Decrease in beta-subunit phosphotyrosine correlates with internalization and activation of the endosomal insulin receptor kinase.

In a previous study, we showed that the rat hepatic insulin receptor (IR) kinase of endosomes (ENs) was transiently activated to levels exceeding those of plasma membrane (PM) receptors following insulin injection. Phosphatase treatment of EN receptors abolished IR kinase activation implicating beta-subunit autophosphorylation as a mediator of the activation process (Khan, M. N., Baquiran, G., Brule, C., Burgess, J., Foster, B., Bergeron, J. J. M., and Posner, B. I. (1989) J. Biol. Chem. 264, 12931-12940). In the present study, the phosphotyrosine (PY) content of the IR beta-subunit in PM and ENs was estimated by two different methods. In one method, direct in vivo labeling with 32Pi followed by receptor immunoprecipitation was carried out. In the second method, immunoblotting with antibodies against the submembrane domain of the IR beta-subunit, encompassing residue 960 (alpha 960), and with antibodies against PY (alpha PY) was used to determine the content of PY/beta-subunit in PM and ENs following injection of insulin. By both methods, it was found that the PY content of PM IR was significantly greater than that of IR in ENs. With doses of 1.5 micrograms of insulin/100 g body weight (50% receptor occupancy) or 15 micrograms/100 g body weight (receptor saturation), the PY/beta-subunit of PM IR attained a level 2.0 to 2.5-fold of that observed for the IR of ENs. Surprisingly, the IR of ENs incorporated 3 to 5 times more PY/beta-subunit than those of PM consequent to autophosphorylation. Exogenous IR kinase activity (poly(Glu:Tyr)) in PM changed only slightly with insulin dose. In contrast, EN receptors exhibited a dose-dependent increase in kinase activity coincident with the decrease in PY/beta-subunit levels. A comparison of the proportion of receptor and kinase activity immunoprecipitated by alpha PY both before and after autophosphorylation indicated that ENs but not PM contained a small population of lightly phosphorylated but highly activated receptors. Since Thr12-Lys (IR kinase residues 1142-1153) efficiently inhibited IR autophosphorylation of both PM and EN receptors, Tris phosphorylation of beta-subunit regulatory tyrosines was unlikely. These results may be explicable by a dephosphorylation-dependent activation of IR kinase, as seen with the src family of tyrosine kinases.

In a previous study, we showed that the rat hepatic insulin receptor (IR) kinase of endosomes (ENS) was transiently activated to levels exceeding those of plasma membrane (PM) receptors following insulin injection. Phosphatase treatment of EN receptors abolished IR kinase activation implicating @-subunit autophosphorylation as a mediator of the activation process (Khan, M. N., Baquiran, G., Brule, C., Burgess, J., Foster, B., Bergeron, J. J. M., and Posner, B. I. (1989) J. Bioi. Chem. 264, 12931-12940). In the present study, the phosphotyrosine (PY) content of the IR Bsubunit in PM and ENS was estimated by two different methods. In one method, direct in vivo labeling with carried out. In the second method, immunoblotting with antibodies against the submembrane domain of the IR &subunit, encompassing residue 960 (a960), and with antibodies against PY (aPY) was used to determine the content of PY/B-subunit in P M and ENS following injection of insulin. By both methods, it was found that the P Y content of PM IR was significantly greater than that of IR in ENS. With doses of 1.5 pg of insulin/100 g body weight (50% receptor occupancy) or 15 pg/lOO g body weight (receptor saturation), the PY/b-subunit of PM IR attained a level 2.0 to 2.5-fold of that observed for the IR of ENS. Surprisingly, the IR of ENS incorporated 3 to 5 times more PY/b-subunit than those of PM consequent to autophosphorylation. Exogenous IR kinase activity (poly(G1u:Tyr)) in PM changed only slightly with insulin dose. In contrast, EN receptors exhibited a dose-dependent increase in kinase activity coincident with the decrease in PY/& subunit levels. A comparison of the proportion of receptor and kinase activity immunoprecipitated by a P Y both before and after autophosphorylation indicated that ENS but not PM contained a small population of lightly phosphorylated but highly activated receptors. Since Thr12-Lys (IR kinase residues 1142-1153) efficiently inhibited IR autophosphorylation of both P M and EN receptors, Tris phosphorylation of &subunit regulatory tyrosines was unlikely. These results may be explicable by a dephosphorylation-dependent 32 Pi followed by receptor immunoprecipitation was * These studies were supported by grants from the Medical Research Council of Canada and Grant RO1-DI 19573 from the United States Public Health Service. 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 U.S.C. Section 1734 solely to indicate this fact. activation of IR kinase, as seen with the src family of tyrosine kinases.
Polypeptide hormones such as prolactin and insulin, and growth factors such as epidermal growth factor, bind to their receptors on target cells and are rapidly internalized with their cognate receptors (reviewed in Refs. 1 and 2). The relationship between the rapid ligand-mediated internalization of receptor, bioresponse, desensitization of receptor activity, and degradation of ligand and cognate receptor has been particularly difficult to unravel. In part, this has been due to the extraordinary rapidity of receptor internalization in target cells such as liver (2, 3) and the only recent uncovering of novel intracellular compartments (collectively called the endosomal apparatus) which transiently concentrate internalized ligands and their receptors (1-4). It is within this apparatus where the sorting, degradation of ligand-receptor complexes, and regulation of receptor activity may be initiated, extended, and/or completed.
Studies of insulin mimetic agents (20)(21)(22), kinase inhibitory antibodies (23,24), and kinase-impaired receptor mutants (25)(26)(27)(28)(29)(30)(31) have emphasized the requirement for receptor kinase activation in realizing the biological effects of insulin. In previous studies, we observed that insulin administration in vivo resulted in the accumulation within ENS of activated insulin receptors (3,32). Similar observations were made in isolated rat adipocytes (33). A relationship between receptor activation and internalization is supported by the observation that kinase-impaired receptors appear to be internalization defective (27,34). The above considerations and our recent observation that peak autophosphorylation activity of the endosomal receptor was 3 to 5 times the peak activity observed in plasmalemma fractions (3) supports an initial suggestion that internalized receptors may be involved in signal transduction (17).
