SH2-containing inositol phosphatase 2 predominantly regulates Akt2, and not Akt1, phosphorylation at the plasma membrane in response to insulin in 3T3-L1 adipocytes.

SH2-containing inositol phosphatase 2 (SHIP2) is a physiologically important negative regulator of insulin signaling by hydrolyzing the phosphatidylinositol (PI) 3-kinase product PI 3,4,5-trisphosphate in the target tissues of insulin. Targeted disruption of the SHIP2 gene in mice resulted in increased insulin sensitivity without affecting biological systems other than insulin signaling. Therefore, we investigated the molecular mechanisms by which SHIP2 specifically regulates insulin-induced metabolic signaling in 3T3-L1 adipocytes. Insulin-induced phosphorylation of Akt, one of the molecules downstream of PI3-kinase, was inhibited by expression of wild-type SHIP2, whereas it was increased by expression of 5'-phosphatase-defective (DeltaIP) SHIP2 in whole cell lysates. The regulatory effect of SHIP2 was mainly seen in the plasma membrane (PM) and low density microsomes but not in the cytosol. In this regard, following insulin stimulation, a proportion of Akt2, and not Akt1, appeared to redistribute from the cytosol to the PM. Thus, insulin-induced phosphorylation of Akt2 at the PM was predominantly regulated by SHIP2, whereas the phosphorylation of Akt1 was only minimally affected. Interestingly, insulin also elicited a subcellular redistribution of both wild-type and DeltaIP-SHIP2 from the cytosol to the PM. The degree of this redistribution was inhibited in part by pretreatment with PI3-kinase inhibitor. Although the expression of a constitutively active form of PI3-kinase myr-p110 also elicited a subcellular redistribution of SHIP2 to the PM, expression of SHIP2 appeared to affect the myr-p110-induced phosphorylation, and not the translocation, of Akt2. Furthermore, insulin-induced phosphorylation of Akt was effectively regulated by SHIP2 in embryonic fibroblasts derived from knockout mice lacking either insulin receptor substrate-1 or insulin receptor substrate-2. These results indicate that insulin specifically stimulates the redistribution of SHIP2 from the cytosol to the PM independent of 5'-phosphatase activity, thereby regulating the insulin-induced translocation and phosphorylation of Akt2 at the PM.

SH2-containing inositol phosphatase 2 (SHIP2) is a physiologically important negative regulator of insulin signaling by hydrolyzing the phosphatidylinositol (PI) 3-kinase product PI 3,4,5-trisphosphate in the target tissues of insulin. Targeted disruption of the SHIP2 gene in mice resulted in increased insulin sensitivity without affecting biological systems other than insulin signaling. Therefore, we investigated the molecular mechanisms by which SHIP2 specifically regulates insulin-induced metabolic signaling in 3T3-L1 adipocytes. Insulin-induced phosphorylation of Akt, one of the molecules downstream of PI3-kinase, was inhibited by expression of wild-type SHIP2, whereas it was increased by expression of 5-phosphatase-defective (⌬IP) SHIP2 in whole cell lysates. The regulatory effect of SHIP2 was mainly seen in the plasma membrane (PM) and low density microsomes but not in the cytosol. In this regard, following insulin stimulation, a proportion of Akt2, and not Akt1, appeared to redistribute from the cytosol to the PM. Thus, insulin-induced phosphorylation of Akt2 at the PM was predominantly regulated by SHIP2, whereas the phosphorylation of Akt1 was only minimally affected. Interestingly, insulin also elicited a subcellular redistribution of both wild-type and ⌬IP-SHIP2 from the cytosol to the PM. The degree of this redistribution was inhibited in part by pretreatment with PI3kinase inhibitor. Although the expression of a constitutively active form of PI3-kinase myr-p110 also elicited a subcellular redistribution of SHIP2 to the PM, expression of SHIP2 appeared to affect the myr-p110-induced phosphorylation, and not the translocation, of Akt2. Furthermore, insulin-induced phosphorylation of Akt was effectively regulated by SHIP2 in embryonic fibroblasts derived from knockout mice lacking either insulin receptor substrate-1 or insulin receptor substrate-2. These results indicate that insulin specifically stimulates the redistribution of SHIP2 from the cytosol to the PM independent of 5-phosphatase activity, thereby regulating the insulin-induced translocation and phosphorylation of Akt2 at the PM.
