Genistein differentially inhibits postreceptor effects of insulin in rat adipocytes without inhibiting the insulin receptor kinase.

Genistein, an isoflavone putative tyrosine kinase inhibitor, was used to investigate the coupling of insulin receptor tyrosine kinase activation to four metabolic effects of insulin in the isolated rat adipocyte. Genistein inhibited insulin-stimulated glucose oxidation in a concentration-dependent manner with an ID50 of 25 micrograms/ml and complete inhibition at 100 micrograms/ml. Genistein also prevented insulin's (10(-9) M) inhibition of isoproterenol-stimulated lipolysis with an ID50 of 15 micrograms/ml and a complete effect at 50 micrograms/ml. The effect of genistein (25 micrograms/ml) was not reversed by supraphysiological (10(-7) M) insulin levels. In contrast, genistein up to 100 micrograms/ml had no effect on insulin's (10(-9) M) stimulation of either pyruvate dehydrogenase or glycogen synthase activity. We determined whether genistein influenced insulin receptor beta-subunit autophosphorylation or tyrosine kinase substrate phosphorylation either in vivo or in vitro by anti-phosphotyrosine immunoblotting. Genistein at 100 micrograms/ml did not inhibit insulin's (10(-7) M) stimulation of insulin receptor tyrosine autophosphorylation or tyrosine phosphorylation of the cellular substrates pp185 and pp60. Also, genistein did not prevent insulin-stimulated autophosphorylation of partially purified human insulin receptors from NIH 3T3/HIR 3.5 cells or the phosphorylation of histones by the activated receptor tyrosine kinase. In control experiments using either NIH 3T3 fibroblasts or partially purified membranes from these cells, genistein did inhibit platelet-derived growth factor's stimulation of its receptor autophosphorylation. These findings indicate the following: (a) Genistein can inhibit certain responses to insulin without blocking insulin's stimulation of its receptor tyrosine autophosphorylation or of the receptor kinase substrate tyrosine phosphorylation. (b) In adipocytes genistein must block the stimulation of glucose oxidation and the antilipolytic effects of insulin at site(s) downstream from the insulin receptor tyrosine kinase. (c) The inhibitory effects of genistein on hormonal signal transduction cannot necessarily be attributed to inhibition of tyrosine kinase activity, unless specifically demonstrated.

Genistein, an isoflavone putative tyrosine kinase inhibitor, was used to investigate the coupling of insulin receptor tyrosine kinase activation to four metabolic effects of insulin in the isolated rat adipocyte. Genistein inhibited insulin-stimulated glucose oxidation in a concentration-dependent manner with an IDe0 of 25 pgfml and complete inhibition at 100 pgfml, Genistein also prevented insulin's (lo-' M) inhibition of isoproterenol-stimulated lipolysis with an ID60 of 15 pgfml and a complete effect at 50 pg/ml. The effect of genistein (25 pg/ml) was not reversed by supraphysiological (IO-? M) insulin levels. In contrast, genistein up to 100 pg/ml had no effect on insulin's ( M) stimulation of either pyruvate dehydrogenase or glycogen synthase activity. We determined whether genistein influenced insulin receptor &subunit autophosphorylation or tyrosine kinase substrate phosphorylation either in vivo or in vitro by anti-phosphotyrosine immunoblotting. Genistein at 100 pg/ml did not inhibit insulin's (10" M) stimulation of insulin receptor tyrosine autophosphorylation or tyrosine phosphorylation of the cellular substrates pp185 and pp60. Also, genistein did not prevent insulin-stimulated autophosphorylation of partially purified human insulin receptors from NIH 3T3/HIR 3.5 cells or the phosphorylation of histones by the activated receptor tyrosine kinase. In control experiments using either NIH 3T3 fibroblasts or partially purified membranes from these cells, genistein did inhibit platelet-derived growth factor's stimulation of its receptor autophosphorylation. These findings indicate the following: (a) Genistein can inhibit certain responses to insulin without blocking insulin's stimulation of its receptor tyrosine autophosphorylation or of the receptor kinase substrate tyrosine phosphorylation. ( b ) In adipocytes genistein must block the stimulation of glucose oxidation and the antilipolytic effects of insulin at site(s) downstream from the insulin receptor tyrosine kinase. (c) The inhibitory effects of genistein on hormonal signal transduction cannot necessarily be attributed to inhibition of tyrosine kinase activity, unless specifically demonstrated.