The contribution of receptor occupancy by ligand to continued receptor activation in vivo is not known presently. It has been demonstrated that internalized insulin receptors persist in a tyrosine-phosphorylated state after the dissociation of insulin (35). We observed that the augmentation of endosomal insulin receptor kinase activity was abolished by alkaline phosphatase treatment indicating that it was contingent on some variation of @-subunit autophosphorylation (3). Autophosphorylation of the insulin receptor occurs at several sites in the @-subunit involving at least six tyrosyl groups (36-40). Five of these tyrosines are located in two domains, one of which is the putative regulatory region involving Tyr1146, Tyr115', and Tyr'151 (36-40)' whose phosphorylation may lead to full kinase activation (36, 37, 40). The second domain includes the C terminus containing Tyr1316 and Tyr1322. Although phosphorylation in this region is not kinetically correlated with activation of substrate phosphorylation, a possible prerequisite role for this process cannot be excluded.
In the present work, we have used in vivo 32P-labeling and immunoblotting techniques to evaluate the phosphotyrosine (PY)z content of the insulin receptor during its cellular itinerary. Surprisingly, internalization was associated with rapid dephosphorylation of @-subunit tyrosine residues which paradoxically was associated with an augmentation of endosomal receptor kinase activity.
Materials-Porcine insulin was generously supplied by the Connaught-Novo Laboratories, Willowdale, Ontario. Carrier-free [1251]iodine and [y3'P]ATP (1000-3000 Ci/mmol) were purchased from New England Nuclear Radiochemicals (Lachine, Quebec). Wheat germ agglutinin-Sepharose 6MB and Protein A-Sepharose were from Pharmacia LKB Biotechnology Inc. Adenosine 5'-triphosphate, disodium salt, was from Boehringer Mannheim. Chemicals for electrophoresis were from Bio-Rad Laboratories. Kodak X-Omat, AR films were purchased from Picker International Canada (Montreal, Quebec). Most other chemicals were supplied by Sigma. A peptide (peptide 960) identical in sequence with a portion of the juxtamembrane region of the insulin receptor P-subunit encompassing residues 942 to 969 (KRQPDGPLGPLYASSNPEYLSASDVFPC) was synthesized with a cysteine at the amino-terminal end. The amino-terminal cysteine permitted cyclization to facilitate antibody preparation against the midportion of the loop. The peptide Thr"-Lys (TRDIYETDYYRK) encompassing residues 1142 to 1152 of the 8-subunit was purchased from Dr. David Coy (Tulane University Medical School, New Orleans).
Antibody Preparations-Antibodies to peptide 960 and to phosphotyrosine (PY) were raised in New Zealand White rabbits. Phosphotyrosine (18 mg) was coupled to keyhole limpet hemocyanin (90 mg) with 288 mg of l-ethyl-3-(3-dimethylaminopropyl)carbodiimide and dialyzed overnight at room temperature against three changes of 50 volumes of PBS. The solution was concentrated and diluted with normal saline to a concentration of 4.5-5.0 mg of conjugate/ml. Complete Freund's adjuvant (Difco) supplemented with 12.5 mg/ml heat-killed Mycobacterium butyricum (Cuti-BCG, Institut Armand Frappier, Laval, Quebec) was mixed with an equal volume of conjugate solution and thoroughly emulsified before injecting a total of 1.7 ml a t multiple (50-100) intradermal sites in the dorsolateral and cervical regions of each animal. Animals were boosted with one-third the amount of conjugate in incomplete Freund's adjuvant 6 weeks after the primary injection and at 3-week intervals thereafter; they were bled 2 weeks after each booster injection. PY antibody (aPY) titers were estimated as the dilution immunoprecipitating 50% of 32Pautophosphorylated insulin receptor generated as previously described (32). Affinity-purified aPY was prepared by adsorption to and elution from a PY-Affi-Gel column as follows. Approximately 25 ml of antiserum was adjusted to a final concentration of 50 mM NaF, Residue number is from the inferred proreceptor sequencL deduced from the cDNA sequence of Ullrich et al. (10).
The abbreviations used are: PY, phosphotyrosine; 01960, antibody to peptide 960; aPY, antibody to PY; PMSF, phenylmethylsulfonyl fluoride; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; BSA, bovine serum albumin; PBS, phosphate-buffered saline; Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; WGA, wheat germ agglutinin; PM, plasma membrane; EN, endosome. 100 pM vanadate and incubated at 4 "C for 30 min. The mixture was centrifuged at 100,000 X g for 1 h, and the supernatant was recycled three times over a PY-Affi-Gel 15 column pre-equilibrated with 0.1 M NaCl, 10 mM potassium phosphate, pH 7.4. The column was washed sequentially with 50 ml of 0.2 M KC], 10 mM potassium phosphate, pH 7.4,250 ml of 10 mM potassium phosphate, pH 7.4, and finally 50 ml of 50 mM Hepes, pH 7.4. Affinity-purified antibodies were eluted with 50 ml of 0.2 M p-nitrophenyl phosphate. The eluate was dialyzed extensively against 0.15 M NaCl and then concentrated in a Centriprep 30 unit (Amicon). The concentrate was brought to 50% glycerol and stored at -20 "C. A comparable procedure was employed to prepare peptide 960 antibodies (01960) except that peptide 960 (5.5 mg) was conjugated to BSA (8.0 mg) with glutaraldehyde as described by Benoit et al. (41). Antibody titer was estimated by assessing binding of lZ5I-peptide 960 by different antiserum dilutions. Peptide 960 was iodinated using NalZ5I and chloramine-?' as previously described (42). The peptide was purified from the reaction by adsorption to an activated Sep-Pak C18 cartridge (Waters Associates, Milford, MA) followed by a step gradient elution with 0-8075 methanol.
Subcellular Fractionation-Animals were anesthetized with ether, and insulin (0, 1.5, or 15 pg/lOO g body weight) was injected into the jugular vein in 0.4 ml of PBS containing 0.1% BSA. At the noted times after insulin injection (0-45 min), the rats were killed by decapitation, and the livers were excised rapidly and minced in icecold 0.25 M sucrose, 5 mM Tris-C1, pH 7.5,l mM MgC12, 1 mM PMSF, 1 mM benzamidine, 2 mM sodium fluoride, and 2 mM sodium orthovanadate (solution A). Combined endosome (EN) and plasma membrane (PM) fractions were prepared, as previously described (3), using sucrose solutions containing phosphatase and protease inhibitors as noted above.
I n Vivo Labeling of Insulin Receptors-Animals received 5 mCi of [32P]orthophosphate via the portal vein 1 h prior to insulin injection. Insulin (15 pg/lOO g body weight) was administered via the portal vein, and the livers were removed at varying times following the injection and homogenized immediately in ice-cold solution A. PM and EN fractions were prepared, solubilized, and subjected to immunoprecipitation (PM, 700 pg; EN, 350 pg of protein) with anti-insulin receptor antibodies, followed by SDS-PAGE as described before (3,32). The gels were exposed to x-ray film for 7 days (untreated gels) or 28 days following treatment with 1 N KOH for 2 h at 56 'C (alkalitreated gels).
