␥RIIa clustering (fourth lane) in comparison with the resting sample (third lane). Overexpression of wild-type SHIP-2 down-regulated Fc␥RIIa-induced Akt phosphorylation (compare the fifth and sixth lanes). In contrast, overexpression of catalytically deficient SHIP-2 resulted in enhanced phosphorylation of Akt in response to Fc␥RIIa clustering (eighth lane), suggesting that catalytically deficient SHIP-2 functions in a dominant-negative manner, overcoming the inhibition of Akt by endogenous SHIP-2 (fourth lane). Akt phosphorylation was quantitated by normalizing phospho-Akt (GST) signals to total Akt (GST) signals as described under "Experimental Procedures." The results are expressed as -fold increase in Akt phosphorylation in activated cells in comparison with that in resting cells. Fig. 3A (lower panel) represents the mean Ϯ S.D. of values obtained from three independent experiments. The inhibition of Akt phosphorylation by wildtype SHIP-2 was statistically significant (p ϭ 0.02). To ensure that transfected Fc␥RIIa was expressed to comparable levels in the different transfectants, WCLs were probed with anti-Fc␥RIIa antibody (Fig. 3B). Likewise, expression of Xpresstagged wild-type and catalytically deficient SHIP-2 was comparable in the transfected cells as shown in Fig. 3C. These data suggest that SHIP-2 down-regulates Fc␥RIIa-induced Akt phosphorylation.
SHIP-2 Associates with the Phosphorylated ITAM of Fc␥RIIa-SHIP-2 is a cytoplasmic protein whose activity is regulated by subcellular localization, i.e. SHIP-2 must translocate to the membrane to access its lipid substrates (20). We therefore tested whether Fc␥RIIa provides a docking site for SHIP-2 in the following experiments. First, we used biotinylated peptides derived from the ITAM of Fc␥RIIa that were either non-phosphorylated or doubly phosphorylated at the tyrosine residues. These peptides were applied to THP-1 lysates, and the peptide-bound material was analyzed by Western blotting with anti-SHIP-2 antibody. As shown in Fig. 4A (upper panel), SHIP-2 associated with the doubly phosphorylated Fc␥RIIa ITAM peptide, but not with the non-phosphorylated peptide. A WCL from THP-1 cells is shown in the third lane. To ensure that the binding of SHIP-2 to the phosphorylated ITAM peptide was specific, the same membrane was reprobed with anti-Syk antibody (Fig. 4A, lower panel). As previously reported, the phosphorylated ITAM of Fc␥RIIa bound Syk (9).
In a second set of experiments, we investigated whether native Fc␥RIIa associates with SHIP-2 upon receptor cluster- ing. Here, THP-1 cells were activated by clustering Fc␥RIIa for 5 min. Fc␥RIIa was immunoprecipitated from resting and activated cells, separated by SDS-PAGE, and analyzed for the presence of SHIP-2 by Western blotting. The results show that SHIP-2 was capable of associating with Fc␥RIIa in an activation-dependent manner (Fig. 4B, upper panel). A reprobe of the same membrane indicated equal loading of Fc␥RIIa in both resting and activated samples. Taken together, the results suggest that SHIP-2 translocates to the membrane upon Fc␥RIIa clustering by associating with the phosphorylated ITAM of Fc␥RIIa.
The SH2 Domain of SHIP-2 Is Required for Association with the Fc␥RIIa ITAM-The above finding that SHIP-2 associates with the phosphorylated ITAM of Fc␥RIIa suggests that the interaction may be mediated via the SH2 domain of SHIP-2. To formally test this possibility, COS-7 fibroblasts were transfected to express epitope (Xpress)-tagged wild-type SHIP-2, R47L SHIP-2 (SH2 domain mutant), or D608A SHIP-2 (catalytically deficient mutant). The tagged SHIP-2 proteins were assessed for their ability to bind the Fc␥RIIa ITAM in a peptide binding assay. Thus, the peptide-bound material was separated by SDS-PAGE and immunoblotted with anti-Xpress antibody. The results indicate that the SH2 domain mutant of SHIP-2 was unable to associate optimally with the phosphopeptide (Fig. 5A, third lane). In contrast, both wild-type SHIP-2 and catalytically deficient SHIP-2 were able to efficiently bind the phosphorylated ITAM peptide (second and fourth lanes). No signal was seen in the first lane, as these cells were not transfected with the Xpress-tagged SHIP-2 proteins. To ensure that the lack of binding observed with the SH2 domain mutant was not due to lack of expression of the transfected protein, parallel immunoprecipitations were performed with anti-Xpress antibody and analyzed by Western blotting with anti-SHIP-2 antibody (Fig. 5B). Fig. 5C is a quantitative measure of the percent of Xpress-tagged SHIP-2 that bound to the phosphorylated ITAM of Fc␥RIIa. The data represent the mean Ϯ S.D. of values obtained from three independent experiments.
