Characterization of insulin-mediated phosphorylation of the insulin receptor in a cell-free system.

Insulin stimulates phosphorylation of both alpha- and beta- subunits of its own receptor in a cell-free system. A solubilized lectin-purified preparation of insulin receptors from rat liver membranes was preincubated with or without insulin at 4 degrees C and labeled for 10 min with Mn[gamma- 32P]ATP; the receptor subunits were isolated by specific immunoprecipitation with anti-receptor antibodies, followed by gel electrophoresis in sodium dodecyl sulfate. In gels run under reduced conditions, two bands (Mr = 135,000 and 95,000) were selectively labeled. These correspond exactly to the position of the alpha- and beta-subunits of the insulin receptor. Labeling of the Mr = 95,000 band was approximately 5-fold that of the Mr = 135,000 band. No labeled bands were detected when identical samples were immunoprecipitated in control serum. Phosphorylation of the receptor subunits required the presence of the divalent cation Mn2+ or Co2+; other cations such as Mg2+, Cr3+, Ca2+, and Zn2+ were ineffective. [gamma- 32P]ATP served as the 32P donor, whereas [gamma- 32P]GTP was ineffective. Phosphorylation of both subunits was stimulated 4-6-fold after a 60-min exposure to 10(-7) M pork insulin. Insulin-stimulated phosphorylation was half-maximal after 5 min of incubation with 10(-7) M insulin or after 18 h with 3 X 10(-10) M hormone. The enhanced phosphorylation was specific for insulin and its analogs; guinea pig insulin was about 2% as potent as pork insulin, whereas epidermal growth factor, adrenocorticotropic hormone, and glucagon, as well as cAMP, were ineffective. The rapidity and specificity of this reaction, as well as the presence of all necessary components in the plasma membrane, suggest that insulin-mediated receptor phosphorylation is one of the earliest biochemical steps following insulin binding.

In a target cell, insulin bound to its receptor ultimately leads to a large array of molecular events (1)(2)(3)(4). While the binding of insulin to the receptor as well as some of the metabolic and growth-promoting effects of the hormone are well characterized, the biochemical events that link these early and late events are unclear. Among the intermediate steps noted are both phosphorylation and dephosphorylation of several cell proteins (2, 3, 5), as well as generation of a * 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   peptide-like soluble second messenger (6,7 ) .
It has been previously demonstrated that EGF (8) and acetylcholine (9) stimulate the phosphorylation of their own receptors. Recently, we showed that insulin stimulates phosphorylation of its own receptor in intact cells (10) as well as in a cell-free system (11, 1.2). In the intact cell, this stimulation resulted from both the increase in the content of phosphoserine and the appearance of phosphotyrosine (13); whereas, in the broken cells, there was an increase in phosphotyrosine only (11). Phosphorylations of tyrosine groups in proteins are known to be involved in both virus transformation (14) and growth stimulation with polypeptide growth factors. For example, EGF stimulates the phosphorylation of a tyrosine residue of its own receptor (15), and platelet-derived growth factor induces the phosphorylation of tyrosine residues of membrane proteins (16).
In order to gain a better insight into the mode of insulin's action to stimulate phosphorylation of the insulin receptor, we have further characterized the effect using solubilized lectin-purified rat liver membranes. We find that

Effect of Insulin on
Receptor Phosphorylation-As previously described (11,12), incubation of the partially purified insulin receptors for 10 min at 4 "C in the presence of Mn[y-"2P]ATP followed by specific immunoprecipitation of the receptor with anti-receptor antibodies revealed two major ."P-   The precipitates were adsorbed to Protein A, washed, and subjected to electrophoresis in 7.5% polyacrylamide slab gel in the presence of 0.1% sodium dodecyl sulfate. The gel was stained, destained, dried, and autoradiographed as described under "Experimental Procedures." 95K ( Fig. 2, lane D). Preincubation of the receptors for 2 h at 4 "C with 1 X 10" M pork insulin produced an approximately 5-fold increase in the "lP incorporation into these two bands3 (Fig. 2, lanes E and F). The position of the 135K-and 95Klabeled bands was indistinguishable from the position on similar gels of the a-and P-subunits, respectively, of the insulin receptor, as determined previously with receptor subunits labeled by various techniques (17,18). No 32P labeling of these bands was detected when identical phosphorylated samples (treated with or without insulin) were immunoprecipitated by a control serum (11).

