Phosphatidylinositol 3-kinase (PI 3-kinase) plays an important role in a variety of hormone and growth factor-mediated intracellular signaling cascades and has been implicated in the regulation of a number of metabolic effects of insulin, including glucose transport and glycogen synthase activation. In the present study we have examined 1) the association of PI 3-kinase with the insulin receptor kinase (IRK) in rat liver and2) the subcellular distribution of PI 3-kinase-IRK interaction. Insulin treatment promoted a rapid and pronounced recruitment of PI 3-kinase to IRKs located at the plasma membrane, whereas no increase in association with endosomal IRKs was observed. In contrast to IRS-1-associated PI 3-kinase activity, association of PI 3-kinase with the plasma membrane IRK did not augment the specific activity of the lipid kinase. With use of the selective PI 3-kinase inhibitor wortmannin, our data suggest that the cell surface IRK β-subunit is not a substrate for the serine kinase activity of PI 3-kinase. The functional significance for the insulin-stimulated selective recruitment of PI 3-kinase to cell surface IRKs remains to be elucidated.
- insulin signaling
- subcellular compartments
the insulin receptor kinase (IRK) is a disulfide-linked heterotetrameric glycoprotein composed of two extracellular insulin-binding α-subunits and two transmembrane β-subunits that contain tyrosine kinase activity in their cytosolic domains (see review in Ref. 16). Upon insulin binding, rapid tyrosine autophosphorylation on the intracellular portion of the β-subunit leads to IRK activation. After activation, the insulin-receptor complex is rapidly internalized into endosomes (ENs), which mediate the sorting and processing of a number of ligand-receptor complexes and have been implicated in both signal transduction and signal termination (reviewed in Refs. 2 and 8).
The activated IRK effects tyrosine phosphorylation of a number of intracellular proteins, including insulin receptor substrates-1 and -2 (IRS-1 and IRS-2) and, in certain cell types, SHC (16). These molecules, rather than the IRK itself, appear to be the predominant proteins coupling to and activating downstream signaling pathways by providing specific phosphotyrosine binding sites for a number of src homology 2 (SH2)-containing proteins, including phosphatidylinositol 3-kinase (PI 3-kinase).
PI 3-kinase is involved in mediating numerous effects of insulin, including glucose transport, glycogen synthase activation, and the inhibition of lipolysis (1). The best-characterized member of the PI 3-kinase family exists as a heterodimer composed of a regulatory 85-kDa subunit (p85), containing two SH2 domains, and a 110-kDa catalytic subunit (p110) that possesses both lipid and serine kinase activity. After binding of the p85 SH2 domains to specific tyrosine-phosphorylated motifs (pYXXM) on molecules such as IRS-1 and IRS-2, the p110 catalytic subunit becomes activated (see review in Ref.43). However, in addition to these adapter molecules, mounting evidence indicates that PI 3-kinase can also associate with the activated IRK β-subunit (4, 21, 29,35, 44, 47). In vitro studies indicate that the SH2 domains of p85 can interact directly with phospho-Tyr1322 (which lies in a Y1322THM motif) of the IRK β-subunit,1 although other investigations suggest that additional regions of the β-subunit may also be required (3, 46, 47). PI 3-kinase activity is detected in anti-IRK immunoprecipitates from insulin-stimulated cells (4, 35). However, the physiological significance of this potential alternative pathway of insulin-stimulated PI 3-kinase activation, or indeed the subcellular location of PI 3-kinase-IRK interaction, remains to be established.
In the present study we examined the association of PI 3-kinase with the IRK in rat liver, a major insulin target tissue, in vivo. Because internalization of activated IRKs into ENs plays an important role in insulin signal transduction (reviewed in Refs. 8 and 18), PI 3-kinase interaction with the IRK was investigated in different subcellular compartments of the liver. Our studies reveal that insulin treatment promotes a rapid and selective recruitment of PI 3-kinase to IRKs located at the plasma membrane. However, in contrast to cytosolic IRS-1-associated PI 3-kinase activity, this association does not augment the specific activity of the lipid kinase. Furthermore, using the selective PI 3-kinase inhibitor wortmannin, we show that the cell surface IRK β-subunit is not a substrate for the serine kinase activity of PI 3-kinase.
