Growth Factors and Insulin Stimulate Tyrosine Phosphorylation of the 51C/SHIP2 Protein

Antibodies raised against the 51C/SHIP2 inositol polyphosphate 5′-phosphatase were used to examine the effects of growth factors and insulin on the metabolism of this protein. Immunoblot analysis revealed that the 51C/SHIP2 protein was widely expressed in fibroblast and nonhematopoietic tumor cell lines, unlike the SHIP protein, which was found only in cell lines of hematopoietic origin. The 51C/SHIP2 antiserum precipitated a protein of approximately 145 kDa along with an activity which hydrolyzed phosphatidylinositol 3,4,5-trisphosphate to phosphatidylinositol 3,4-bisphosphate. Tyrosine phosphorylation of the 51C/SHIP2 protein occurred in response to treatment of cells with epidermal growth (EGF), platelet-derived growth factor (PDGF), nerve growth factor (NGF), insulin-like growth factor-1 (IGF-1), or insulin. EGF and PDGF induced transient tyrosine phosphorylation of 51C/SHIP2, with maximal tyrosine phosphorylation occurring at 5–10 min following treatment and returning to near basal levels within 20 min. In contrast, treatment of cells with NGF, IGF-1, or insulin resulted in prolonged tyrosine phosphorylation of 51C/SHIP2 protein, with 40–80% maximal phosphorylation sustained for up to 2 h following agonist treatment. The kinetics of activation of the Akt/PKB protein kinase by the various factors correlated well with the kinetics of tyrosine phosphorylation of 51C/SHIP2. EGF, NGF, and PDGF stimulated the association of 51C/SHIP2 protein with the Shc adapter protein; however, no Shc could be detected in 51C/SHIP2-immune precipitates from cells treated with IGF-1 or insulin. The data suggest that 51C/SHIP2 may play a significant role in regulation of phosphatidylinositol 3′-kinase signaling by growth factors and insulin.

The production of some forms of phosphoinositides depends on the activities of inositol polyphosphate-5-phosphatases. Many different species of these enzymes exist, with a recently described member, SHIP (for SH2 domain-containing inositol 5-phosphatase), being implicated in receptor signaling in hematopoietic cells) (19 -21). SHIP is tyrosine phosphorylated in response to treatment of cells with erythropoietin, interleukin-2, interleukin-3, macrophage colony-stimulating factor, B cell receptor cross-linking, and T cell activation (22); however, in most cases, phosphorylation does not appear to affect activity. SHIP has an SH2 domain, multiple proline-rich sites representing possible sites of interaction with SH3 domains, two NPXY motifs, and associates with Shc, Grb2, and SHP-2 under certain conditions (23)(24)(25). The 5Ј-phosphatase activity of SHIP is specific for phosphatidyinositols and inositol which are phosphorylated at the 3Ј position (23). In some systems, SHIP appears to negatively regulate cell growth (21) or induce apoptosis (25), perhaps counteracting growth factor signals. It is not clear whether tyrosine phosphorylation of SHIP promotes or diminishes these inhibitory effects.
Since PI 3-kinase appears to play a fundamental role in signal transduction in all mammalian cell types and SHIP expression is limited to hematopoietic cells, we examined whether the more widely expressed SHIP-related protein 51C/ SHIP2 (26,27) is involved in signaling from receptor tyrosine kinases in other cells.
Immune Precipitation-Cells were cultured in 100-mm dishes, treated with agonist for the indicated times, washed twice with 10 ml of cold phosphate-buffered saline, and lysed in 1 ml of buffer containing 50 mM HEPES, pH 7.5, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1 mM EGTA, 10 mM sodium pyrophosphate, 100 mM NaF, 30 mM p-nitrophenol phosphate, 100 M sodium orthovanadate, 10 M aprotinin, 10 M leupeptin, and 1 mM benzamidine (modified HNTG buffer as described by Margolis et al. (30)). Extracts were centrifuged for 10 min at 10,000 ϫ g, and supernatants were incubated for 60 min at 4°C with the desired antiserum. Immune complexes were collected on protein A-Sepharose beads, washed three times with 1 ml of lysis buffer, and analyzed on SDS-polyacrylamide gels as described previously (28).
