Enhancement by adenosine of insulin-induced activation of phosphoinositide 3-kinase and protein kinase B in rat adipocytes.

The role of adenosine receptor in regulation of insulin-induced activation of phosphoinositide 3-kinase (PI 3-kinase) and protein kinase B was studied in isolated rat adipocytes. Rat adipocytes are known to spontaneously release adenosine, which in turn binds and stimulates the adenosine A1 receptors on the cells. In the present study, we observed that degradation of this adenosine by adenosine deaminase attenuated markedly the insulin-induced accumulation of phosphatidylinositol 3,4,5-trisphosphate (PtdIns(3,4,5)P3), a product of PI 3-kinase. p-Aminophenylacetyl xanthine amine congener (PAPA-XAC), an inhibitor of the adenosine A1 receptor, also inhibited the insulin-induced PtdIns(3,4,5)P3 accumulation. When extracellular adenosine was inactivated by adenosine deaminase, phenylisopropyladenosine, an adenosine A1 receptor agonist, potentiated the insulin-induced accumulation of PtdIns(3,4,5)P3. Insulin-induced activation of protein kinase B, the activity of which is controlled by the lipid products of PI 3-kinase, was also potentiated by adenosine. Prostaglandin E2, another activator of a pertussis toxin-sensitive GTP-binding protein in these cells, potentiated the insulin actions. Thus, the receptors coupling to the GTP-binding protein were found to positively regulate the production of PtdIns(3,4,5)P3, a putative second messenger for insulin actions, in physiological target cells of insulin.

The role of adenosine receptor in regulation of insulin-induced activation of phosphoinositide 3-kinase (PI 3-kinase) and protein kinase B was studied in isolated rat adipocytes. Rat adipocytes are known to spontaneously release adenosine, which in turn binds and stimulates the adenosine A 1 receptors on the cells. In the present study, we observed that degradation of this adenosine by adenosine deaminase attenuated markedly the insulin-induced accumulation of phosphatidylinositol 3,4,5-trisphosphate (PtdIns(3,4,5)P 3 ), a product of PI 3-kinase. p-Aminophenylacetyl xanthine amine congener (PAPA-XAC), an inhibitor of the adenosine A 1 receptor, also inhibited the insulin-induced PtdIns-(3,4,5)P 3 accumulation. When extracellular adenosine was inactivated by adenosine deaminase, phenylisopropyladenosine, an adenosine A 1 receptor agonist, potentiated the insulin-induced accumulation of PtdIns(3,4,5)-P 3 . Insulin-induced activation of protein kinase B, the activity of which is controlled by the lipid products of PI 3-kinase, was also potentiated by adenosine. Prostaglandin E 2 , another activator of a pertussis toxin-sensitive GTP-binding protein in these cells, potentiated the insulin actions. Thus, the receptors coupling to the GTPbinding protein were found to positively regulate the production of PtdIns(3,4,5)P 3

, a putative second messenger for insulin actions, in physiological target cells of insulin.
Adenosine deaminase has been shown to impair the insulin sensitivity for glucose transport and antilipolysis by inactivating extracellular adenosine, which adipocytes release spontaneously (1)(2)(3). The sensitivity can be restored by treatment of the cells with an adenosine analogue phenylisopropyladenosine (PIA) 1 (4). Because adenosine exerts its effect through the adenosine A 1 receptor coupling to a pertussis toxin-sensitive GTP-binding protein, another activator of the GTP-binding protein, prostaglandin E 2 also affects the insulin sensitivity (5). The effect of adenosine is not accompanied by changes in the cAMP-dependent protein kinase activity and can be observed even when cellular cAMP was maintained at high concentrations (6,7). Thus adenosine action on the insulin sensitivity is considered to be independent of its inhibitory effect on adenylyl cyclase.