MATERIALS AND METHODS
Female Sprague-Dawley rats, 10 wk of age and 160–180 g body weight, were purchased from Charles River Canada (St. Constant, PQ, Canada), housed in an animal facility with 12:12-h light-dark cycles at 25°C, and fed ad libitum on Purina chow. Animals were fasted overnight (16–18 h) before use.
All studies herein cited were performed with the approval of the McGill University Animal Care Committee.
Porcine insulin was a gift from Eli Lilly (Indianapolis, IN). Phenylmethanesulfonyl fluoride, HEPES (free acid), sodium orthovanadate, rabbit γ-globulin, Tris, radioimmunoassay grade BSA, and most other chemicals were purchased from Sigma (St. Louis, MO). Protein A-Sepharose CL-4B (PAS) was from Pharmacia LKB Biotechnology (Uppsala, Sweden). ATP (disodium salt) was obtained from Boehringer Mannheim (Laval, PQ, Canada). [γ-32P]ATP (3,000 Ci/mmol) and 125I-labeled goat anti-mouse (GAM) and125I-labeled goat anti-rabbit (GAR) secondary antibodies were purchased from Du Pont-NEN Radiochemicals (Lachine, PQ, Canada). Reagents for electrophoresis were from Bio-Rad (Richmond, CA) with the exception of 14C-labeled protein standards, which were supplied by GIBCO/BRL Canada (Burlington, ON, Canada). Kodak X-OMAT AR film was from Picker International (Montreal, PQ, Canada). Polyvinylidene fluoride Immobilon-P transfer membranes were from Millipore (Mississauga, ON, Canada). The protein tyrosine phosphatase (PTP) inhibitor bisperoxo(1,10-phenanthroline)oxovanadate anion [bpV(phen)] was prepared by Dr. Jesse Ng of the Department of Chemistry, McGill University, as described previously (32). l-α-Phosphatidylinositol (sodium salt) was purchased from Avanti Polar Lipids (Alabaster, AL). Merck Silica Gel 60 and cellulose thin-layer chromatography (TLC) plates were purchased from EM Separation Technology (Gibbstown, NJ). An antibody raised against a peptide corresponding to residues 942–969 of the juxtamembrane region of the IRK β-subunit (α960) was prepared and purified on a PAS column as previously described (11). Polyclonal αp85 (recognizing both p85α and p85β) for immunoprecipitation and Western blotting and a polyclonal αIRS-1 for Western blotting were purchased from Upstate Biotechnology (Lake Placid, NY). A monoclonal anti-phosphotyrosine antibody (αPY) and a polyclonal anti-SHP-2 antibody for Western blotting were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Antibodies recognizing the human IRK α-subunit (αIRKα) were purified from a patient's serum as previously described (24). A polyclonal antibody raised against a peptide corresponding to the fourteen carboxy-terminal residues of IRS-1 was used for immunoprecipitation.
Preparation of subcellular liver fractions.
After ether anesthesia, rats received an intrajugular injection of insulin (1.5 μg/100 g body wt) in PBS containing 0.1% BSA for the times indicated. Hepatic plasma membrane (PM), endosomal (EN), and cytosolic fractions were prepared as described previously (7). The protein content of these fractions was measured using a modification of Bradford's method (10), with BSA as standard (24).
Immunoprecipitation and immunoblotting.
Cell fractions (300 μg protein), in a final volume of 550 μl, were incubated in the presence of 1% (vol/vol) Triton X-100 and 0.5% (wt/vol) sodium deoxycholate at 4°C for 1 h. After centrifugation at 10,000 g AV for 2 min, the supernatant was precleared for 30 min with 1.0 μg of nonimmune IgG. Twenty microliters of a 50% slurry of PAS were added, and after 30 min of incubation, the solution was centrifuged as above. Supernatants were incubated for 2 h with 20 μl of αIRS-1, αp85, or αIRK at 4°C, after which 50 μl of a 50% slurry of PAS were added, and the solution was incubated for a further 1 h. After centrifugation as above, 100 μl of the supernatant were added to 50 μl of 3× Laemmli sample buffer (2.3% SDS, 10% glycerol, 100 mM dithiothreitol, and 0.37 M Tris · HCl, pH 6.8: final concentration, 1×). The pellet was washed three times with 1 ml of wash buffer (50 mM HEPES, pH 7.4, containing 1% Triton X-100, 0.1% SDS, 150 mM NaCl, 100 mM NaF, and 2 mM sodium orthovanadate) followed by boiling in 210 μl of Laemmli sample buffer. Eighty-microliter samples were subjected to SDS-PAGE (7.5% gel) and then transferred to Immobilon-P membranes for Western blotting. Either 125I-GAR or 125I-GAM was used as secondary antibody and, after autoradiography at −80°C, appropriate bands were quantified using a Bio-Rad GS-700 Imaging Densitometer.