* This work was supported by National Institutes of Health Grants DK 28144 and DK 08484. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. T To whom all correspondence should be addressed Dept. of Pathology and Laboratory Medicine, University of Pennsylvania School of Medicine, 3400 Spruce St., 6 Gates, Philadelphia, PA 19104. Tel.: 215-662-6880;Fax: 215-349-5039. The predominant theory of insulin action proposes that the interaction of insulin with its receptor causes tyrosine autophosphorylation of the receptor /3-subunit resulting in activation of the receptor tyrosine kinase (1-3). The tyrosine kinase then initiates the activation and inactivation of various kinases and phosphatases, leading to the pleiotropic effects of insulin. Evidence that the tyrosine kinase activity of the insulin receptor was required for insulin action came primarily from two types of studies (4-7). First, site-directed mutagenesis of the insulin receptor at either the ATP binding site (4, 5) or selected tyrosine residues (6) yielded mutant insulin receptors which failed to autophosphorylate and to mediate certain responses to insulin when expressed in Chinese hamster ovary cells or Rat1 fibroblasts. Second, insulin receptors with defective tyrosine kinase activity have been detected in insulin-resistant patients (7). Analysis of the biological function of such tyrosine kinase-defective receptors demonstrated that the insulin receptor tyrosine kinase domain must be intact for hormone-dependent signal transduction.
Although the data obtained from biochemical and mutational analysis of insulin receptors clearly indicate that mutations in the receptor kinase domain impair insulin receptor function, concluding that activation of the receptor tyrosine kinase activity alone triggers all responses to insulin may be an overinterpretation of the data. Such an interpretation does not consider the possibility that mutant receptors are nonfunctional because of conformational changes that prevent noncovalent coupling of the hormone receptor complex to effector proteins. That changes in receptor conformation may be important for receptor signal transduction is suggested by the observation that anti-insulin receptor antibodies can mimic insulin action without stimulating receptor autophosphorylation or tyrosine kinase activity (8-10) and that conformational changes of the insulin receptor /3-subunit occur after insulin binding (11,lZ). Additionally, two insulin receptors with defective tyrosine kinase activity have been isolated from insulin-resistant diabetics, but only one of these receptors fails to mediate certain insulin bioeffects when expressed in Chinese hamster ovary cells whereas the other kinasedefective receptor signals normally (13). Finally, Gottschalk (14) recently compared the effect of insulin on activation of pyruvate dehydrogenase in cell lines expressing normal human insulin receptors or those deficient in tyrosine kinase activity. He found that both normal and kinase-deficient receptors could mediate the activation of pyruvate dehydrogenase to similar levels. His findings show that the insulin signaling pathway that activates pyruvate dehydrogenase bypasses the insulin receptor tyrosine kinase.