Phosphatidylinositol (PI) 1 3-kinase plays a central role in the metabolic actions of insulin. PI(3,4,5)P 3 produced by activated PI3-kinase is thought to function as a key lipid second messenger for signaling to further downstream molecules including Akt and atypical PKC (1)(2)(3)(4). We and others (5,6) have recently cloned SH2-containing inositol phosphatase 2 (SHIP2), which has 5Ј-phosphatase activity toward the PI3-kinase product, PI(3,4,5)P 3 , in the target tissues of insulin. Overexpression of SHIP2 inhibited insulin-induced metabolic signaling leading to glucose uptake and glycogen synthesis via 5Ј-phosphatase activity hydrolyzing the PI3-kinase product PI(3,4,5)P 3 to phosphatidylinositol 3,4-diphosphate in 3T3-L1 adipocytes and L6 myotubes (7,8). Importantly, targeted disruption of the SHIP2 gene in mice increased insulin sensitivity without affecting other biological systems (9). These reports indicate that SHIP2 is a physiologically important negative regulator relatively specific to the insulin signaling. This prompted us to clarify the molecular mechanism by which SHIP2 specifically regulates the metabolic actions of insulin.
Among the effector molecules downstream of PI3-kinase, Akt is strongly implicated in the metabolic action of insulin including glucose uptake and glycogen synthesis (10 -12). Upon insulin treatment, Akt is known to translocate from the cytosol to the plasma membrane where it is primarily activated by phosphorylation at Thr 308/309 and Ser 473/474 (13)(14)(15)(16). Because Akt1 and Akt2 are the predominant isoforms expressed in 3T3-L1 adipocytes (17), the role of SHIP2 in the insulin-induced phosphorylation of Akt1 and Akt2 at various subcellular locations was examined by expressing the wild-type SHIP2 (WT-SHIP2) and a 5Ј-phosphatase-defective SHIP2 (⌬IP-SHIP2) into 3T3-L1 adipocytes using adenovirus-mediated gene transfer (7). Although PI3-kinase is activated by a number of growth factors, only insulin elicits the physiologically important metabolic action via the PI3-kinase pathway (18 -20). In this regard, we investigated the impact of SHIP2 expression on the translocation and phosphorylation of Akt induced by the constitutively active form of PI3-kinase, myr-p110 (7,21). Furthermore, to clarify whether SHIP2 specifically or non-specifically regulates the metabolic signaling of insulin mediated via IRS-1 and IRS-2, the effect of SHIP2 expression on the insulin-induced phosphorylation of Akt was studied in embryonic fibroblasts lacking either IRS-1 or IRS-2 (22). Here, we show that * This work was supported in part by a grant-in-aid for Scientific Research from the Japan Society for the Promotion of Science (to T. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  SHIP2 negatively regulates the insulin-induced translocation and phosphorylation of Akt2 at the PM mediated via both the IRS-1 and IRS-2 pathway.

EXPERIMENTAL PROCEDURES
Materials-Human crystal insulin was provided by Novo Nordisk Pharmaceutical Co., (Copenhagen, Denmark). The two polyclonal anti-SHIP2 antibodies were described previously (5). The anti-SHIP2 antibodies raised against the C terminus and N terminus were used for the immunoprecipitation and immunoblotting, respectively. A monoclonal anti-phosphotyrosine antibody (PY20) was purchased from Transduction Laboratories (Lexington, KY). A polyclonal anti-Thr 308 phosphospecific Akt antibody and a polyclonal anti-Ser 473 phospho-specific Akt antibody were obtained from New England Biolabs, Inc. (Beverly, MA). A polyclonal anti-Akt antibody and a polyclonal anti-Akt1-specific antibody were from Santa Cruz Biotechnology (Santa Cruz, CA). A polyclonal anti-Akt2-specific antibody was from Calbiochem. Enhanced chemiluminescence reagents were from Amersham Biosciences. Dulbecco's modified Eagle's medium (DMEM), minimum essential medium vitamin mixtures, and minimum essential medium amino acid solutions were from Invitrogen. All other reagents were of analytical grade and purchased from Sigma or Wako Pure Chemical Industries, Ltd. (Osaka, Japan).
Construction of Adenoviral Vectors-cDNAs encoding rat WT-SHIP2 and ⌬IP-SHIP2 were subcloned into the vector pAxCAwt and transferred to recombinant adenovirus by homologous recombination utilizing an adenovirus expression vector kit (Takara Biomedicals, Tokyo, Japan) as described previously (7). The adenoviral vector encoding the constitutively active form of bovine p110 with a Src myristration signal sequence at the N terminus (myr-p110) was reported previously (21).