IRS-1 and IR-associated PI 3-kinase activity.
Immunoprecipitates (either αIRKα or αIRS-1) were extensively washed (20), and the PAS pellet was resuspended in 50 μl of kinase assay buffer (in mM: 20 Tris · HCl, pH 7.5, 100 NaCl, and 0.5 EGTA) containing 0.5 mg/ml PI and assayed for PI 3-kinase activity, as previously described (6).
Phosphoamino acid analysis of the IRK and p85: effect of wortmannin.
IRK or p85 immunoprecipitates were washed as described above and resuspended in 20 μl of assay buffer (in mM: 20 HEPES, pH 7.4, 3 MnCl2, and 10 MgCl2). Samples were preincubated for 20 min at 22°C with 500 nM wortmannin in DMSO, or with DMSO alone, and kinase reactions were started by the addition of 20 μCi of [γ-32P]ATP (5 μM final concentration) as described in Ref. 27. After 10 min of incubation at 22°C, the reaction was stopped by the addition of 500 μl of ice-cold PBS. Samples were centrifuged at 10,000 g AV, and pellets were washed twice with 500 μl PBS, resuspended in 90 μl of Laemmli sample buffer, and boiled for 5 min. Samples were subjected to SDS-PAGE (7.5%) and transferred onto Immobilon-P membrane, and phosphobands were visualized by autoradiography at −80°C.
Phosphoamino acid analysis was performed on amino acids liberated by partial acid hydrolysis. Briefly, membrane segments corresponding to the IRK β-subunit or p85 were excised and boiled in 200 μl of 6 N HCl for 1 h at 110°C. After centrifugation at 10,000g AV for 5 min, supernatants were evaporated to dryness in vacuo. Digests were resuspended in 10 μl of pH 1.9 buffer (formic acid and acetic acid, pH 1.9), spotted onto TLC plates, and subjected to two-dimensional phosphoamino analysis using a Hunter thin-layer peptide mapping apparatus (CBS Scientific, Del Mar, CA), as described in Ref. 23. Phosphoamino acids were visualized by autoradiography and identified by comparison with ninhydrin-stained phosphoamino acid standards.
Initial studies evaluated whether insulin treatment promoted a change in the level of the PI 3-kinase regulatory subunit p85 associating with rat hepatic PM and EM fractions. Figure1 shows the distribution of p85 induced by insulin treatment of overnight-fasted rats as determined by subcellular fractionation and immunoblotting with anti-p85. Insulin treatment led to an ∼50% increase in the level of p85 protein present in both PM and EN fractions, which reached a peak at 1 min postinjection in both subcellular fractions. Whereas PM-associated p85 content declined rapidly to basal levels, EM p85 content declined more gradually and remained elevated above basal at 15 min postinjection. At the peak time of insulin-stimulated p85 recruitment to membrane fractions, p85 content in the PM (expressed per μg PM protein) was approximately threefold greater than that observed in ENs (inset, Fig. 1).
Effect of insulin treatment on p85 subunit association with PM and EN IRKs.
