An Extracellular Domain of the Insulin Receptor @-Subunit with Regulatory Function on Protein-Tyrosine Kinase*

Anti-insulin receptor monoclonal antibody MA- 10 inhibits insulin receptor autophosphorylation of purified rat liver insulin receptors without affecting insulin binding (Cordera, Goldfine, (1987) Endocrinology 121,2007-2010). The effect of MA-10 on insulin receptor auto- phosphorylation and on two insulin actions (thymidine incorporation into DNA and receptor down-regulation) was investigated in rat hepatoma Fao cells. MA-10 inhibits insulin-stimulated receptor autophosphorylation, thymidine incorporation into DNA, and insulin- induced receptor down-regulation without affecting insulin receptor binding. We show that MA-10 binds to a site of rat insulin receptors different from the insulin binding site in intact Fao cells. Insulin does not inhibit MA-10 binding, and MA-10 does not inhibit insulin binding to rat Fao cells. Moreover, MA-10 bind- ing to down-regulated cells is reduced to the same extent as insulin binding. In rat insulin receptors the MA-10 binding site has been tentatively localized in the extracellular part of the insulin receptor @-subunit based on the following evidence: (i) MA-10 binds to

Anti-insulin receptor monoclonal antibody MA-10 inhibits insulin receptor autophosphorylation of purified rat liver insulin receptors without affecting insulin binding (Cordera, R., Andraghetti, G., Gherzi, R., Adezati, L., Montemurro, A., Lauro, R., Goldfine, I. D., and De Pirro, R. (1987) Endocrinology 121,[2007][2008][2009][2010]. The effect of MA-10 on insulin receptor autophosphorylation and on two insulin actions (thymidine incorporation into DNA and receptor down-regulation) was investigated in rat hepatoma Fao cells. MA-10 inhibits insulin-stimulated receptor autophosphorylation, thymidine incorporation into DNA, and insulininduced receptor down-regulation without affecting insulin receptor binding. We show that MA-10 binds to a site of rat insulin receptors different from the insulin binding site in intact Fao cells. Insulin does not inhibit MA-10 binding, and MA-10 does not inhibit insulin binding to rat Fao cells. Moreover, MA-10 binding to down-regulated cells is reduced to the same extent as insulin binding. In rat insulin receptors the MA-10 binding site has been tentatively localized in the extracellular part of the insulin receptor @-subunit based on the following evidence: (i) MA-10 binds to insulin receptor in intact rat cells; (ii) MA-10 immunoprecipitates isolated insulin receptor B-subunits labeled with both ["S]methionine and "P; (iii) MA-10 reacts with rat insulin receptor @-subunits by the method of immunoblotting, similar to an antipeptide antibody directed against the carboxyl terminus of the insulin receptor @-subunit. Moreover, MA-10 inhibits autophosphorylation and protein-tyrosine kinase activity of reduced and purified insulin receptor @-subunits. The finding that MA-10 inhibits insulin-stimulated receptor autophosphorylation and reduces insulin-stimulated thymidine incorporation into DNA and receptor down-regulation suggests that the extracellular part of the insulin receptor @-subunit plays a role in the regulation of insulin receptor protein-tyrosine kinase activity.
The insulin receptor is a transmembrane glycoprotein composed of distinct subunits linked by disulfide bonds: two a-subunits (Mr 135,000 by SDS-PAGE),' comprising the insulin binding site and located entirely at the extracellular face of the plasma membrane, and two @-subunits (Mr 95,000 by SDS-PAGE) that are protein-tyrosine kinases and span the plasma membrane. Insulin binding to insulin receptor @subunit causes rapid phosphorylation of tyrosine residues on the receptor @-subunit (for a review, see Ref. 1).
Although the structure of insulin receptor is characterized (2,3) and the relevance of its protein-tyrosine kinase activity is defined (4)(5)(6)(7)(8)(9), less information is available on the relative importance of different domains of the molecule in the regulation of its enzymatic activity (10)(11)(12). The following data suggest that the insulin receptor is composed of functionally independent domains: first, a-subunits of chimeric receptors both lacking the cytoplasmic portion of the @-subunit and presenting substitutions in this domain bind insulin with normal affinity (13,14); second, in the absence of the asubunit, the cytosolic part of insulin receptor P-subunit, expressed in Chinese hamster ovary cells, is a protein-tyrosine kinase more active than the entire receptor (15); third, the protein-tyrosine kinase of the cytoplasmic domain of insulin receptor can be activated by autophosphorylation independently of its organization within the native receptor oligomer (16); fourth, some agents (sodium orthovanadate, hydrogen peroxide, anti-@-subunit monoclonal antibodies) activate insulin receptor protein-tyrosine kinase directly acting on @subunits (17- 19).