Cell Culture and Infection with Adenovirus-3T3-L1 fibroblasts were grown and passaged in DMEM supplemented with 10% newborn calf serum. Cells at 2 to 3 days post-confluence were used for differentiation. The differentiation medium contained 10% fetal calf serum (FCS), 250 nM dexamethasone, 0.5 mM isobutyl methylxanthine, and 500 nM insulin. After 3 days, the differentiation medium was replaced with postdifferentiation medium containing 10% FCS and 500 nM insulin. After 3 more days, the post-differentiation medium was replaced with DMEM supplemented with 10% FCS (7). Preparation of IRS-1(Ϫ/Ϫ) and IRS-2(Ϫ/Ϫ) embryonic fibroblasts from IRS-1-and IRS-2-deficent mice was described previously (22). Embryonic fibroblasts were cultured with ␣-minimum essential medium supplemented with 10% FCS. WT-SHIP2, ⌬IP-SHIP2, and myr-p110 were transiently expressed in differentiated 3T3-L1 adipocytes and embryonic fibroblasts by means of adenovirus-mediated gene transfer. A multiplicity of infection (m.o.i.) of 10 -40 pfu/cell was used to infect 3T3-L1 adipocytes and embryonic fibroblasts in DMEM containing 2% FCS, with the virus being left on the cells for 16 h prior to removal. Subsequent experiments were conducted 24 to 48 h after initial addition of the virus. The efficiency of the adenovirus-mediated gene transfer of WT-SHIP2, ⌬IP-SHIP2, and myr-p110 was ϳ95%.
Subcellular Fractionation-3T3-L1 adipocytes were washed twice with phosphate-buffered saline and once with HES buffer (255 mM sucrose, 20 mM HEPES, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM Na 3 VO 4 , 2 g/ml aprotinin, and 50 ng/ml okadaic acid, pH 7.4) and immediately homogenized by 20 strokes with a motordriven homogenizer in HES buffer at 4°C. The homogenates (two 10-cm-diameter dishes per condition) were subjected to subcellular fractionation as described previously to isolate PM, high density microsomes, low density microsomes (LDM), and cytosol (23,24). In brief, the homogenates were centrifuged at 19,000 ϫ g for 20 min. The resulting supernatant was centrifuged at 41,000 ϫ g for 20 min, yielding a pellet of high density microsomes. The supernatant from this spin was centrifuged at 250,000 ϫ g for 90 min, yielding a pellet of LDM. Remaining supernatant was concentrated by Centricon-30 (Amicon Inc., Beverly, Mass.) and used as cytosol. The pellet obtained from the initial spin was resuspended in HES buffer, layered onto a 1.12 M sucrose cushion, and centrifuged at 100,000 ϫ g in a swing rotor for 60 min. A white fluffy band at the interface was collected and resuspended in HES buffer and centrifuged at 40,000 ϫ g for 20 min, yielding a pellet of PM. All fractions were adjusted to a final protein concentration of 1 to 3 mg/ml, which was measured by the Bradford method, and stored at Ϫ80°C until use.
Immunoprecipitation and Western Blotting-3T3-L1 adipocytes and embryonic fibroblasts grown in 6-well multiplates were serum-starved for 16 h in DMEM. The cells were treated with 17 nM insulin at 37°C for various periods. They were then lysed in a buffer containing 20 mM Tris, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 2.5 mM sodium deoxycholate, 1 mM ␤-glycerophosphate, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 1 mM Na 3 VO 4 , 50 mM sodium fluoride, 10 g/ml aprotinin, and 10 M leupeptin, pH 7.4, for 15 min at 4°C. Lysates obtained from the same number of cells were centrifuged to remove insoluble materials. The supernatants (100 g of protein) were immunoprecipitated with antibodies for 2 h at 4°C. The precipitates or whole cell lysates were then separated by 7.5% SDS-PAGE and transferred onto polyvinylidene difluoride membranes using a Bio-Rad Transblot apparatus. The membranes were blocked in a buffer containing 50 mM Tris, 150 mM NaCl, 0.1% Tween 20, and 2.5% bovine serum albumin or 5% non-fat milk, pH 7.5, for 2 h at 20°C. The membranes were then probed with antibodies for 2 h at 20°C or for 16 h at 4°C. After the membranes were washed in a buffer containing 50 mM Tris, 150 mM NaCl, and 0.1% Tween 20, pH 7.5, blots were incubated with a horseradish peroxidase-linked secondary antibody and subjected to enhanced chemiluminescence detection using ECL reagent according to the manufacturer's instructions (Amersham Biosciences) (5,7).
Statistical Analysis-The data are represented as means Ϯ S.E. p values were determined using a Student's t test, and p Ͻ 0.05 was considered statistically significant.