Until recently, the role of insulin receptor tyrosine phosphorylation and substrate tyrosine phosphorylation in signal transduction was studied in cells transfected with cDNAs encoding wild-type or mutant receptors. By necessity, these studies utilized recipient cell lines that were insensitive to insulin and expressed low levels of endogenous receptors (1-3946 Genistein and Insulin Action 3947 3), two properties that distinguish these cell lines from bonafide targets of insulin action such as muscle and adipose tissue. In addition, although the assumption was made that the recipient cell lines expressed the signal transduction machinery necessary to couple an occupied insulin receptor to a cellular response, it is now known that these cell lines do not express insulin-sensitive glucose transporter (GLUT4), which is a molecular marker of muscle and adipose tissue (15). Therefore, data obtained from the study of tyrosine kinasedeficient insulin receptors in transfected cells may not be directly applicable to physiological targets of insulin. The identification and synthesis of inhibitors of protein tyrosine kinases (16-18) allow a new biochemical approach to study the role of tyrosine phosphorylation in insulin action. This approach permits assessment of the potential anti-insulin effects of these inhibitors in physiologically meaningful target cells such as the rat adipocyte. We undertook such a study with the antibiotic genistein, which inhibits both EGF' (18) and PDGF (19) receptor tyrosine kinase activity. We found that genistein inhibited insulin-stimulated glucose oxidation and blocked insulin's inhibition of isoproterenol-stimulated lipolysis but had no effect on insulin's stimulation of pyruvate dehydrogenase or glycogen synthase activity. However, genistein did not block insulin receptor or substrate tyrosine phosphorylation either in vivo or in vitro. These results indicate that beyond the insulin receptor there are multiple pathways for carrying out the pleiotropic effects of insulin and that in the adipocyte, genistein works downstream from the insulin receptor tyrosine kinase and not on the enzyme itself. These studies show that the inhibitory effects of genistein on hormonal signal transduction cannot necessarily be attributed to inhibition of tyrosine kinase activity, unless specifically demonstrated.

EXPERIMENTAL PROCEDURES
Chemicals-Phenylmethanesulfonic acid, leupeptin, pepstatin, dithiothreitol, Triton X-100, Nonidet P-40, and prestained molecular weight standards were from Sigma. Collagenase was from Worthington. Bovine serum albumin was from U. S. Biochemical Cop. Porcine insulin was from Lilly. Genistein, '251-protein A, and Cytoscint-ES were from ICN. 14C-Labeled glucose and ['4C]pyruvate were from Amersham Corp. Male Sprague-Dawley rats were from Hilltop Laboratories. Me,SO and immobilized protein A beads (Trisacryl) were from Pierce Chemical Co. SDS and reagents for SDS-PAGE were from Bio-Rad. Nitrocellulose (BA85, 0.2 pm) was from Schleicher & Schuell. Anti-phosphotyrosine antibody, affinity purified from rabbit, was prepared as described previously by Rothenberg et al. (20). All other chemicals were of reagent grade or better and were purchased from standard vendors.
Isolation of Adipocytes-Male Sprague-Dawley rats (100-150 g) were fed ad libitum with Purina Laboratory Rodent Chow (5012). The epididymal fat pads were removed from rats killed by cervical dislocation and the cells isolated by the method of Rodbell (21) as modified by Jarett et al. (22).
Metabolic Responses to Insulin-Glucose oxidation, lipolysis, pyruvate dehydrogenase activity, and glycogen synthase activity were assayed as described previously (22). A stock solution of genistein was prepared in MezSO and refrigerated. All whole cell incubations were performed in a final concentration of 0.1% MezSO, which was found to have no effect on the insulin responsiveness or basal metabolism of the adipocytes (data not shown). Genistein was added to suspensions of adipocytes either simultaneously with, or 15 min prior to, the addition of other agents. Incubations were performed at 37 'C in the buffers indicated in the figure legends. Incubations were stopped as described previously (22).
(128 mM NaCl, 25 mM MOPS, 5 mM KCl, 5 mM NaH,PO,, 1.5 mM MgS04, 1.5 mM CaCl2) buffer, pH 7.4, with 1% BSA, 200 mg % dextrose, and 0.1% MezSO for 15 min at 37 "C with or without 100 pg/ml genistein. Insulin was added for 2 min, and the total incubation mixture was extracted by the method of Rothenberget al. (20). Briefly, the cell suspension was added to solubilization buffer (2% SDS, 100 mM HEPES, pH 7.8, at 22 "C, 100 mM NaCl, 10 mM EDTA, 50 mM dithiothreitol, and 2 mM Na3V04) at 100 'C. The mixture was homogenized in a Brinkmann Polytron for 10 s at setting 10. The homogenate was heated at 100 "C for 3 min. It was cooled to 22 "C and then centrifuged for 1 h in a Beckman type 45 rotor (143,000 X g,,,) at 18 "C. The supernatant was acidified with 100% trichloroacetic acid to a final concentration of 10%. After 15 min on ice, the precipitate was recovered by centrifugation at 5,000 rpm in a Beckman JA-20 rotor for 5 min at 4 "C. Each pellet was washed three times with ethanol diethyl ether (1:l v/v) at 4 "C. The precipitate was dried in vacuo overnight and then pulverized into a fine powder and used immediately for immunoprecipitation.