The SH2 Domain of SHIP-2 Is Necessary for Fc␥RIIa-induced SHIP-2 Tyrosine Phosphorylation-As an additional approach to determine the importance of the SH2 domain of SHIP-2 in its association with Fc␥RIIa and thereby its translocation to the membrane, we next assessed whether SH2 domain mutants of SHIP-2 would become phosphorylated upon Fc␥RIIa clustering. Here, COS-7 cells were transfected to express Fc␥RIIa along with wild-type SHIP-2, R47L SHIP-2 (SH2 domain mutant), or D608A SHIP-2 (catalytically deficient mutant). SHIP-2 proteins were immunoprecipitated from resting and activated cells, which had been activated by clustering Fc␥RIIa, using anti-Xpress antibody and analyzed by Western blotting with anti-phosphotyrosine antibody. The results indicate that, although both wild-type SHIP-2 and catalytically deficient SHIP-2 became tyrosine-phosphorylated upon Fc␥RIIa clustering, the SH2 domain mutant of SHIP-2 failed to become phosphorylated (Fig. 6A, upper panel). A reprobe of the membrane with anti-SHIP-2 antibody showed the presence of SHIP-2 in all lanes (Fig. 6A, middle panel). The phosphorylation level of SHIP-2 was quantitated as described under "Experimental Procedures," normalized for the amount of total SHIP-2 present in each lane, and expressed as -fold increase in SHIP-2 phosphorylation in the activated samples over that in resting samples. The graph in Fig. 6A (lower panel) represents the mean Ϯ S.D. of values obtained from three independent experiments. To ensure that the lack of phosphorylation of the SH2 domain mutant of SHIP-2 was not due to lack of expression of Fc␥RIIa in these cells, WCLs from the transfectants were probed with anti-Fc␥RIIa antibody (Fig. 6B). Taken together, these data indicate that the SH2 domain of SHIP-2 is necessary for the association of SHIP-2 with Fc␥RIIa.
SHIP-2 Is Phosphorylated by Fc␥RI Clustering in THP-1 Cells-In additional experiments, THP-1 cells were activated by clustering Fc␥RI with monoclonal antibody 197, followed by goat anti-mouse IgG. SHIP-2 phosphorylation was assessed by Western blotting with anti-phosphotyrosine antibody (Fig. 7A,  upper panel). Fig. 7A (lower panel) is a reprobe with anti-SHIP-2 antibody to ensure equal loading of SHIP-2 in all lanes.
The results indicate that SHIP-2 was indeed invoked upon Fc␥RI clustering in THP-1 cells.
Because our data indicated that SHIP-2 associated with the phosphorylated ITAM of Fc␥RIIa, we next analyzed whether a similar association might exist between SHIP-2 and the phosphorylated ITAM of the Fc␥R ␥-chain. For this, biotinylated peptides corresponding to the ITAM of the ␥-chain that were either non-phosphorylated or doubly phosphorylated were applied to THP-1 lysates. The peptide-bound material was analyzed by Western blotting with anti-SHIP-2 antibody. The results shown in Fig. 7B (upper panel) indicate that the phosphorylated ITAM of the ␥-chain was incapable of association with SHIP-2. To ensure that the phosphopeptide used in these experiments is functional, the same membrane was reprobed with anti-Syk antibody (Fig. 7B, lower panel). As shown, the phosphopeptide efficiently bound Syk, consistent with earlier reports (2). These data suggest that SHIP-2 translocation to the membrane in response to Fc␥RI clustering might involve adapter molecules or membrane-associated proteins other than the ␥-chain. Further analysis is needed to resolve these issues. DISCUSSION Fc␥R-mediated phagocytosis is regulated by a complex set of phosphatases, each using a distinct molecular mechanism for mediating its inhibitory function. Thus, recent reports have implicated the inositol phosphatase SHIP-1 in modulating Fc␥R-mediated macrophage function, including phagocytosis of IgG-opsonized particles (4,29) and induction of gene transcription (3). Likewise, Durden and co-workers have demonstrated a role for the protein-tyrosine phosphatase SHP-1 (5) and for PTEN (6) in down-regulating phagocytosis. Although it is clear that to limit the tissue damage caused by the reactive oxygen and nitrogen radicals and the inflammatory cytokines that accompany phagocytosis, this process must be subject to tight regulation, the mechanisms involved in this regulation are not fully understood.