Structures of SHIP2 Constructs and the Expression in 3T3-L1
Adipocytes-SHIP2 is a 140-kDa protein composed of an SH2 domain at the N terminus, a central 5Ј-phosphatase catalytic domain, and a proline-rich region including the phosphotyrosine binding domain binding consensus at the C terminus. Three amino acids, located within the catalytic domain of SHIP2, that are highly conserved among known 5Ј-phosphatases were mutated to generate ⌬IP-SHIP2 (7) (Fig. 1A). WT-SHIP2 and ⌬IP-SHIP2 were transiently expressed in 3T3-L1 adipocytes by adenovirus-mediated gene transfer. Endogenous SHIP2 was seen in control 3T3-L1 adipocytes transfected with LacZ alone. On transfection with either WT-SHIP2 or ⌬IP-SHIP2 at an m.o.i. of 40 pfu/cell, we observed similar levels of expression of WT-SHIP2 and ⌬IP-SHIP2, which were 5-fold greater than the levels of endogenous SHIP2. Insulin treatment did not affect the expression of WT-SHIP2 and ⌬IP-SHIP2 (Fig. 1B).

Effect of SHIP2 Expression on Insulin-induced Phosphorylation of Akt in Whole Cell
Lysates-Akt is a downstream target of PI3-kinase important for mediation of the metabolic actions of insulin (1)(2)(3)(4). Because Akt is primarily activated as a result of its phosphorylation at the Thr 308 (Akt2 at Thr 309 ) and Ser 473 (Akt2 at Ser 474 ) residues (13, 16 -20, 25), we examined the effect of SHIP2 expression on the insulin-induced phosphorylation of Akt in 3T3-L1 adipocytes. Treatment with insulin induced phosphorylation of Akt at Thr 308 and Ser 473 in a timedependent manner in LacZ-transfected control 3T3-L1 adipocytes. Transfection of WT-SHIP2 decreased insulin-induced phosphorylation of Akt at both Thr 308 and Ser 473 . In contrast, insulin-induced phosphorylation of Akt at Thr 308 and Ser 473 was increased by transfection with ⌬IP-SHIP2 (Fig. 2, A and  B). These results are summarized in Fig. 2, E and F. Following 15 min of insulin treatment, the phosphorylation of Akt at Thr 308 was significantly decreased 30.1 Ϯ 4.9% by the expression of WT-SHIP2 and increased 34.6 Ϯ 5.7% by the expression of ⌬IP-SHIP2. Similarly, the phosphorylation of Akt at Ser 473 was decreased 27.9 Ϯ 3.4% by the expression of WT-SHIP2 and increased 31.8 Ϯ 4.4% by the expression of ⌬IP-SHIP2 following 5 min of insulin stimulation compared with that in control 3T3-L1 adipocytes transfected with LacZ. To assure equal amounts of protein were loaded among the samples, the cell lysates were immunoblotted with anti-Akt antibody (Fig. 2C). Similar expression levels of WT-SHIP2 and ⌬IP-SHIP2 were detected on the immunoblotting of the cell lysates with anti-SHIP2 antibody (Fig. 2D).
Effect of SHIP2 Expression on Insulin-induced Phosphorylation of Akt at Subcellular Locations-Because it is known that Akt is localized in the cytosol, PM, and LDM fractions (10), we next examined the effect of SHIP2 expression on the insulin-induced phosphorylation of Akt at subcellular locations (Fig. 3). Insulin induced the phosphorylation of Akt at Thr 308 and Ser 473 in the cytosol, PM, and LDM fractions. Although a large amount of Akt resides in the cytosol, the insulin-induced phosphorylation of Akt at Thr 308 and Ser 473 in the cytosol was not significantly affected by the expression of either WT-SHIP2 or ⌬IP-SHIP2. In contrast, the phosphorylation of Akt in the PM and LDM was affected by the expression of SHIP2. Notably, the insulin-induced phosphorylation of Akt at both Thr 308 and Ser 473 in the PM fraction was markedly decreased by the expression of WT-SHIP2, whereas it was increased by the expression of ⌬IP-SHIP2. Densitometric analysis revealed that insulin-induced phosphorylation of Akt at Thr 308 and Ser 473 was decreased by 47.3 Ϯ 1.2% and 45.7 Ϯ 3.1%, respectively, in WT-SHIP2expressing cells, whereas it was enhanced by 44.3 Ϯ 5.6% and 45.3 Ϯ 6.6% in ⌬IP-SHIP2-expressing cells.
