Enzymatic activity of the Src homology 2 domain-containing inositol phosphatase is regulated by a plasma membrane location.

The negative regulatory role of the Src homology 2 domain-containing inositol 5-phosphatase (SHIP) has been invoked in a variety of receptor-mediated signaling pathways. In B lymphocytes, co-clustering of antigen receptor surface immunoglobulin with FcgammaRIIb promotes the negative effects of SHIP, but how SHIP activity is regulated is unknown. To explore this issue, we investigated the effect of SHIP phosphorylation, receptor tyrosine engagement by its Src homology 2 domain, and membrane recruitment of SHIP on its enzymatic activity. We examined two SHIP phosphorylation kinase candidates, Lyn and Syk, and observed that the Src protein-tyrosine kinase, Lyn is far superior to Syk in its ability to phosphorylate SHIP both in vitro and in vivo. However, we found a minimal effect of phosphorylation or receptor tyrosine engagement of SHIP on its enzymatic activity, whereas membrane localization of SHIP significantly reduced cellular phosphatidylinositol 3,4, 5-triphosphate levels. Based on our results, we propose that a membrane localization of SHIP is the crucial event in the induction of its phosphatase effects.

Clustering of the B cell surface immunoglobulin (sIg) 1 antigen receptor by binding of foreign antigen initiates a set of biochemical events termed positive signaling which activate B lymphocytes to proliferate and secrete soluble antigen specific Ig (reviewed in Refs. [1][2][3]. The most proximal signaling event is the stimulation of the Src family of protein-tyrosine kinases (PTKs), which phosphorylates tyrosine residues within conserved immunoreceptor tyrosine-based activation motifs (ITAMs), found in receptor-associated proteins (4). Once the tyrosines in the ITAM are phosphorylated, they serve as docking sites for numerous proteins and enzymes containing Src homology 2 (SH2) domains including the PTK Syk and the p85 adapter subunit of phosphatidylinositol 3-kinase (PtdIns 3-ki-nase; reviewed in Refs. 5 and 6). These activation signals are then propagated through additional tyrosine phosphorylation and protein-protein interactions and result in changes in B cell biology.
In contrast to activating signals generated upon sIg clustering, co-clustering of sIg with the B cell Fc receptor for IgG (Fc␥RIIb) aborts B cell activation. It has been proposed (14,15) that co-clustering of sIg and Fc␥RIIb occurs late in the humoral immune response to block continued Ig production. We have termed sIg-Fc␥RIIb co-clustering "negative signaling," to contrast with positive signaling initiated by sIg clustering alone and that promotes B cell proliferation. Phillips and Parker (16 -18) described an in vitro model using a F(abЈ) 2 fragment and intact anti-Ig reagents of the IgG class to study biochemical events associated with positive and negative signaling.
Earlier studies using this model demonstrated phosphorylation of a tyrosine residue contained within a 13-amino acid motif of the Fc␥RIIb cytoplasmic tail (19). Additional experiments showed that certain structural features of the Fc␥RIIb 13amino acid motif were shared in common with other receptors that likewise conferred an inhibitory function; accordingly, the motif has been termed the immunoreceptor tyrosine-based inhibitory motif (ITIM; reviewed in Ref. 20). Like signaling through the ITAM, the phosphorylated ITIM recruits SH2 domain-containing molecules to carry out negative signaling.
The SH2 domain-containing inositol 5-phosphatase, SHIP, was identified as one of several proteins that bind to the tyrosine-phosphorylated ITIM of Fc␥RIIb (21,22). SHIP is a 145-kDa cytosolic protein that contains a single SH2 domain, a central catalytic region, and two tyrosine phosphorylation sites in the C-terminal region (23)(24)(25). In addition to the B cell Fc␥RIIb, SHIP is recruited to and inhibits cellular activation by a variety of other receptors, including numerous cytokine receptors and the mast cell Fc⑀ receptor (reviewed in Refs. 26 and 27). Recently, in the B cell model, we demonstrated that SHIP was tyrosine-phosphorylated to high stoichiometry and associated with Ras adapter protein Shc only upon co-clustering sIg with Fc␥RIIb (28). These events were because of the direct recruitment of the SH2 domain of SHIP to the phosphorylated cytoplasmic tyrosine residue of the Fc␥RIIb ITIM and hence induces SHIP recruitment to the B cell plasma membrane (29). Similar findings of SHIP membrane translocation have been made in T cells stimulated via CD28 (30). The kinase-phosphorylating SHIP is unknown in any cellular system.