Insulin Receptor Phosphorylation
Effect of Duration of Preincubation with Insulin on the Extent of Receptor Phosphorylation-In our previous studies (ll), preincubation of receptors with insulin was for 18 h. As seen in Fig. 3, insulin M) stimulated incorporation of into both the 135K and 95K subunits of the receptor after as little as 1 min of preincubation at 4 "C, with half-maximal stimulation by 4 min; maximal phosphorylation was approached with a 30-min preincubation, and the effect of the hormone was changed little when the preincubation with insulin was extended for as long as 18 h at 4 "C. In a separate experiment (data not shown), insulin (or buffer) and [y-"PI ATP was added simultaneously. In the presence of insulin, phosphorylation of the receptor was higher than the basal level but not as high as that observed with a 1-min preincubation. That 1 min or less of preincubation with insulin was sufficient to enhance receptor phosphorylation suggests to us that effects of insulin on phosphorylation were largely direct and early. Furthermore, it argues strongly against the possi-"Two additional polypeptides (210K and 240K) were frequently detected, which were also specifically immunoprecipitated by antireceptor antibodies and whose phosphorylation was stimulated by insulin. These were not characterized during this study. Water (15 pl) was added to A. Ten min following addition of the salts, the phosphorylation was initiated by adding 135 pl of [y-'"PIATP to a final concentration of 5 p~. Phosphorylation was carried out for 10 min at 4 "C and terminated by adding 250 pl of stopping solution.
The samples were further processed as described in the legend to Fig.  2. Similar results were obtained when Mn(CH3COO)z was substituted for MnCI2, or when the MgCh was omitted from Buffer I. bility that insulin acted solely to prevent denaturation or destruction of the receptor, i.e. to preserve the receptor as a substrate for an insulin-independent phosphorylating activity.
Time Course of the Phosphorylation Reaction-The incorporation of 32P into both receptor subunits (Fig. 4) was rapid even a t 4 "C, similar to the phosphorylation of other membrane proteins (19); the 32P content of receptor reached halfmaximal levels by 5 min and leveled off after 10 min but showed no fall even after 180 min (data not shown). The reaction was linear for only 3 min, possibly because it was impossible to ensure that the reaction was conducted under saturating concentrations of the substrate (i.e. dephosphoform of receptor). Characterization of the Nucleotides a n d Cations-When we measured the rate of [y-'"PIATP hydrolysis during the phosphorylation reaction, we found that only 20% of the total [y-32P]ATP was released as free "P within the initial 10 min of the reaction, both in the presence and absence of insulin (Fig. 5). Therefore, depletion of ATP during the fist 10 min of the reaction did not seem to limit the extent of phosphorylation. These findings also suggest that the enhanced receptor phosphorylation in the presence of insulin was not a result of inhibition of an ATPase activity by insulin. The reason for the sharp leveling off of the time curve (Fig. 5) is not clear yet. Incorporation of "*P into the receptor subunits was markedly inhibited when the specific activity of the [y-"lP]ATP was reduced 2000-fold (Fig. 2, lane G). Furthermore, a linear increase in "P incorporation occurred when we used increasing concentrations of [y-"'P]ATP a t a constant specific activ-it^.^ These findings argue strongly against the possibility that some labeled contaminant was the '12P source rather than ATP. When we compared [y-"PIATP and [y-:'2P]GTP as phosphate donors, it could be demonstrated (Fig. 2) that, in both basal and insulin-stimulated states, [y-,"2P]GTP was not utilized as a "'P donor. This is in contrast to the EGF-mediated phosphorylation of EGF receptors where GTP can substitute for ATP (19).

FIG. 7.
Comparing the effect of insulin and insulin analogs on receptor phosphorylation. The experiment was carried out as described in the legend to Fig. 2. Purified soluble receptors (0.2 mg) were preincubated for 18 h (right) or 1 h (left) with the indicated stimulant at the given final concentration, and the amount of 32P incorporated into the 135K band (top) or 95K band (bottom) was determined as described in the legend to Fig. 3  As with the EGF-stimulated phosphorylation of the EGF receptor (19), Mn2+ is the most potent cation in stimulating the phosphorylation of the insulin receptor, with Co2+ being considerably less potent (Fig. 6). In contrast to the EGF (19) and acetylcholine (9) systems, Mg2+ (even at 20 mM) was totally ineffective, as were Zn'+, Ca2', and Cr3+.
The Dose Response of the Insulin-mediated Phosphorylation-The effect of insulin on receptor phosphorylation was dose-dependent (Fig. 7). Following an 18-h preincubation of hormone with the receptor, half-maximal stimulation occurred with 3 X 10"' M insulin, a concentration which is within the physiological range of insulin concentrations in uiuo. Shortening the duration of the preincubation from 18 to 1 h at 4 "C shifted the dose response curve 20-fold to the right.
This shift of the dose response curve may be explained by the slow binding of insulin to its receptor at 4 "C; as seen in Figure  8, approximately 10% of the amount of insulin bound to the receptor by 24 h is bound after 1 h of preincubation.
Note that receptor phosphorylation is stimulated at insulin concentrations which are more than 1 order of magnitude less than those required for occupancy of the receptor. Half-maximal binding (18 h, 4 "C) occurred in the presence of 60-100 n g / d (1-1.6 X M) of insulin (Fig.  l ) , whereas halfmaximal phosphorylation (under similar conditions) occurred after preincubation with 2-3 ng/ml (3-5 X 10"' M) of insulin (Fig. 7). A remarkably similar phenomenon, which has been attributed to "spare receptors", is associated with biological responses to insulin (20, 21) or other hormones (22).
Hormonal Specificity for Receptor Phosphorylation-The effect of insulin on the phosphorylation of both receptor subunits was specific for insulin ( Figs. 7 and 9). CAMP, as well as other hormones, including EGF, had no effect on the phosphorylation of the insulin receptor. Guinea pig insulin, which is 2% as potent metabolically as pork insulin (23), was 2% as potent in stimulating receptor phosphorylation (Fig. 7). Furthermore, guinea pig insulin manifested the same timeand dose-dependent shifts as pork insulin. That guinea pig insulin at 10" M did not inhibit the stimulatory effect of 10" M pork insulin when both were preincubated together with the receptor (Fig. 7) suggests that effects of guinea pig insulin, as in other systems, are related solely to its reduced affinity for the insulin receptor. Desoctapeptide insulin, which has a potency that is 1% or less than that of pork insulin in stimulating glucose metabolism (23) and in competing for the binding of 1251-insulin to the receptors, produced only minimal stimulation of phosphorylation (Fig. 7). Similarly, multipli-cation stimulating activity (an insulin-like growth factor (23)) had only minimal effect (11). Thus, the specificity for the phosphorylation was consistent with the affinity of the hormone for the insulin receptor.