Immunoblotting-For analysis of whole cell levels of 51C/SHIP2 and SHIP, cells were grown in 35-mm dishes, washed with phosphatebuffered saline, and lysed in SDS sample buffer. Samples containing 25 g of protein were loaded onto 4 -15% polyacrylamide-SDS gels and transferred to nitrocellulose membranes, and the membranes were probed with preimmune serum, 51C/SHIP2 antiserum, or SHIP antiserum. For other experiments, immunoprecipitated proteins were resolved on SDS-gels, transferred to nitrocellulose, and probed with antiphosphotyrosine antibody (4G10, Upstate Biotechnology) (1:1000 dilution in 2% bovine serum albumin) or with a 1:1000 dilution of Shc antibodies. Blots were developed with a chemiluminescent detection system (New England Biolabs), except for anti-Shc immunoblots which were developed with 125 I-protein A.
The lipids were suspended in 200 l of reaction buffer consisting of 50 mM Tris-HCl, pH 7.4, and 10 mM MgCl 2 by sonication with a probe sonicator. 30 l of lipid suspension were added to each sample of washed beads, and the reaction mixtures were incubated with rocking at room temperature for 20 min. Samples were extracted with 100 l of chloroform:methanol 1:1 and 100 l of 2 M KCl, and the products were separated by TLC (21).

RESULTS
Antiserum was raised against the 51C/SHIP2 protein to examine the role of this protein in growth factor signaling. The region of 51C/SHIP2 chosen as antigen showed no similarity to any region of SHIP. Initial experiments were performed to test the specificity of the antiserum and compare the distribution of 51C/SHIP2 with that of SHIP in a variety of cell lines. As shown in Fig. 1A, the 51C/SHIP2 antiserum specificity recognized a protein of 145 kDa in a number of adherent human and rodent cell lines. Only the HL-60 and A20 hematopoietic cell lines failed to express the protein. In contrast, 145-kDa SHIP reactivity was seen only in HL-60 and A20 cells, with a somewhat larger prominent band present in PC-12 cells. The 51C/ SHIP2 antiserum also specifically immunoprecipitated a protein of 145 kDa from [ 35 S]methionine-labeled SH-SY5Y cells (Fig. 1B).
Since SHIP becomes tyrosine-phosphorylated following cytokine treatment of hematopoietic cells, we examined the effects of growth factor treatment on tyrosine phosphorylation of 51C/ SHIP2 (Fig. 3). EGF, PDGF, and IGF-1 all stimulated tyrosine phosphorylation of 51C/SHIP2 in SH-SY5Y cells, as did NGF in PC-12 cells and insulin in 3T3-L1 adipocytes. Tyrosine-phosphorylated bands of about 52 and 66 kDa co-precipitated with 51C/SHIP2 from EGF-treated SH-SY5Y cells and a band of 52 kDa co-precipitated with 51C/SHIP2 from NGF-treated PC-12 cells. 51C/SHIP2 was constitutively tyrosine-phosphorylated in Rous sarcoma virus-transformed rat fibroblasts and several other tyrosine-phosphorylated proteins co-precipitated with 51C/SHIP2 from these cells.
Since elements of PI 3-kinase function such as the extent and duration of Akt/PKB activation vary with different growth factors (33), the effects of growth factors and insulin on the time course of tyrosine phosphorylation of the 51C/SHIP2 protein were determined (Fig. 4). EGF and PDGF caused transient increases in tyrosine phosphorylation of 51C/SHIP2 protein in SH-SY5Y cells, with maximal effects at 5-10 min. Tyrosine phosphorylation returned to near control levels after 20 min. Transient tyrosine phosphorylation of 51C/SHIP2 protein was also seen in EGF-treated PC-12 cells (not shown). In contrast, IGF-1, insulin, and NGF induced long term tyrosine phosphorylation of 51C/SHIP2, with 40-80% maximal levels maintained for up to 2 h. Again, 52-and 66-kDa tyrosine phosphorylated proteins were present in 51C/SHIP2 immunoprecipitates from EGF-and PDGF-treated cells and the 52-kDa tyrosine phosphorylated protein was seen in 51C/SHIP2 immunoprecipitates from NGF-treated PC-12 cells.