The method of Rothenberg et al. (20) was used for the immunoprecipitation of phosphotyrosyl proteins. Briefly, the dry protein precipitate was dissolved in 0.1 N NaOH, neutralized to pH 7.8 with 100 mM Tris-HC1, then made to 1 mM EDTA and 0.02% NaN, and filtered through a 0.45-pm-pore diameter polyvinyl chloride filter (Millex-HA, Millipore Corp.). Anti-phosphotyrosine antibody was added and incubated at 4 "C for 4-16 h. The antibody was absorbed to protein A beads for 2 h at 4 "C. The immunocomplexes were washed twice in 1% Triton X-100, 0.1% SDS, 100 mM NaC1, 50 mM Tris, pH 7.8. The pellet was rewashed in the same buffer but without NaC1. The immunocomplexes in the pellet were solubilized in 50 pl of SDS-PAGE sample buffer (23) containing 50 mM dithiothreitol at 100 "C for 3 min.
Solubilized immunoprecipitated proteins were separated by SDS-PAGE by the method of Laemmli (23) using 0.75-mm, one-dimensional (5 or 6% T acrylamide) gels in a Bio-Rad miniature slab gel apparatus (Mini-Protean) at constant voltage (175 V). Electrotransfer of the separated proteins to a nitrocellulose membrane was performed for 2 h at constant voltage (100 V) and between 5 and 15 "C in the Mini-Protean transfer apparatus as described previously (20). After blocking in 5% fatty acid-free BSA, 1% ovalbumin, 10 mM Tris, pH 7.2, 100 mM NaC1, 0.02% NaN3 the blot was incubated with polyclonal rabbit anti-phosphotyrosine antibodies diluted in blocking buffer (1.5 pg/ml) for 2 h at 22 "C and washed. The blot was then incubated with 0.5 pCi/ml lZ5I-protein A (6-30 pCi/pg) for 1 h at 22 "C and then washed as described previously (20). Bound antiphosphotyrosine antibodies were detected by autoradiography using Kodak XAR film with a Cronex Lightning Plus intensifying screen at -70 "C for 12-72 h.
Determination of Insulin Receptor Tyrosine Kinase Activity in Vitro-A wheat germ agglutinin-enriched fraction of insulin receptor was prepared by the method of Cuatrecasas et al. (24) as modified by Soler et al. (25) from murine fibroblasts transfected with an expression vector containing human insulin receptor cDNA (NIH 3T3/HIR 3.5) (26). Receptor prepared from 3 X lo5 cells was incubated for 10 min at 30 "C in 30 p1 containing a final concentration of 33.3 mM HEPES, pH 7.4, 10 mM MgC12, 3 mM MnCIZ, 100 ~L M Na3V0,, 1% MezSO, 50 mM N-acetyl-D-glucosamine, 25 mM NaCl, 0.017% Triton X-100, 0.017% BSA, and in the presence or absence of 25 mg of histone H2b, lo-? M insulin, in the presence or absence of 10-100 pg/ ml genistein. At 22 "C, Na,ATP was added to a final concentration of 1 mM, and after 1 min the reaction was stopped by the addition of an equal volume of SDS-PAGE sample buffer (2 X ) and heating to 100 "C for 2 min. SDS-PAGE and immunoblotting with anti-phosphotyrosine antibodies were carried out as described above.