Studies of SHIP enzymatic activity revealed an exclusive preference for the hydrolysis of 3-phosphoinositides. As such, SHIP can reverse the action of PtdIns 3-kinase by consuming the PtdIns 3-kinase products (31). Other experiments revealed that co-clustering of sIg and Fc␥RIIb leads to a dramatic reduction of cellular PtdIns-(3,4,5)P 3 (32,33) and reduced activity of distal PtdIns-(3,4,5)P 3 -responsive enzymes such as Btk (32,34) and Akt (35,36). Experiments by our lab (35) and others (37) showed that the reduction of PtdIns-(3,4,5)P 3 is not due to inactivation of PtdIns 3-kinase, because we observed no defect in p85 protein association or in membrane translocation of PtdIns 3-kinase. Therefore, stimulation of SHIP enzymatic activity and the resulting hydrolysis of PtdIns-(3,4,5)P 3 are most likely the cause of PtdIns-(3,4,5)P 3 -dependent enzyme inhibition.
Despite the fact that SHIP enzymatic activity is induced by various cytokines and immunoreceptors, it is unclear how the enzymatic activity of SHIP is regulated. Because SHIP was highly phosphorylated upon co-clustering sIg with Fc␥RIIb (21) and recruited to the plasma membrane through the engagement with the phosphorylated receptors (29,30,35,38), we have formulated three distinct hypotheses that can account for the observed increased SHIP enzymatic activity. First, SHIP enzymatic activity may be directly stimulated by tyrosine phosphorylation. Other enzymes such as Vav (39,40) or phospholipase C␥ (41) respond in this way to tyrosine phosphorylation. Second, the engagement of a phosphorylated ITIM may promote the elevated activity of SHIP. The SHP-1 tyrosine phosphatase, which binds to the same phosphorylated ITIM of Fc␥RIIb, exhibits an ϳ50-fold enhancement of phosphatase activity upon ITIM engagement by its SH2 domain (42). Third, SHIP membrane recruitment may lead to induced enzymatic activity. Many enzymes involved in signal transduction including Akt (43), and PtdIns 3-kinase (13) are activated by changing their subcellular location from the cytosol to the membrane.
To investigate the effect of phosphorylation on SHIP enzymatic activity, we sought the identity of the kinase capable of SHIP phosphorylation. We considered two candidate PTKs, Src family kinases (44) and Syk (45), because members of these two PTK families are present and associated with the ITAMs of sIg and thus are capable of promoting SHIP phosphorylation upon sIg-Fc␥RIIb co-clustering. We observed that Src family PTKs are superior to Syk regarding SHIP phosphorylation, both in vitro and in vivo. Next, we addressed the three proposed mechanisms of SHIP enzymatic regulation and found a minimal effect on SHIP 5-phosphatase activity by either phosphorylation or engagement of the phosphorylated ITIM of Fc␥RIIb by the SH2 domain of SHIP. However, enforced plasma membrane localization of SHIP decreased the total amount of cellular PtdIns-(3,4,5)P 3 in cells expressing a chimera of human CD8-SHIP, indicating that the subcellular localization directly contributes to SHIP-mediated substrate hydrolysis. Based on these results, we propose that SHIP negatively affects PtdIns-(3,4,5)P 3 -dependent enzymes by altering its subcellular location from cytosol to the plasma membrane, which is accomplished by recruitment to phosphorylated cytoplasmic tyrosines of receptors.