Phosphoinositide 3-kinase (PI 3-kinase), which phosphorylates the 3-OH position of phosphoinositides, is a key signaling enzyme in insulin-induced activation of glucose uptake (8). Studies using either inhibitors of PI 3-kinase such as wortmannin and LY294002 or a dominant negative mutants of PI 3-kinase have demonstrated that PI 3-kinase is necessary for the metabolic action of insulin (9 -12). It has also been demonstrated that constitutively active mutants of PI 3-kinase are sufficient to induce the translocation of glucose transporter (GLUT4) to the plasma membrane (13,14). Although the precise mechanism by which PI 3-kinase regulates the glucose transport system is not completely understood, lipid products of PI 3-kinase, phosphatidylinositol 3,4-bisphosphate (PtdIns-(3,4)P 2 ) and phosphatidylinositol 3,4,5-trisphosphate (PtdIns-(3,4,5)P 3 ), are expected to function as second messengers. In this regard, the lipid products are reported to activate protein kinase B (PKB) directly (15), or indirectly through activation of 3-phosphoinositide-dependent protein kinases (16). Expression of a constitutively activated mutant of PKB has been shown to increase glucose uptake activity of adipose cells (17)(18)(19).
Because the mechanism by which adenosine changes insulin sensitivity has not been understood, we examined whether adenosine exerts any effect on insulin-induced activation of PI 3-kinase in rat adipocytes. We observed that adenosine enhanced both the insulin-induced accumulation of PtdIns-(3,4,5)P 3 and the insulin-induced activation of PKB by a mechanism independent of its inhibitory action on adenylyl cyclase.

EXPERIMENTAL PROCEDURES
Materials-The sources of materials used in this work were as follows: 32 P i and [␥-32 P]ATP from NEN Life Science Products Inc.; paminophenylacetyl xanthine amine congener (PAPA-XAC) from Research Biochemicals; 2Ј,5Ј-dideoxyadenosine (DDA) from BIOMOL Research Laboratories; (R)-N 6 -(phenylisopropyl)-adenosine (PIA), adenosine deaminase, and histone H2B from Roche Molecular Biochemicals; anti-PKB␣(Akt-1) antibody from Santa Cruz; protein G-Sepharose from Pharmacia; and collagenase (type I) from Worthington. Wortmannin was a gift of Dr. Y. Matsuda, Kyowa Hakko Kogyo Co. The reagents for determination of cAMP were kindly supplied by Dr. K. Saito, Yamasa Shoyu Co. All other reagents from commercial sources were of analytical grade.
Isolation of Rat Adipocytes-Epididymal fat pads from male Wistar rats weighing 100 -120 g were cut into small pieces and incubated at 37°C for 30 min in a medium consisting of 130 mM NaCl, 4.7 mM KCl, 1 mM CaCl 2 , 1.2 mM MgSO 4 , 1.2 mM KH 2 PO 4 , 10 mM glucose, 25 mM HEPES-NaOH (pH 7.4), 2% (w/v) bovine serum albumin, and 0.2% (w/v) collagenase. The digested tissues were filtered through a double-layered nylon stocking and the filtrate was centrifuged at 600 ϫ g for 3 min. The floated cells were then washed three times by suspension and flotation in buffers used for the assays.
Measurement of the PtdIns (3,4,5)P 3 Production-The isolated adipocytes were washed with an incubation medium consisting of 130 mM NaCl, 4.7 mM KCl, 1 mM CaCl 2 , 1.2 mM MgSO 4 , 25 mM HEPES-NaOH (pH 7.4), and 0.5% (w/v) bovine serum albumin and suspended at a density of 1 ϫ 10 6 cells/ml in the same medium supplemented with 0.2 mCi/ml of carrier-free 32 P i . The cells (0.2 ml) were incubated at 37°C for 10 min with or without adenosine deaminase, PAPA-XAC, or wortmannin and then for 10 min with or without insulin. When the effects of PGE 2 , DDA, or PIA were examined, the chemicals were added 5 min before the insulin addition. The reaction was allowed to proceed in a volume of 0.35 ml before being stopped by mixing with 50 l of 8% (w/v) HClO 4 . The mixture was mixed with 1.5 ml of CHCl 3 /methanol (1: 2), stirred vigorously, and then mixed successively with 0.5 ml of CHCl 3 and 0.5 ml of 8% (w/v) HClO 4 . The organic phase was washed once with a CHCl 3 -saturated solution containing 0.5 M NaCl and 1% HClO 4 before being concentrated to 0.1 ml. An aliquot (5 l) of the extract was spotted onto a Silica Gel 60 plate, which had been impregnated with 1.2% (w/v) potassium oxalate in methanol/water (2: 3) and heated at 110°C for 20 min. The plate was developed in CHCl 3 and then in CHCl 3 /methanol/ acetic acid/acetone/water (70:50:20:20:20) to the same direction. The radioactivity in the PtdIns(3,4,5)P 3 fraction was determined using a Fuji BAS2000 analyzer.