RESULTS
The Effect of Genistein on Insulin Action-Experiments were designed to determine if genistein, a known tyrosine kinase inhibitor (18,19), blocked insulin action in isolated adipocytes. Four acute responses to insulin were studied stimulation of glucose oxidation, inhibition of isoproterenolstimulated lipolysis, activation of pyruvate dehydrogenase, and activation of glycogen synthase.
The effect of genistein on the stimulation of glucose oxidation was the first response examined. As shown in Fig. 1, genistein caused a dose-dependent inhibition of insulin's abil- ity to stimulate glucose oxidation in adipocytes. This effect of genistein occurred with an IDso of approximately 25 wg/ml and was maximal at a concentration of 100 pg/ml.
As illustrated in Fig. 2, genistein did not affect the rate of isoproterenol-stimulated lipolysis. Genistein, however, did block insulin's counterregulatory effect on the lipolytic response of adipocytes to isoproterenol. Genistein caused a dosedependent inhibition of insulin action with a maximum at 50 pg/ml and at an IDsrr of 15-20 pg/ml, similar to that described for glucose oxidation in Fig. 1. The data presented in Fig. 3 demonstrate that increasing the concentration of insulin to a level that should saturate cell surface receptors did not overcome the inhibitory effect of genistein on insulin action. The results shown in Fig. 3 also demonstrate the biphasic nature of the insulin dose-response curve and show that the apparent insulin sensitivity of the isolated adipocytes was unaltered by an intermediate dose of genistein.
Two lines of evidence argue against general cytotoxicity as an explanation for genistein's inhibition of insulin's stimulation of glucose oxidation or inhibition of isoproterenol-stimulated lipolysis. Adipocytes incubated in the presence or absence of genistein showed no difference in cellular ATP levels (data not shown). In addition, genistein failed to inhibit the  At the end of the incubation, the adipocytes were lysed, and pyruvate dehydrogenase activity in the lysates was determined as described (19). The data from two experiments that were performed in triplicate were pooled.
Each bur represents the mean * S.D. of six samples. ability of adipocytes to respond to isoproterenol, which signals through a protein kinase A-dependent pathway.
We next tested the effect of genistein on the ability of insulin to stimulate the activity of pyruvate dehydrogenase and glycogen synthase. As shown in Fig. 4, the antibiotic at concentrations up to 100 wg/ml had no effect on insulindependent activation of pyruvate dehydrogenase. Concentrations of genistein as high as 250 pg/ml did not block the activation of pyruvate dehydrogenase by insulin in isolated adipocytes (data not shown). This result initially suggested that the activation of pyruvate dehydrogenase was independent of the insulin receptor tyrosine kinase activity. Since transfection studies had established that glycogen synthesis was dependent on insulin receptor tyrosine kinase activity (27), experiments were performed to determine if genistein would block the activation of glycogen synthase. Fig. 5 shows that genistein did not block insulin-dependent activation of glycogen synthase.
Although genistein did not inhibit insulin-dependent activation of either pyruvate dehydrogenase or glycogen synthase, incubation of cells with the drug did lower the basal level of activity for each of these enzymes. The decreased basal activity had little effect on the magnitude of the insulin response for either enzyme. The observation that genistein did not inhibit either insulin-dependent activation of pyruvate dehydrogenase or glycogen synthase also indicated that genistein did not block insulin binding, eliminating this possible mech- At the end of the incubation, the cells were lysed, and the activity of glycogen synthase was determined as described under "Experimental Procedures." Total enzyme activity was determined in the presence of 10 mM glucose 6-phosphate. Active enzyme activity was determined in the presence of 0.1 mM glucose 6-phosphate. The effect of insulin was indicated by an increase in the ratio of active to total enzyme activity. The data from two experiments, which were performed in triplicate, were pooled. Each bur represents the mean f S.D. of six samples.
anism as an explanation for the results described above for the studies of glucose oxidation and isoproterenol-stimulated lipolysis.