Plasmids-SHIP cDNA was obtained from Dr. G. Krystal (University of British Columbia), and the EcoRI-PvuII fragment was subcloned into the yeast expression vector, pPICZB-His (Invitrogen, San Diego, CA). The subcloning removes a region encoding the C-terminal 34 amino acids of SHIP; this region contains no phosphorylation site or protein interaction domain and is not part of the catalytic region. Human cDNA encoding CD8␣ was obtained from Dr. J. Parnes (Stanford University, Palo Alto, CA) and inserted into the EcoRI site of pEF-SHIP (35), containing the full-length SHIP cDNA. The resulting material encodes the extracellular and transmembrane domain of CD8␣ and SHIP as the intracellular domain of the chimera. cDNA encoding murine Lyn was obtained from Dr. A. DeFranco (University of California, San Francisco); cDNA encoding Syk was obtained from Dr. K. Zoller (Ariad Pharmaceuticals, Cambridge, MA).
Generation of Recombinant SHIP from Yeast-Cloned pPICZB-SHIP-His was transformed to yeast pichia strain, GS115. Transformants expressing high levels of SHIP protein were selected and induced by methanol according to manufacturer's directions. His-tagged SHIP protein was purified by Ni 2ϩ -agarose affinity chromatography and subsequent size exclusion chromatography. The resulting material produced a single band upon SDS-PAGE of ϳ145 kDa, the expected size of the recombinant protein.

5-Phosphatase Activity Measurements of SHIP Phosphorylated by in Vitro and in Vivo Kinase
Reaction-Lyn from pervanadate (1 mM sodium orthovanadate, 0.6% H 2 O 2 )-stimulated A20 B cells and Syk from pervanadate-stimulated THP-1 monocytic cells were immunoprecipitated as described earlier (46) and resuspended in kinase buffer (20 mM HEPES (pH 7.4), 10 mM MgCl 2 , 5 mM MnCl 2 ). To each reaction, 10 M cold ATP, 2-4 Ci of [␥-32 P]ATP (3000 Ci/mmol), and 2 g of recombinant SHIP were incubated with the immunoprecipitated kinases for 10 min at 30°C in a total volume of 25 l to permit SHIP phosphorylation. The reaction was stopped by adding 5ϫ SDS sample buffer (0.6 M Tris (pH 6.8), 50% glycerol, 12% SDS) and incubating the samples at 95°C for 5 min. The phosphorylated products, including the autophosphorylated kinases themselves, were analyzed by 7.5% SDS-PAGE, identified by autoradiography, and quantified by a Molecular Dynamics Storm system. To assess the effect of phosphorylation on enzymatic activity, the in vitro kinase reaction was performed as above. Because the Mn 2ϩ cation was inhibitory toward SHIP 5-phosphatase activity (see "Results"), we lowered [Mn 2ϩ ] by mixing 23 l of supernatant containing recombinant, phosphorylated SHIP with 1 and 10 mM EDTA and MgCl 2 , respectively. After incubation on ice for 10 min, 0.1 g of recombinant SHIP was subjected to 5-phosphatase assay. For transfection into COS-7, 10 g of cDNA encoding Lyn or Syk PTKs were co-transfected with 10 g of cDNA encoding SHIP. Cells were harvested after 48 h, stimulated with pervanadate, and analyzed by antiphosphotyrosine blotting as described (21). In parallel, Lyn and Syk were immunoprecipitated from the SHIP co-transfected cells and assessed for in vitro kinase activity by autophosphorylation and by antiphosphoty-rosine blotting. For measurements of SHIP 5-phosphatase activity after tyrosine phosphorylation in vivo, COS-7 cells co-transfected with SHIP and/or Lyn were stimulated with or without pervanadate for 5 min and immunoprecipitated with anti-SHIP antibody. For measurements of SHIP 5-phosphatase activity after stimulation of B cells, lysates of resting or intact anti-Ig-stimulated A20 B cells were incubated with 2.4 M biotinylated, phosphorylated ITIM peptide, as earlier described (29). The bound material was eluted by incubating streptavidin beads with 100 mM phenylphosphate in SHIP assay buffer, and the eluted material was applied to 5-phosphatase assays.