Effect of SHIP2 Expression on Insulin-induced Phosphorylation of Akt1 and Akt2 Isoforms-Because Akt1 and Akt2 are the main isoforms expressed in 3T3-L1 adipocytes (17), we next examined the effect of SHIP2 expression on the insulin-induced phosphorylation of Akt1 and Akt2. The cell lysates were immunoprecipitated with anti-Akt1 antibody, and the precipitates were immunoblotted with anti-phosphospecific Akt antibody. As shown in Fig. 4C, Akt1 is efficiently immunoprecipitated by this procedure, and the Akt2 isoform is not present in the precipitates. Insulin induced phosphorylation of the Akt1 isoform at Thr 308 and Ser 473 in anti-Akt1 immunoprecipitates, and this phosphorylation was not affected by the expression of either WT-SHIP2 or ⌬IP-SHIP2 (Fig. 4, A and B). Because an anti-Akt2 antibody was not available for the immunoprecipita-tion, we performed an immunodepletion experiment. After the cell lysates were effectively immunoprecipitated with anti-Akt1 antibody, the supernatants were used for the experiment with Akt2. As can be seen in Fig. 4F, only Akt2, not Akt1, is present in the sample obtained by this procedure. Importantly, insulin-induced phosphorylation of Akt2 at Thr 309 and Ser 474 was markedly decreased by the expression of WT-SHIP2, whereas it was increased by the expression of ⌬IP-SHIP2 (Fig.  4, D and E). These results indicate that SHIP2 regulates the insulin-induced phosphorylation of Akt2, and not Akt1, in 3T3-L1 adipocytes.
Effect of SHIP2 Expression on the Insulin-induced Subcellular Distribution of Akt Isoforms-It is known that growth factor induces a subcellular relocalization of Akt to the plasma membrane to be phosphorylated (10,20,25). Although SHIP2 negatively regulates insulin-induced Akt2 phosphorylation, it is unclear whether SHIP2 affects the phosphorylation of Akt directly or via its translocation to the PM. To address this issue, we next examined the effect of SHIP2 expression on the insulin-induced subcellular redistribution of Akt1 and Akt2. The Akt1 isoform mainly resides in the cytosol fraction, and insulin treatment did not appear to induce apparent subcellular redistribution. In addition, overexpression of neither WT-SHIP2 nor ⌬IP-SHIP2 appeared to affect the subcellular localization of Akt1 (Fig. 5A). Thus, the amount of Akt1 in the cytosol did not significantly alter in response to insulin. The Akt2 isoform is also mainly localized in the cytosol fraction in the basal state. Compared with the results obtained with Akt1, insulin efficiently elicited a subcellular redistribution of the Akt2 isoform from the cytosol and LDM to the PM. Importantly, the redistribution was markedly decreased by the expression of WT-SHIP2, whereas it was enhanced by that of ⌬IP-SHIP2 (Fig.  5B). These results indicate that SHIP2 appears to regulate the subcellular redistribution of Akt2, and not Akt1, in 3T3-L1 adipocytes.
Insulin-induced Subcellular Redistribution of SHIP2-Our previous study (26) indicated that the membrane localization of SHIP2 is important for its functioning via the 5Ј-phosphatase activity. Expression of SHIP2 with the myristoylation signal efficiently inhibited insulin-induced phosphorylation of Akt in Rat1 fibroblasts (26). Given this, we reasoned that SHIP2 might elicit this function by changing the subcellular localization to efficiently regulate the phosphorylation of Akt in the PM fraction. We examined whether insulin induces the subcellular redistribution of SHIP2 (Fig. 6A). WT-SHIP2 resides largely in the cytosol and partly in the LDM and PM fractions in the basal state. Insulin treatment significantly induced a redistribution of some of the expressed WT-SHIP2 and ⌬IP-SHIP2 to the PM fraction. We further assessed the role of PI3-kinase in the insulin-induced redistribution of SHIP2. Pretreatment of the cells with the PI3-kinase inhibitor LY294002 partly, but significantly, inhibited the insulin-induced redistribution of both WT-SHIP2 and ⌬IP-SHIP2 to the PM. Densitometric analysis demonstrated that the redistribution of WT-SHIP2 and ⌬IP-SHIP2 to the PM was inhibited 41.3 Ϯ 7.2 and 51.7 Ϯ 6.6%, respectively, by treatment with LY294002. A similar degree of inhibition was obtained on treatment with wortmannin (data not shown). To assure that the LY294002 used in the experiment effectively inhibited the PI-kinase activity, the effect of pretreatment on insulin-induced phosphorylation of Akt was examined. Pretreatment with LY294002 effectively inhibited the insulin-induced phosphorylation of Akt (Fig. 6B). These results indicate that insulin induces a redistribution of SHIP2 from the cytosol to the PM fraction independent of the 5Јphosphatase activity of SHIP2 and that this redistribution is partly dependent on the PI3-kinase activity.