We sought to determine the impact of insulin stimulation on the association of p85 with IRKs in vivo. In agreement with previous studies (11), insulin treatment promoted a time-dependent decrease in IRKs located at the PM, whereas an approximate 10-fold increase in EM IRK levels was observed that was maximal at 2–5 min postinjection (Fig. 2, top). Insulin treatment promoted an approximate fivefold increase in p85 associating with the PM IRK that was maximal between 30 s and 2 min postinjection (P < 0.01).2 In contrast, a 50% decrease in p85 associating with the EN IRK was observed between 1 and 5 min after insulin treatment (2 min postinsulin injection;P < 0.05), with the association returning to basal levels by 15 min postinjection. It is noteworthy that in the basal non-insulin-stimulated state (t = 0), association of p85 with EN IRKs was fourfold less than that observed for PM receptors (Fig. 2, top). IRS-1 was detected in anti-IRK immunoprecipitates from PM fractions (Fig.3 A), but not IRS-2 (results not shown). This raised the possibility that the observed association between IRKs and p85 in the PM could be the result of a ternary complex containing IRS-1. To test for this, the kinetics of insulin-stimulated IRS-1 and p85 association with PM IRKs were compared. IRS-1 association with PM IRKs reached a peak at 30 s postinsulin injection, after which levels rapidly declined to basal by 10 min postinjection (Fig.3 A). In contrast, although levels of p85 associating with PM IRKs also peaked at 30 s after insulin administration, levels remained elevated and substantially above basal at 10 min postinjection (Fig. 3 A). Also noteworthy is the finding that, after insulin treatment, the p85-to-IRS-1 ratio in IRK immunoprecipitates from PM increased by fivefold between 30 s and 2 min postinsulin injection and declined thereafter (Fig. 3, A andC). In contrast, the ratio of cytosolic p85 to IRS-1 increased only twofold after insulin administration and remained elevated at 10 min postinjection (Fig. 3, B andC). These data are consistent with the possibility that a population of p85 associates with the PM IRK in an IRS-1-independent manner. Because the PTP, SHP-2 (PTP-1D, Syp), has been reported to associate with the tyrosine-phosphorylated β-subunit of the IRK via an SH2-mediated interaction (26,33, 38, 41), the presence of this enzyme in IRK immunoprecipitates was also assessed. Immunoblotting with anti-SHP-2 antibodies failed to detect SHP-2 associating with rat hepatic PM and EN IRKs (results not shown).
Effect of bpV(phen) pretreatment on insulin-stimulated p85 association with PM and EN IRKs.
Through an SH2-mediated interaction, the p85 subunit is reported to associate directly with the IRK via its β-subunit phospho-Tyr1322 (29, 44). Previous studies identified that IRK internalization in liver was associated with a rapid (within 2 min postinsulin injection) and partial dephosphorylation of β-subunit phosphotyrosine residues due to the action of an endosomally located protein tyrosine phosphatase(s) (PTP) (11, 19). The comparatively low level of p85 associating with EN IRKs (cf., PM IRKs) may therefore be the result of IRK dephosphorylation at this intracellular locus. To test for this possibility, EN IRK dephosphorylation was specifically blocked by pretreating rats with the PTP inhibitor bpV(phen) (32), as described by Drake et al. (19). Although we cannot directly demonstrate the phosphorylation status of the carboxy-terminal tyrosine residue (Tyr1332) in the present study, previous work (19) has demonstrated that at 15 min after a single injection of bpV(phen), EN IRK β-subunit dephosphorylation is inhibited by ∼100%. As shown in Fig.4, a 15-min bpV(phen) pretreatment had negligible effect on the time course or magnitude of insulin-stimulated association of p85 with either EN or PM IRKs.
Effect of insulin treatment on PI 3-kinase lipid kinase activity associated with PM and EN IRKs.
We determined whether the lipid kinase activity of the PI 3-kinase associating with PM and EN IRKs was modulated by insulin treatment in vivo. Insulin treatment led to a rapid and approximate sixfold increase in lipid kinase activity associated with the PM IRK that reached a peak at 30 s postinjection and declined thereafter (Fig.5). In contrast, at 2 min after insulin treatment, an ∼50% decrease in PI 3-kinase activity associated with EN IRKs was observed. Because PM IRK-associated lipid kinase activity closely mirrored insulin-stimulated p85 recruitment to IRKs in this cell fraction, we sought to establish whether the intrinsic (or specific) activity of PI 3-kinase was modified by p85-IRK association. Figure 6 shows the intrinsic activity of PI 3-kinase associated with PM IRKs at 30 s postinjection, the time of maximal recruitment of p85 to the PM IRK. For comparison, the specific activity of PI 3-kinase associated with rat liver cytosolic IRS-1 was determined. Although insulin treatment promoted a fivefold increase in the amount of p85 subunit associating with PM IRKs at 30 s postinjection (Fig. 2), this elevated association did not increase the specific activity of the enzyme (Fig. 6). In contrast, insulin treatment led to an approximate fivefold increase in the intrinsic activity of PI 3-kinase associated with cytosolic IRS-1 (Fig.6). It has been reported that insulin treatment stimulates only the activity of p85α-associated PI 3-kinase (5). The difference in insulin-stimulatable PI 3-kinase activity associated with PM IRKs and cytosolic IRS-1 in rat liver may therefore lie with the nature and/or level of the p85 isoform that is associated with these two molecules. To test for this, we examined whether the levels of insulin-stimulated p85α association with the PM IRK and with cytosolic IRS-1 were similar. Table 1shows levels of p85α and total p85 (both p85α and p85β; PAN-p85) in PM IRK and cytosolic IRS-1 immunoprecipitates after insulin treatment. The ratios of p85α to PAN-p85 associated with either PM IRK (0.15 ± 0.06) or cytosolic IRS-1 (0.20 ± 0.05) were very similar (Table 1). This suggests that the lack of insulin-stimulated PI 3-kinase activity in PM IRK immunoprecipitates, in contrast to that associated with cytosolic IRS-1, cannot be explained by an absence, or indeed a reduced proportion, of p85α associating with the IRK.