Immunofluorescence Studies-COS-7 cells were transfected with cDNA encoding wild-type SHIP or the CD8-SHIP chimera. 48 h after transfection, the viable cells were detached by cell dissociation solution TM (Sigma) and stained with fluorescein isothiocyanate-conjugated anti-CD8 and biotinylated cholera toxin B subunit (to visualize the plasma membrane), followed by Alexa 568 (Molecular Probes, Eugene, OR)-conjugated streptavidin. Additionally, transfected COS-7 cells were permeabilized to identify intracellular proteins with 0.5% Triton X-100 in Tris-buffered saline and immunostained with Alexa 568-conjugated anti-SHIP antibodies. Cells were analyzed by confocal microscopy, and digital images were prepared.

Generation, Purification, and Assay of Recombinant SHIP-
Purified SHIP enzyme was prepared from yeast transformed with cDNA encoding SHIP and containing a 6x Histidine tag at the C terminus. About 1 mg of SHIP enzyme was purified from yeast lysates derived from a 1-liter culture (Fig. 1A). We tested the purified recombinant enzyme in a phosphatase assay, using [ 32 P]PtdIns-(3,4,5)P 3 prepared from commercial PtdIns-(4,5)P 2 and immunoprecipitated PtdIns 3-kinase. The results, shown in Fig. 1B, indicated that SHIP generated [ 32 P]PtdIns-(3, 4)P 2 from [ 32 P]PtdIns-(3,4,5)P 3 , indicating that the recombinant enzyme had 5-phosphatase activity, as expected. The enzymatic activity was lost in the presence of EDTA, revealing a need for a divalent cation to support SHIP hydrolytic activity toward [ 32 P]PtdIns-(3,4,5)P 3 . We examined several divalent cations for their ability to support SHIP activity and found ( Fig. 2A) that only Mg 2ϩ was active in this regard. SHIP enzymatic activity was greatly reduced in the presence of Mn 2ϩ and the oxidant, H 2 O 2 (Fig. 2B). SHIP displayed a relatively broad pH optimum from pH 6 to pH 8 (Fig. 2C).
The Src Family PTK Lyn Is Superior to Syk in Its Ability to Phosphorylate SHIP-To investigate whether tyrosine phosphorylation of SHIP affects its enzymatic activity, it was necessary to identify the PTK capable of phosphorylating SHIP. Both Src family PTKs and Syk associate with the B cell antigen receptor and either kinase family member are thus properly positioned for carrying out SHIP phosphorylation upon sIg-Fc␥RIIb co-clustering. Accordingly, we investigated the ability of each of these candidate PTKs to phosphorylate SHIP. Using immunoprecipitated Lyn (as a representative Src family kinase member) or Syk in an in vitro kinase reaction with the recombinant SHIP prepared as described above, we found (Fig. 3A) that Lyn was far superior to Syk in SHIP phosphorylation; indeed, there was no detectable increase in SHIP tyrosine phosphorylation in the presence of Syk, despite the fact that Syk displayed high autophosphorylating activity (Fig. 3A, in-set 2 ). As a second approach, we applied the Src family PTK inhibitor PP2 (48) and the Syk inhibitor piceatannol (49) to A20 B cells stimulated under negative signaling conditions with intact anti-Ig, which promotes high stoichiometry SHIP tyrosine phosphorylation, as described earlier (21). We observed (Fig. 3B) that the Src family PTK inhibitor PP2 reduced SHIP tyrosine phosphorylation about 60%, whereas the Syk inhibitor piceatannol did not affect SHIP tyrosine phosphorylation. Lastly, we transfected COS-7 fibroblasts with cDNA encoding SHIP and co-transfected the cells with cDNA encoding either Lyn or Syk. After pervanadate stimulation, SHIP phosphotyrosine content was assessed by antiphosphotyrosine Western blots of SHIP immunoprecipitates from the transfectants. The results (Fig. 3C) indicated that although SHIP was phosphorylated upon pervanadate stimulation in all conditions, cotransfection with Lyn but not Syk induced potent and prominent SHIP tyrosine phosphorylation. Both kinases displayed activity when obtained from identically transfected cells and applied to an in vitro kinase assay and as measured by antiphosphotyrosine immunoblotting (Fig. 3C, inset 2 ). Together, these findings reveal that Src family PTKs but not Syk are capable of phosphorylating SHIP both in vitro and in vivo.
