THEJOURNAL OF BIOLOGICAL CHEMISTRY
Vol. 262, No. :30, Issue of October 25. pp. 14663-14671,1987 Printed in U.S.A .
A Mutant Insulin Receptor with Defective Tyrosine Kinase Displays No Biologic Activity andDoes Not Undergo Endocytosis* (Received for publication, March 24, 1987)
Donald A. McClainSjll, Hiroshi MaegawaSj,James Lee[[, Thomas J. Dull[[,Axel Ulrichll , and JerroldM. OlefskySj From the §Department of Medicine, University of California, San Diego, La Jolla, California 92093, the $Veterans Administration Medical Center, Medical Research Seruice, San Diego, California 92161, and (IGenentech, South SanFrancisco, California 94080
3) This loss of kinase The cDNAs encoding the normal human insulin re- any tyrosine kinase activity. activity wasaccompanied by a nearly complete lack of ceptor (HIRc) and a receptor that had lysine residue 1018 replacedbyalanine(A/K1018)were used to both endocytosis and biologic activity. transfect Rat 1 fibroblasts. Lysine 1018 is a critical residue in the ATP binding site of the tyrosine kinase domain in the
[email protected] Rat 1 The insulin receptor is aglycoprotein consisting of two cells express 1700 endogenous insulin receptors. Expressed HIRc receptors had levels of insulin-stimulable extracellular a-subunits (135 kDa) and two membrane-spanning @-subunits(95 kDa). The cytoplasmic portion of the pautophosphorylation in vitro comparabletonormal of other receptors, whereas A/KlO18 receptors ~ had1 of% that subunit has homology to the tyrosine kinase domain After activity. Stimulation by insulin of HIRc receptors in growth factor receptors and transforming proteins (1,2). situ in intact cells led to phosphorylation of @-subunit insulin binds to its receptor, thereceptor’s tyrosine kinase is tyrosine residues and activationof tyrosine kinase ac- activated and the receptor is autophosphorylated (3, 4). This tivity that could be preserved and assayed in vitro tyrosine kinase activity is thought to be central to insulin after receptor purification. In contrast, A/K1018 re- action. The cloning of the insulin receptor cDNA (1, 2) has ceptors showed no such activation, either of autophos- made it possible to test this hypothesis using in vitro, sitephorylation or of kinase activity towardhistone. Cells directed mutagenesis. Twomutatedinsulinreceptorsexexpressing HIRc receptors display enhanced sensitiv- pressed in mammalian cells, one with replacement of two key ity to insulin of 2-deoxyglucose transport andglycogen tyrosine residues (5) and one with replacement of a lysine in synthase activity. This increased sensitivity was pro- the ATP binding site of the p-subunit(6,7), have been shown portional to insulin receptor numberat low but not at to bedefective both in tyrosine kinase activity and in mediathigh levels of receptor expression. A/K1018 receptors were unable to mediate these biologic effects and ac- ing the biologic responses to insulin. Inadditionto possessing kinaseactivityandinitiating tually inhibited insulin’s ability to stimulate glucose transport and glycogen synthase through the endoge- insulin’s action, the insulin receptoralso participates in a complex intracellular itinerary. After insulin binding, occunous Rat 1receptors. pied receptors are internalized both via coated pits and by Expressed HIRc receptors mediated insulin internalcoated pit-independentprocesses (8,9). Internalized receptorization and degradation, whereas A/K1018 receptors mediated little, if any. Endocytotic uptake of the ex- ligand complexes dissociate in endosomesprobably due to pressed A/K1018 insulin receptors was also markedly acidification of the vesicle (10). The receptorslargely recycle cell surface (11); whereas most of the internalized depressed compared to normal receptors. Unlike HIRc intact to the receptors,A/K1018receptors also failtoundergo insulin is degraded,a portionis released intact, aprocess down-regulation after long (24 h) exposures to high termed retroendocytosis (12). The receptor itinerary is still (170 nM) concentrations of insulin. incompletely understood thespecific biochemical signalsthat We conclude the following. 1)Normal human insulin control the intracellular routing of receptors are not known receptors expressed in Rat1fibroblasts display active nor is thebiologic function of the itineraryclear. Endocytosis tyrosine-specific kinase, normal intracellular itinermay functionprimarilytomediatehormonedegradation; ary after endocytosis, and normal coupling to insulin’s alternatively, the presenceof insulin and receptorsinside the biologic effects. 2) A receptor mutated toalter the ATP cell raises the possibility that internalization of receptors or binding site in the tyrosine kinase domain had little if hormone may be necessary for some aspects of insulin action (13). * This work was supported by a Career Development Award from Based on analysesof other tyrosine kinases andnucleotide the Veterans Administration (to D.A. M.) and by grants from the Life and Health Insurance Medical Research Fund (to D. A. M.), the binding proteins (14), lysine 1018 of the insulin receptor pUpJohn Company (to J. M. O . ) ,the Medical Research Service of the subunit represents a key component of the ATP binding site. Veterans Administration Hospital, San Diego, and by National In- Therefore, by means of site-directed mutagenesis we have stitutes of Health Grants DK 33649 and DK 33651. The costs of replaced lysine 1018 with alanine andexpressed this receptor publication of this article were defrayed in part by the payment of (A/K1018) in Rat 1 fibroblasts. These transfected fibroblasts page charges. This article must therefore be hereby marked “aduer- were then used to investigate the controlof receptor endocytisement” in accordance with 18U.S.C. Section 1734 solelyto indicate tosis and the interrelationships among the tyrosine kinase, this fact. ll To whom correspondence should be addressed Veterans Admin- biologic activity, and intracellular itinerary of insulin recepistration Medical Center, Medical Research Service (V-lllG), 3350 tors. We have found that the kinase-defective A/K1018 inLa Jolla Village Dr., San Diego, CA 92161. sulin receptor does not mediate insulin’s biologic actions and
14663
14664
A Kinase-defective Mutant Receptor Insulin
has a markedly reduced ability to enter the endocytotic path- stepwise using methotrexate at various concentrations. Methotrexate-resistant cell lines expressinghigh levels of human insulin recepway.