Effect of SHIP2 Expression on myr-p110-induced Subcellular
Distribution and Phosphorylation of Akt-A number of growth factors stimulate PI3-kinase activity in addition to insulin (2, 3). We next examined the effect of SHIP2 expression on the PI3-kinase-induced subcellular distribution and phosphorylation of Akt. To this end, we employed myr-p110, which is a constitutively active from of the catalytic subunit of PI3-kinase (21). Similar to insulin, myr-p110 induced a recruitment of SHIP2 to the PM fraction (Fig. 7A). In addition, myr-p110 also induced phosphorylation of Akt at Thr 308 in the cytosol, PM, and LDM fractions. Similarly, the myr-p110-induced phosphorylation of Akt in the cytosol fraction was apparently not affected by the expression of WT-SHIP2. SHIP2 regulated the insulin-induced phosphorylation of Akt at Thr 308 in the PM fraction, but the degree to which the myr-p110-induced phosphorylation of Akt was regulated by SHIP2 relatively mild (Fig.  7B). Thus, expression of WT-SHIP2 inhibited myr-p110-induced phosphorylation of Akt by 27.3 Ϯ 4.2%. Similar results were obtained concerning the effect of SHIP2 expression on myr-p110-induced phosphorylation of Akt at Ser 473 (data not shown). Interestingly, the myr-p110-induced redistribution of Akt2 to the PM fraction was not affected by SHIP2 expression (Fig. 7C).

Effect of SHIP2 Expression on Insulin-induced Phosphorylation of Akt in Embryonic Fibroblasts Derived from IRS-1 and IRS-2 Knockout
Mice-The metabolic action of insulin is mediated via IRS-1 and/or IRS-2 (1,4,(27)(28)(29)(30)(31). Although IRS-1 and IRS-2 have structural similarities, each coordinates, at least in part, different insulin actions (22,27,(32)(33)(34). We further investigated whether IRS-1-and IRS-2-mediated insulin signaling can be regulated by SHIP2. To this end, we employed embryonic fibroblasts derived from IRS-1 and IRS-2 knockout mice (22). Insulin induced the phosphorylation of Akt at Thr 308 and Ser 473 in control embryonic fibroblasts expressing both IRS-1 and IRS-2. The phosphorylation of Akt was also seen in IRS-1(Ϫ/Ϫ) and IRS-2(Ϫ/Ϫ) embryonic fibroblasts. The insulininduced phosphorylation of Akt at Thr 308 and Ser 473 was decreased by the expression of WT-SHIP2, whereas it was enhanced by the expression of ⌬IP-SHIP2 in control, IRS-1(Ϫ/Ϫ), and IRS-2(Ϫ/Ϫ) embryonic fibroblasts (Fig. 8, A and B). These results indicate that SHIP2 is involved in the negative regulation of both IRS-1-and IRS-2-mediated insulin signaling. To assure equal amounts of protein were loaded among the samples, the cell lysates were immunoblotted with anti-Akt antibody. Similar expression levels of WT-SHIP2 and ⌬IP-SHIP2 were confirmed on immunoblotting of the cell lysates with anti-SHIP2 antibody (Fig. 8C).

DISCUSSION
The SHIP family is composed of SHIP1 and SHIP2, and overall, the SHIP2 protein exhibits about 40% amino acid identity to SHIP1 (5,6). Despite their similarities, differences between SHIP1 and SHIP2 have been identified. First, SHIP2 is expressed relatively ubiquitously including in the targets of insulin such as skeletal muscles and fat cells, whereas the expression of SHIP1 is restricted to hematopoietic and spermatogenetic cells (6,35,36). Second, these two SHIP isozymes appear to have different substrate specificities. SHIP2 may have greater activity than SHIP1 for the hydrolysis of the PI3-kinase product PI(3,4,5)P 3 (37). These reports suggest that SHIP1 and SHIP2 regulate different inositol-mediated pathways and/or interact differently with effector molecules. Along this line, our previous studies (7,8) with cultured cells showed that overexpression of SHIP2 negatively regulated insulininduced metabolic signaling via 5Ј-phosphatase activity in 3T3-L1 adipocytes and L6 myocytes. Experiments with knockout mice revealed that disruption of SHIP2 expression caused hyperinsulin sensitivity without affecting biological systems other than insulin signaling (9). In addition, the expression of SHIP2 was enhanced leading to an attenuation of insulin signaling distal to PI3-kinase in an animal model of type 2 diabetes (38). Furthermore, mutations in the SHIP2 gene may contribute, at least in part, to the genetic susceptibility to type 2 diabetes in humans (39). Taken together, SHIP2 appears to be a physiologically important negative regulator that is relatively specific to insulin signaling and has a fundamental impact on the pathological state of type 2 diabetes.