The Effect of Genistein on Insulin Receptor Tyrosine Kinase-The experimental survey of genistein's effects on four metabolic responses to insulin revealed that genistein differentially inhibited two metabolic responses that are normally coordinately regulated in adipocytes. These results raised the possibility that the activation of pyruvate dehydrogenase and glycogen synthase was not mediated through the hormonedependent tyrosine kinase activity of the receptor. Alternatively, these results might be explained if genistein acted downstream of the insulin receptor, along a pathway(s) necessary for activation of glucose oxidation and inhibition of isoproterenol-stimulated lipolysis but unnecessary for activation of pyruvate dehydrogenase and glycogen synthase. To resolve these two possibilities, experiments were performed to assess whether genistein inhibited the insulin receptor tyrosine protein kinase and to examine the effect of genistein on insulin-stimulated tyrosine phosphorylation of cellular proteins.
T o investigate the effect of genistein on insulin-dependent protein tyrosine phosphorylation in isolated adipocytes, antiphosphotyrosine antibody was used to immunoprecipitate the phosphotyrosyl proteins in adipocytes that had been incubated with or without insulin (lo-: M) in the presence or absence of 100 pg/ml genistein. Immunoprecipitated phosphotyrosyl proteins were separated by SDS-PAGE and detected by anti-phosphotyrosine immunoblotting. As shown in Fig. 6, insulin stimulated protein tyrosine phosphorylation in isolated adipocytes. Three distinct insulin-sensitive protein bands were detected, and these bands migrated with apparent molecular masses of 185, 95, and 60 kDa. The 95-kDa band is identified as the @-subunit of the insulin receptor (20). The 185and 60-kDa proteins have been observed previously, and both are likely direct cellular substrates of the insulin receptor tyrosine kinase (20). The 120-kDa band constitutively contains phosphotyrosine and is unaffected by insulin in other tissues (20). The preincubation of adipocytes with genistein a t concentrations that completely inhibited insulin-dependent activation of glucose oxidation and insulin-dependent inhibition of isoproterenol-stimulated lipolysis had no significant 97 -5s.
~~6 0 FIG. 6. The effect of genistein on insulin-dependent protein tyrosine phosphorylation in isolated adipocytes. Lanes are shown from an autoradiogram of a Western blot of samples immunoprecipitated with anti-phosphotyrosine antibody and immunostained with the same antibody. Samples represent phosphoprotein from -2.5 X IO6 cells. C, no additions; G, 100 pg/ml genistein; G + I, 100 pg/ml genistein plus 10" M insulin; I, 10" M insulin. Adipocytes (-IOfi cells/ml) were incubated in Krebs-Ringer MOPS buffer, pH 7.4, with 1% BSA, 200 mg % dextrose for 15 min at 37 "C with or without 100 pg/ml genistein or an equal volume of M e 8 0 in the control (0.1% final). Insulin was added for 2 min, and the total incubation mixture was extracted and processed as described under "Experimental Procedures." Samples were run on a 5% T gel. ZR, insulin receptor. effect on insulin-stimulated receptor @-subunit (95-kDa) autophosphorylation. With both genistein and insulin present, the insulin receptor @-subunit band was 83.4% & 9.2% (mean f S.E., n = 8) as intense as this same band measured in the presence of insulin alone. Similarly, no consistent inhibitory effects of genistein were observed on the degree of tyrosine phosphorylation of pp185 or pp60. These results suggest that in the intact adipocyte, genistein has little, if any, effect on insulin activation of the insulin receptor tyrosine kinase or the tyrosine phosphorylation of the insulin receptor @-subunit and other endogenous substrates such as pp185 and pp60. As a positive control, parallel procedures were used to test the effect of genistein on PDGF receptor tyrosine kinase activity in NIH 3T3 mouse fibroblasts, a cell whose growth is regulated by PDGF (28). PDGF at 100 ng/ml markedly stimulated the tyrosine phosphorylation of the receptor. Genistein, in a dosedependent manner, inhibited PDGF-stimulated tyrosine phosphorylation of the PDGF receptor with almost complete inhibition a t 100 pg/ml genistein (data not shown). These findings are similar to those reported previously (19).