Monoclonal antibodies are convenient probes to understand the functional topography of receptor molecules (20)(21)(22)(23). In previous reports we demonstrated that protein-tyrosine kinase activity, stimulated in the entire insulin receptor by insulin and insulin mimickers in vitro, can be inhibited by the anti-insulin receptor monoclonal antibody MA-10 (24,25). Based on this finding we suggested that MA-10 recognizes a district, conserved in human and rat insulin receptor, involved in the regulation of protein-tyrosine kinase activity but different from the insulin binding site (24).
In this paper we investigate: (i) whether MA-10 recognizes insulin receptors in intact rat cells, (ii) whether the portion of insulin receptor molecule, recognized by antibody MA-10, is important to regulate receptor autophosphorylation in intact cells and whether the inhibition of the enzyme by MA-* This work was supported in part by grants from M.P.I. (to R. DP.) and by Grant 870008204 from Consiglio Nazionale delle Richerche and grants from M.P.I. and NOVA Italia (to R. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore he hereby marked "aduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

8627
10 has biological significance, and (iii) whether MA-10 is able to inhibit the protein-tyrosine kinase activity of the isolated P-subunit of insulin receptor.
Data presented herein demonstrate that MA-10 inhibits insulin receptor autophosphorylation and the transduction of two insulin actions in insulin-responsive rat hepatoma cells binding to a region of the receptor molecule which is different from the insulin binding site and which is localized in the protein backbone of the extracellular portion of insulin receptor P-subunit.
Antibodies-Monoclonal antibody MA-10 is an IgG2b raised in mice against highly purified human placental receptors as previously described (27). MA-10 is devoid of any phosphatase or ATPase activity, it is not able to bind iodinated insulin, although an antiinsulin antiserum did (24), and it does not react with IGF I receptors (27). Antibody PC is a polyclonal antibody raised in rabbit against a synthetic peptide corresponding to the deduced sequence of the carboxyl-terminal 17 amino acids of the human insulin receptor. This antibody immunoprecipitates both human and rat insulin receptors but is unable to immunoprecipitate IGF I receptors.233 Anti-insulin receptor antibody ARA (a kind gift of Dr. Lawrence Mandarino) is a human polyclonal antibody that recognizes both human and rat insulin receptors (28). Antibody H 65.6 is a monoclonal IgG2b that recognizes a monomorphic epitope present on HLA class I1 DQ molecules.
Cells-Insulin-sensitive rat hepatoma cells Fao (kindly provided by Dr. Bernard Rossi, University of Nice, France) and human hepatoma cells Hep G2 (a generous gift of Dr. Ora Rosen, Sloan-Kettering Cancer Center, New York) were grown in DME with 10% fetal calf serum.