Determination of cAMP-The isolated adipocytes were suspended at a density of 1 ϫ 10 6 cells/ml in the incubation medium described above. The cells (0.2 ml) were incubated at 37°C for 6 min with 5 M PAPA-XAC and then for 6 min with or without PGE 2 or DDA before the addition of 1 M isoproterenol. The mixture (0.35 ml) was incubated further for 5 min, mixed with 0.35 ml of 0.2 N HCl, and then heated at 100°C for 3 min. The concentration of cAMP in the sample was determined by radioimmunoassay (20).
Measurement of PKB Activity-The isolated adipocytes were suspended in 0.3 ml of the incubation medium at a density of 1 ϫ 10 6 cells/ml. After the incubation, cell extracts were prepared by lysing the cells in 600 l of buffer A (50 mM Tris-HCl (pH 7.4), 50 mM sodium fluoride, 0.27 M sucrose, 1% (w/v) Nonidet P-40, 1 mM EGTA, 1 mM EDTA, 10 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 0.5 mM phenylmethylsulfonyl fluoride, 18 g/ml aprotinin, and 1 mM dithiothreitol). The lysates were centrifuged for 15 min at 12,000 ϫ g and the supernatants were incubated with 2 g of anti-PKB␣ antibody at 4°C for 2 h. Immune complexes were collected using protein G-Sepharose beads and the beads were washed 4 times with buffer A containing 0.5 M NaCl and twice with buffer B (50 mM Tris-HCl (pH 7.4), 0.1 mM EGTA, 1 mM dithiothreitol). The immunoprecipitates were resuspended in 15 l of buffer B containing 0.4 Ci/l [␥-32 P]ATP, 0.1 mM ATP, 10 mM MgCl 2 , and 0.5 g/l histone H2B and incubated at 30°C for 10 min. The reaction mixtures were cooled in an ice-bath and then centrifuged at 12,000 ϫ g for 2 min. The supernatants (15 l) were transferred to tubes containing an equal volume of 2 ϫ sample buffer. After being heated at 100°C for 2 min, aliquots of the sample were spotted onto nitrocellulose filters. The filters were washed 8 times with 2 ml of 50 mM Tris-HCl (pH 8.0) containing 100 mM NaCl, and radioactivity remained on the filters was determined.

Suppression by Adenosine Deaminase of Insulin-induced PtdIns(3,4,5)P 3 Accumulation in Rat Adipocytes-Stimulation
of the 32 P i -labeled adipocytes with 0.1 M insulin caused an increased incorporation of radioactivity into a phospholipid fraction corresponding to PtdIns(3,4,5)P 3 , a product of PI 3-kinase ( Fig. 1). Treatment of the cells with 0.3 M wortmannin, an inhibitor of PI 3-kinase, abolished the insulin-induced production of [ 32 P]PtdIns(3,4,5)P 3 . LY294002, another inhibitor of PI 3-kinase, also prevented the [ 32 P]PtdIns(3,4,5)P 3 production at a concentration of 0.3 mM (data not shown). Rat adipocytes spontaneously release adenosine, which in turn binds to the A 1 receptors on the cells. Because degradation of this adenosine by addition of adenosine deaminase is reported to modulate the insulin action on glucose uptake (1), we examined whether the insulin-induced production of [ 32 P]PtdIns(3,4,5)P 3 is affected by the enzyme. Incubation of the adipocytes with 2 units/ml adenosine deaminase before the insulin addition effectively suppressed the insulin-induced production of [ 32 P]PtdIns(3,4,5)-P 3 (Fig. 1). In this experiment, the radioactivity in the PtdIns(3,4,5)P 3 spot from the cells treated with adenosine deaminase was 30.2% of that from the untreated cells. Fig. 2A shows typical dose-dependent curves of insulin in the presence or absence of adenosine deaminase. Insulin induced a dose-dependent increase in the [ 32 P]PtdIns(3,4,5)P 3 accumulation with the lowest effective concentration around 0.3 nM. Treatment of cells with adenosine deaminase caused a 62 Ϯ 14.5% decrease in the insulin action (mean Ϯ S.E. from four separate experiments). Half-maximal effects of insulin in the presence or absence of adenosine deaminase were not significantly different. As shown in Fig. 2B, the accumulation of [ 32 P]PtdIns(3,4,5)P 3 peaked within 1 min after stimulation with insulin. The increase could still be observed at 10 min after the start of stimulation. The addition of 0.3 M wortmannin at 1 min after insulin caused a rapid decrease in [ 32 P]PtdIns(3,4,5)P 3 , indicating a rapid turnover of this lipid product. Adenosine deaminase suppressed the accumulation without causing marked changes in these kinetic properties.