tors were screened by binding of radiolabeled insulin in the presence and absence of unlabeled insulin. Rat 1 cells were grown in F12/ Dulbecco's modified Eagle's medium (50:50) supplemented with 10% Materials-Porcine insulin and '251-insulin, monoiodinated at the fetal bovine serum. When placed under methotrexate selection, cells alanine 14 position (300-400 pCi/pg), were kindly provided by Eli were grown in the same media containing 7% extensively dialyzed Lilly, Inc. Cell culture reagents and dialyzed fetal calf serum were (against PBS) fetal bovine serum. Cultures were maintained in a purchased fromGIBCO, and cells were cultured in dishes from Costar humidified atmosphere of 95% air and 5% CO, at a temperature of (Cambridge, MA). Wheat germ agglutinin (WGA)' coupled to agarose 37 "C. was from VectorLabs, Inc. (Burlingame, CA). Protein concentrations Insulin BindingAssay-Insulin binding to whole cells was quantiwere determined using Bio-Rad protein assay dye. [32P]ATPlabeled tated using A-14 monoiodinated 'Z51-insulin as described (12). Cells in the y position (-5000 Ci/mmol, 10 mCi/ml) was from Amersham in 35 mm dishes were exposed to 0.03 nM '251-insulin and various COT. [3H]2-deoxyglucose (30.2 Ci/mmol) and uridine disphospho-D- concentrations of unlabeled insulin for 2.5 h a t 16 "C, and specific [U-'4C]glucose (322 mCi/mmol) were from Du Pont-New England binding was determined by subtracting the amount of '2sII-insulin Nuclear. Electrophoresis reagents were from Bio-Rad. All other rou- bound in the presence of excess (330 nM) unlabeled insulin. Under tine reagents were purchased from Sigma. A monoclonal antibody these conditions, bound insulin is >98% acid-extractable and thus specificfor humaninsulinreceptors waskindlyprovidedby Dr. still at the cell surface. Insulin binding to solubilized receptors was Steven Jacobs (Wellcome Research Laboratories, Research Triangle measured using polyethylene glycol precipitation as described (22). Park, NC). A polyclonal human anti-insulin receptor antibody with Insulin Receptor Purification-Insulin receptors were purified from cross-species specificity was provided by Dr. Larry Mandarin0 (Uni- 1-5 X lo7cells grown in 150-mm culture dishes. Cells were solubilized versity of California, San Diego). A monoclonal anti-receptor anti- in 5 ml of 1%Triton X-100 with bacitracin(0.5 mg/ml) and phenylbody with insulin-agonist properties (15)was kindly provided by Dr. methylsulfonyl fluoride (0.2 mM) for 30 min at 0 "C. The solution Kenneth Siddle (Universityof Cambridge, Cambridge, United King- was centrifuged at 80,000 X g for 15 min at4 "C and the supernatant dom). diluted to50 ml in elutingbuffer, containing 0.5 % Triton X-100,120 ExpressionPlasmids-Expression plasmids (pCVSVEHBVE400, mM NaCl, 5mM KCI, 0.8 mM MgSO,, 1mM CaClZ, 25 mM HEPES, European Publication Number 117060) contained the origin of rep- pH 7.4, with 10% glycerol. This was applied slowly over a period of lication and the ampicillin resistance gene of the Escherichia coli 1-2 h a t 4 "Cto 0.5 ml of WGA-agarose, washed with 50ml of eluting plasmid pBR322 (16), DNA encodingmouse dihydrofolate reductase buffer, 10% glycerol, and eluted with 0.3 M N-acetyl glucosamine in (17) under the controlof the SV40 early promoter, and a 2279-5231 eluting buffer-10% glycerol. 250-pl fractions were collected and probase pair XbaI-DraI fragment containing the complete or modified tein assayed on 10-pl samples in 200 p l of protein assay dye (BioHIR coding sequence (1) under SV40 early promoter control. This Rad). The single protein peak was pooled and assayed for insulin fragment was inserted into the expression vectorwhich had been binding activity. digested first with BamHI.The BamHI expression vector sticky ends Insulin Receptor Tyrosine Kinase Actiuity-To study receptor autowere filled in with Klenow Pol I and subsequently the plasmid was phosphorylation 200-800 fmol of WGA-agarosepurified receptors digested with XbaI. Thus, insertion of the XbaI-DraI was only pos- were assayed as previously described (23). Receptors were also stimsible in the orientation necessary for expression of the HIR mRNA. ulated and preactivated while still in intact cells as previously deThe resultinginsulinreceptorexpressionplasmid was designated scribed (24). To do this, 10-cm dishesof confluent cells were treated pCVSVE-HIRc. All constructs share the pre-insulin receptor initia- with insulin (170 nM in MEM, 1% BSA, pH 7.4) for 30 min a t 37 "C. tion codon, as well as a50-nucleotide-longsequence from the 5'The cells were rinsed in ice-cold PBS and solubilized in 1 ml/dish of untranslated region and a 10-base pair polylinker region from the eluting buffer, 10% glycerol, 0.4% Triton X-100 containing protease cDNA cloning linker, and pUC12 joined to an SV40 early promoter and phosphatase inhibitors: 0.5 ng/ml bacitracin, 0.2 mM phenylfragment at an XbaI site. This XbaI site had been generated down- methylsulfonyl fluoride, 1000kIU/ml Trasylol (aprotinin,FBA Pharstream of the SVE by replacing the SV40 HindIII site a t position maceuticals, New York), 50 mM NaF, 2 mM EDTA, 2 mM NaaV04, 5171 with a synthetic oligonucleotide containing a HindIII, ClaI, and 10 mM sodium pyrophosphate, 15 mM benzamidine, and 2 mM dichloroacetic acid. Dishes (typically five 10-cm dishes/condition) were XbaI site (5'-3') directly joined to each other. HIR 3"untranslated scraped with a rubber policeman, and the cells collected into ultrasequences continue 1018 nucleotides downstream from the stop codon ending a t a DraI restriction site, where they were combined at an centrifuge tubes. After 30 min of solubilization with occasional vorend-filled BamHI restriction site with 3"untranslated sequences from texing, the solutions were centrifuged (80,000 X g, 15 min, 4 "C). the hepatitis virus surface antigen gene. Plasmids were amplified in Receptors were purified on WGA-agarose as described above except that all bufferscontained 2 mM EDTA,50 mM NaF, and 2 mM E. coli and purified (18). Na3V04.The purified receptors were assayed for 1251-insulinbinding Plasmidfortheexpression of modified insulinreceptorswas generated by oligonucleotide-directedinvitromutagenesis. A as described above. 200 fmol of receptor in 100 p1 of column eluting buffer were then immunoprecipitated, always a t 0 "C, using antiBglII-Hind111 fragment from CVSVHIRc was ligated to M13mp19 andmutagenesiscarriedoutusingtheprimer 5'insulin receptor antibodies and protein A as described above. The CCGCGTGGCGGTGGCCACGGTCAACGAGT-3' (19, 20). This washed protein A pellets were resuspended in 40 p1 of eluting buffer, mutation, pCVSVHIRA/K1018, results in a substitution of alanine 10% glycerol, 2 mM Na3V04. Then 20 p1of the reaction buffer was for lysine in position 1018 of the insulin receptor. E.coli JMlOl were added togive final concentrations of 1 p~ ["PIATP (-500 Ci/mmol), 5 mM MnCl,, 12 mM MgCl,, 500 p~ CTP, and 1 mg/ml histone 2B. transformedwithdouble-strandedcircularDNAand placques screened using theprimer as probe. Themutatedsequence was After 4 min, the reaction was terminated by adding SDS-polyacrylamide gel electrophoresis samplebuffer (23). The samples were boiled confirmed by sequencing (21) and reintroduced into the CVSVHIR and analyzed by SDS-polyacrylamide gel electrophoresis on14% gels. vector. Histone and insulin receptor @-subunitswere visualized by autoraTransfection, Amplification, and Culture of Rat I Cell Lines-Rat embryo fibroblasts (Rat 1) cells (IO6cells/dish) were cotransfected diography, excised, and counted by scintillation as described above. Insulin Internalization and Degradation-The culture medium was with CVSV HIRc (10 fig) or CVSV A/K1018 (10 pg) and pSVEneo (500 ng) expression plasmids by the calcium phosphate precipitation aspirated from 35-mmdishes of cells (1-4 X lo5 cells/dish) and method with the addition of a glycerol shock after 4-h exposure to replaced with 1 ml of MEM(bicarbonate-free), 1% BSA, 10 mM HEPES, pH7.4. Cultures were held a t 37 "C in a shaking water bath. the precipitate. Following 48-h exposure to the DNA, the cells were '251-insulin (0.07 nM) and cold insulin at various concentrations were trypsinized, replatedat lower density (1:4), and placed under selection using the antibiotic G418 (400 pglml). Following 2 weeks of G418 added, and a t various times, cultures were assayed for insulin degradation and internalization as described (12). Degraded insulin was selection, independent colonies were picked, cultured, and amplified estimated by measuringtrichloroacetic acid-soluble material,and internalized insulinwas defined as that notremoved by pH 4.0 buffer I The abbreviations used are: WGA, wheat germ agglutinin; HIR, human insulin receptor; PBS, phosphate-buffered saline; HEPES, 4- (12). Photoaffinity Labeling Studies-Iodination, binding, and photolysis (2-hydroxyethyl)-l-piperazineethanesulfonicacid; MEM, minimum of '25I-B2(2-nitro-4-azidophenylacetyl)-des-PheB'-insulin and its use Eagle's medium; BSA, bovine serum albumin; SDS, sodium dodecyl sulfate; HIRc, complete human insulin receptor; KRP, Krebs-Ringer in measuringreceptor internalization have been described previously (25). phosphate solution. EXPERIMENTALPROCEDURES