To understand the novel control mechanisms in the regulation of insulin signaling by the lipid phosphatase SHIP2, it is important to elucidate how SHIP2 functions following stimulation with insulin. The total 5Ј-phosphatase activity of SHIP2 is known not to be altered following stimulation with growth factors including insulin (40). Instead, the relocalization of SHIP2 to the vicinity of the plasma membrane would appear to be critical for the functioning based on experiments with SHIP1 (41). Targeting of SHIP2 to the plasma membrane on addition of the myristoylation signal efficiently inhibited the insulin-induced phosphorylation of Akt in Rat1 fibroblasts expressing insulin receptors, although the level of SHIP2 expression in the plasma membrane was low (26). Given this, we reasoned that insulin treatment changes the subcellular relocalization of SHIP2. In fact, insulin induced a redistribution of WT-SHIP2 from the cytosol to the plasma membrane fraction. The insulin-induced localization of SHIP2 to the membrane appears to be critical for the functioning possibly by providing appropriate access to the substrate PI(3,4,5)P 3 . In addition, treatment with the PI3-kinase inhibitor LY29004 partly inhibited the insulin-induced redistribution of SHIP2 to the plasma membrane fraction. These results indicate that insulin causes SHIP2 to be redistributed to the plasma membrane where it functions to hydrolyze the PI3-kinase product, and the activation of PI3-kinase itself is required, at least in part, for the appropriate targeting of SHIP2 to the plasma membrane. It is possible that other signaling molecules, in addition to PI3kinase, are involved in the subcellular redistribution of SHIP2. Along this line, the association of SHIP2 with Shc appears to be required for the efficient negative regulation of insulin-induced phosphorylation of Akt, at least in part, in Rat1 fibroblasts (26). The association of SHIP2 with Cbl is also reported in Chinese hamster ovary cells (42). Because Shc and Cbl are known to be tyrosine-phosphorylated and undergo a subcellular redistribution to the plasma membrane following insulin stimulation, the tyrosine phosphorylated molecules associated with SHIP2 may have a critical role in the targeting of SHIP2 to the plasma membrane (26,42). The function of SHIP2 in the negative regulation of insulin signaling remains to be more precisely elucidated.
We have previously reported (7) that SHIP2 inhibits the insulin-induced activation of Akt via 5Ј-phosphatase activity. The main isoforms of Akt in 3T3-L1 adipocytes are Akt1 and Akt2 (17). Although the two have structural similarities, their physiological roles appear to differ (10,(43)(44)(45)(46). Mice with a targeted disruption of Akt1 revealed a defect in growth. However, Akt1-deficient mice are normal with respect to glucose tolerance and insulin-stimulated disposal of blood glucose (43,44). In contrast, mice lacking Akt2 demonstrated insulin resistance in target tissues of insulin (45,46). It is also known that Akt2 is abundant in insulin-responsive tissues, and insulin increases the association of Akt2 with Glut4-containing vesicles (10). These results indicate that Akt2 is more important than Akt1 to the metabolic actions of insulin. In this regard, SHIP2 appeared to predominantly regulate the insulininduced phosphorylation of Akt2, and not Akt1, in 3T3-L1 adipocytes as shown in Fig. 4. Concerning the subcellular localization of Akt, both isoforms were mainly concentrated in the cytosolic fraction of 3T3-L1 adipocytes in the basal state as reported previously (10,20). Although insulin treatment stimulated phosphorylation of Akt1 in the cytosolic fraction, it did not appear to induce the subcellular redistribution of Akt1. It is of note that, because lesser levels of Akt1 than Akt2 are expressed in differentiated adipocytes (20), lack of the translocation of Akt1 to the plasma membrane may be below the detection limit of our Western blot analysis. In contrast, insulin clearly induced the subcellular redistribution of Akt2 from the cytosol to the plasma membrane fraction and stimulated phosphorylation of Akt2 in all subcellular fractions. The issue of an insulin-induced redistribution of Akt2 to the plasma membrane is controversial, as there is a report that no such redistribution of Akt2 was found in rat fat cells (10). However, our findings are consistent with reports that insulin treatment induces the translocation of Akt2 to the plasma membrane dependent on the PI3-kinase activity in human ovarian cells and 3T3-L1 adipocytes (20,46,47). The difference may arise from the experimental conditions used or cells employed among previous reports. Interestingly, the insulin-induced phosphorylation of Akt1 in the cytosolic fraction and the localization of Akt1 did not appear to be affected by the expression of either WT-SHIP2 or ⌬IP-SHIP2. In contrast, the insulin-induced phosphorylation of Akt2 was apparently regulated by SHIP2. Furthermore, the degree of inhibition was greater in the plasma membrane fraction than cytosolic fraction. The SHIP2 regulation of Akt phosphorylation and the isoform specificity observed with maximal insulin stimulation (17 nM) appeared to be similarly maintained at a lower concentration of insulin (1.7 nM) stimulation (data not shown). These results indicate that SHIP2 preferentially regulates insulin-induced phosphorylation of Akt2 in the plasma membrane rather than cytosolic fraction. Our results do not rule out the possibility that SHIP2 regulates phosphorylation of Akt2, albeit to a lesser extent, in the cytosolic fraction. It is also possible that the insulin-induced phosphorylation of Akt2 is regulated by SHIP2 exclusively in the plasma membrane fraction, and the weak effect seen in the cytosol fraction was derived from Akt2 returning from the plasma membrane.