Tyrosine phosphorylation could regulate the activity of 51C/ SHIP2 leading to changes in intracellular levels of PtdIns(3,4,5)P 3 and PtdIns(3,4)P 2 which could influence the activity of the Akt/PKB protein kinase (12)(13)(14)(15)(16). For this reason, the effects of the various growth factors on the extent and duration of Akt/PKB activation were determined. (Fig. 5). EGF stimulated a transient activation of Akt/PKB which returned to near base-line levels after 20 min of treatment. IGF-1 and insulin treatment resulted in a somewhat higher maximal level of activation which was sustained for at least 2 h. NGF and PDGF treatment produced equivalent maximal activation of AKT/PKB relative to IGF-1 and insulin but of shorter duration.
The 52-and 66-kDa tyrosine-phosphorylated proteins coprecipitating with 51C/SHIP2 were similar in size to two forms of Shc, and the 52-kDa form of Shc has been shown to associate with SHIP, leading us to directly investigate whether the coprecipitating proteins were forms of Shc. In these experiments, lysates from cells treated with growth factors and insulin were immune-precipitated with 51C/SHIP2 antiserum and immunoblotted with Shc antibodies (Fig. 6). The 52-and 66-kDa forms of Shc co-precipitated with 51C/SHIP2 protein from SH-SY5Y cells treated with EGF and PDGF, and the 52-kDa form coprecipitated with 51C/SHIP2 from NGF-treated PC-12 cells. No Shc was detected in 51C/SHIP2 immunoprecipitates from insulin-treated 3T3-L1 cells or IGF-1-treated SH-SY5Y cells. Even though the 66-kDa form of Shc was equally abundant in SH-SY5Y cells and PC-12 cells, it associated with 51C/SHIP2 from EGF-and PDGF-treated SH-SY5Y cells, but not with 51C/SHIP2 from NGF-treated PC-12 cells. DISCUSSION The 51C/SHIP2 protein is a prominent substrate for tyrosine phosphorylation in response to treatment of cells with a number of growth factors and insulin. These results suggest that 51C/SHIP2 may be a general modulator the PI 3-kinase response to growth factors and insulin by regulating levels of PtdIns(3,4)P 2 and PtdIns(3,4,5)P 3 . As with SHIP, the question arises whether 51C/SHIP2 activity would function to positively or negatively regulate growth factor signaling. For example, the PH domain of Akt/PKB binds both PtdIns(3,4)P 2 and PtdIns(3,4,5)P 3 , however, only binding of PtdIns(3,4)P 2 has been reported to result in activation of the enzyme (11,15,16). On the other hand, PDK1, which phosphorylates and partially activates Atk/PKB, is more strongly activated by PtdIns(3,4,5)P 3 than by PtdIns(3,4)P 2 (13,14). If 51C/SHIP2 is activated by growth factors and plays a positive role in signaling, it could increase levels of PtdIns(3,4)P 2 without totally depleting PtdIns-(3,4,5)P 3 and bring about increased activation of Akt/PKB or other downstream effectors of PI 3-kinase. Alternatively, 51C/ SHIP2 activity could function to deplete PtdIns(3,4,5)P 3 levels and attenuate PI 3-kinase signaling and the effect of growth factors would be to inhibit 51C/SHIP2 activity and potentiate PI 3-kinase signaling. In support of the latter hypothesis, overexpression of SHIP in the FD hematopoietic cell line inhibited mitogenic signaling by macrophage colony-stimulating factor (22) and microinjection SHIP into Xenopus oocytes inhibited insulin-induced germinal vesicle breakdown (34).