Insulin and MA-10 Binding-Insulin binding to Fao and Hep G2 cells was measured for 4 h at 4 "C as described in Ref. 6. Scatchard analysis of insulin binding to both human and rat hepatoma cells, measured under these conditions, indicated that Fa0 and Hep G2 cells have approximately 31,000 and 26,000 high affinity insulin receptors per cell, respectively. The affinity constants of '"I-insulin for high affinity receptors of Fao and Hep G2 are both about 1 nM. The binding affinities of both human and porcine '"I-insulin to rat (Fao) and human (Hep G2) hepatoma cells are superimposable. To measure MA-10 binding, cells in six-well dishes were incubated with different concentrations of MA-10 or mAb H 65.6 together with 1.5 x IO5 cpm '251-protein A/well for 4 h at 4 "C. Then cells were washed, solubilized, and counted for radioactivity as described in Ref. 6 M) (in some experiments, normal mouse IgG (10" M)) was added, and the incubations were continued for 10 min. Cells were then rapidly cooled in ice, washed three times with ice-cold PBS containing 5 mM EDTA, 5 mM EGTA, and 1 mM sodium orthovanadate and dissolved in 0.5 ml of 50 mM Hepes buffer, pH 7.4, containing 2% Triton X-100, 5 mM EDTA, 5 mM EGTA, 20 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 20 mM NaF, 1 mM phenylmethylsulfonyl fluoride, and 20 pg/ ml each of aprotinin and leupeptin. The cell extracts were clarified by centrifugation at 100,000 X g for 30 min at 4 "C, and the supernatant fluid was applied to 0.3-ml columns of WGA-agarose; columns were washed with 20 volumes of 50 mM Hepes buffer, pH 7.4, containing 0.1% Triton X-100,20 mM sodium pyrophosphate, 20 mM NaF, 1 mM sodium orthovanadate, 5 mM EDTA, and 5 mM EGTA. The insulin receptor was then eluted in the same buffer containing 0.3 M N-acetylglucosamine. The eluate was incubated with the antipeptide polyclonal antibody AbPC or with the polyclonal anti-insulin receptor antibody ARA for 16 h at 4 "C and the immune complex precipitated by protein A-Sepharose (8). Immunoprecipitated receptors were analyzed by SDS-PAGE and subjected to autoradiography. Gels were then rehydrated in 1 M KOH and heated at 55 "C for 1 h. This treatment hydrolyzes most of phosphates incorporated into serine and threonine residues of the receptor but leaves phosphotyrosine residues relatively intact (30). Following incubation in KOH, gels were washed in 10% acetic acid, 40% methanol, 1% glycerol for 1 h, neutralized, dried, and subjected again to autoradiography. Both autoradiograms were scanned with an Ultroscan LKB laser densitometer.
PSIMethionine Labeling of the Insulin Receptor-Confluent monolayers of Fao or Hep G2 cells in 100-mm dishes were incubated with methionine-free MEM containing 10% dialyzed fetal calf serum for 2 h. [35S]Methionine (200 pCi/ml) was added, and the incubation was continued further for 30 min. The medium was changed to DME containing 10% fetal calf serum and 10 mM unlabeled L-methionine, and the incubation was continued for up to 8 h.
Cells were then solubilized in 50 mM Hepes buffer, pH 7.4, containing 2% Triton X-100,5 mM EDTA, 5 mM EGTA, and 10 pg/ml each of aprotinin, leupeptin, and soybean trypsin inhibitor. Cell lysates were centrifuged at 150,000 x g for 30 min at 4 "C. The supernatant fluid was diluted 20-fold into 50 mM Hepes buffer, pH 7.4 containing 0.1% Triton X-100,150 mM NaCl, 5 mM EDTA, 5 mM EGTA, and the protease inhibitors (10 Fg/ml each) and subjected to chromatography on WGA-agarose as described in Refs. 6 and 24. The insulin receptor was eluted with 0.3 M N-acetylglucosamine. Receptors were immunoprecipitated with MA-10, AbPC or mAb H 65.6 and 35S-labeled proteins analyzed by SDS-PAGE. Gels were stained, destained, soaked (30 min, room temperature) in Amplify, and dried. In experiments designed to study whether MA-10 recognizes nonglycosylated insulin receptors Fao cells were incubated in methioninefree MEM (+lo% dialyzed fetal calf serum) in the presence of 20 fig/ ml tunicamycin. Tunicamycin (20 pg/ml) was maintained in the incubation medium of cells during [35S]methionine pulse and during chase time. The viability of tunicamycin-treated cells was superimposable on that of control cells at the end of the incubation. Cells were then solubilized as described above and, to reduce the concentration of Triton X-100 from 2 to 0.1%, the clarified supernatant was chromatographed through a Bio-Beads SM-2 column. The eluted material was concentrated by precipitation with 12.5% polyethylene glycol 6000 and 0.2% 7-globulins (26,31). The pellet was resuspended in 50 mM Hepes buffer, pH 7.4, containing 0.1% Triton X-100, 150 mM NaC1, aprotinin, leupeptin, and soybean trypsin inhibitor (10 pg/ ml each) and samples subjected to immunoprecipitation.