Insulin alone can effectively transmit the signaling leading to glucose uptake, although PI3-kinase is activated by a number of growth factors (1)(2)(3)(4). Interestingly, platelet-derived growth factor was more effective than insulin in stimulating Akt phosphorylation in fibroblasts, whereas platelet-derived growth factor did not stimulate Akt2 phosphorylation to any significant extent in adipocytes (20). Moreover, insulin, but not platelet-derived growth factor, induced the translocation of Akt2 to the plasma membrane and high density microsome fractions of 3T3-L1 adipocytes (20). These results prompted us to examine the notion of another role for SHIP2, in addition to the regulation of Akt2 phosphorylation, in the control of insulin-induced metabolic signaling. In this regard, we further focused on the part of SHIP2 in the insulin-induced subcellular redistribution of Akt2. Our results showed that SHIP2 regulates myr-p110-induced phosphorylation of Akt mainly in the plasma membrane fraction. However, the degree of the inhibition was relatively low compared with that for insulin. Furthermore, expression of SHIP2 affected the insulin-induced, and not myr-p110-induced, translocation of Akt2 to the plasma membrane. These results further indicate the molecular mechanism by which SHIP2 relatively specifically regulates insulininduced metabolic signaling in 3T3-L1 adipocytes.
The insulin-induced activation of PI3-kinase is mediated via insulin receptor substrates (1,4). Tyrosine-phosphorylated IRS binds to the p85 regulatory subunit of PI3-kinase leading to an activation to generate PI(3,4,5)P 3 (1-4). IRS-1 and IRS-2 are the two most ubiquitously expressed members of the IRS family of proteins, which regulates the metabolic actions of insulin (1,4,27). Indeed, disruption of the genes for IRS-1 and IRS-2 was shown to result in insulin resistance in mice (27). Although the structures of these two proteins are very similar, targeted disruption of the IRS-1 and IRS-2 gene in mice produced distinct phenotypes (28 -31). Some previous studies demonstrated that IRS-1 has a major role in adipocyte differentiation (22,33). In addition, the IRS-1-PI3-kinase, but not IRS-2-PI3-kinase, pathway was essential for SREBP1c and glucokinase gene expression in rat hepatocytes (32). Furthermore, IRS-2, but not IRS-1, was predominantly involved in Glut4 translocation and glucose uptake in brown adipose tissue (34). Therefore, IRS-1 and IRS-2 regulate the specific biological actions in addition to the redundant metabolic actions of insulin. It is therefore important to clarify whether SHIP2 specifically or non-specifically regulates IRS-1-and IRS-2-mediated insulin signaling. Our results provide direct evidence of a role for SHIP2 in the regulation of IRS-1-and IRS-2-mediated phosphorylation of Akt in IRS-1-or IRS-2-deficient cells. Because the insulininduced phosphorylation of Akt was similarly regulated by the expression of WT-SHIP2 and ⌬IP-SHIP2 among control cells, IRS-1-deficient cells, and IRS-2-deficient cells, SHIP2 appears to have a similar impact on the regulation of insulin signaling by both IRS-1 and IRS-2.
In summary, our results clarified the molecular mechanism by which SHIP2 relatively specifically regulates the metabolic actions of insulin seen in mice with a targeted disruption of SHIP2 (9). First, SHIP2 appears to specifically regulate the metabolic functions of insulin, at least in part, by inhibiting the insulin-induced phosphorylation of Akt2, and not Akt1, in the plasma membrane in 3T3-L1 adipocytes. Second, upon insulin stimulation, SHIP2 is translocated to the plasma membrane, where it inhibits the insulin-specific subcellular redistribution of Akt2. Third, SHIP2 regulates the insulin-induced phosphorylation of Akt mediated via both IRS-1 and IRS-2. Although we clarified the mechanisms of SHIP2 in the regulation of Akt, atypical PKCs including PKC and PKC are another downstream target of PI3-kinase important for the metabolic actions of insulin. Because the insulin-induced activation of aPKC is also known to be regulated by SHIP2, further study will be required to clarify the molecular mechanism of the regulation of aPKC by SHIP2. Clarification of the novel control mechanisms for insulin could provide a new insight into the development of therapeutic drugs. Thus, inhibition of an endogenous amount and/or function of SHIP2 would be an important therapeutic target of insulin resistance in type 2 diabetes.