Insulin Receptor Autaphosphorylation-Aliquots of intact EN (5 pg of protein) or PM (50 pg of protein) cell fractions, diluted to 50 p1 with solution A, or WGA-Sepharose eluates (20 fmol of insulin binding in 20 pl) were added to autophosphorylation buffer containing, at a final concentration, 50 mM Hepes, pH 7.4, 10 mM MnC12, 270 nM dithiothreitol, and 100 pg/ml BSA in a volume of 90 pl. Phosphorylation was initiated by the addition of 10 pl of 250 pM unlabeled ATP or [Y-~'P]ATP (28 pCi/nmol). The reaction was terminated after 15 min by immunoprecipitation with a960 (see below) or by adding 50 pl of 3x concentrated electrophoresis buffer (see below) followed immediately by boiling for 5 min prior to electrophoresis.
Electrophoresis and Immunoblotting-For analysis of 8-subunit PY content, 100 pg of either PM or EN cell fraction protein was adjusted to 100 pl with solution A and 50 pl of 3X concentrated electrophoresis sample buffer (6% SDS, 30% glycerol, 300 mM dithiothreitol, and 1.1 M Tris-HC1, pH 6.8) was added. The samples were boiled for 5 min prior to electrophoresis. Boiled samples were subjected to SDS-PAGE (4% stacking and 7.5% resolving gel) under reducing conditions as previously described (3). Electrophoretic transfer of phosphoproteins from SDS gels to nitrocellulose membranes was performed essentially as described by Burnette (43). The transfer was carried out overnight at 4 "C at 200 mA in a buffer containing 20% methanol, 50 mM glycine, 50 mM Tris-HC1, pH 8.8. The nitrocellulose paper with transferred proteins was blocked with 50 ml of PBS containing 20% fetal calf serum for 1 h at room temperature. The blocking solution was then exchanged for 50 ml of affinity-purified aPY (1:lOO dilution in PBS containing 3% BSA) and gently shaken for 2 h at room temperature. This was followed by three 10-min washes with 50 ml of PBS containing 0.1% Tween 20. The blots were then transferred to 50 ml of 1z51-goat anti-rabbit antibody (700,000 cpm/electrophoretic lane transferred) diluted in PBS containing 3% BSA for 1 h at room temperature, followed by three additional washes in PBS containing 0.1% Tween 20. The nitrocellulose papers were mounted on Whatman No. 3MM paper squares and allowed to air-dry. Labeled proteins were visualized by autoradiography at -70 "C using enhancing screens and Kodak X-Omat AR films. Labeling intensity of the @-subunit (94-kDa band) was quantitated by densitometry of the autoradiograms with a Zeineh soft laser scanning densitometer (model SL-504-XL) interfaced with an IBM-PC using a GS350 Data System (Hoefer Scientific Instruments).
Estimation of receptor content in membrane fractions was achieved by immunoblotting with antibodies against peptide 960 (a960). Immunoblots with a960 were obtained in a manner similar to that described for those with aPY but with the following changes: electrophoretic separations were performed on 60 pg of PM and 20 pg of EN cell fraction protein and blocking of the nitrocellulose membranes, dilution of a960 (l:lOO), and suspension of the '"1-goat anti-rabbit antibody were done in PBS containing 2-3% powdered milk. Scanning of these autoradiograms as well as those with aPY were done at immunoblot intensities that fell within the linear range of the signal intensity standard curve (Fig. 2). Thus, signals generated from aPY immunoblots of freshly prepared PM (100 pg of protein) and ENS (100 pg of protein) were substantial and within the linear range after 48 h of x-ray exposure. aPY immunoblots of autophosphorylated fractions (PM, 50 pg; EN, 5 pg of protein) were satisfactory after 22-24 h of x-ray exposure. a960 immunoblots (PM, 60 pg; EN, 20 pg of protein) generated appropriate signals after 24 h of x-ray exposure.
Immunoprecipitation of Insulin Receptors-Following autophosphorylation with [y3'P]ATP, the reaction was terminated by adding 3 X concentrated stopping buffer (150 mM Hepes, pH 7.4,0.1% Triton X-100,150 mM EDTA, 30 mM sodium pyrophosphate, 6 mM vanadate, 120 mM NaF, 1 mM PMSF, 1 mM benzamidine, 20 p M leupeptin, 20 p M pepstatin A, 1 trypsin inhibitory unit aprotinin/ml, and 0.3% BSA) to give a final volume of 150 pl. Receptor immunoprecipitation was achieved by incubation (4 h at 4 "C) with 10 p1 of protein Apurified a960 followed by a 1-h incubation with protein A-Sepharose. The immunoprecipitate was washed twice with 50 mM Hepes, pH 7.4, containing 0.1% Triton X-100 and 0.1% SDS, followed by a final washing with the same solution but omitting SDS. SDS-polyacrylamide electrophoresis of the immunoprecipitates and quantitation of @-subunit phosphorylation were performed as described previously (3).
In some studies (Table 111), lectin-purified receptors were immunoprecipitated with aPY or a960 before and after autophosphorylation. WGA-Sepharose eluates (500 pl containing 100 fmol of insulin binding activity) were added to autophosphorylation buffer (see above) in a final volume of 900 pl. To this was added 100 pl of 50 mM Hepes, pH 7.4, with or without 250 p~ ATP. Immediately (-ATP) or after a 15-min incubation at 4 "C (+ATP), 500 pl of 3X concentrated stopping buffer was added followed by 50 pl of antibody (control rabbit IgG or aPY or a960). Samples were shaken for 4 h at 4 "C, and protein A-Sepharose (250 pl, 50% v/v) was added followed by continued shaking for 1 h at 4 "C. Immunoprecipitates and clear supernatants were collected by centrifugation.
After 10 min at room temperature, the reaction was terminated by spotting 50-pl aliquots on Whatman No. 3MM paper disks with immersion in ice-cold 10% trichloroacetic acid, 10 mM sodium pyrophosphate. The disks were washed as previously described (3), and radioactivity was determined using Universol scintillation solution (ICN Biomedicals).