Reduction of Insulin Receptors into a-and @-Subunits-Insulin receptor disulfide bonds were reduced by incubating unlabeled, 35Sor 32P-labeled WGA-purified receptor preparations at 22 "C with 100 mM DTT and 75 mM Tris (pH 8.5) for 30 min using a modification of the method described by Boni-Schnetzler et al. (32). The reaction was stopped by adding NEM (to 100 mM final concentration) and transferring tubes on ice. Then reduced receptors (indicated as "DTT/ Tris-reduced") were chromatographed on a 2-ml GF5 desalting column equilibrated with HTG buffer (50 mM Hepes, 0.05% Triton X-100, 0.5% glycerol, 5 mM NEM) and eluted in the same buffer. This procedure results in a near complete removal of DTT and Tris. Insulin receptor @-subunits were immunoprecipitated with the antipeptide polyclonal antibody PC, with MA-10, or with mAb H 65.6 as described in other sections. Insulin Receptor and Histone Phosphorylation in Vitro-Insulin receptors were partially purified from human placenta and rat liver membranes using WGA-agarose as described (24). Insulin activation and phosphorylation of WGA-purified receptors were conducted as previously reported (24,26). When WGA-purified receptors were phosphorylated prior to being subjected to disulfide bonds reduction, the phosphorylation reaction was stopped by the addition of 10 volumes of a 50 mM Hepes buffer, pH 7.4, containing 5 mM EDTA, 5 mM EGTA, 1 mM unlabeled ATP, 1 mM sodium orthovanadate, 20 mM NaF, 20 pg/ml aprotinin and leupeptin, and 1 mM phenylmethylsulfonyl fluoride. To evaluate the effect of MA-10 on autophosphorylation and protein-tyrosine kinase activity of isolated insulin receptor @-subunit, aliquots of DTT/Tris-reduced receptors were incubated with AbPC and the immune complexes, comprising insulin receptor @-subunit, precipitated with protein A-Sepharose CL4B. The immunoprecipitates were washed, resuspended in HTG buffer, and incubated with 10-j M MA-10 or M normal mouse IgG (in some experiments 10-7 M mAb H 65.6) for 4 h a t 4 "C. [y-32P]ATP (50 mM (final concentration) MnCl2 were then added for 30 min a t 4 "C. The reaction was stopped by the addition of SDS sample buffer and samples subjected to SDS-PAGE followed by autoradiography (24,26). Protein kinase activity of immunoprecipitated insulin receptor @-subunit on histone was assayed in the same reaction mixture as above, except that 0.5 mg/ml histone HFPB was added and incubated 4 h at 4 "C before adding [y-3ZP]ATP. The reaction was stopped by the addition of SDS sample buffer, and histone phosphorylation was analyzed as described in Ref. 24.
Immunoblotting-WGA-purified rat liver receptors were subjected to SDS-PAGE under reducing conditions. The proteins were electrophoretically transferred to Hybond-N nylon membranes and reacted first with a 1:lOO dilution of antipeptide polyclonal antibody AbPC, loF6 M antibody MA-10, or 1:lOO dilution of nonimmune rabbit serum and then with '251-protein A as described in Ref. 33, with modifications according to the filter manufacturer's instructions (Amersham). Filters were washed and subjected to autoradiography.

RESULTS
Antibody MA-10 Binding to Human and Rat Cells-Antibody MA-10 inhibits '251-insulin binding to human hepatoma Hep G2 cells, 50% inhibition occurring at lo-' M, but it does not inhibit '251-insulin binding to rat hepatoma Fao cells even at M concentration both at 4 and 37 "C (Table I)  Binding of MA-10-'251-protein A complexes to the two hepatoma cell lines, rat Fao (0) and human Hep G2 (O), was measured as described under "Experimental Procedures." Cells in six-well dishes were incubated with increasing amounts of MA-10 (from 10"' to 5 X 10-7 M ) and 1.5 X IO5 cpm/well 1251-protein A in the absence (-) or presence

Regulation of Insulin Receptor
Kinase by a Monoclonal Antibody insulin, MA-10 binding to Hep G2 cells is similar to MA-10 binding to Fao cells in the absence or in the presence of insulin (Fig. 1). Superimposable results were obtained performing MA-10 binding experiments at 37 "C instead of 4 "C (data not shown). When Fao and Hep G2 cells are probed with MA-10 in indirect immunofluorescence experiments, in the absence of insulin, both Fao and Hep G2 cells exhibit membrane-associated fluorescence, while coincubation with 10"' M insulin reduces staining of Hep G2 cells but not staining of Fao cells (data not shown). When insulin receptors were down-regulated in both Fao and Hep G2 cells, l2'1-insulin and MA-lO-"'I-protein A binding were reduced to the same extent in both Fao and Hep G2 cells, thus suggesting that MA-10 binds to insulin receptor in these cell lines (Table 11).