Phosphorylation Has a Minimal Effect on the Enzymatic Activity of SHIP-To explore the effect of in vivo SHIP tyrosine 2 A summary of the findings appears here; the original data were provided for review. phosphorylation on its 5-phosphatase activity, SHIP enzyme was collected from lysates of resting B cells or B cells stimulated with intact anti-mouse IgG, as we have described (38) using the phosphorylated ITIM peptide of Fc␥RIIb. SHIP obtained in this way was tyrosine-phosphorylated under these stimulation conditions (Fig. 4A, upper panel). The enzyme was eluted by the addition of 100 mM phenylphosphate, and equal amounts of phosphorylated or nonphosphorylated forms, confirmed by Western blot, were applied to the 5-phosphatase assay described above. We found (Fig. 4B) that the rate of PtdIns-(3,4,5)P 3 hydrolysis by SHIP was not affected by its tyrosine phosphorylation. To confirm the data, we took advantage of the fact that Lyn phosphorylates SHIP to high stoichiometry in vitro. Thus, recombinant SHIP was phosphorylated by immunoprecipitated Lyn in vitro, as shown in Fig. 3A, and the phosphorylated material, or sham-phosphorylated control, was subjected to the 5-phosphatase assay. We observed no significant difference in enzymatic activity between the nonphosphorylated and phosphorylated forms of SHIP (Fig. 4, C  and D). Lastly, SHIP was phosphorylated in vivo by co-transfection into COS-7 cells with Lyn kinase and stimulated by pervanadate treatment. The activity of the phosphorylated and nonphosphorylated forms of SHIP was assessed after immunoprecipitation. The results (Fig. 4, C and D) indicated no significant change in SHIP 5-phosphatase activity upon its phosphorylation, despite efficient tyrosine phosphorylation by cotransfected Lyn. Thus, SHIP catalytic activity is not affected by tyrosine phosphorylation.
The Engagement of the Phosphorylated ITIM of Fc␥RIIb with SHIP Has No Effect on the Enzymatic Activity of SHIP-Earlier studies indicated stimulation of SHP-1 phosphotyrosine phosphatase activity upon engagement of a phosphorylated peptide corresponding to the ITIM of Fc␥RIIb (42). To test the possibility that SHIP activity was similarly regulated, recom-binant SHIP was incubated with phosphorylated ITIM peptide or an unrelated phosphorylated peptide for 1 h, and the mixture was subjected to the 5-phosphatase assay (Fig. 5). Control experiments using these peptides in a pull-down assay (Fig.  5A) indicated that the phosphorylated Fc␥RIIb ITIM peptide but not other peptides were capable of engaging SHIP. However, despite its ability to bind SHIP, the ITIM peptide did not affect SHIP 5-phosphatase activity (Fig. 5, B and C). Thus, unlike SHP-1 phosphatase, SHIP is not activated by engagement of its SH2 domain.