A Kinase-defective Insulin Mutant
Receptor
14665
disphospho-~-[U-~~C]glucose incorporated into glycogen per h / l mg Studies of Receptor Internalization and Down-regulation-Cells of cell extract. were used a t concentrations of approximately 5 X 105/60-mm dish. Insulin was added to cells at 37 "C, and at various times, cultures were removed from the incubator and immediately rinsedtwice with RESULTS ice-cold MEM, 1% BSA at p H 4.0 and incubated on ice in the same Insulin Binding to Control and Transfected Cell Linesbuffer for 3 min. This treatment arrested theprogression of receptor metabolism and removed greater than 98% of the cell surface-bound Binding of '251-insulin to Rat 1 fibroblasts and to Rat1 cells insulin. Cells, still attached to the dish, were then rinsed with ice- transfected with the complete human insulin receptor (HIRccold PBS. Cells to be used forquantitation of the total insulin receptor B) or the mutated insulin receptor (A/K1018-B) isshown in pool were solubilized in 740 pl of PBS with soybean trypsin inhibitor Fig. 1. At tracer concentrations of insulin (0.03 nM), 0.13,90, (2.5 mg/ml), Triton X-100 (0.4%), bacitracin (0.5 mg/ml), and phenand 6 fmol insulin/106 cells were bound to Rat 1, HIRc-B, ylmethylsulfonylfluoride (0.2 mM). The dishes were incubated at 0 "C for 30 min with occasional swirling, after which time the cells and A/KlO18-B cells, respectively. When bindingis expressed were dislodged with vigorous pipettingand collected into 1.9-ml as a percent of the maximum (Fig. lB), it can be seen that microfuge tubes. Cell cultures tobe used for quantitation of intracel- the ED5" values for insulin binding are the same for native lular receptors were treated with trypsin before solubilization: after rat insulin receptors and transfected normal insulin receptors, rinsing in PBS, thecells were treated for 30 min at 0 "C with 1 mg/ (0.5-0.67 nM); the A/K1018 receptor has a slightly but statisml trypsin (Worthington, L-1-tosyl-amino-2-phenylethyl chlorometicallysignificantly lower relative affinity (ED50 1.07 nM; thy1 ketone).Afterthistime,concentrates of trypsininhibitors, detergent, and inhibitorsof degradation were added to the same final Table I). Scatchard plotsof the datain Fig. 1 and binding datausing concentrations andvolume as described above for the nontrypsinized cells. other clones were used to estimate receptor number various on Solubilized receptorsfromone60-mmdish were made 0.2% in transfected cell lines (TableI). For further analysis,we chose human y-globulin as a carrier protein, and polyethylene glycol 8000 four clones of cells transfected by normal and mutant recepwas added as a 35% (w/v) solution in PBS to a final concentration of 17%, all a t 4 "C. The polyethylene glycol-receptor solution was tors, one each of relatively high (HIRc-B and A/K1018-B) thoroughly mixed by vortexing and centrifugedfor 2.5 min at 4 "C in and relatively low (HIRc-A andA/K1018-A) receptor number. Purification of Transfected Receptors-Detergent-solubia microfuge (Eppendorf). The pellet was rinsed once with PBS and resuspended in 800 p1of binding buffer (MEM, pH 7.6, 1% BSA, lized receptors from the HIRc and A/K1018 cell lines with 0.025% Triton X-100, 2 mM EDTA). Insulin binding was assayed in high levels of receptor expression were purified by affinity solution as described above. chromatography on WGA-agarose. As can be seen in Fig. 2, Insulin-stimulated 2-Deoxyglucose Uptake-Insulin-stimulated uptake of 2-deoxy-~-glucose by control and transfected rat1 fibroblasts solubilized receptors from A/K1018 cells and HIRc cells have was measured as described previously (26) exceptthat cells were refed very similar ED5,, values for insulin binding (0.52 nM). Thus, with complete medium 18 h before the uptake studies to lower the the apparent difference in binding affinity between the A/ basal2-deoxy-~-glucoseuptake.Confluentmonolayersin60 mM K1018 and HIRc receptors in intact cells is not exhibited by culture dishes were rinsed three times with 2 ml of KRP-HEPES the same receptorsin solution. Scatchard plots of the binding buffer, pH 7.45, containing 131.2 mM Nacl, 4.71 mM KCI, 2.47 mM data in Fig. 2 reveal similar nonlinear curves for both HIRc CaCl,, 1.24 mM MgSO,, 2.48 mM NaH2P04, 10mM HEPES, and0.5% and A/K1018 receptors (not shown). BSA and then preincubatedfor 40 min a t 37 "C in 10ml of the buffer Autophosphorylation and Tyrosine Kinase Activity of Puricontainingdifferentconcentrations of insulin. After 40 min,the buffer was replaced with 1.9 ml of fresh medium containing insulin. fied Receptors-We next examined whether the purified reThe initial incubation in a large volume of insulin-containing buffer ceptors were capable of undergoing autophosphorylation in and the replenishment of insulin at 40 min were done to minimize vitro. Purified HIRc receptors (400 fmol) or A/K1018 recepthe influence of insulin degradation. Under these conditions, insulin tors (800 fmol) were exposed to saturating concentrations of degradation at the lowest insulin concentration could be kept less than 10%. as determined by precipitation with 7.5% trichloroacetic insulin and then allowed to react with ["'PIATP in varying concentrations of constant specific activity. As can be seen in acid. The transport reaction was initiated by the addition of 0.1 ml Fig. 3, this resulted in significant incorporation of ["P]phosof ["H]2-deoxy-~-glucose toa final concentration of 0.1 mM with 0.4 pCi/dish. After 5 min, the reaction was terminated by rapidly aspir- phate into normal receptors butvery little into the A/K1018 ating the buffer and washing 3 times with 3 ml of ice-cold PBS. The cells were solubilized with 1 ml of 1 N NaOH, neutralized with HC1, B and counted for 'H radioactivity in liquid scintillant (ASCII, Amer100 10oC" sham Corp.). Under these conditions, simple diffusion measured by L-glucose uptake was less than 5% of 2-deoxy-~-glucose uptake. 80 Glycogen Synthase Actiuation-Confluent monolayers in 100-mm culture dishes were rinsed with KRP-HEPES buffer and then prein60 cubated for 60 min a t 37 "C in20 ml of the buffer containing different concentrations of insulin. After aspirating the buffer, each dish was washed 4 times with 5 ml of ice-cold PBS. The cells were harvested 40 using a rubber policeman, suspended in0.5 ml of 50 mM glycylglycine, 2 mM EDTA, and 25 mM NaF, homogenized with a Polytron homog20 enizer for 15 s at a setting of 6, and then centrifuged for 5 min a t 10,000 rpm in a Beckman model B centrifuge. - 0 Glycogen synthase activitywas assayed in the supernatant fraction .03 .1 1.0 10 .03.1 1.0 10 of the cell homogenates by a modification of the method of Thomas [Insulin]. nM et al. (27) and Mandarin0 et al. (28).Reactions were initiated by adding 30 pl of a cell extract to 60 p1 of a reaction mixture at 30 "C FIG. 1. Insulin binding to transfected cell lines. HIRc-B composed of 50 mM Tris-HCI, pH 7.8, 20 mM EDTA, 25 mM NaF, (O----*), A/KlO18-B (O-----O), and Rat 1 (E---=) cells in 601%glycogen, and 0.2 mM uridine diphospho-~-[U-'~C] glucose in the mm dishes (-lo6 cells/dish) were exposed to 0.07 nM "'1-insulin and presence of either 0.3 or 6 mM glucose 6-phosphate. Reactions were varying concentrationsof unlabeled insulin for 3 ha t 15 "C. The cells terminated after 60 min by precipitatingaliquots of the reaction were then washed and counted as described. Counts bound in the mixture on 2 X 2-cm squares of filter paper dropped into cold 66% presence of 170 nM unlabeled insulin (nonspecific binding) were ethanol. The filter paperswere washed four times for 30 min each in subtractedto give specificbinding. In A, absolutebinding (fmol cold ethanol, once for 5 min in acetone, dried, and placed in Omni insulin/1O6 cells) is plotted asa function of insulin concentration. In counting solution (Amersham Corp.)for determination of the radio- B, the ordinate is percent maximal binding in order to better compare activity.Insulin-stimulable glycogen synthaseactivity (glucose 6- the relative affinitives of the receptors. Results are the mean of four phosphate-independent, I form) is expressed as nmol of uridine experiments, each done in duplicate.