Hormone Binding and Protein Determimtion-1251-Insulin prepared using chloramine-T as previously described (42) had a specific activity of 130-200 pCi/pg. Specific binding of 1251-insulin to lectinpurified receptor or subcellular fractions, pretreated with 0.1% Triton X-100, was determined after overnight incubation at 4 "C, as previously described (32). In subcellular fractions isolated from rats preinjected with insulin, the binding of 1251-insulin was corrected for the concentration of unlabeled insulin present in the fraction as previously described by Khan et al. (3). Protein content in the fractions was determined by a modification of Bradford's method, using bovine serum albumin as a standard (44).

32P-Labeling of the Insulin Receptor 8-Subunit in PM and
ENS-Initial attempts to determine the extent of @-subunit tyrosine phosphorylation in PM and EN receptor populations involved the injection of 32Pi into rats to label radioactively the intracellular ATP pool. Employing the method of England and Walsh (45), we determined that from 30-75 min following the injection of 5 mCi of 32Pi the specific radioactivity of [32P]ATP in rat liver parenchyma was constant at 0.74 Ci/ mol (data not shown). Consequently, 60 min after 32Pi injection, insulin was administered (15 pg/lOO g body weight), and hepatic PM and EN fractions were prepared following death of the rats at 0, 30 s, and 5 min postinjection of insulin. Following solubilization of the cell fractions, insulin receptors were immunoprecipitated with insulin receptor antibody and subjected to SDS-PAGE with or without alkali treatment prior to radioautography. As seen in Fig. 1, 32P-labeling of the insulin receptor 8-subunit of PM was observed even under basal conditions (zero time) with an increase noted after the injection of ligand; much lower labeling was found in ENS (lanes 4-6). After alkali treatment of the gels, labeling of the 8-subunit in PM fractions was absent at zero time, abruptly augmented at 30 s, and diminished at 5 min. Maximal labeling in EN fractions was found at 5 min, but the intensity was much lower than that found at 30 s in PM. Receptor content, evaluated by Scatchard analysis, was rapidly elevated in ENS as a consequence of ligand injection (Table I). However, internalization coincided with a diminished 32P-labeling of receptor and especially alkali-resistant 32P-labeling. Analysis of receptor PY content by this method was cumbersome due to the large amount of 32Pi being injected per rat, the low level of radioactivity incorporated per 8-subunit, and the long time required to generate radioautographic signals. Furthermore, Scatchard analysis of receptor content required corrective with 5 mCi of 32Pi in 0.9% saline. Insulin (15 pg/lOO g body weight) was injected 60 min later, and the animals were killed at 30 s and 5 min postinjection. Control rats (zero time) were killed 60 min after 32Pi injection. Plasmalemma (PM) and endosomes (EN) were isolated, solubilized, immunoprecipitated with insulin receptor antibody, and subjected to SDS-PAGE and radioautography as described under "Experimental Procedures." '251-Insulin binding was determined in aliquots of Triton X-100-solubilized PM and EN. The gels were exposed to x-ray film for 7 days (untreated gels) or 28 days (alkalitreated gels).

TABLE I Effect of insulin administration on 32P-labeling of the insulin receptor in PM and EN
Rats received 5 mCi of [82P]orthophosphate via the portal vein, and insulin (15 pg/100 g body weight) was injected 1 h later by portal vein. At the noted times, rats were killed and hepatic PM and EN fractions were prepared, solubilized, assayed for '2sI-insulin binding, and immunoprecipitated with insulin receptor antibody (PM, 700 pg; EN, 350 pg), followed by SDS-PAGE and autoradiography as described under "Experimental Procedures." The gels were exposed to x-ray film for 7 days or for 28 days following treatment with 1 N KOH for 2 h at 56 "C. The radiolabel in each band was measured by densitometer. Receptor number was determined in each cell fraction by Scatchard analysis of '251-insulin binding data corrected for the concentration of unlabeled insulin present in the fraction as described previously (3). - Obtained by dividing data in column 3 by that in column 1.
factors for ligand content in the membrane fractions (3). Hence, we attempted to evaluate these observations by a more rapid and direct method.

Qwlntitation of Insulin Receptor P-Subunit and Its P Y Content in Cell Fractions by Western Blotting Analysis-To
determine the receptor content of PM and EN cell fractions, we employed the method of immunoblotting using antibody against peptide 960 (a960). This antibody recognizes virtually all solubilized insulin receptors in PM and EN as evidenced by its efficiency in immunoprecipitating insulin receptors from these fractions (see below). An estimate of P-subunit PY content was derived following Western blotting with affinitypurified aPY.
To determine the linear range of immunoblot signal intensity, increasing quantities of hepatic ENS from insulin-injected rats were autophosphorylated in vitro3 and subjected to SDS-PAGE followed by immunoblotting with antibodies to peptide 960 (a960) and PY (aPY). Fig. 2 (top panels) demonstrates autoradiograms of these immunoblots. Densitometric scanning of these autoradiograms permitted graphic presentation of these data (Fig. 2, bottompanels) and showed that the maximum immunoblotting signals generated from 20 pg of EN and 60 pg of PM were within the linear range of signal intensity for a960 Western blots; 100 pg of EN or PM protein were within the linear range of signal intensity for aPY Western blots.
Estimation of relative receptor content using an '251-insulin binding assay was in close agreement with determinations by immunoblotting with antibody to peptide 960 when 20 pg of EN and 60 pg of PM protein were used (Table 11).
Effect of in Vivo Insulin on Receptor Content of PM and ENS-We conducted these studies with doses of injected insulin which resulted in 50% hepatic insulin receptor occupancy (1.5 pg/100 g body weight) and near-saturation of receptors (15 pg/lOO g body weight) (17). Rats were killed at different times after insulin administration. PM and EN cell fractions were prepared and subjected to SDS-PAGE and immunoblotting with a960 (Fig. 3, right). The time-dependent changes of PM and EN receptor content relative to control fractions were determined by densitometric scanning of the autoradiograms (Fig. 3, left). Following insulin administra-3ENs from such animals contain high concentrations of insulin receptor per mg of protein (Fig. 3) and display much higher autophosphorylation activities than corresponding PM fractions ( Ref. 3;  Fig. 6). In order to generate a range of immunoblot intensities, insulin receptors heavily phosphorylated on tyrosine residues were prepared. Thus, ENS were isolated from rat livers 2 min after insulin injection (1.5 pg/lOO g body weight), autophosphorylated with unlabeled ATP (2 mg of EN protein/ml), subjected to SDS-PAGE over a range of volumes (lanes 1-7; 2.5-30 pl), immunoblotted with aPY or a960, and subjected to autoradiography and densitometric scanning as described under "Experimental Procedures." Typical immunoblot data points for PM and ENS from rat liver are noted. For immunoblots with a960 (left panel), the PM sample (60 pg of protein) was from uninjected rats and the EN sample (20 pg of protein) from rats 5 min postinjection of 15 pg of insulin/100 g body weight. For immunoblots with aPY (right panel), the PM sample (100 pg of protein) was from rats 30 s postinjection of 1.5 pg of insulin/100 g body weight, and the EN sample (100 pg of protein) from rats 2 min postinjection of 1.5 pg of insulin/100 g body weight.
tion, there was a rapid and marked increase in EN receptor concentration which peaked at 2-5 min postinjection at about 13-fold over control for both doses of 1.5 and 15 pg of insulin/ 100 g body weight. At later times, EN receptor concentration returned to control levels for the 1.5-pg dose but remained at 8-fold above control at 45 min postinjection for the 15-pg dose.