Membrane Localization of SHIP Is the Major Mechanism Promoting Increased Activity upon Co-clustering of sIg-Fc␥RIIb-Enzymes involved in signal transduction are frequently regulated by redistributing their subcellular location to the plasma membrane, which likely promotes enzyme access to substrates or interaction with other molecules. SHIP can be recruited to the membrane by binding of its SH2 domain to the phosphorylated ITIM of Fc␥RIIb or cytoplasmic tyrosines of other receptors, as shown earlier (29,30,38). To investigate whether a membranous location of SHIP leads to enhanced hydrolysis of PtdIns-(3,4,5)P 3 , the SHIP substrate, we measured cellular PtdIns-(3,4,5)P 3 levels in cells harboring SHIP in the cytosol or in the plasma membrane. A membrane-targeted chimera of SHIP was generated by fusing the enzyme with the C-terminal region of human CD8␣. COS-7 fibroblasts were transfected with a plasmid encoding SHIP, the CD8-SHIP chimera, or empty vector. Expression of CD8-SHIP was confirmed by fluorescence-activated cell sorter analysis and indicated that ϳ17% of transfected cells expressed CD8 on their surface (Fig. 6A). Additionally, cells transfected with the CD8-SHIP chimera displayed a slower migrating form, relative to COS-7 cells transfected with wild-type SHIP, as revealed by SDS-PAGE and immunoblotting with anti-SHIP sera (Fig. 6B). The slower migrating form is because of the presence of the ϳ30 kDa CD8␣ fused to the ϳ150-kDa SHIP gene product. To assess the membrane localization of the CD8-SHIP chimera, we performed immunofluorescence studies using confocal microscopy. Cells transfected with wild-type SHIP showed no CD8 on their surface upon immunostaining with anti-CD8, in contrast to those transfected with CD8-SHIP (Fig. 6C). Lastly, cells transfected with CD8-SHIP showed SHIP present in the plasma membrane when immunostained with anti-SHIP antibodies, whereas those transfected with wild-type SHIP showed a diffuse cytosolic location of the enzyme (Fig. 6D).
To measure the effect of membrane-targeted SHIP on 3-phosphoinositide levels, the transfected COS-7 cells were labeled to equilibrium with [ 32 P]orthophosphate for 14 h, trypsinized, and stimulated with pervanadate for 3 min. Equilibrium labeling was done so that changes in 32 P counts in 3-phosphoinositides represent changes in their mass levels, rather than rates of label incorporation. These experiments revealed an ϳ9% decrease in cellular PtdIns-(3,4,5)P 3 levels in cells transfected with wild-type SHIP localized to the cytosol, as compared with vector only transfectants (Fig. 6, E and F). However, similar measurements of CD8-SHIP-transfected cells showed a nearly 30% decline in total cellular PtdIns-(3,4,5)P 3 levels, despite the fact that only a small fraction of the cells express the transfected gene. These findings argue that PtdIns-(3,4,5)P 3 consumption by SHIP is activated upon redistribution of the active enzyme to the plasma membrane. DISCUSSION It has been proposed that co-clustering of Fc␥RIIb and sIg on B cells results in an inhibitory signal that provides a negative feedback mechanism for antibody production. Previous findings (32) indicated that sIg-Fc␥RIIb co-clustering decreased B cell PtdIns-(3,4,5)P 3 levels; moreover, this reduction of PtdIns-(3,4,5)P 3 was reversed by a catalytically inactive SHIP mutant. These data strongly suggest that the decreased PtdIns-(3,4,5)P 3 was because of an increase in the enzymatic activity of SHIP under a sIg-Fc␥RIIb co-clustering condition, but how the enzymatic activity of SHIP is regulated is unknown. We investigated various known features of the SHIP enzyme that occur under sIg-Fc␥RIIb co-clustering conditions as potential means to implement activation of 5-phosphatase activity. Our results indicated that cells harboring a membrane-localized form of SHIP showed significantly reduced PtdIns-(3,4,5)P 3 levels. We conclude that SHIP-mediated consumption of PtdIns-(3,4,5)P 3 and inhibition of PtdIns-(3,4,5)P 3 -dependent enzymes is initiated by and a consequence of recruitment of the constitutively active enzyme to a phosphorylated receptor within the plasma membrane.