Effect of wortmannin on IRK β-subunit phosphotyrosine, -serine, and -threonine content.
Because the intrinsic lipid kinase activity of PI 3-kinase did not increase after binding to rat hepatic IRKs, we examined whether recruitment to PM-located IRKs might serve as a mechanism by which the serine kinase moiety of PI 3-kinase could phosphorylate the IRK β-subunit. This was assessed using a procedure described by Lam et al. (27) and Tanti et al. (40). Briefly, IRKs immunoprecipitated from rat liver PM fractions at 30 s after insulin treatment were preincubated with wortmannin,3 a selective inhibitor of both the lipid and serine kinase activities of PI 3-kinase (see review in Ref. 42). Immunocomplexes containing both IRK and PI 3-kinase were subsequently incubated with [32P]ATP in the presence of Mn2+ and Mg2+, and IRK β-subunits were subjected to phosphoamino acid analysis (Fig.7), as described in materials and methods. As shown in Fig. 7, wortmannin pretreatment was without effect on the phosphoserine, -threonine, or -tyrosine content of PM IRKs, either in controls or in animals that were killed at 30 s postinsulin injection. In contrast, wortmannin pretreatment promoted an ∼50% decrease in the phosphoserine content of the p85 regulatory subunit in anti-p85 immunoprecipitates from PM fractions isolated at 30 s postinsulin injection (results not shown). Thus, despite a rapid recruitment of PI 3-kinase to rat hepatic PM IRKs at 30 s postinsulin treatment, the IRK β-subunit does not appear to be a substrate for the serine kinase activity of PI 3-kinase.
To our knowledge, this is the first study that has 1) assessed the insulin-stimulated recruitment of PI 3-kinase to insulin receptors in a major insulin target tissue (that is, rat liver) in vivo and 2) examined this interaction in different subcellular compartments. A major finding of the study was the rapid recruitment of PI 3-kinase to rat hepatic IRKs located at the PM after injection of a single physiological dose of insulin. Reports on the nature of the interaction between p85 and the IRK are in conflict with some studies indicating that this association is direct and mediated via binding of the SH2 domains of p85 to phospho-Tyr1322 of the IRK β-subunit (29, 44), whereas others suggest an indirect mechanism involving the formation of a ternary complex with IRS-1 (3, 22). The reasons for these discrepant results are unclear but may reflect the use of different cell lines, where the relative levels of IRK, p85, and IRS-1 may differ. Although IRS-1 was detected in anti-IRK immunoprecipitates in the present study, the different time courses of insulin-stimulated IRS-1 and p85 association with PM IRKs suggest that at least a proportion of p85-IRK association is not mediated via a ternary complex containing IRS-1. Whether this p85-IRK association is direct or mediated via another signaling molecule remains to be established.