There were corresponding rapid decreases in PM receptor concentration with an initial maximal decline by 2-5 min to 80% and 67% of control following the 1.5 and 15 pg/lOO g body weight doses of insulin, respectively. At 45 min postinjection, the PM receptor concentration returned to control values for the 1.5-pg dose but remained markedly decreased at 54% of control for the 15-pg dose of insulin.
Effect of in Vivo Insulin on PM and EN @-Subunit P Y Concentration-The studies of 32P-labeled insulin receptors following 32Pi administration suggested that internalization Effect of in vivo insulin on relative receptor content in PM and EN measured by immurwblotting and '25Z-insulin binding Rats were injected with insulin (15 pg/lOO g body weight), and hepatic PM and EN fractions were prepared and solubilized. An aliquot was subjected to SDS-PAGE and immunoblotting with a960, and a second aliquot was submitted to 1251-insulin binding assays and Scatchard analysis as described under "Experimental Procedures" and elsewhere (3). The data of each experiment were expressed as a percent of these measurements on PM and EN from uninjected (0 min) rats. The depicted data are the mean f S.D. of three separate experiments. of receptors to EN was associated with a significant decrease in PY concentration. More detailed studies of this phenomenon were carried out by immunoblotting with aPY as depicted in a typical study (Fig. 4). It can be seen that, at both doses of administered insulin, the maximum PY content per 100 pg of cell fraction protein in the 94-kDa band was at 30 s and 2 min postinjection for PM and ENS, respectively. In PM fractions, a tyrosine-phosphorylated band was often observed at 90 kDa. This band however did not react with a960 (see Fig. 3) and was not included when densitometrically scanning the 94-kDa band.

Postinjection
Densitometric analyses of autoradiograms of immunoblots with a960 and aPY were performed on the same cell fractions. The ratio of the density with aPY and that with a960 (i.e. (zPY/a960) reflects the average PY content of the insulin receptor population in a given cell fraction. Fig. 5 depicts the  4. Immunoblots with aPY of insulin receptor @-subunit at different times following injected insulin. Rats were injected with insulin (1.5 or 15 pg/100 g body weight) and killed at the noted times. PM and EN fractions were prepared, and aliquots of 100 pg of protein were subjected to SDS-PAGE followed by immunoblotting with aPY as described under "Experimental Procedures." Autoradiograms for these aPY immunoblots were routinely developed after 45-to 50-h exposure time. A typical experiment at each dose of injected insulin is shown. In the autoradiograms of PM fractions, only the 94-kDa band density was measured; that at 90 kDa being excluded from the determination of 0-subunit PY content.

Change in Insulin Receptor PY Content Following in Vitro Autophosphorylation-We then attempted to
correlate the PY content per receptor with receptor autophosphorylation activity. Fig. 6 demonstrates the time-dependent changes in aPY/ a960 ratios in PM and ENS at both doses of insulin following in vitro autophosphorylation in the absence of added insulin. Peak autophosphorylation activity was observed at 30 s postinsulin injection for PM and at 2 min for ENS. Comparison of autophosphorylation activity (aPY/a960 ratio after autophosphorylation ( Fig. 6) uersus basal aPY/a960 ratio (Fig.  5)) at peak times demonstrated that EN receptors were 4.0and 4.2-fold more active than PM receptors at 1.5 and 15 pg of insulin/100 g body weight, respectively. This is in close agreement with our previous studies in which 32P-labeling was measured (3). Of interest was the inverse relationship between the amounts of PY incorporated per @-subunit in uiuo and those obtained after autophosphorylation in ~i t r o .~ PM fractions exhibited higher PY/@-subunit ratios than ENS following in uiuo insulin treatment but were less active in the autophosphorylation assay than corresponding ENS.
It is also evident that maximum EN insulin receptor autophosphorylation occurred at 2 min postinjection (Fig. 6) when the in uiuo PY content per receptor had decreased from maximum levels (Fig. 5).

Effect of Insulin Dose on Exogenous Kinase Actiuity of kctin-purified PM and EN Receptors-
The exogenous substrate kinase activities of WGA-purified insulin receptors from control and insulin-treated rats were evaluated as a function of insulin dose. PM and EN cell fractions prepared 30 s and 2 min, respectively, following insulin injection were solubilized and purified by WGA affinity chromatography in the presence of protease and phosphatase inhibitors as noted under "Experimental Procedures." The exogenous substrate assay was designed to eliminate insulin-dependent autophosphorylation/activation of the receptor tyrosine kinase by removal of receptor-associated insulin during lectin chromatography (pH 6.0 wash) and by the use of a high concentration of poly(G1u:Tyr) (5 mg/ml) in the assay as noted previously (46). As seen in Fig. 7, the kinase activities of PM and EN receptors, prepared after a dose of 1.5 pg of insulin/100 g body weight, were comparable. At the 15 pg of insulin dose, a marginal decrease occurred in PM tyrosine kinase activity (7.8 rt 0.4 pmol/lO min/5 fmol of insulin binding) even as the relative PY content/@-subunit (i.e. aPYla960) decreased significantly (Fig. 5)  pmol/lO min/5 fmol of insulin binding) although the PY content/@-subunit (Fig. 5) showed a significant decrease.