If the inhibitory action of adenosine deaminase is really via degradation of adenosine, prevention of adenosine binding to the A 1 receptor by pharmacological tools is also expected to suppress the insulin action. PAPA-XAC and 8-cyclopentyl-1,3dipropylxanthine, inhibitors of the adenosine A 1 receptors, inhibited the insulin-induced [ 32 P]PtdIns(3,4,5)P 3 production at concentrations of 5 and 10 M, respectively (data not shown, but see Fig. 5 for PAPA-XAC). Thus adenosine deaminase was considered to modulate the insulin-induced accumulation of [ 32 P]PtdIns(3,4,5)P 3 by eliminating the adenosine action on the A 1 receptors on the adipocytes.
Enhancement of Insulin-induced PtdIns(3,4,5)P 3 Accumulation by GTP-binding Protein-coupled Receptors-The above results suggested that the spontaneously released adenosine in the absence of adenosine deaminase or the A 1 receptor antagonists potentiated the insulin action on the accumulation of [ 32 P]PtdIns(3,4,5)P 3 . Thus we next intended to directly show that stimulation of the A 1 receptors potentiates the insulininduced accumulation of [ 32 P]PtdIns(3,4,5)P 3 under the condition where the extracellular adenosine was depleted. For this purpose, PIA, a poorly hydrolyzable analogue of adenosine, was utilized. When the spontaneously released adenosine was inactivated by adenosine deaminase, PIA effectively enhanced the insulin-induced [ 32 P]PtdIns(3,4,5)P 3 production in a dosedependent manner (Fig. 3). The effect of PIA on insulin action could not be observed when adenosine deaminase was not included in the incubation mixture, indicating that the spontaneously released adenosine caused nearly full activation of the A 1 receptors in the absence of adenosine deaminase.
Adenosine A 1 receptor is a member of the receptors coupling to pertussis toxin-sensitive GTP-binding proteins. Thus we next examined the effect of pertussis toxin on insulin-induced accumulation of PtdIns(3,4,5)P 3 . After treatment of rat adipocytes with 1 g/ml pertussis toxin for 45 min, the action of insulin was markedly reduced (by 66 and 73% in two separate experiments). Thus endogenous adenosine affected the insulin action via GTP-binding protein-linked A 1 receptors on the cells. We next examined whether PGE 2 , which also activates the GTP-binding protein in the cells, induced any effect on the insulin action. Fig. 4 is a typical result showing that PGE 2 potentiated the insulin-induced [ 32 P]PtdIns(3,4,5)P 3 accumulation. PGE 2 at its concentration of 1 M potentiated the action of 0.1 M insulin by 3.6 Ϯ 1.5-fold (results from six separate experiments). In these experiments, adenosine deaminase was included in the suspension to eliminate the action of endogenous adenosine. PGE 2 showed little effect without the addition of adenosine deaminase (data not shown) as was the case of PIA (see Fig. 3). The effect of PGE 2 on the insulin-induced [ 32 P]PtdIns(3,4,5)P 3 accumulation could be observed when the action of endogenous adenosine was blocked by PAPA-XAC or 8-cyclopentyl-1,3-dipropylxanthine (data not shown, but see Fig. 5 for PAPA-XAC).