Although the current study does not provide an explanation for the difference in p85 association between PM and EN IRKs, dephosphorylation of key IRK β-subunit tyrosine residues necessary for p85 binding, by an EN-located PTP, appears to be ruled out. Thus pretreatment of rats with bpV(phen), a specific inhibitor of endosomal IRK tyrosine dephosphorylation (19, 32), had a negligible effect on p85-IRK interaction at this intracellular locus (Fig. 4). It is possible, therefore, that p85 interacts with the IRK in a phosphotyrosine-independent manner, although the fact that the time course of insulin-stimulated p85 association with the PM IRK closely mirrors IRK β-subunit phosphotyrosine content (11) suggests otherwise. Because bpV(phen) is without effect on IRK dephosphorylation at the PM (19), dephosphorylation of IRK β-subunit tyrosine residues, through the action of a PM-located PTP, may have occurred before internalization. Alternatively, the activated IRK may, via a phosphotyrosine-dependent mechanism, interact with another molecule in ENs that occludes an association with p85 in this subcellular compartment. A number of studies have described the direct interaction of the SH2-containing PTP, SHP-2, with the IRK β-subunit (26, 38,41) via an interaction of the SH2 domains of this enzyme with both the kinase regulatory domain and Tyr1322 of the IRK β-subunit (33). Our studies, however, did not detect SHP-2 in IRK immunoprecipitates from either PM or EN fractions, thus excluding the possibility that this PTP may impede p85 interaction with the IRK in the endosomal compartment.
On the basis of the observation that the association between p85 and the IRK decreases in ENs after insulin treatment (Fig. 2), it would appear that activated IRKs entering the EN apparatus from the cell surface are not complexed with p85, thus “diluting” the fraction of p85-associated IRKs that are present in this subcellular fraction. It is possible, therefore, that a selective internalization of IRKs takes place from the PM, with those receptors binding p85 being retained at the cell surface. This may allow a subset of IRKs to stimulate a PM-specific insulin-signaling cascade. It should be noted, however, that we cannot rule out the possibility that transfer of p85 from internalized p85-IRK complexes to another binding partner could also account for the observed decrease in p85-IRK association in ENs.
PI 3-kinase activity is necessary for ligand-mediated internalization of both the c-Kit receptor (22) and the CD28 receptor (13), and for efficient trafficking of the platelet-derived growth factor (PDGF) receptor from peripheral compartments to juxtanuclear vesicles (37). Such a role for insulin-stimulated PI 3-kinase association with cell surface IRKs is unlikely, because the maximal interaction between p85 and PM IRKs occurs while IRK content in PM is decreasing. Although this suggests that p85-IRK association is not a principal effector of IRK endocytosis, we cannot rule out the possibility that the association of p85 with the IRK is responsible for its selective retention at the cell surface.
A second important finding of our study was that, although PI 3-kinase was recruited to PM IRKs after insulin treatment, the intrinsic activity of the lipid kinase did not increase. This is in marked contrast to cytosolic IRS-1-associated PI 3-kinase activity, where insulin treatment led to a rapid recruitment and activation of the enzyme (Figs. 2 and 6). It has been reported that only p85α-associated PI 3-kinase activity is stimulated by insulin (5). Thus the difference in insulin-stimulatable PI 3-kinase activity in PM IRK and cytosolic IRS-1 immunoprecipitates may lie with a difference in the type or proportion of p85 isoforms associating with these two molecules. However, the present study revealed that a very similar proportion of p85α associates with PM IRK and cytosolic IRS-1 (Table 1). Previous studies have recognized that full activation of PI 3-kinase is achieved when both SH2 domains of the p85 regulatory subunit bind phosphotyrosine (34) in the context of pYXXM (see review in Ref. 43). The IRK β-subunit contains only a single YXXM motif (at Tyr1322), implying that only one p85 SH2 domain is occupied upon binding to the IRK. The PDGF receptor, by contrast, contains two YXXM motifs (at Tyr740 and Tyr751) and has been shown to potently stimulate PI 3-kinase activity through receptor binding (see review in Ref. 9). Because IRS-1 has nine YXXM motifs, the potential to bind both p85 SH2 domains and hence fully activate PI 3-kinase is high. It is therefore possible that the inability of the IRK to stimulate PI 3-kinase activity through p85-IRK interaction may result from the absence of a second YXXM site on the β-subunit. Thus, although it has been proposed that the p85 subunit may be able to span identical sites in activated receptor homodimers (31), spatial or structural constraints may prevent this for the IRK.