Effect of Antibody MA-10 on Insulin Receptor Phosphorylation, Thymidine Incorporation into DNA, Insulin Receptor
Internalization, and Down-regulation-MA-10 antibody decreases insulin-stimulated protein tyrosine kinase activity of WGA-purified rat insulin receptors without affecting insulin binding (24). To understand whether this phenomenon occurs also when insulin receptors are located in the cell membrane and to define the mechanism by which MA-10 reduces insulin receptor autophosphorylation, the effect of this antibody on insulin receptor phosphorylation in intact cells was studied.
M, for 10 min at 37 "c) stimulates 32P incorporation into the insulin receptor @-subunit 9-fold compared to control cells incubated with both normal mouse IgG (Fig. 2, A and B) or with the irrelevant monoclonal IgG2B H 65.6 (Fig. 2C). MA-10 reduces insulin-stimulated '*P incorporation into insulin receptor @-subunit in a dose-dependent manner, half-maximal effect being a t IO-" M (Fig. 2). *' P incorporated into the insulin receptor @-subunit was partially resistant to alkaline hydrolysis, and MA-10 reduces the incorporation of both alkalisensitive and -resistant '"P into the insulin receptor suggesting that the content of phosphotyrosine as well as phosphoserine residues is reduced in cells incubated with MA-10 ( Fig.  2, A and R). It is noteworthy that MA-10 M ) does not affect basal (not insulin-activated) insulin receptor autophosphorylation (data not shown) as already demonstrated in partially purified receptors (Ref. 24 and Fig. 8).
In order to verify whether the inhibition of insulin receptor autophosphorylation produced by MA-10 is relevant in terms of insulin action, the effect of MA-10 on insulin-stimulated thymidine incorporation into DNA of Fao cells was studied. In rat hepatoma Fao cells insulin (lo-" M ) stimulates thymidine incorporation into DNA 2.5-fold and MA-10 (10" M) inhibits insulin action by 75% (Table 111). MA-10 inhibition of insulin-stimulated thymidine incorporation into DNA is dose-dependent with half-maximal effect at lo-" M (data not shown). On the contrary MA-10 does not affect basal thymidine incorporation into DNA of Fao cells (Table 111). The functional relevance of MA-10 inhibition of insulin receptor phosphorylation was investigated also measuring the effect of the antibody on insulin-induced receptor downregulation. Fao cells were exposed to lo-' M insulin in the presence of M mAb H 65.6 (control cells) or lo-' MA-10, for 20 h in complete medium, following which the residual binding activity was measured. The insulin binding activity on Fao cells exposed to insulin was reduced by about 40%, while that of cells incubated with insulin plus MA-10 was reduced by IO%, 10" M MA-10 reduces the insulin-induced receptor down-regulation by 75%. Control experiments showed that MA-10 does not down-regulate insulin receptor in Fao cells per se but, at the same concentration, inhibits insulin effect on receptor down-regulation (Table 111). As previously reported (6), MA-10 induces insulin-independent internalization of human insulin receptors. However MA-10 fails to produce the same effect in rat cells: in fact the amount of internalized MA-10-12'I-protein A complex is less than 10% of the total antibody bound to Fao cells after 15 min a t 37 "C, thus suggesting that the epitope of insulin receptor molecule, to which MA-10 binds in rat cells, is not involved in antibody-

TABLE I11 Effect of MA-IO on thymidine incorporation into DNA and on insulin receptor down-regulation