Failure of Adenylyl Cyclase Inhibition to Potentiate the Insulin-induced PtdIns (3,4,5)P 3 Accumulation-We reported that the accumulation of [ 32 P]PtdIns(3,4,5)P 3 in U937 cells was inhibited by treatment of the cells with a high concentration of dibutyryl cAMP (21). Because activation of pertussis toxinsensitive GTP-binding protein in rat adipocytes suppresses adenylyl cyclase of the cells, the effects of adenosine and PGE 2 might be explained by the decreased adenylyl cyclase activity. We therefore examined whether direct inhibition of adenylyl cyclase enhances the insulin-induced [ 32 P]PtdIns(3,4,5)P 3 accumulation.
In the experiment shown in Fig. 5, the effect of endogenous adenosine was prevented by the addition of 5 M PAPA-XAC, an inhibitor of adenosine A 1 receptors. The treatment effectively augmented the isoproterenol-induced accumulation of cAMP (Fig. 5A), in agreement with a previous finding showing that adenosine deaminase potentiates the cAMP production of rat adipocytes by inactivating endogenous adenosine. Under these conditions, PGE 2 attenuated the cAMP production in a dose-dependent manner (Fig. 5A). DDA, which directly inhibits adenylyl cyclase without activating GTP-binding protein, also inhibited the cAMP production (Fig. 5A). As shown in Fig. 5B, PAPA-XAC attenuated the insulin-induced accumulation of [ 32 P]PtdIns(3,4,5)P 3 as adenosine deaminase did. In contrast, direct inhibition of adenylyl cyclase by DDA did not affect the accumulation of the lipid product even at high concentrations.
One possible explanation for the failure of DDA to increase

FIG. 2. Inhibition by adenosine deaminase of the insulin-induced [ 32 P]PtdIns(3,4,5)P 3 production in rat adipocytes.
A, isolated adipocytes, labeled with 32 P i , were incubated at 37°C for 10 min with (q) or without (E) 2 units/ml adenosine deaminase and then for 10 min in the absence or presence of various concentrations of insulin. B, isolated adipocytes, labeled with 32 P i , were incubated at 37°C for 10 min with (q) or without (E) 2 units/ml adenosine deaminase and then for the indicated times with 0.1 M insulin. At the time indicated by arrows (1 min after the insulin addition), the suspension was further added with 0.3 M wortmannin (broken lines) or vehicle (solid lines). After extraction and separation of [ 32 P]PtdIns(3,4,5)P 3 , the radioactivity in the lipid fraction was determined using a Fuji BAS2000 analyzer. The results are shown as percent of the radioactivity in the PtdIns(3,4,5)P 3 fraction from the cells treated with 0.1 M insulin alone for 5 min. Each point indicates the mean of duplicate determinations. the insulin-induced [ 32 P]PtdIns(3,4,5)P 3 accumulation is that the compound possesses dual actions; DDA has an ability to inhibit adenylyl cyclase but the compound also inhibits the putative cAMP-dependent mechanism leading to the increased lipid production. This might not be the case because DDA did not affect the augmented accumulation of [ 32 P]PtdIns(3,4,5)P 3 observed in the presence of both insulin and PGE 2 (data not shown).
Synergistic Activation of Protein Kinase B by Insulin and Adenosine-Insulin is known to activate rapidly a serine/threonine kinase PKB in rat adipocytes (22). Because the activity of PKB is increased directly (15) or indirectly (16) by lipid products of PI 3-kinase, it is intriguing to examine whether GTPbinding protein-mediated enhancement of [ 32 P]PtdIns(3,4,5)P 3 production is accompanied by enhanced activation of PKB.