Immunoprecipitation of in Vivo-activated Receptors with aPY and a960"The above observations indicate that, following insulin treatment, PM receptors contain more PY per receptor than do EN receptors. The latter, however, manifest greater autophosphorylation activity and equal or greater kinase activity against exogenous substrate. This discrepant relationship between @-subunit PY content and receptor kinase activity may reflect a different pattern of tyrosine residue phosphorylation in PM compared to EN receptors. Thus, tyrosine phosphorylation in EN receptors might be limited to those residues involved in kinase activation (i.e. Tyr1146, T~rl'~', and/or Tyr"") (36), whereas PM receptors might be phosphorylated at these residues but also at other sites such as T Y P O , Tyr1316, Tyr13", etc. As a consequence, EN receptor preparations, although reduced in PY content, might contain a larger fraction of activated receptors than those in PM. To test this hypothesis, we attempted to determine the percentage of activated receptor in lectin-purified extracts of cell fractions (PM, 30 s; EN, 2 min) following injection of 1.5 pg of insulin/100 g body weight.
In the first instance, we determined the efficacy with which a960 immunoprecipitated 1251-insulin binding and exogenous kinase activities from PM and EN cell fractions prepared from insulin-treated rats (Table 111). For receptors not subjected to in vitro autophosphorylation (-MnATP), 86.1% of PM and 95.4% of EN insulin binding and 73.6% of PM and 90.2% of EN kinase activity were immunoprecipitated by a960. When these solubilized lectin-purified receptors were subjected to in uitro autophosphorylation (+MnATP) prior to immunoprecipitation with a960, the extent to which insulin binding and kinase activity were immunoprecipitated did not change significantly. Thus, a marked increase in @-subunit PY content (Fig. 6) did not affect the interaction of a960 with the @-subunit. Of interest is the observation that in vitro autophosphorylation without added insulin of in uiuo-activated insulin receptors produced no (PM) or little (EN, 2.7 to 3.6 (+33%)) change in the exogenous kinase activity of the receptors. Thus, in vitro autophosphorylation largely occurred on receptors activated in vivo so that further in vitro receptor activation was absent or modest. One exdanation for the before and after autophosphorylation Rats were killed at 30 s (PM) or 2 min (EN) post-insulin (1.5 pg/100 g body weight). The cell fractions were solubilized and chromatographed on WGA-Sepharose, and the eluates split for incubation with or without ATP as described under "Experimental Procedures." These solutions were then immunoprecipitated with antisera and protein A-Sepharose as described under "Experimental Procedures," and supernatants were assayed for kinase activity (K.A.) or '251-insulin binding (B). The percentage of receptor immunoprecipitated was determined by subtracting supernatant values for aPY and a960 from those derived with control IgG. All data are the mean f S.D. of five separate experiments. derive from tyrosine kinase(s) other than the insulin receptor.
To measure the percentage of activated receptor in PM and EN cell fractions, we assessed the efficacy with which aPY immunoprecipitated insulin binding and exogenous kinase activity. For partially purified receptors not subjected to in vitro autophosphorylation (Table III; -MnATP), aPY immunoprecipitated substantially less insulin binding and kinase activity than that seen with a960. Subsequent to autophosphorylation (+MnATP), the percentage of kinase activity immunoprecipitated by aPY approximated that immunoprecipitated by a960 (PM, 70.6 uersus 74.0%; EN, 84.0 uersus 87.9%). This indicates that the increase in @subunit PY content consequent to autophosphorylation rendered the insulin receptor a better antigen for aPY and hence more readily immunoprecipitated. The fraction of insulin receptors containing PY is given by the percent of insulin binding immunoprecipitated by aPY. Following autophosphorylation, this amounted to 22.2% of PM and 18.3% of EN receptors. Thus, the augmented autophosphorylation activity of EN receptors was not due to a larger fraction of tyrosinephosphorylated receptors in ENS.
It is noteworthy that the in uitro autophosphorylation of PM receptors in the absence of added insulin increased the immunoprecipitation by aPY of both insulin binding (13.9 to 22.2%) and tyrosine kinase activity (43.9 to 70.6%) by 1.6fold. Autophosphorylation of EN receptors in the absence of added insulin increased the immunoprecipitation by aPY of insulin binding by about 1.1-fold (16.4 to 18.3%) and that of kinase activity by 2.2-fold (38.8 to 84.0%). This suggests that in ENS a substantial proportion of kinase activity was associated with a small fraction of receptors which was lightly phosphorylated in uiuo.
The validity of the immunoprecipitation studies was tested. Autophosphorylation of receptors with [Y-~'P]ATP, under identical conditions employed in the studies of Table 111, demonstrated that complete immunoprecipitation of 32Pphosphorylated receptors occurred with either aPY or a960 (data not shown). In addition, a second round of protein A-Sepharose treatment removed no additional receptor as assessed by changes in the insulin binding activity, nor did it change the specific tyrosine kinase activities. Thus, removal of antibody with protein A-Sepharose was complete, and the persistence of unbound antibody in the supernatants was not affecting these results.

Effect of Thr"-Lys on the Autophosphorylating Activity of Lectin-purified PM and EN Insulin Receptors-
The immunoprecipitation studies suggested that differences in the proportion of activated receptor in PM and EN preparations could not account for the increased tyrosine kinase activity of EN receptors. Rather, it would appear that EN receptors maintain greater intrinsic tyrosine kinase activity in a manner that is not reflected by net @-subunit tyrosine phosphorylation. To probe the extent of phosphorylation of key tyrosine residues within the regulatory region, PM and EN receptors (activated in uiuo) were examined in autophosphorylation assays in the presence of the synthetic peptide Thr"-Lys (TRDIYETDYYRK, residues 1142-1153 of the insulin receptor @-subunit). This peptide is identical with the region of the @-subunit that includes the major autophosphorylation/kinase activation sites (Tyr1l4'j, Tyr115', and Tyr1151). Concentrations of peptide exceeding 1 mM were reported to inhibit conversion of the bis-phosphorylated forms of the receptor to fully activated tris-phosphorylated forms (37). Exogenous substrate phosphorylation and autophosphorylation activities of the tris-phosphorylated form of the insulin receptor were largely unaffected by the peptide (37). Fig. 8 demonstrates the effect of Thr"-Lys on autophosphorylation activity of lectin-purified control and in uiuo-activated PM (30 s) and EN (2 min) insulin receptors. Autophosphorylation of control PM (lanes 1 and 2; upper panel) or control ENS (lanes 1 and   2; lowerpanel) in the presence of in vitro insulin was markedly inhibited in the presence of 5 mM Thr"-Lys (lane 2, both panels) during the assay. Following activation in uiuo (1.5 pg of insulin/100 g body weight), the autophosphorylation activity of PM (lanes 3-6; upper panel) or EN receptors (lanes 3-6; lower panel) was also markedly inhibited by inclusion of Thr"-Lys in the assay system (compare lanes 3 uersus 4 or 5 uersus 6 in both panels). Inclusion of insulin in uitro (lanes 3 and 4 in upper and lower panel) stimulated PM and EN receptor autophosphorylation activity (compare lanes 3 versus 5 in both panels) but did not lessen the inhibitory effects of ThP-Lys. Prephosphorylation of PM (lanes 7 and 8; upper panel) or EN (lower panel) receptors with unlabeled ATP dramatically reduced the inhibitory effects of Thr"-Lys during autophosphorylation, suggesting that tris-phosphorylation of PM and EN receptors has occurred (37). The data suggest that tris-phosphorylated forms of the insulin receptor are absent, or present in very small amounts, in both PM and prepared from rats killed at 30 s and 2 min, respectively, following insulin injection (1.5 pg/100 g body weight) and also from uninjected animals. The cell fractions were solubilized, partially purified by WGA-Sepharose chromatography, and assayed for autophosphorylation activity of PM (upperpanel) or EN (lowerpanel) in the absence (lanes 1, 3, 5, 7) or presence (lanes 2, 4, 6, 8)  EN preparations following activation in vivo. However, we cannot formally rule out the possibility that some level of posthomogenization dephosphorylation has occurred which might alter the properties of the insulin receptor kinase. This possibility seems unlikely in view of the observation that ENS containing [32P]phosphotyrosine-labeled EGF receptors were added to liver homogenates and showed no loss of 32P upon subsequent resolution of the ENS (47), and since the conditions of cell fractionation (4 "C throughout, 2 mM sodium vanadate) have been shown to inhibit markedly in situ dephosphorylation of the insulin receptor. Thus, tris-phosphorylation of the kinase regulatory domain probably does not account for the augmented kinase activity of EN receptors.