A Kinase-defective Mutant Insulin Receptor
14666 TABLE I
Insulin binding and internalization in control and transfected cell lines T o determine the ED, for insulin binding cells in 60-mm dishes (-10' cells/dish) were exposed to '2SI-insulin (0.07 nM) and varying concentrations of cold insulin for 3 h at 15 "C. The cells were washed and cell-bound radioactivity determined as described. Receptor number was calculated from Scatchard plots (not shown)of binding data such as thatshown in Fig. 2. Insulin internalization was measured in 35-mm dishes of cells (2-5 X lo5 cells/dish) as described under "Experimental Procedures" except that a larger incubation volume wasused (2-4 ml) to avoid thedeclineininternalizationdueto hormone depletion seen in Fig. 5. All determinations were done after exposure to 0.07 nM insulin for 30 min a t 37 "C. Results are the mean (? S.E.) of four (Rat 1, HIRc-A, A/K1018-A) or six (HIRc-B, A/ K1018-B) independent experiments. Results areexpressed as fmol of insulin internalized per 10' cells as well as fmol insulin internalized per 10'' receptors.
Autophosphorylation of HIRc and A/K1018 Receptor
B~A/KIOIS
5
10
25
50
100 200 500 1000
[ATPI, PM FIG.3. Autophosphorylation of HIRc and Am1018 recepnM
Rat 1 1,700 (1) 0.61 f 0.06 0.03 ? 0.01 17.1 f 3.8 HIRc-A 6,400 (3.8) 0.66 f 0.14 0.12 f 0.02 18.8 f 3.1 6.1 26 k 4.9 HIRc-B 1,250,000 (735) 0.50 f 0.01 32.5 A/KlO18-A 5,700 (3.4) 0.67 f 0.11 0.05 f 0.01 8.8 f 0.2 A/KlOlS-B 220,000 (129) 1.07 f 0.02h 0.22 f 0.04 1.0 0.2 Numbers in parentheses refer to the relative receptor number compared to Rat 1 cells. * Different from HIRc-B, p < 0.01.
tors. Purified HIRc receptors ( A , 500 fmol/assay point)and A/ K1018 receptors ( R , 800 fmol/assay point) were exposed to 170 nM insulin (1 h, 0 'C) and then varying concentrations of ['"PIATP a t constant specific activity (2.3 pCi/pmol) for 5 min ( H I R c ) or 15 min (AIKZOZB)a t 0 "C. The reactionwas stopped as described and receptorsimmunoprecipitatedusing a human insulinreceptor-specific monoclonal antibody.Immunoprecipitates were fractionatedon a 7.5% reducing SDS gel. The gel was dried and exposed for 24 h a t -70 "C on XAR film using an intensifyingscreen. Labeled gel bands were excised and counted in scintillant for quantitation (see text).
may do so in intact cells. For this reason, we also examined the activity of receptors purified from cells that had been 0 HlRc exposed to insulin prior solubilization to and purification. We 0AIK1018 have previously shown that this procedure, when carried out in thepresence of protease and phosphatase inhibitors, allows one to purify receptors whose state of activation and phosphorylation is maintained as itexisted in the intactcell (24). HIRc-B or A/KlOlS-B cells were incubated in the presence 40 or absence of high concentrations of insulin. The receptors were extracted, purified, and then assayed for their ability to 20 autophosphorylate and to phosphorylate histone in vitro. It should be emphasized that insulin is not added to in thevitro assay, nor is the concentration of ATP in this assay (1 p M ) 6.07 0 1 1 .o 10 sufficient to support significant activation of the receptor tyrosine kinase beyond what has already occurred in the cell [Free Insulin]. nM FIG. 2. Insulin binding to purified HIRc and A/K1018 re- (24). As can be seen in Fig. 4, only the receptors from HIRc ceptors. Receptors were purified from solubilized HIRc-B (*) and cells thathad beenexposed toinsulin led to significant A/K1018-R (0)cells using WGA-agarose affinity chromatographyas phosphorylation of histone and autophosphorylation. Counts described. Binding of insulin (0.07 nM '"I-insulin andvarying of the radioactive bands seen in Fig. 4 show that insulin led amounts of cold insulin) was assayed using 10 fmol HIRc-receptors to a 9-fold increase in kinase activity of HIRc receptors, but and 6 fmol of A/K1018 receptors incubated with ligand for 18 h a t 4 "C. Results areexpressed as percent maximal bindingas a function no increase in that of the A/K1018 receptors. Furthermore, an antibody to phosphotyrosine precipitated no insulin-bindof insulin concentration and represent four experiments, each assayed activated insitu, alin triplicate. Tracer insulinbinding, when normalized to receptor ing activity fromA/K1018receptors number, was equal for HIRc and A/K1018 receptors. Receptor num- though the antibodydid immunoprecipitate 80% of the HIRc ber was calculated from Scatchard plotsof these data. receptor from insulin-stimulated cells (not shown). Insulin Internalization andDegradation in Transfected receptors. At 1 mM ATP, HIRc receptors incorporate-2 mol Cells-Transfected cells were exposed to tracer concentraof phosphate/mol of receptor, whereas A/K1018 receptors tions (0.07 nM)of"'1-insulin to investigate whetherthe insulin incorporate only 0.016 mol of phosphate/mol of receptor. The transfectednormal or mutatedreceptorsmediated receptors visualized in Fig. 3 were precipitated by a human internalization and degradation. As canbe seen in Fig. 5, transfected normal receptorsin HIRc-B cells mediate insulin specific anti-receptor antibody and control studies using normal Rat 1 fibroblasts showed no detectable precipitation of metabolism. Significant amountsof insulin were internalized phosphorylated rat receptors (not shown). Thus, the minimal and degraded by these cells, far greater amounts than could phosphorylationseen in Fig. 3B can be attributed to the be attributed to the endogenous Rat 1 receptors. In fact, by transfected A/K1018 receptors and not to the endogenous 30 min insulin degradation is so extensive that the extracelRat 1 receptors. lular insulin concentration falls, leading to a decrease in the It ispossible that in intactcells, conditions may be different amount of intracellular insulin. A/KlOlS-B cells, on the other from those used in in vitro assays of kinase activity. Thus, hand,internalizedand degradedonly 5-7 timesas much in vitro insulin as the untransfected Rat 1 cells despite the fact that receptors that do not autophosphorylate and activate 100
-
A Kinase-defective Mutant Insulin Receptor
14667
"