The role to tyrosine phosphorylation of SHIP and 51C/SHIP2 in the function of these proteins is not well understood. Osbourne et al. (24) have shown that tyrosine phosphorylation of SHIP by Lck inhibits SHIP activity, suggesting that SHIP is normally active and that one effect of agonist treatment would be to inhibit its activity in order to increase PI 3-kinase-generated levels of PtdIns(3,4,5)P 3 . Constitutive phosphorylation of SHIP in bcr-Abl-transformed cells and of 51C/SHIP2 in Rous sarcoma virus-transformed cells could support this model, since transformation by these oncogenes activates PI 3-kinase and results in elevated basal levels of PtdIns(3,4,5)P 3 (35,36). Our data showing prolonged tyrosine phosphorylation of 51C/ SHIP2 induced by IGF-1, NGF, and insulin could also support this model. IGF-1 and NGF are important survival factors which protect cells from apoptosis, and activation of Akt/PKB through the PI 3-kinase appears to represent a significant pathway in the anti-apoptotic activity of these growth factors (7,32,37). Insulin signaling has also been shown to be dependent on PI 3-kinase signaling through Akt/PKB, which has been shown to be involved in both GLUT4 translocation (9) and activation of glycogen synthesis (38). Inhibition of PtdIns(3,4,5)P 3 hydrolysis to PtdIns(3,4)P 2 by relatively long term tyrosine phosphorylation of 51C/SHIP2 could yield ratios of these mediators, which would potentiate PI 3-kinase signal- ing and contribute to the prolonged activation of Akt/PKB seen with IGF-1 and insulin. However, thrombin stimulation of platelets leads to translocation of SHIP and PI 3-kinase to the actin cytoskeleton, coincident with prolonged tyrosine phosphorylation of SHIP and with the accumulation of PtdIns(3,4)P 2 (39,40). In this case, agonist-stimulated tyrosine phosphorylation of SHIP appears to increase its activity (i.e. increase generation of PtdIns(3,4)P 2 ) and SHIP appears to play a positive role in thrombin signaling. The transient tyrosine phos-phorylation of 51C/SHIP2 by EGF and PDGF may correlate with a shorter duration of activation of PI 3-kinase by these factors in the cell types we have examined.
Another point of divergence for signaling by the different growth factor receptors could be the ability to induce association of 51C/SHIP2 with Shc. EGF, PDGF, and NGF clearly stimulate the association of 51C/SHIP2 with Shc. On the other hand, IGF-1 and insulin stimulate comparable tyrosine phosphorylation of 51C/SHIP2 relative to EGF and NGF, but fail to promote formation of 51C/SHIP2-Shc complexes. The binding of Shc to 51C/SHIP2 and SHIP may serve to attenuate signaling through the Ras pathway by competing with Grb2/Sos complexes for binding to Shc (41). The differences in Shc binding to 51C/SHIP2 found for EGF, PDGF, and NGF versus insulin and IGF-1 may reflect the relative importance of the Ras pathway in cellular responses generated by these growth factors. These differences in interaction with Shc may also influence the capacity of the different factors to stimulate Akt/ PKB activity.
In addition, the 66-kDa form of Shc associated with 51C/ SHIP2 from EGF-and PDGF-treated SH-SY5Y cells, but not from NGF-treated PC-12 cells, even though similar levels of 66-kDa Shc were present and tyrosine phosphorylated by respective agonists in both cell types. This form of Shc has recently been reported to limit activation of mitogen-activated protein kinase by EGF (42), perhaps by competing Grb2/Sos complexes away from the 52-kDa form of Shc. The capacity to stimulate interaction of 66-kDa Shc with 51C/SHIP2 could contribute to differences in the kinetics of activation of mito- gen-activated protein kinase observed for EGF versus NGF (4).
In summary, the 51C/SHIP2 protein is a novel substrate tyrosine phosphorylated in response to treatment of cells with growth factors and insulin and which may serve varied functions for different growth factors and cell types. Future efforts will be directed toward more detailed delineation of the processes regulated by 51C/SHIP2.