Fao cells, in 24-well dishes, were incubated in DME containing 0.5% dialyzed USA for 30 h a t 37 "C. The irrelevant monoclonal IgG Next, the possibility that MA-10 immunoprecipitates isolated insulin receptor @-subunits was investigated. Interchain disulfide bonds of insulin receptors, WGA-purified from human or rat hepatoma cells, were reduced using a modification of the DTT/Tris method reported by Boni-Schnetzler et al. (32). As depicted in Fig. 4 this procedure yields isolated insulin receptor @-subunits that are immunoprecipitated by an antipeptide antibody directed against the carboxyl-terminal 17 amino acids of the insulin receptor (AbPC). As shown in Fig.  4 antibody MA-10 immunoprecipitates reduced insulin receptor @-subunits of insulin receptors WGA-purified from [' "SS] methionine-labeled Fao cells. In experiments performed using reduced receptors WGA-purified from human Hep G2 cells, MA-10 immunoprecipitates both a-and @-subunits (Fig. 4). Reduction of class 2 disulfide bonds is not 100% complete: in fact discrete amounts of a/@ dimers are present and are immunoprecipitated by MA-10. The small amount of rat insulin receptor a-subunit evident in the MA-10 immunoprecipitates of reduced rat receptors (Fig. 4) probably depends on the coprecipitation of a-subunits linked to @-subunits through hydrogen bonds or other noncovalent bonds. How-

66-
ever the ratio of a to @ subunits immunoprecipitated from rat DTT-reduced receptors by MA-10 and AbPC (as measured by densitometer quantitation of autoradiograms) was similar (0.21 uersus 0.23), while the ratio of a-to @-subunits immunoprecipitated from human DTT-reduced receptors by MA-10 was -3-fold higher (0.70) (Fig. 4).
In order to exclude the possibility that MA-10 immunoprecipitates rat insulin receptor @-subunits by cross-reacting with the small amount of noncompletely reduced a-subunits, also present in our DTT-reduced preparations, we performed immunoblotting experiments probing WGA-purified rat liver receptors with MA-10 or with the anti-insulin receptor @subunit antipeptide antibody AbPC. As demonstrated in Fig.  5, both MA-10 and AbPC react with a major band of approximately M, 95,000 that corresponds to insulin receptor @subunits. No immunoreactivity is evident at the molecular weight corresponding to insulin receptor a-subunits (Fig. 5).
Finally MA-10 immunoprecipitates in vitro phosphorylated insulin receptor @-subunits purified from both human placenta and rat liver (data not shown).
Antibody MA-10 Immunoprecipitates Nonglycosylated Insulin Receptors-Fao cells were incubated with tunicamycin to inhibit protein glycosylation and labeled with ["S]methionine. Cells were solubilized and the clarified supernatant, methionine for 30 min, washed, and exposed to medium containing 10 mM unlabeled methionine as described under "Experimental Procedures." Insulin receptors were WGA-purified and subjected to the disulfide bond reduction procedure detailed under "Experimental Procedures." The neutralized and alkylated receptors were immunoprecipitated using the polyclonal antibody AbPC (1:lOO) or a preimmune rahbit serum (as indicated in the figure) and the immunoprecipitated material subjected to SDS-PAGE (6% polyacrylamide) under nonreducing conditions. The autoradiogram was exposed for 36 h at -70 "C. The autoradiogram of AbPC immunoprecipitate was scanned densitometrically. The absorbance of the -130 kDa band was 825 arbitrary units, the absorbance of the -95 kDa band was min, washed, and exposed to DME containing 10 mM methionine for 8 h, as described under "Experimental Procedures," in the presence of tunicamycin (20 pglml). Cells were solubilized, Triton X-I00 concentration in the clarified supernatant was lowered by chromatography on a SM-2 Bio-Beads column, and the "S-labeled proteins concentrated by polyethylene glycol precipitation as described under "Experimental Procedures." Insulin receptors were then immunoprecipitatedwith MA-IO (2 X M) or with AbPC (1:lOO final dilution), as indicated in the figure, and analyzed by SDS-PAGE (6% polyacrylamide) under reducing conditions. In order to localize the exact position of mature (glycosylated) insulin receptor a-and &subunits, an aliquot of insulin receptors immunoprecipitated by MA-10, as shown in Fig. 4C, was run in parallel (lost lane on the right). The autoradiogram was exposed 24 h at -70 "C, molecular mass markers are shown on the left of the figure. munoprecipitated with AbPC. Autophosphorylation and histone kinase activity of immunoprecipitated material was measured in the presence of normal mouse IgG (1O"j M) or antibody MA-10 (10" M). As shown in Fig. 7, MA-10 inhibits both insulin receptor beta subunit autophosphorylation and phosphorylation