DISCUSSION
In previous work, we documented the distinctiveness of PM and EN fractions and reported that following insulin injection PM receptor kinase was maximally activated at 30 s followed by internalization of insulin-receptor complexes to ENS wherein maximal activation was at 2 min postinjection (3). In the present investigation, we have examined the level of receptor PY content in both PM and EN insulin receptors in an attempt to gain insight into our previous observation that internalized receptors demonstrated augmented kinase activity compared to PM (3).

Internalized Receptors Have Reduced PY Content-In our
initial studies, we used 32P-labeling to assess @-subunit PY content and demonstrated that, following an insulin dose of 15 pg/lOO g body weight, there was less @-subunit PY content in EN than in PM receptors. This observation was more extensively explored at doses of 1.5 pg and 15 pg of insulin/ 100 g body weight using immunoblotting to measure @-subunit PY content. An antibody to the submembrane domain of the @-subunit (a960) was used to measure @-subunit (94 kDa) content, while an anti-PY antibody (aPY) was used to measure the PY content in the @-subunit. The relative PY content per @-subunit is reflected in the aPY/(u960 ratio. Such measurements confirmed in detail that PM contained more PY per @-subunit than EN receptors (Fig. 5). This difference was as high as 2-to %fold at both insulin doses. These measurements of the differences in PY content between PM and EN were less than those based on 32P labeling (Table I). We are uncertain about the basis for this discrepancy, but perhaps loss of 32P content during the immunoprecipitation step prior to SDS-PAGE may in part account for this (see "Experimental Procedures"). It is also possible that closely spaced phosphorylated tyrosine residues may be unable to bind multiple antibodies due to steric interference. Since PM receptors are more PY-phosphorylated than EN receptors, steric interference may lead to underestimation of the PY content in the PM compared to EN receptors. In any event, both methods indicate that internalization was associated with a rapid and significant reduction of insulin receptor PY content. This may reflect the selective internalization of less phosphorylated PM receptors or dephosphorylation of PM receptors during internalization. This may be related to the endosomal degradation of insulin (48) as well as an insulin receptorassociated phosphotyrosine phosphatase(s) in intact ENS which rapidly reduced the PY content of the in situ receptor.s Autophosphorylation and Exogenous Kinase Activities-In previous studies we described the persistence of activated insulin receptors following internalization and, indeed, the attainment in ENS of an autophosphorylation activity which was 3 to 5 times the maximal activity found in PM. Autophosphorylation of the endosomal receptor was shown to be key to this activated state since treatment of EN cell fractions with alkaline phosphatase abolished endosomal receptor kinase activity (3). In agreement with our previous studies, maximal autophosphorylation activity, as reflected in aPY/ a960 ratios (Fig. 6), was 3-to 4-fold greater in ENS (2-min values) than in PM (30-s values). Therefore, the earlier observations, which employed 32P labeling, did not reflect accelerated turnover of phosphate on antecedant PY residues but the net accumulation of PY residues in the P-subunit. The tyrosine kinase assay, employing an exogenous substrate (poly(Glu:Tyr)), demonstrated similar kinase activities in PM and ENS at an insulin dose of 1.5 pg/100 g body weight and a greater level of kinase activity in ENS than PM at an insulin dose of 15 pg/lOO g body weight ( p < 0.01). Comparison of Figs. 6 and 7 demonstrate a poor correlation between exogenous kinase and autophosphorylation activities. This was more evident at the 1.5 pg of insulin dose where exogenous kinase activities were similar in PM and ENS, whereas autophosphorylation was much greater in the latter. We noted this discrepancy in an earlier study (3) and discrepancies between exogenous kinase and autophosphorylation activities have been described by others (46). Although the molecular basis for this discrepancy remains uncertain, it is quite possible that autophosphorylation activity might reflect a physiologically relevant consequence of activation. Actiuated Receptor Populations in PM and ENS-The immunoblotting analyses provided the PY content of the average receptor in the cell fractions. It is possible that, despite a lower PY content/@-subunit in EN than PM receptors, the proportion of activated receptors in ENS could be higher than that observed in PM. To examine this possibility, the proportion of activated receptor in each cell fraction was determined by immunoprecipitation of receptors with aPY . Following in vitro autophosphorylation, it was possible to immunoprecipitate almost all activated insulin receptors with aPY (Table  111). Using this approach, it was found that the fraction of activated receptors in EN (18.3%) was very comparable to that in PM (22.2%). The fact that the fraction of activated receptors in EN was not greater than that in PM suggests that activated receptors were not selectively internalized. This agrees with the conclusion reached by Backer et al. (49)  found that treatment of Fao cells with 2,4-dinitrophenol severely reduced intracellular ATP levels and insulin receptor autophosphorylation without substantively impeding the internalization of insulin-receptor complexes. Similar findings have been reported in H35 hepatoma cells following the reduction of intracellular ATP pools with 2,4-dinitrophenol, azide, cyanide, or oligomycin (50).