After IRK activation there is an increase in the phosphorylation of the IRK β-subunit on serine and threonine residues (reviewed in Refs. 16and 36). Although the functional significance of IRK serine/threonine phosphorylation is unclear, it is thought to serve to inhibit or dampen IRK activity and, hence, insulin signaling. To date, however, the identity of the insulin-stimulated serine/threonine kinase(s) that phosphorylates the IRK in vivo remains unclear. Previous studies have noted that IRK β-subunit serine/threonine phosphorylation was restricted to cell surface receptors (24). Because the present study showed that p85 was rapidly recruited to PM IRKs after insulin treatment, our observations raised the possibility that PI 3-kinase may phosphorylate the IRK in vivo. However, we did not observe a wortmannin-sensitive serine phosphorylation of the IRK β-subunit in PM immunocomplexes containing both IRK and PI 3-kinase activity. Moreover, pretreatment of primary cultured hepatocytes with wortmannin before insulin stimulation was without effect on total cellular IRK activation (P. G. Drake and V. M. Dumas, unpublished observations). This is in agreement with a study of Lam et al. (27), who report that wortmannin has no effect on the in vitro kinase activity of adipocyte IRKs. The lack of an effect of PI 3-kinase inhibition on IRK activation or on the β-subunit phosphoamino acid profile of rat liver PM IRKs (Fig. 6) suggests that the PI 3-kinase serine kinase is not a direct regulator of IRK function. However, it should be noted that we cannot rule out the possibility that a potential in vivo phosphorylation of all available serine residues by PI 3-kinase before animals are killed and before subcellular fractionation could prevent further phosphorylation in vitro. Recent studies have observed a PI 3-kinase association with a 76-kDa wortmannin-insensitive serine kinase, termed a PI 3-kinase-associated kinase, that can phosphorylate IRS-1 (14, 15). It is possible, therefore, that although the PI 3-kinase does not phosphorylate the IRK itself, it may form part of a complex that brings this novel serine kinase into contact with the PM IRK or, indeed, other PM-located proteins. Such an adaptor function of the p85 regulatory subunit has been proposed by Sung et al. (39), who report that p85 can form a complex linking the activated IRK to a GTPase-activating protein (GAP) and a 62-kDa GAP-associated protein.
In summary, this study has identified a major difference in the association of PI 3-kinase with IRKs in different subcellular compartments of rat liver. Whereas insulin treatment promoted a rapid and pronounced increase in the amount of p85 and PI 3-kinase activity associating with cell surface IRKs, no increase was observed for EN IRKs. Potential roles for an insulin-stimulated association of PI 3-kinase with the PM IRK include a regulation of insulin signal transduction from the cell surface and/or targeting of a molecule to the PM and/or IRK. Because the subcellular location of activated IRKs has been shown to be an important element of normal insulin signal transduction in vivo, future studies will evaluate these possibilities.
We thank Gerry Baquiran for excellent technical assistance and Dr. Louise Larose for insightful comments.
↵* P. G. Drake and A. Balbis contributed equally to this work.
This work was supported by a grant from the Medical Research Council of Canada (to B. I. Posner).
Address for reprint requests and other correspondence: B. I. Posner, Polypeptide Hormone Lab., Strathcona Anatomy and Dentistry Bldg., 3640 University St., Rm. W3.15, Montreal, Quebec, Canada H3A 2B2 (E-mail:).
1 The IRK amino acid numbering system used in this paper is based on the human insulin proreceptor without the 12 amino acids encoded by exon 11 (HIR-A or Ex11) (28).
↵2 By analyzing the amount of p85 immunoprecipitated by phosphotyrosine antibodies from cytosol and PM (data not shown), and on the basis of previously determined recoveries of PM (18), we calculated that the proportion of PI 3-kinase activity in PM is ∼10% of that in cytosol. In PM the proportion of p85 precipitated by αIRK was not >10% (data not shown). Thus the PI 3-kinase associated with the IRK is in the range of 1% of the total activity in cytosol.
↵3 A concern that wortmannin, at the concentration used to inhibit PI 3-kinase activity, might directly effect IRK activity (both IRK autophosphorylation and activity toward exogenous substrates) was tested by preincubating wheat germ agglutinin-purified PM IRKs with 500 nM wortmannin for 20 min at room temperature before measurement of IRK activity, as previously described (11). Under these conditions, wortmannin was without effect on either IRK autophosphorylating activity or exogenous activity toward polyglutamic acid-tyrosine (4:1).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 2000 the American Physiological Society