of histone HFZB. Similar results were ob-  (lanes 3 and 6). Insulin receptors, reduced and immunoprecipitated, were incubated for 4 h a t 4 "C with normal mouse IgG M) (lanes 1 and 4 ) , lo-' M MA-10 (lanes 2 and 5) in the absence (lanes 1-3) or presence (lanes [4][5][6] of histone HFLB. Then phosphorylation was conducted as detailed under "Experiment a l Procedures," and samples were analyzed by SDS-PAGE (6 and 15% polyacrylamide in the right and left panels, respectively) under nonreducing conditions. Autoradiograms were exposed for 30 h a t -70 "C with an intensifying screen. Arrows mark the position of insulin receptor &subunit and histone. tained using @-subunits purified from rat liver receptors as source of kinase. In some experiments the nonrelevant monoclonal antibody H 65.6, used as negative control, did not affect either insulin receptor @-subunit autophosphorylation or histone kinase activity. As expected, insulin does not stimulate autophosphorylation of reduced (?-subunits (data not shown).
Finally it is noteworthy that MA-10 reduces basal (not insulin-stimulated) autophosphorylation of insulin receptor @-subunits while it does not decrease, a t any concentration, basal autophosphorylation of the native receptor (Fig. 8). DISCUSSION We previously reported that anti-insulin receptor monoclonal antibody MA-10 affects both insulin binding and insulin-stimulated receptor protein-tyrosine kinase activity in partially purified human insulin receptors, but that it inhibits this enzymatic activity without inhibiting insulin binding in partially purified rat insulin receptors (24).
The aim of this study was to investigate the mechanism by which MA-10 inhibits insulin receptor protein-tyrosine kinase activity. First the MA-10 binding site to rat insulin receptor was characterized. Data presented herein indicate that antibody MA-10 binds to rat hepatoma Fao cells. By both MA-lO-'*'I-protein A complex binding and indirect immunofluorescence MA-10-specific binding to rat as well as to human cells was demonstrated. The fact that MA-10 does not reduce insulin receptor binding and, conversely, insulin does not inhibit MA-10 binding to Fao cells suggests that, in rat cells, MA-10 does not recognize the insulin binding site of insulin receptor molecule. In human hepatoma Hep G2 cells, MA-10 interacts with the insulin binding site since it inhibits insulin binding as insulin (on a molar basis) (27), and furthermore saturating concentrations of insulin reduce MA-10 binding.
In the presence of M insulin, MA-10 binding to Hep G2 cells is reduced to the same amount found in Fao cells both in the absence and in the presence of insulin. The demonstration that insulin receptor down-regulation is associated with a reduction of both '2'I-insulin and MA-lO-'*'I-protein A complex binding to Fao cells suggests that MA-10 binds to an. extracellular portion of insulin receptor in rat cells.
MA-10 also immunoprecipitates both phosphorylated and native, '?3-labeled rat insulin receptors. This finding is only apparently in contrast with data reported by Forsayeth et al. (27). These authors reported that MA-10 does not immunoprecipitate purified rat insulin receptors. In fact using a relatively low antibody concentration (lo-' M), the amount of immunoprecipitated rat receptors is 10-15-fold less than the amount of immunoprecipitated human insulin receptors. Optimizing immunoprecipitation conditions and exposing gels for a convenient period of time (up to 3 days a t -70 "C), it was possible to detect discrete immunoprecipitated material a t a molecular weight exactly corresponding to insulin receptor @-subunit using WGA-purified rat liver insulin receptors. It is noteworthy that MA-10 interactions (binding and immunoprecipitation) with the native rat insulin receptor a t submaximal antibody concentrations occur with a 10 times lower efficiency compared to MA-10 interactions with native human insulin receptors. Using saturating antibody concentrations (2 X M), MA-10 immunoprecipitates approximately the same amount of insulin receptors from [3sSS]methionine-labeled rat and human hepatoma cells.