Membrane localization of cytosolic proteins is a common means by which enzymes involved in signaling pathways promote their activation. In this report, we have shown that expression of a CD8-SHIP chimera resulted in depletion of nearly 30% of total cellular PtdIns-(3,4,5)P 3 , the SHIP substrate, de- spite the fact that only 17% of COS-7 cells expressed the chimeric protein. Based on this finding, we propose that SHIP exerts its negative influence on signaling pathways by membrane localization upon SH2 domain engagement of cytoplasmic tyrosine residues within cytokine receptors or immunoreceptors. In earlier reports (29,38,50), we proposed a model regarding SHIP-protein interactions in which the phosphoryl-ated ITIM of Fc␥RIIb or cytoplasmic tyrosines of cytokine receptors formed a docking site for the SH2 domain of SHIP. The SH2 domain engagement thus serves two purposes: first, to bring SHIP within the range of receptor-associated PTKs to allow SHIP phosphorylation and second, to bring SHIP within the range of its cellular substrate, PtdIns-(3,4,5)P 3 . Consistent with this interpretation of events, earlier experiments revealed that overexpression of the SH2 domain of SHIP blocked the negative signal delivered by the cytoplasmic tail of Fc␥RIIb (51). It is likely in those cells that the overexpressed SH2 domain bound to the phosphorylated ITIM and prevented membrane recruitment of endogenous SHIP. However, and in contrast to this view, experiments of Xenopus oocyte maturation (31) revealed that an inactivating mutant of the SH2 domain of SHIP did not affect its inhibitory influence on extracellular signal-regulated kinase kinase induction by insulin. These findings suggest that SHIP carries out its inhibitory function without recruitment to any receptor and may therefore have a distinct mechanism in this particular cellular context.
Interestingly, we did not see a large decrease in total cellular PtdIns-(3,4,5)P 3 of resting cells expressing CD8-SHIP but only in those cells stimulated with pervanadate. The reasons for this observation may be technical; we found only very small levels of PtdIns-(3,4,5)P 3 in resting cells, and it may not be possible to detect a decrease in the extremely low background levels. We interpret the need for pervanadate stimulation as indicating the need for induction of PtdIns 3-kinase to elevate PtdIns-  (IgH). B, SHIP was obtained using a synthetic phosphopeptide corresponding to the phosphorylated ITIM of Fc␥RIIb from lysates of B cells stimulated with intact anti-IgG or from nonstimulated B cells. The phosphopeptide-bound material was eluted with 100 mM phenylphosphate, and equal amounts of the eluted enzyme were applied to the SHIP 5-phosphatase assay. Samples were obtained from resting (open squares) or activated (closed triangles) B cells or activated B cells using beads only and no peptide (closed circles). The eluted samples were subjected to SHIP 5-phosphatase assay. The reaction was stopped at the indicated times, and the products were separated by oxalate TLC. Shown is the ratio of product to substrate plus product. C and D, recombinant SHIP was phosphorylated using the in vitro kinase (IVK) reaction with anti-Lyn, anti-Syk, or normal rabbit serum immunoprecipitates, as described in Fig. 2A. Following the phosphorylation reaction, the enzymatic activity of SHIP in the IVK was measured as described under "Experimental Procedures." Alternatively, SHIP immunoprecipitates from COS-7 cells transfected with SHIP and vector or SHIP and Lyn, as indicated, were assessed for 5-phosphatase activity. Shown are the TLC analyses of the reaction products (C) and quantified activities as the ratio of product to substrate plus product after background levels of product (D). These results are representative of three separate and similar assays.
(3,4,5)P 3 in the transfected fibroblasts to a level in which it is detectable and sensitive to SHIP-mediated hydrolysis. However, this phenomena may also reflect a requirement to induce SHIP phosphorylation; i.e. membrane-associated SHIP may not be active in vivo until it is phosphorylated. Because pervanadate induces SHIP phosphorylation, as well as increased PtdIns-(3,4,5)P 3 levels through PtdIns 3-kinase in B cells and fibroblasts, we cannot distinguish between these possibilities.
Activation of SHIP 5-phosphatase activity upon its tyrosine phosphorylation was proposed by several groups. However, reports of SHIP immunoprecipitates from cytokine-activated B6SYtA1 (23) or FDC-P1 cells (25) indicated no significant difference in hydrolysis of PtdIns-(3,4,5)P 3 or Ins-(1,3,4,5)P 4 , implying that phosphorylation per se was not causally related to activation of the enzyme. However, it is possible that only a very small amount of phosphorylated SHIP was present in these samples, which might preclude detection of enhanced activity in an in vitro assay. Other studies in yeast indicated that co-expression of the tyrosine kinase Lck and SHIP resulted in the tyrosine phosphorylation of SHIP, accompanied by a 2-3-fold reduction in the level of 5-phosphatase activity (52), suggesting that SHIP enzymatic activity might be reduced upon its phosphorylation.