insulin concentrations (not shown). Internalization and Down-regulation of Transfected Receptors-When internalization of insulin receptors, rather than of insulin itself, was measured, a similar pattern of results emerged (Fig. 6). HIRc-B or A/KlOlS-B cells were exposed to insulin and a t various times treated with anice-cold wash with an acidicbuffer to arrest receptormovement and to remove cell surface-bound insulin (29). The cells were then solubilized so that insulin receptors could be quantitated by their ability to bind '2sI-insulin. Before solubilization, the cells were trypsinized or not so that intracellular (trypsinresistant) or total receptors could be quantitated (22). Over the time course of these experiments, total insulin receptor number remains constant. However, in HIRc cells the addiHistone 2Btion of insulin rapidly leads to the internalization of insulin receptors. By 15 min the intracellularreceptors have reached a new steady state which represents a balance of ongoing internalizationand recycling (22).At15min ininsulinInsulin - + - + treated HIRc-B cells, -30% of the total receptors are intracellular, compared to9% in untreatedcells (time 0). A smaller Cell Type AM1018 HIRc A/K1018 HlRc " proportion (p< 0.025) of A/K1018 receptors are intracellular in basal untreated cells (4.6% at time 0), and with insulin Antibody NHS (Y-HIR treatment few additional receptors are internalized. Almost FIG. 4. Activation of HIRc and Am1018 receptors in situ in living cells. HIRc-B or A/K1018-B cells (5 X 10' cells/lO-cm no receptor internalization was seen up to 24 h after insulin assay, the relatively small dish, 5 dishes/condition) were exposed to 170 nM insulin (+) or no treatment(notshown).Inthis insulin (-) for 30 min a t 37 "C. Receptors were then purified in the number of endogenous Rat 1 receptors ( 4 % ) is undetectable presence of protease and phosphatase inhibitors as described. Recep- and does not contribute to the observed values. The experitor number was determined by '2sI-insulin binding, and 200 fmol of ments in Fig. 6 were performed a t 17 nM insulin, although receptors/gel lane were immunoprecipitated using a human insulin lower (0.17 nM) insulin receptor-specific antibody or normal serum. Immunoprecipitateswere similarresults were seenusinga exposed to 1 PM ["'PIATP and 1 mg/ml histone (2B) for 5 min at concentration. Another method of visualizing insulin receptors is to use 0 "C as described. The reaction was stopped by addition of gel loading buffer and the samples fractionated on a 14% acrylamide gel. Auto- photoaffinity labeling with a radioiodinated photoactive inradiography proceeded for 2 h at -70 "C. sulin derivative (25). Cells are first treated with photoactive insulin at low temperatures such that surface receptors are occupied while internalization isprevented. The probe is then photoactivated, and noncovalently bound insulin is washed off. The cells are then warmed to allow internalization to proceed. At different times, cells are trypsinized so that receptors remaining on thecell surface are degraded but internalized receptors remain intact.Shown inFig. 7 are the results of suchexperimentsusingHIRc-Band A/K1018-B cells. Insulin receptors arevisualized as an'2sI-labeled band of -M, 430,000 (Fig. 7, inset). In HIRc cells, progressively more of the 430-kDa species becomes trypsin-resistant (i.e. internal-
P-Subunit
B
+
+
5 1 . /
2
C L
0
10
20
30
10
20
30
Tlme (rnin)
FIG. 5. Insulin metabolism by transfected cell lines. HIRc-B
(O-----O), A/K1018-B (O-----O), or Rat 1 (m-----B) cells (2-5 X lo" cells/:15-mm dish) were exposed to 0.07 nM 'ZsI-insulin a t 37 "C
as described. A t various times afterinsulin addition, the medium was removed to assay for insulin degradation, and the cells were rinsed with ice-cold PBS. Surface-bound insulin was removed by treatment with pH 4.0 buffer and residual (internalized) insulin determined as described. Nonspecific, non-receptor-mediatedinternalization was determined in the presence of 170 nM insulin. This value, -5% of the total counts/mininternalized at 0.07 nM and -40% of thetotal counts/min degraded at 0.07 nM, was subtracted to yield specific internalization ( A ) and degradation ( B ) of insulin. Results of four experiments 5 S.E. were normalized to fmol of insulin/l0' cells.
they expressed 130 times as many insulinreceptors. A summary of insulin internalization by the different cell lines (low and high receptor expression) is given in Table I. Internalization was measured a t 30 min and is expressed in terms of internalization in absolute amounts as well as on a per receptor basis. Both A/K1018 cell lines internalize insulin to only a slightly greater degree than do the untransfected Rat 1 cells. This effect is seen at both low and high (7 nM)
01
'
5
'
10
I
20
40
Tlme (rnin)
FIG. 6. Internalization of HIRc and A/K1018 receptors after exposure to insulin. Cultures of HIRc-B (O-----O) or A/
K1018-B (O-----O) cells (-lofi cells/60-mm dish) were exposed to 16 nM insulin at 37 "C. At different times, duplicate cultures were rinsed with ice-cold PBS, andcell surface-bound insulin was removed using acid treatment asdescribed. Cells were either exposed to trypsin or not and then solubilized. Total (non-trypsin-treated) or intracellular (trypsin-resistant) receptors were quantitated as described. Total receptor number did not vary over the period of this experiment. Therefore, results were normalized to the percent of total receptors that were intracellular and plotted as function of time. Results are the mean (5S.E.) of four experiments.
A Kinase-defective Mutant InsulinReceptor
14668
I B: A/K1018
A:.HIRc 430 k D -
0- "me Total
5' 10' 20' 30' 45' 60'
Total
5' 10' 20' 30' 45' 60'
Time (min) FIG. 7. Internalization of photoaffinity-labeled HIRc and A/K1018 receptors. Receptors on HIRc-B ( A ) orA/K1018-B ( B ) cells at 0 "C were labeled with ~'sII-B2(2-nitro-4-azidophenylacetyl)des-PheH1-insulin as described. Cells were warmed to 37 "C to allow internalization to proceed, and at various times cultures were treated with trypsin in order to quantitate trypsin-resistant (intracellular)receptors. Control cells not warmed and not treated with trypsin ("total," insets) wereused to quantitate total receptors. In cells treated with trypsin without warming (not shown),95-98% of the receptors were degraded bytrypsin. Solubilized cells were fractionated by nonreducing SDS-polyacrylamide gel electrophoresis and subjected to autoradiography. The insulin receptor bands at 430-kDa (inset) from three independent experiments wereexcised and counted in a y-counter. The trypsin-resistant (intracellular)counts/min are plottedas a percent of total counts/min from non-trypsinized cells as a function of time.