MA-10 immunoprecipitates both a-and @-subunits in DTT-reduced human receptor preparations but reacts only with @-subunits of reduced rat receptors in both immunoprecipitation and immunoblotting experiments. This fact indicates that MA-10 recognizes human and rat insulin receptor @-subunits. While species specificity of the insulin binding site of the receptor is well documented (37-39), data presented here give further evidence to our previous observation (24) that a site involved in insulin receptor protein-tyrosine kinase activity regulation, recognized by MA-10, is conserved through rat and human species.
Data discussed above suggest that MA-10 binds to a region of insulin receptor molecule common to both human and rat cells and that this region is not the receptor insulin binding site. The insulin receptor region recognized by MA-10 belongs to the protein backbone of the molecule because MA-10 immunoprecipitates not completely glycosylated receptors from tunicamycin-treated rat hepatoma Fao cells. Experiments were carried out to explore the functional relevance of the insulin receptor region bound by MA-10 in intact rat cells. MA-10 inhibits insulin-stimulated receptor autophosphorylation in intact rat cells without affecting insulin receptor binding. In intact cells a 10-min exposure to insulin leads to the insulin receptor autophosphorylation in both tyrosine and serine residues (40). Recent evidence suggests that after tyrosine autophosphorylation insulin receptors become substrates for a putative serine kinase (1) and that only insulin receptors phosphorylated in tyrosine residues can undergo to serine phosphorylation (6). Our data, demonstrating that MA-10 similarly reduces 32P incorporation into both tyrosine and serine residues of insulin receptor, @-subunits are concordant with this last finding. The inhibition of insulin-stimulated receptor phosphorylation has biological significance: at the same concentrations, MA-10 reduces to a similar extent insulin-stimulated receptor autophosphorylation, thymidine incorporation into DNA, and insulin-induced receptor down-regulation in rat hepatoma Fao cells. Recent data, obtained in cells expressing human insulin receptor partially (13) or completely (4-9) devoid of protein-tyrosine kinase activity by site-directed mutagenesis, demonstrate that insulin receptor enzymatic activity is essential for cell transduction of insulin receptor signal, thus confirming previous indirect data obtained by injection of antiinsulin receptor monoclonal antibodies into cells (36). Our present findings further substantiate this evidence utilizing nonengineered receptors and avoiding any cell manipulation.
Interacting with its binding site insulin induces internalization of its receptors (41). In human cells MA-10 binding to insulin receptor induces receptor internalization (6). In this paper we demonstrate that MA-10 binding to rat insulin receptors is not followed by insulin receptor internalization. This finding suggests that the interaction of the same ligand (MA-10) with different sites of the insulin receptor molecule can result in different biological effects.
MA-10 also inhibits autophosphorylation and protein-tyrosine kinase activity of the isolated insulin receptor @-subunit. This last finding indicates that the protein-tyrosine kinase activity of isolated insulin receptor @-subunit can be modulated independently of a-subunit, as suggested by Herrera et al. (16) and that MA-10 binding to isolated @-subunit reduces the basal activity of the enzyme. Recently it has been suggested that the a-subunit of the insulin receptor inhibits protein-tyrosine kinase activity, and, consequently, the physiological responses mediated by the enzyme and the insulin relieves this inhibition (15). The demonstration that the digestion of insulin receptor a-subunit by trypsin leads to the activation of insulin receptor autophosphorylation in intact cells agrees with this hypothesis (42, 43). Data presented in this report give more support to this hypothesis: MA-10 does not reduce kinase activity of the @-subunit in the native receptor, that is under the inhibitory effect of a-subunit, but has an inhibitory effect when the @-subunit (protein-tyrosine kinase) is isolated. These findings suggest that MA-10 can substitute the a-subunit for the inhibition of insulin receptor @-subunit protein-tyrosine kinase activity.
In conclusion, MA-10 antibody recognizes the extracellular portion of the insulin receptor @-subunit in rat cells or tissues. Data presented above indicate that (i) the extracellular part of the insulin receptor &subunit recognized by MA-10 is more conserved, through human and rat species, than the insulin binding domain, and (ii) the extracellular part of the insulin receptor @-subunit is involved in the regulation of insulin receptor phosphorylation and in the transduction of insulin biological actions.