To more rigorously assess the effect of SHIP tyrosine phosphorylation on its enzymatic activity, it was necessary to identify a kinase capable of phosphorylating SHIP. We examined the Src family member Lyn and Syk in this capacity, because both tyrosine kinases are present upon co-clustering sIg-Fc␥RIIb on the B cell surface. Src family PTKs associate with sIg in the resting state and to a greater extent upon sIg stimulation (53). In addition, Fc␥RIIb was phosphorylated by Src family PTKs in vitro (54), and Fc␥RIIb phosphorylation was deficient in lyn (Ϫ/Ϫ) cells (55). Additionally, both Lyn and Syk bind with high affinity to the phosphorylated ITAM-containing proteins associated with sIg (56).
We present three distinct arguments to indicate that SHIP is phosphorylated by a Src family kinase. First, the Src family member Lyn but not Syk was capable of phosphorylating recombinant SHIP in vitro. Second, SHIP tyrosine phosphorylation in B cells was sensitive to the Src inhibitor, PP2, but not the Syk inhibitor, piceatannol. Third, co-transfection of SHIP with Lyn promoted strong SHIP phosphorylation in fibroblasts, whereas co-transfection with Syk failed to induce SHIP tyrosine phosphorylation. Although these findings are consistent with a direct role for a Src family PTK in SHIP phosphorylation, we cannot rule out the possibility of an indirect role mediated through another Src-dependent PTK.
We tested the effect of phosphorylation on SHIP enzymatic activity by three methods. First, kinetic measurements indicated no change in 5-phosphatase activity when SHIP was obtained under conditions leading to its high stoichiometry phosphorylation upon stimulation of cells with intact anti-Ig. Second, phosphorylation of SHIP by in vitro kinase reactions did not significantly change its 5-phosphatase activity. Third, SHIP activity was unaffected when derived from in vivo phosphorylation conditions of SHIP and Lyn co-transfected COS-7 fibroblasts. These data argue that phosphorylation of SHIP by co-clustering of sIg and Fc␥RIIb contributes to SHIP protein interactions to affect other cellular functions (38), rather than regulating SHIP enzyme activity.
FIG. 6. Cells harboring membrane-targeted SHIP have decreased PtdIns-(3,4,5)P 3 levels. A, COS-7 cells were transfected with vector or CD8-SHIP chimera. 48 h after transfection, the cells were stained with antibodies to human CD8 and analyzed by flow cytometry. The untransfected cells displayed less than 1% in the M2 gate. B, COS-7 cells were transfected with vector encoding wild-type SHIP or the CD8-SHIP chimera. 48 h after transfection, the cells were lysed and analyzed by immunoblotting with anti-SHIP. Untransfected COS-7 cells do not express SHIP protein (not shown). The arrowheads indicate the ectopically expressed SHIP or CD8-SHIP chimera. C, COS-7 cells were transfected with the indicated cDNA and stained with anti-CD8 to reveal the CD8-SHIP chimera in the plasma membrane and the B subunit of cholera toxin to outline the plasma membrane. D, COS-7 cells were transfected with the indicated cDNA, permeabilized with detergent, and immunostained with anti-SHIP antisera. E, COS-7 cells were transfected with vector only or vector encoding SHIP or the CD8-SHIP chimera. After 30 h, the transfectants were metabolically labeled to equilibrium with [ 32 P]inorganic phosphate to label pools of inositol phospholipids. The cells were stimulated with pervanadate for 3 min, and extracted phospholipids were analyzed by TLC. The arrowhead indicates the spot corresponding to PtdIns-(3,4,5)P 3 , as determined by migration of authentic standards. F, PtdIns-(3,4,5)P 3 levels for the TLC shown in C were quantitated by Molecular Dynamics Storm System and normalized to total phospholipids to control for cell number. The data are shown as a percentage of the PtdIns-(3,4,5)P 3 amount in the pervanadate-stimulated, vector only transfected cells. The data are the average of two identical experiments.