It is interesting to note that the extent to which PY contentlo-subunit was increased by autophosphorylation (Fig. 5  uersus 6; about 24-fold in ENS and 4-to %fold in PM) suggests phosphorylation occurs on both @-subunits in each insulin receptor heterotetramer as indicated by a recent study of transphosphorylation (51). Finally, the maximum proportion of exogenous kinase activity precipitated by a960 was 90.2% in ENS but only 74% in PM. This may reflect the presence in PM cell fractions of tyrosine kinase(s) other than the insulin receptor (52). Mechanism of Augmented EN Receptor Kinase Activity- The mechanism accounting for the greater autophosphorylating activity of EN than PM receptors remains elusive. This was not due to the selective accumulation of activated receptors in ENS so that they constituted a higher proportion of the insulin receptor pool in ENS compared to PM. A more intimate state of association (i.e. oligomerization of receptors within ENS compared to PM) might facilitate receptor crossphosphorylation leading to greater PY accumulation on a higher proportion of receptors. In this case, autophosphorylation (+MnATP) of EN preparations should considerably augment the proportion of insulin receptors immunoprecipitated with aPY compared to that seen in PM. The data of Table I11 indicate that this did not occur. Another mechanism which could account for the greater autophosphorylating activity of EN than PM receptors would be a different extent of phosphorylation of tyrosine residues within the kinase regulatory domain (i.e. T Y~"~~, Tyr"", Tyr1I5l). Thus, EN receptors might be restrictively phosphorylated on all three such tyrosines (tris-phosphorylated species), whereas the less active PM receptors may be bisphosphorylated within this region as well as being phosphorylated on various other 0-subunit tyrosine residues. Studies with the peptide Thr"-Lys were designed to probe the state of phosphorylation of key tyrosines within the regulatory region of EN and PM receptors following activation in vivo. These studies indicate that the autophosphorylation activity of PM or EN receptors following activation in viuo was sensitive to inhibition by Thr"-Lys. The generation, by prephosphorylation in uitro, of tris-phosphorylated insulin receptors as documented by White et al. (37) reduced the ability of Thr"-Lys to inhibit autophosphorylation activity in both PM and ENS (Fig. 8). Our results are thus compatible with those of White et al. who found that Thr"-Lys-inhibitable forms of the receptor (ie. mono-or bisphosphorylated) predominate in Fao hepatoma cells incubated with insulin (37).
Several observations raise the possibility that a modulated P Y dephosphorylation of the 0-subunit could augment receptor kinase activity. Maximal EN receptor autophosphorylation activity occured at 2 min postinjection when the PY/Psubunit content of native receptor had declined from its maximum at 30 s postinjection (Fig. 5). There was a dosedependent increase in EN exogenous kinase activity coincident with decreases in P-subunit PY content (Figs. 5 and 7). Most strikingly, there was a small proportion of EN receptors which was lightly phosphorylated on tyrosine residues but which accounts for the majority of activated insulin receptor kinase in ENS (Table  111). Recent evidence from our laboratory5 and the work of Meyerovitch et al. (53) suggest the presence in rat liver of insulin-stimulated PY phosphatase activity. Thus, the insulin receptor @-subunit may be regulated in a manner comparable to the src family of tyrosine protein kinases where dephosphorylation of pp60""" at TyrSZ7 (54,55) or ~~5 6 ' "~ at Tyr505 (56,57) stimulated protein tyrosine kinase activity. Thus, autophosphorylation of the insulin receptor at the cell surface could involve tyrosine residues whose phosphorylation is both activating and restrictive to tyrosine kinase activity. Internalization to ENS may thus be associated with dephosphorylation of inhibitory PY residues and further activation of the intrinsic receptor kinase activity; hence, the appearance of a highly activated form of the insulin receptor. Such a phenomenon might explain why a mutant insulin receptor lacking the 43 COOH-terminal amino acids is more active in promoting mitogenesis in Rat 1 fibroblasts than the fully intact insulin receptor (58). Further dephosphorylation, mediated by endosomal and/or soluble PY phosphatase(s), would lead to inactivation of kinase activity prior to recycling to the cell surface. Thus, PY dephosphorylation may play a dual role in regulating the insulin receptor kinase.
Although dephosphorylation of internalized receptors may have a regulatory role regarding kinase activity, the situation in PM seems less clear. For example, at the 15 pg/lOO g body weight dose of injected insulin, there was a greater net decrease in @-subunit PY content of PM than ENS with no corresponding increase in PM exogenous kinase activity (Figs. 5 and 7). It may be that PM receptors are rendered insensitive to the regulatory impact of modulated dephosphorylation, perhaps via association with membrane-specific modulator protein(s). In this regard, it is noteworthy that immunoblotting with aPY revealed the presence of a 90-kDa phosphoprotein in PM, but not in ENS, whose appearance in the former was insulin-dependent (Fig. 4).
Internalization and Cell Signaling-Studies on mutant receptors (25)(26)(27)(28)(29)(30)(31) and the impact of kinase inhibitory antibodies (23,24), as well as the demonstration that peroxovanadium, a potent insulin receptor kinase activator, mimics insulin action (21,22), highlight the key role of the receptor kinase in realizing insulin action. Our work to date has indicated that an activated receptor kinase is internalized and concentrated within ENS (3, 32). We have suggested that internalization to the endosomal system could be a process whereby activated receptor kinase is brought into contact with intracellular substrates in cellular domains other than those adjacent to the cell surface (3). Recently, several groups observed that deletion of the submembrane domain of the insulin receptor @-subunit (encompassing residues 954-965 including the putative consensus sequence NPEY) significantly reduced insulin-stimulated internalization (59,60).
In the studies of Thies et al. (60), the insulin receptor mutant lacked exon 16 (IR Aex16) which encodes a 22-amino acid sequence encompassing residues 944-965. Although this receptor mediated the stimulation of glycogen synthetase and mitogenesis in transfected cells, the sensitivity of these responses, especially the mitogenic response, was considerably reduced (61). As initially suggested, a particular insulin response need not absolutely require receptor kinase internalization (17, 32), but may be rendered more sensitive to insulin as a consequence of the endocytic process. The recent observations by McClain on the IR Aex16 mutant (61) may reflect a requirement for receptor kinase endocytosis for at least some actions of insulin to occur within the physiologic range of hormone concentration.
In summary, these studies have shown that, at both a subsaturating and saturating dose of insulin, the internalized insulin receptor was dephosphorylated on tyrosine residues