HIRc -
A:
6:A/K1018
time- and dose-dependent loss of total and surface receptors such that after 24 h of treatment, 72% of the total cellular insulin receptors have been lost, presumably due to internal= 10ization and degradation. Similar treatment led to no loss of u A/K1018 receptors. Furthermore, after exposure to insulin b 025for long periods, HIRc receptors redistribute such that 30e 60% of the total receptor number is intracellular a t 4 and 24 a h; in contrast, no appreciable redistribution of receptors ocw 05P curs in A/K1018 cells even after 24 h in the presence of 170 0 nM insulin. 025Biological Activities of Transfected Receptors-Transfected HIRcreceptors have all theproperties of normalinsulin 0tl 4 t, 14 h Oh 4h 24h receptors, whereas, the A/K1018 receptors do not have tyrodo not participate in the normal FIG. 8. Down-regulation of insulin receptors. HIRc-B cells sinekinaseactivityand (left panel)or A/K1018-B cells (right panel)were exposed to 17 nM insulin receptor itinerary. We next asked whether these re(dotted bars) or170 nM (hatched bars) insulin. After 4 and 24 h, ceptors mediated thebiological actions of insulin. In Fig. 9A cultures were washedand treated with pH 4.0 buffer to remove insulin it canbe seen thatin untransfected Rat 1cells, insulin causes as described. The cells were trypsinized or not, solubilized, and total stimulation of 2-deoxy-~-glucose transport with an EDSO of as well as intracellular (trypsin-resistant, solid bar) receptors were 0.45nM. Introduction of normal insulin receptors leads to quantitated. Results are the mean of two experiments. increased insulin sensitivity. In HIRc-A cells, with 3.8 times as many receptors as untransfected Rat 1 cells, the EDSO for izes) with time a t 37 "C. The kinetics of internalization are is to 0.125 very similar to the results presented in Fig. 6 with a steady stimulation of 2-deoxy-~-glucose uptake decreased nM, or by a factor of 3.6. Introduction of more insulin recepstate reached a t -20 min. InA/K1018 cells, very little receptor internalizes with time: even after 60 min, only 8% of the A/ tors (HIRc-B) leads to even greater insulin sensitivity (ED5, K1018 receptors have escaped trypsin sensitivity. In labeled = 0.065 nM), although not in direct proportion to the number cells that are not warmed to37 "C, that is, cells in which no of expressed receptors as was the case for the HIRc-A cells. A/KlOlS-B cells, on the other hand, show insulin-induced endocytosis is allowed, -5% of the receptors remain after trypsin treatment. Thus, the 8% trypsin-resistant A/K1018 stimulation of 2-deoxy-~-glucose transportonly a t very high receptorsseen in Fig. 7 can belargely accountedfor by insulin concentrations, with an EDso of -230 nM. Thus, not incomplete trypsinization (i.e. assay background) asopposed only is there no leftward shift of the dose-responsecurve (increased insulin sensitivity), but curve the is actually shifted to true internalization. fibroblasts, We next examined the insulin-induced loss of insulin re- to the right compared to the nontransfected 1Rat ceptors (down-regulation). As can be seen in Fig. 8, trans- suggesting that thepresence of expressed A/K1018 receptors fected normal receptors down regulate but the A/K1018 re- inhibits the function of the normal rat insulin receptors. A/ ceptors do not. In HIRc-B cells, insulin (170 nM) causes a K1018-A cells, with fewer inactivereceptors,alsodisplay 125-
I
\
1
A Kinase-defective Mutant Insulin Receptor
14669
another cell line transfected with the same mutant. Wehave also shown that this kinase defective mutant receptor does 2-DOG not undergo endocytosis and does not mediate ligand interUptake nalization or degradation. (% maximal) The A/K1018 receptor has defective kinase activity by a number of criteria. I n vitro autophosphorylation is depressed in the mutant receptor. Even at very high ATP concentraGlycogen Synthase tions, the amount of autophosphorylation of the mutant reActivity ceptor is -1% that of the normal receptor. Most likely, the (% small amountof phosphorylation seenin Fig. 3 is attributable maximal) to the transfected mutant receptors opposed as to endogenous Rat 1 receptors, because the assay was done using material ,003 .01 .1 1.0 10 100 1000 [Insulin\, nM immunoprecipitated by a monoclonalantibody specific for the FIG. 9. Insulin stimulation of 2-deoxyglucose uptake and humaninsulin receptor. The antibody will notdetectably glycogen synthase activity in control and transfected fibro- precipitate insulin receptors from equivalent numbers of unblasts. Insulin-stimulable 2-DOG uptake ( A ) and glycogen synthase transfected Rat 1 cells. Because the mutated receptor has I activity ( B )were measured in Rat 1 (H HIRc-A ), (A-A), some kinase activity, and such activity canbe modulated by HIRc-B (c-“.), A/KLOlS-A (U-- - U),and A/K 1018-B (0-- - - -0)as described. Each value, the mean of five experiments, divalent cations as well as substrate availability (24), it was is plotted as thepercentage of the maximal insulin effect as a function theoretically possible that the submembranous milieu could of insulin concentration. Absolute levels of 2-deoxy-D-glucoseuptake have supported significant kinase activity that might not be (nmol/106 cells/5 min) under basal (B) and maximally insulin-stim- seen i n vitro. For this reason, we purified receptors thatwere ulated (M) conditions were as follows: rat, B = 0.51, M = 0.70; HIRc- activated by insulin i n situ in living cells. In the presenceof A, B = 1.78, M = 2.23; HIRc-B, B = 1.11, M = 1.97; A/K1018-A, B phosphatase and protease inhibitors, these receptors retain = 2.18, M = 2.45; A/KlOlS-B, B = 1.66, M = 1.97. Absolute levels of glycogen synthase activity (nmol/mg protein/h) measured at 0.3 mM their “in vivo’’ state of activation when assayed in solution glucose 6-phosphate were: rat 1, B = 161.2, M = 264.4; HIRc-A, B = (24). A/K1018 receptors prepared in this way do not exhibit 83.8, M = 123.7; HIRc-B, B = 249.4, M = 468.0; A/KlOlS-B, B = significant insulin-stimulable kinase activity assayed either 184.1, M = 268.1.2-DOG, P-deoxy-~-glucose. by autophosphorylation or by phosphorylation of histone. In contrast, a 9-fold increase inreceptor kinase activityof HIRcdecreased insulin sensitivity (EDBo= 0.81 nM), but the mag- B cells was observed in insulin-treated cells. In agreement we alsofound that a phosphotyrosine nitude of the inhibitory effect is less than in thehigh expres- withtheseresults, sion A/KlO18-B cells. The basal level of 2-deoxy-~-glucose antibody immunoprecipitated -80%of the HIRc receptors uptake in Rat 1 fibroblasts was 0.5 nmol/106cells/5min. but failed to precipitate any A/K1018 receptors preparedfrom Basal levels of the transfected lines ranged from 1.1 to 2.2 insulin-stimulated cells indicating a lack of endogenous innmol/106cells/5 min. Thus, basal transport varied in the sulin-mediated tyrosine phosphorylation of A/K1018 recepdifferent clones but was not lower in the A/K1018 cells as tors. These results are consistent with the findings of Chou has been reported for other cell lines with inactive receptors et al. (6) who observed neither insulin stimulated phospho(6, 7). A monoclonal antibody that acts as an insulin agonist rylation of the mutant receptor nor phosphorylation of enbut is specific for human insulin receptors andwill not stim- dogenous substrates mediatedby the mutant receptor in intact ulate rat (endogenous) receptors did not stimulate 2-deoxy- cells. D-glucose transport in the A/K1018cells but did enhance We also found that the A/K1018 receptor did not mediate transport in HIRccells (data not shown). It should be noted, either of two biologic effects of insulin, i.e. the stimulation of however, that A/K1018-Acells still expressed therodent hexose transport and glycogen synthase activity. An imporinsulin receptor. Thus, a polyclonalantibody which recognizes tantfindinginthesestudiesisthatthe A/K1018 insulin both rat and human receptors reproducibly immunoprecipi- receptors did not simply behave as inactive receptorsdevoid tated 30% more photoaffinity labeled &-subunit than did a of biologic activity. In fact, when these receptors were exhuman-specific monoclonal antibodythatimmunoprecipipressed in cells at either low (A/K1018-A) or high (A/K1018tated >90% of the human receptor (data not shown). B) levels, a decrease in insulin sensitivity was observed comRat 1 cells also have insulin stimulable glycogen synthase pared to normal Rat 1 fibroblasts. The dose-response curves activity (Fig. 9B, EDso 0.38 nM), and expression of greater for stimulation of glucose transport and glycogen synthase numbers of normal insulin receptors leads to greater insulin activity were shifted to the right relative to normal Rat 1 sensitivity (HIRc-A, ED,,, 0.08 nM; HIRc-B, ED50 0.06 nM). fibroblasts. Thus, introduction of kinase-defective A/K1018 Expression of A/K1018 receptors, on the other hand, leads toinsulin receptors interfered with the biologic signaling of the markedly reduced insulin sensitivity (A/K1018-B, ED,, 170 normalcomplement of nativereceptorspresentinRat 1 