In the experiments in Table I, cell lysate was incubated with anti-PKB␣ antibody and the immune complex was subjected to kinase assay with histone H2B as substrate. In agreement with a previous report (22), insulin caused a marked activation of PKB. This activation of PKB was completely blocked by treatment of rat adipocytes with 0.3 M wortmannin. Incubation of the cells with 2 units/ml adenosine deaminase attenuated the insulin-induced activation of PKB, suggesting that endogenous adenosine potentiated the insulin-induced activation of PKB. In agreement with this expectation, a further addition of 0.1 M PIA to the incubation medium reversed the PKB activity to the level of insulin alone. PGE 2 also increased the PKB activity in the presence of both insulin and adenosine deaminase (data not shown, but see below).
Inhibition of adenosine A 1 receptors by PAPA-XAC attenuated the insulin-induced activation of PKB (Table II). Under the conditions, PGE 2 increased the PKB activity to the level without PAPA-XAC. Inhibition of adenylyl cyclase by DDA had no effect on the PKB activity. The results suggested again that activators of pertussis toxin-sensitive GTP-binding proteins acted as a positive regulator of insulin actions by a mechanism independent of their inhibitory action on adenylyl cyclase. DISCUSSION In the present study, we showed that activation of a pertussis toxin-sensitive GTP-binding protein enhanced the insulininduced incorporation of radioactivity to the fraction of PtdIns(3,4,5)P 3 in 32 P i -labeled rat adipocytes. The increased incorporation is considered to accompany the increased intracellular concentration of the second messenger because insulin-

TABLE I Inhibition by adenosine deaminase of the insulin-induced PKB
activation Isolated adipocytes were incubated at 37°C for 5 min with or without 2 units/ml adenosine deaminase, and then for 5 min with or without 1 M PIA. The cells were then incubated for further 5 min with or without 0.1 M insulin. Cell lysate was subjected to immunoprecipitation with anti-PKB␣ antibody, and the immune complex was assayed for kinase activity with histone H2B as substrate. The results are shown as the mean Ϯ S.E. of triplicate determinations. induced activation of PKB, the activity of which is under the control of the lipid product of PI 3-kinase, was also enhanced by activation of the GTP-binding protein. Thus, the receptors coupling to the GTP-binding protein were found to positively regulate insulin-signaling systems by affecting the production of the second messenger of insulin. It is reported that adenosine modulates the insulin action for glucose transport in adipocytes (2,3). Adenosine exerts this effect through the adenosine A 1 receptor coupling to a pertussis toxin-sensitive GTP-binding protein. Although the GTP-binding protein is a negative regulator of adenylyl cyclase, the adenosine effect on the insulin action has been regarded as mostly independent of the change in the adenylyl cyclase activity (6,7). In the present paper, we showed that adenosine has an ability to enhance the insulin-signaling cascade mediated by PI 3-kinase. This effect of adenosine was through the GTP-binding protein but was not due to the decreased production of cAMP, because the direct inhibition of adenylyl cyclase did not affect the insulin actions on PtdIns(3,4,5)P 3 and PKB. Thus the enhanced accumulation of PtdIns(3,4,5)P 3 is another cAMP-independent effect of adenosine. Recent studies demonstrated that PI 3-kinase is enough to stimulate the translocation of glucose transporter from intracellular compartment to plasma membrane (13,14). Constitutively activated mutants of PKB is shown to cause the transporter translocation (17)(18)(19). Thus it is intriguing to discuss a possible effect of the enhanced PtdIns(3,4,5)P 3 accumulation on the cellular glucose transport activity. Fig. 6 shows a correlation between a [ 32 P]PtdIns(3,4,5)P 3 content and a rate of glucose uptake in insulin-treated rat adipocytes. The open symbols represent the results obtained when the cells were treated with the various concentrations of insulin alone. The results from the cells treated with the submaximal concentrations (0.01-0.3 M) of wortmannin and 0.1 M insulin were shown by the closed symbols. The correlation indicates that the rate of glucose uptake is saturated by a small change in the [ 32 P]PtdIns(3,4,5)P 3 content. Because of this saturation characteristic, an enhanced accumulation of the second messenger is expected to cause the increased glucose uptake only when the cells are stimulated by submaximal concentrations of insulin. Previous studies in literature indicated that adenosine increases both insulin sensitivity and maximal response to insulin for glucose transport. One report showed that adenosine increases the glucose uptake in response to a maximal concentration of insulin without changing the extent of the translocation (6). Thus the augmented production of PtdIns(3,4,5)P 3 is not considered to explain the effect of adenosine on the maximal glucose uptake but may contribute in some degree to the increased insulin sensitivity for glucose uptake.