In this study, we have analyzed the molecular details of a novel level of regulation exerted by the inositol phosphatase SHIP-2. Our results indicate that SHIP-2 associates with the ITAM of Fc␥RIIa via the SH2 domain of SHIP-2 and becomes tyrosine-phosphorylated. As in the case of SHIP-1, tyrosine phosphorylation of SHIP-2 does not influence SHIP-2 enzyme activity. However, because membrane translocation of SHIP-2 is necessary for SHIP-2 tyrosine phosphorylation by Src kinases (30) and because membrane translocation puts SHIP-2 in the proximity of its lipid substrates, tyrosine phosphorylation of SHIP-2 serves as a correlate of SHIP-2 activation. Our data also demonstrate that SHIP-2 serves to down-regulate Fc␥RIIa-induced activation of Akt and NFB-dependent gene transcription, as overexpression of wild-type SHIP-2 (but not catalytically deficient SHIP-2) almost completely abrogates these two events. Both SHIP-1 and SHIP-2 mediate their inhibitory effects via the SH2 domains, which are reported to have 54% homology. Thus, it must be noted that the enhanced activation of NFB-dependent luciferase gene induction by the dominant-negative SHIP-2 mutant D608A used in the transfection studies in THP-1 cells (Fig. 2C) may be the result of competition of D608A SHIP-2 with both endogenous SHIP-1 and SHIP-2.
In contrast to the association of SHIP-2 with the Fc␥RIIa ITAM, we were unable to detect any binding of SHIP-2 to the phosphorylated ITAM of the Fc␥RI-associated ␥-chain, despite the fact that SHIP-2 was efficiently phosphorylated upon Fc␥RI clustering (Fig. 7). Co-immunoprecipitation experiments revealed a very weak association of SHIP-2 with the ␥-chain that became apparent only when the Western blot was overexposed (data not shown). Based on the results of Fig. 7, we suggest that SHIP-2 association with the ␥-chain is likely mediated via adapter molecules or that the association may be weak and transient and is disrupted by the detergent-based lysis buffer. In a recent report, Mitchell and co-workers (31) demonstrated an association of SHIP-2 with the actin-binding protein filamin. This association is mediated via the C-terminal proline-rich domain of SHIP-2 and the SH3 domain of filamin. Because filamin is known to be associated with Fc␥RI (reviewed in Ref. 32) in macrophages, it is possible that Fc␥RI clustering recruits SHIP-2 via filamin. Nonetheless, our data indicate that Fc␥RI clustering induces tyrosine phosphorylation of SHIP-2, suggesting a role for SHIP-2 in regulating Fc␥RI-mediated events.
Interestingly, SHIP-2 expression in human PBMs is almost undetectable, but is induced upon activation of these cells with bacterial lipopolysaccharide. Similar findings were recently reported by Cambier and co-workers (24) in murine B cells. It is tempting to speculate that the induction of SHIP-2 by LPS is a mechanism employed by Gram-negative bacteria to suppress the host immune response so that the pathogen may evade the phagocytic machinery. Numerous earlier studies have reported that LPS attenuates Fc␥R-mediated phagocytosis (33,34). However, the molecular mechanism by which LPS mediates its suppressive effect has thus far remained elusive. Further studies are underway to investigate the details of LPS induction of SHIP-2 expression and the functional consequence on Fc␥Rmediated phagocytosis.
Finally, is not clear why myeloid cells express two inositol 5Ј-phosphatases, SHIP-1 and SHIP-2. Our data suggest that these two molecules may not be coexpressed at all times. Furthermore, although the results of this study demonstrate overlapping functions for SHIP-1 and SHIP-2, i.e. inhibition of Akt and NFB activation, it is likely that these two phosphatases mediate additional and non-overlapping effects on signaling pathways. Indeed, previous studies have demonstrated that the C-terminal proline-rich regions of SHIP-1 and SHIP-2 are quite different, leading to association of a distinct sets of SH3 domain-containing molecules (19), which likely influence distinct signaling pathways. We are currently investigating the functional differences between SHIP-1 and SHIP-2 in Fc␥Rmediated activation of monocytes and macrophages.