nM). fibroblasts. This decrease ininsulinsensitivity(rightward shift to thedose-response curve) was dose-responsive and in DISCUSSION proportion to the expression level of A/K1018 receptors in Rat 1 fibroblasts were transfected with cDNAs coding for clonesA and B. This apparent inhibitory effect of kinase either normal (HIRc) insulin receptors or mutated receptors defective receptors has been observed previously (6, 7), although Chou et al. (6) have suggested that the inhibition is withLysine 1018 replaced by alanine. This lysineresidue, highly conserved in several oncogene products and growth restricted only to insulin’s action on glucose transport. Our results clearly show that A/K1018 receptors confer decreased factorreceptors (1, 2), lies in a region homologous to the known ATP binding site of several proteins (14). The same insulin sensitivity for both stimulation of glucose transport mutant receptor, expressed in Chinese hamster ovary cells, and glycogen synthase activity. T h e reasons for the differhas recently been shown to lack tyrosine kinase activity i n ences between the current results and thoseof Chou et al. (6) vitro and to be defective inmediating several biologic re- are not clear, but based on our findings, we suggest that the sponses to insulin (6, 7). We have confirmed these results in inhibitory effects of A/K1018 receptors occur at a relatively
14670
A Kinase-defective Insulin Mutant
Receptor
proximal step in the insulin actionprogram. with insulin treatment. A/K1018-B cells show no such inThe biochemical mechanismsunderlyingthisinhibitory crease. Resting, untreated A/K1018-B cells also have a sigeffect of A/K1018 receptors are currently unknown. Conceiv- nificantly higher proportion of theirreceptors at the cell ably, A/K1018 receptors interact with normal Rat1receptors surface, which may reflect a lack of low level, basal endocyduring biosynthesis and receptor processing to form biologi- tosis andrecycling of receptors. The lack of internalization of cally incompetent mixed heterotetramers. It is also possible A/K1018 insulin receptors was also shown using photoaffinthat receptor dimers, or high order oligomers, are necessary ity-labeled receptors (Fig. 7). Both of thesemethodsare to elicit biologic responses and if A/K1018 receptors interact specific for transfected receptors because of the relative prewith normal receptors, then an inactive unit results. Alter- ponderance of transfected insulin receptors (>99% of total natively, A/K1018 receptors may compete for a cellular sub- receptors) over the endogenous Rat 1 receptors in the A/ strate or in some way bind to or remove an important com- K1018-B cells. ponent of the insulin response sequence. It is unlikely that Since theA/K1018 receptor has only a very limited ability the inhibitory effect of A/K1018 receptors is a result of lack to undergo endocytosis, and the receptor is clearly defective of expression of the endogenous Rat 1 receptors or is a intyrosinekinaseactivity displaying very little (-1%of nonspecific effect of expressing inactive receptors in thiscell normal) ability to autophosphorylate, these results strongly line. Thus, using a combination of human-specific and poly- suggest that receptor kinase activity is critical to internalizaclonal antibodies recognizing both rat and human receptors, tion. The phosphorylated insulinreceptor may assume a more we have been able to directly demonstrate expressionof the favorable conformation allowing endocytosis to occur, faciliendogenous ratreceptorsinthese cells; furthermore,the tating physical interaction of the receptor with other comporodent receptor mRNA can readily be detected in A/K1018 nents of the endocytotic process. Alternatively, the activated cells.’ Additionally, we have found that the dose-response receptor kinase may itself lead to phosphorylation of other curve for insulin-like growth factor I stimulation of glucose cellular components which then mediate receptor internalitransport is completely normal in A/K1018 cells compared to zation.Although it is alwayspossible that the alanine for Rat 1 fibroblasts2 and that expression of high levels of a lysine substitution a t position 1018 alters the receptor’s tertruncated, biologically inactive, insulin receptor in these cells tiary structure ina domain separatefrom the receptor kinase, does not lead to any inhibition of insulin sensitivity.* These the simplest interpretation of our studies is that autophoslatter findings argue strongly that the inhibitory effect ob- phorylation and/or an activatedreceptor kinase in some way served inthe A/K1018 cellsis not simply a nonspecific cellular mediates endocytotic uptake of insulin receptors. This may response to expression of biologically inactivetransfected prove to be a generalized phenomenon regulating internalireceptors. zation of a variety of other peptide hormone receptors and An analogy canbedrawn between the cellular insulin- other surface proteinswhich possess tyrosine kinase activity. resistantstate induced by expression of A/K1018 insulin Insulin-induced receptor loss or down-regulation is seen in receptors and a human insulin-resistant state, i.e. non-insu- a variety of normal cellsina time-anddose-dependent lin-dependent diabetes mellitus. It has been shown that cells fashion after exposure to high insulin concentrations for a from diabeticpatientsexpressinsulinreceptors withdefew hours (31). A/K1018 receptors, however, do not downcreased kinase activity (23) along witha further reduction in regulate after prolonged exposure to insulin (Fig. 8). Interestglucose transport system activity (30). Thus, cellular insulin ingly, the HIRc cells do exhibit down-regulation of the transresistance involves both receptor and post-receptor abnorfected, expressed receptors. Because the cDNA encodingHIRc malities. Likewise, in the A/K1018clones, the presence of receptors does not include genomic, regulatory sequences, it kinase-defective insulin receptors leads to an apparent desen-isprobablethat down-regulation is simplya function of sitization of the insulin action program aatpost-binding (i.e. increased receptor degradation and does not involve changes glucose transport and glycogen synthase activity) level. at thelevel of gene transcription. We assume, therefore, that Our results demonstrate that the tyrosinekinase-defective because A/K1018 receptors do not internalize, the usual fracA/K1018 receptors maintain their ability to bind insulin with tion of endocytosed receptors does not enter the degradative high affinity analogous to that of normal insulin receptors. pathway and therefore the receptors do not down-regulate. The A/K1018 receptors are thus apparently inserted into the The current experimentsshow that the A/K1018 mutated plasma membranewith the appropriate orientation.However, insulin receptor expresses defective kinase activity and does a key finding of the work presented here is that the inactive, not mediateinsulin’s stimulatory actions onglucose transport mutated receptordoes notparticipateintheendocytotic or glycogen synthase.Theseresults, coupledwith earlier intracellular itinerary of normal insulin receptors. This lack reports (6, 7) showing that kinase defective insulin receptors of endocytosis is probably a property of the receptor and not do not mediate stimulation of S6 phosphorylation, endogeof the hostcell because the effect was seen in two independent nous substrate phosphorylation, thymidine incorporation into clones of cells. The A/K1018 receptors mediate much less glycogen synthesis,argue insulin internalizationor degradation thanwould be predicted DNA, glucose transport,and based on the number of expressed receptors (Table I). A/ strongly for an importantrole of the insulinreceptor tyrosine kinaseactivityas a proximalstep in many, if notall, of K1018 receptorsthemselvesalso do notinternalizeinreinsulin’s actions. Although this is the interpretation we cursponse to insulin. This was demonstrated usingtwo independent assays. In the first, solubilized receptors are quanti- rently favor, an alternativeview is possible. Thus, the current fied from cellsthat have been treated or not with trypsin (Fig. studies also show that theA/K1018 kinase defective receptor 6). This trypsin treatment nearly completely abolishes insulin is almost completely unable to enter the normalendocytotic/ binding activity of surface receptors but leaves intracellular recycling pathway. Since endocytotic uptake of insulin and receptors intact (22,25). Normal cells, as well as cells express- the insulin receptor is normally very rapid, it is possible that internalization of either the insulin molecule or the insulin ing transfectednormal receptors, show an increase in the receptor is necessary for insulin to exert its biologic effects, proportion of trypsin-resistant ( i e . intracellular)receptors possibly by allowing physicalaccess of these molecules to intracellular effectorsystems.Indeed, intracellularbinding D. A. McClain, A. Ullrich,and J. M. Olefsky,manuscriptin sites for insulin have been reported (32), and after initial preparation.