The mechanism by which activation of a pertussis toxinsensitive GTP-binding protein enhances the insulin-induced PtdIns(3,4,5)P 3 accumulation is not clear from the present study. The form of PI 3-kinase responsible for the insulininduced activation is a heterodimer consisting of an 85-kDa regulatory subunit and a 110-kDa catalytic subunit. Stimulation of insulin receptors activates tyrosine kinase activity of the receptors leading to a tyrosine phosphorylation of insulin receptor substrate-1 (IRS-1), which in turn stimulates the catalytic activity of the PI 3-kinase by binding to the Src homology 2 (SH2) domains in the regulatory subunit (8). Thus the GTPbinding proteins might act by increasing the tyrosine phosphorylation of IRS-1. In repeated experiments, however, we could not observe any stimulatory effect of adenosine on the insulininduced tyrosine phosphorylation of IRS-1 (data not shown). Another possible mechanism is that adenosine increases the efficiency of binding between PI 3-kinase and IRS-1. We determined repeatedly the PI 3-kinase activity in the anti-IRS-1 immunoprecipitate and observed that adenosine never increased but on the contrary decreased the activity slightly but reproducibly (data not shown). The reason is not clear now but PtdIns(3,4,5)P 3 has been reported to interact with SH2 domains of PI 3-kinase and to modulate the association between the domains and tyrosine-phosphorylated proteins (23). Thus adenosine is suggested to decrease the interaction between p85 and IRS-1 by increasing the PtdIns(3,4,5)P 3 production of cells.
We have reported that one of the PI 3-kinase activities in a cytosolic fraction of THP-1 cells is stimulated by both a tyrosine-phosphorylated peptide derived from IRS-1 and the ␤␥ subunits of GTP-binding proteins in a synergistic manner (24). We also reported that a heterodimer consisting of the ␤-subtype of p110 and p85 was activated by both of these regulators (25). Thus the present observation suggests that such a dual regulation of the lipid kinase is operating in rat adipocytes. In this regard, we have observed that the PI 3-kinase activity in the anti-p85 immunoprecipitate from these cells is increased by the addition of the ␤␥ subunits (data not shown). If adenosine augmented the insulin-induced PtdIns(3,4,5)P 3 production by this mechanism, it is interesting to examine whether a similar interaction could be observed for PDGF which activates PI  3-kinase by recruiting the enzyme to membrane receptors (instead of inducing the association with cytosolic protein IRS-1). In order to address this point, we examined the effect of PDGF on rat adipocytes, because stimulation of PDGF receptors on the cells is reported to increase the PI 3-kinase activity associated with an antibody against the receptors even when the cells have not been transfected with the receptors (26). However, we could not detect any effect of PDGF on the PtdIns(3,4,5)P 3 production and the PKB activation regardless of whether adenosine deaminase is included in the assay mixture (data not shown). Thus further study using other cell systems is necessary and is now under study in our laboratory to clarify if the interaction could be observed for the agonists other than insulin.
In summary, the present paper showed that a second messenger production by insulin is positively regulated by the receptors coupling to GTP-binding protein in physiological target cells of insulin. Although the precise molecular mechanism of this cross-talk is not clear now, the finding is intriguing in view of the central role of PI 3-kinase in insulin-signaling cascades leading to metabolic and mitogenic cellular responses. One possible consequence of the cross-talk is a modified insulin sensitivity for glucose transport as discussed above. In this regard, it is interesting to note that the abnormal function of adenosine A 1 receptors in genetically obese animals and the decreased level of G i in a streptozotocin-induced diabetic rat have been reported (27,28). Furthermore, a study using transgenic mice has demonstrated that a deficiency of G␣ i2 produces impaired glucose tolerance and resistance to insulin (29). A possible relation of the present study to these observations is intriguing to be examined.