Genistein has been shown to inhibit the protein tyrosine kinase activity of the EGF receptor (18) and the PDGF receptor (19), both in uiuo and in uitro. T o determine whether genistein inhibited the hormone-dependent protein tyrosine kinase activity of the isolated insulin receptor, insulin receptors were extracted from NIH 3T3/HIR 3.5 mouse fibroblasts, a cell line stably transfected with a cDNA encoding the normal human insulin receptor. These partially purified receptors were used to study the effect of 100 pg/ml genistein on receptor protein tyrosine kinase activity in uitro. After a 1min incubation with 1 mM ATP, the receptor preparation was separated by SDS-PAGE and immunoblotted with anti-phosphotyrosine antibody. As shown in Fig. 7, insulin (lo-' M ) markedly stimulated tyrosine phosphorylation of the 95-kDa band, the @-subunit of the insulin receptor. Genistein had no effect on this tyrosine phosphorylation. In addition, insulin stimulated the tyrosine phosphorylation of added histones, but genistein at the same concentration had no effect on this exogenous substrate phosphorylation reaction (data not shown). As an additional positive control, a partially purified membrane preparation from NIH 3T3 cells was used to test the effect of genistein on PDGF receptor phosphorylation, using methods similar to those used for partially purified insulin receptors. In this in uitro system, PDGF markedly stimulated PDGF receptor phosphorylation, and this phos- G, 100 pg/ml genistein; G + I, 100 pg/ml genistein plus 10" M insulin; I, 10" M insulin. The receptor was incubated in a final concentration of 33.3 mM HEPES, pH 7.4, 10 mM MgCl?, 3 mM MnC12, 100 p M NanV04, 1% Me2S0, 50 mM N-aCetyl-D-glUCOSamine, 25 mM NaC1, 0.017% Triton X-100, and 0.017% BSA in the presence or absence of insulin and/or genistein at 30 "C. After 10 min Na?ATP was added to 1 mM. After 1 min the reaction was stopped by adding an equal volume of SDS-PAGE sample buffer (2 X), boiled for 3 min, run on a 6% T gel, immunoblotted, and stained with anti-phosphotyrosine antibodies as described under "Experimental Procedures." phorylation was partially inhibited by 10 pg/ml genistein and completely inhibited a t 100 pg/ml (data not shown). These findings are similar to those reported previously (19).

DISCUSSION
The current hypothesis of insulin action proposes that interaction of insulin with its receptor leads to phosphorylation of the 0-subunit of the receptor with subsequent activation of the receptor tyrosine kinase activity. Presumably, the receptor kinase then phosphorylates specific cellular substrates, leading to the regulation of various kinases and phosphatases that, in turn, mediate many of the pleiotropic effects of insulin (1-3). Most of the data supporting this hypothesis are derived from the study of mutated receptors expressed in cultured cell lines that are not normal targets for insulin action. As discussed in the Introduction, several lines of evidence raise questions as to whether all of the actions of insulin involve an active receptor tyrosine kinase. One recent study (14) clearly showed that in cells expressing kinasedeficient mutant human receptors, the insulin signaling pathway for activation of pyruvate dehydrogenase bypasses the insulin receptor tyrosine kinase. Another study (13) showed that a kinase-defective insulin receptor derived from an insulin-resistant diabetic patient responded normally to insulin while another kinase-defective receptor did not. Furthermore, certain anti-insulin receptor antibodies mimic insulin action without stimulating tyrosine kinase activity (8-10).
The present study was designed to determine if in a physiological target cell, the rat adipocyte, all of the actions of insulin were dependent on an active insulin receptor tyrosine kinase. We approached this study using a putative tyrosine kinase inhibitor, genistein, and studied its effects on four well documented actions of insulin on the adipocyte. These effects included stimulation of glucose oxidation as a measure of glucose transport, the antilipolytic effect on isoproterenolstimulated lipolysis, activation of pyruvate dehydrogenase, and activation of glycogen synthase.