A Kinase-defective M Ixtant Insulin Receptor
14671
Ullrich, A., and Rosen, 0. M. (1987) J. Biol. Chem. 262,1842endocytoticuptakethemovement of theinsulin receptor 1847 through various intracellular organelles, including the nucleus 7. Ebina, Y., Araki, E., Taira, M., Shimada, R., Mori, M., Craik, C. (13), is well described (33). Thus, one interpretation of our S., Siddle, K., Pierce, S. B., Roth, R.A., and Rutter, W. J. findings is that itis the lackof internalization which leads t o (1987) Proc. Natl. Acad. Sci. U. S. A. 8 4 , 704-708 8. Smith, R.M., and Jarett, L. (1983) J. Cell. Physiol. 115, 199the absenceof insulin’s biologic effects in the A/K1018 cells. 207 Quite possibly, other insulin receptorswhich have been made 9. Pilch, P. F., Shia, M. A., Benson, R. J. J., and Fine, R. E. (1983) kinase-defective by deletion of putative autophosphorylation J. Cell. B i d . 96, 133-138 sites ( 5 ) are also unable to enter the endocytotic receptor 10. Forgas, M., Cantley, L., Wiedenmann, B., Altstiel, L., and Branpathway. Alternatively, the insulin receptor kinase may diton, D. (1983) Proc. Natl. Acad. Sci. U. s. A. 8 0 , 1300-1303 rectly mediate insulin’s biologic effects as well as endocytotic 11. Marshall, S., Green, A., and Olefsky, J. M. (1981) J. Biol. Chem. 2 5 6 , 11464-11470 receptor uptake through independent mechanisms. With this latter formulation, the absenceof biologic activity and inter- 12. Levy, J. R., and Olefsky, J. M. (1986) Endocrinology 119, 572579 nalization of the A/K1018 receptor would be mediated by a 13. Smith, R.M., and Jarett, L. (1987) Proc. Natl. Acad. Sci. U. S. common defect but not causally related. A . 84,459-463 In conclusion, the resultsof these studies demonstrate that 14. Sternberg, M. J. E., and Taylor, W . R. (1984) FEBS Lett. 175, 387-392 the A/K1018 insulin receptor has severely reduced tyrosine 15. Siddle, K., Soos, M. A., O’Brien, R. N., Ganderton, R. H., and protein kinase activity and fails to mediate insulin’s biologic Taylor, R. (1987) Biochem. SOC. Trans.15, 47-51 signals. Indeed, the presence of A/K1018 receptors appears to 16. Lusky, M., and Botchan, M. (1981) Nature 293, 79-81 inhibit the biologic activity of normal receptors. A/K1018 17. Simonsen, C. C., and Levinson, A.D. (1983) Proc. Natl. Acad. Sci. U. S. A . 8 0 , 2495-2499 receptors are alsoseverely limited in their ability to enter the 18. Birnboim, H. C., and Doly, J. (1979) Nucleic Acids Res. 7, 1513endocytotic receptor itinerary. These results argue that the 1523 tyrosine kinase activity of the insulin receptor, or the auto19. Razin, A., Hirose, T., Itakura, K., and Riggs, A. D. (1978) Proc. phosphorylated receptor itself, is necessary both for biologic Natl. Acad. Sci. U. S. A. 7 5 , 4268-4272 signalling and for endocytosis.Whether these are independent 20. Adelman, J. P., Hayfilch, J. S., Vasser, M., and Seeburg, P. H. (1983) DNA 2 , 183-193 functional defects conferred by the impaired kinase activity F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. or whether impaired endocytosis causes the reduced biologic 21. Sanger, Acad. Sci. U. S. A . 74,5463-5467 signaling remains to be determined. 22. Marshall, S. (1985) J. Biol. Chem. 260,4136-4144 23. Freidenberg, G. R., Henry, R. R., Klein, H. H., Reichart, D, R., Acknowledgments-We wish to thankDrs. Steven Jacobs, Kenneth Siddle, and Larry Mandarino for the antibodies used in this study and Dr. Gary Freidenberg for assistance inthe tyrosine kinase assays. We are grateful to Elizabeth Martinez and Cleon Tate for preparing the manuscript. REFERENCES 1. Ullrich, A., Bell, J. B., Chen, E. Y., Herrera, R., Petruzzelli, L. L. M., Dull, T. J., Gary, A., Coussens, L., Liao, Y.-C., Tsubokawa, M., Mason, A,, Seeberg, P. H., Grunfeld, C., Rosen, 0. M., and Ramachandran, J. (1985) Nature 3 1 3 , 756-761 2. Ebina, Y., Ellis, L., Jarmagin, K., Edery, M., Graf, L., Clauser, E., Ou, J.-H., Masiarz, F., Kan, Y. W . , Goldfine, I. D., Roth, R., and Rutter, W . (1985) Cell 4 6 , 747-758 3. Kasuga, M., Karlsson, F. A., and Kahn, C. R. (1982) Science 215,185-187 4. Roth, R. A,, and Cassell, D. J. (1983) Science 2 1 9 , 299-301 5. Ellis, L., Clauser, E., Morgan, D., Edery, M., Roth, R., and Rutter, W . (1986) Cell 45, 721-732. 6. Chou, C. K., Dull, T. J., Russell, D. S., Cherzi, R., Lebwohl, D.,
and Olefsky, J. M. (1987) J . Clin. Invest. 79, 240-250 24. Klein, H.H., Freidenberg, G. R., Kladde, M., and Olefsky, J. M. (1986) J. Biol. Chem. 2 6 1 , 4691-4697 25. Heidenreich, K. A., Brandenburg, D., Berhanu, P., and Olefsky, J. M. (1984) J. Biol. Chem. 259,6511-6515 26. Berhanu, P., and Olefsky, J. M. (1981) Diabetes 3 0 , 523-529 27. Thomas, J . A., Schlender, K. K., andLarner, , J. (1968) . . Anal. Biochem. 25,486-499 28. Mandarino, L. J., Madar, Z., Kolterman, 0. G., Bell, J . M., and Olefsky, J. M. (1986) Am. J. Physiol. 2 5 1 , E489-E496 29. Caro, J. F., Muller, G., and Glennon, J . A. (1982) J. Bid. Chem. 257,8459-8466 30. Ciaraldi, T. P., Kolterman, 0. G., Scarlett, J. A., Kao, M., and Olefsky, J . M. (1982) Diabetes 3 1 , 1016-1022 31. Gavin, J. R., Roth, J., Neville, D. M., DeMeyts, P., and Buell, D. N. (1974) Proc. Natl. Acad. Sci. U. S. A. 7 1 , 84-88 32. Goldfine, I. D., Jones, A. L., Hradek, G. T., Wong, K. Y., and Mooney, J. S. (1978) Science 202, 760-763 33. Khan, M.N., Posner, B. I., Verma, A. K., Kahn, R. J., and Bergeron, J. J. M. (1981) Proc. Natl. Acad. Sci. U. S. A . 7 8 , 4980-4984