We found that genistein did inhibit both insulin stimulation of glucose oxidation and insulin's ability to block isoproterenol-stimulated lipolysis. In contrast, genistein had no effect on insulin-stimulated pyruvate dehydrogenase or glycogen synthase activity even at the highest concentrations examined. These findings suggested either that the effects of insulin on the latter two processes bypassed the insulin receptor tyrosine kinase or that genistein did not inhibit the insulin receptor tyrosine kinase. The latter finding would suggest that genistein was working downstream from the receptor along a signal transduction pathway(s) that regulates glucose transport and antilipolysis but is not involved in insulin's effects on pyruvate dehydrogenase and glycogen synthase.
The data in the present study clearly show that genistein had no significant effect on insulin receptor 0-subunit autophosphorylation or tyrosine kinase phosphorylation of the pp185 and pp60 substrates either in intact adipocytes or with partially purified insulin receptors in vitro. These observations were surprising since genistein has been shown to inhibit EGF receptor (18) and PDGF receptor (19) tyrosine kinase activities both in intact cells and in vitro. Similarly, genistein blocks CD3 activation of the T cell receptor-associated tyrosine kinase (29). Schecter et al. (16) have also shown that other inhibitors that block certain actions of insulin in adipocytes also blocked the tyrosine kinase activity of isolated liver membrane insulin receptors. Fujita-Yamaguchi and Kathuria (30) reported that two agents, staurosporine and apigenin, which affected biological responses to insulin, blocked the tyrosine kinase activity of insulin receptors purified from human placenta. Genistein has been shown to be a competitive inhibitor of ATP binding to the catalytic domain of tyrosine kinases (18), a highly conserved structural feature among all tyrosine kinases, including the insulin receptor. Perhaps because the insulin receptor forms heterodimers (3), in contrast to the monomeric structure of the other receptors, the isoflavone genistein is sterically hindered from access to the ATP binding site on the insulin receptor. However, the insulin receptor of adipocytes behaves differently from insulin receptors in other cell types (31). For example, it is aggregated prior to and independent of hormone occupancy and undergoes internalization through noncoated invaginations. Therefore, the lack of genistein's effect in the adipocyte may have represented a cell type-specific phenomenon. This possibility is excluded by genistein's lack of inhibition of isolated human insulin receptor tyrosine kinase activity in vitro. The lack of effect of genistein on insulin receptor tyrosine kinase is even more unique since several recent reports have shown that genistein's ability to inhibit responses to growth factors correlates with a decrease in the level of cellular tyrosine phosphorylation (32)(33)(34)(35)(36)(37). In contrast, in a number of other studies (38)(39)(40)(41)(42)(43)(44), genistein has been shown to block biological effects without demonstration that the antibiotic inhibited protein tyrosine phosphorylation. The present data caution against assuming a causal correlation without direct evidence supporting that assumption. Such caution is further warranted by the recent report of Faaland et al. (45) showing that tyrphostin, another putative tyrosine kinase inhibitor, was not a competitive inhibitor of the EGF receptor tyrosine kinase in intact A431 cells, and it blocks hormone action by an indirect mechanism.
Genistein most likely blocks insulin stimulation of glucose oxidation and inhibition of isoproterenol-stimulated lipolysis by inhibiting a process (or processes) downstream from the insulin receptor tyrosine kinase but required for both of these biological effects. Likely targets of genistein's action are protein kinases, in which this antibiotic is known to compete with ATP binding at the catalytic site of these enzymes. Moreover, serine/threonine-specific protein kinases are strongly implicated in postreceptor insulin signaling mechanism (3). Studies have shown recently that genistein can block S6 kinase activity (46) and topoisomerase I and I1 activity (47)(48)(49) by competing with ATP for binding. Thus, genistein as well as other putative tyrosine kinase inhibitors must be used carefully in trying to dissect the various kinases