DISCUSSION
Insulin-regulated protein phosphorylation has been studied extensively in several cell-free systems (24-27). For example, insulin causes dephosphorylation of the a-subunit of mitochondrial pyruvate dehydrogenase in a mixture of mitochondria and plasma membrane (26). Direct addition of insulin and [y3'P]ATP to rat liver plasma membranes triggers CAMP-dependent phosphorylation of three peripheral membrane proteins including the low K,,, CAMP phosphodiesterase (25). In a cell-free system derived from 3T3-Ll adipocytes, insulin stimulates 32P incorporation from [y-"'P]ATP into the ribosomal protein S6 (27).
In this study, using detergent-solubilized partially purified insulin receptors, we show that insulin markedly stimulates phosphorylation of its own receptor. Our results correlate with results of previous studies using intact cells (10). Receptor phosphorylation is rapid, specific, and occurs a t concentrations of insulin comparable to those in the circulation in uiuo. All of the necessary components of this reaction appear to be present in the plasma membrane. These findings lead us to hypothesize that receptor phosphorylation may be an early step in insulin action.
Insulin-stimulated incorporation of 3'P into the receptor subunits represents the balance between the activities of protein kinases and phosphatases. We favor the idea that insulin acts by activation of protein kinase(s) rather than inhibition of protein phosphatase(s) because there was no reduction in the 32P content of the receptor subunits, typical of the action of a protein phosphatase (19), even when the phosphorylation reaction at 4 "C was extended to 3 h. Moreover, we have not detected phosphatase activity when prephosphorylated receptors were incubated in the presence of freshly purified receptors, with or without i n s~l i n .~ Our observations indicate that, even after extensive purification of the insulin receptor (i.e. solubilization and lectin affinity chromatography), the essential components of the phosphorylation machinery remain as a functional unit. The components of the unit are glycoproteins or peptides bound to glycoproteins. They include at least an insulin binding site, a tyrosine protein kinase, and the phosphorylated domain(s) of the receptor subunits.

Insulin Receptor
The phosphorylation reaction described in this report has several features common to other phosphorylation systems. For example, there are many similarities between the insulinstimulated and EGF-stimulated phosphorylations of their respective receptors (8,19): ( a ) occurrence at 4 "C with soluble receptor preparations; ( b ) incorporation of 32P into tyrosine residues; ( e ) independence from CAMP; (d) requirement for divalent cations, with Mn" being the most effective, Co2+ being much less potent, and Ca" being ineffective; and (e) occurrence without detectable effect of the hormone on the rate of ATP hydrolysis. However, note that these two systems differ in their receptor substrates, receptor binding sites, and some key features of their phosphorylation machinery. The cation and nucleotide requirements for the insulin system seem to be narrower since GTP and Mg2' were ineffective as substrates in the insulin-stimulated receptor phosphorylation but were effective in the EGF system. Thus, it seems that EGF and insulin utilize, at least in part, different enzymatic machineries to mediate the phosphorylation of their own receptors.
The possibility must therefore be entertained that an early event following insulin binding is a specific phosphorylation of a tyrosine residue of the insulin receptor. This reaction is catalyzed by a protein kinase(s), presumably distinct from the better known EGF-, Ca2+-, or CAMP-dependent protein kinases. Whether the same or different kinases phosphorylate the a-and ,&subunits of the insulin receptor is still unclear.
Since maximal phosphorylation occurs at relatively low receptor occupancy, one could speculate that, in the solubilized preparation, both the free and insulin-bound receptors serve as substrates for the putative insulin-stimulated protein kinase. In the intact cells, however, this might not necessarily be the case, and the relative locations of the receptors and kinase(s) in the membrane will determine the availability of free receptors for phosphorylation.
Although phosphorylation of proteins as means of regulating their function is a well known phenomenon (28,29), no data are yet available on the functional implications of the phosphorylation of the insulin receptor. Among several possibilities, phosphorylation could alter the affinity of insulin for its binding site, trigger aggregation and internalization of the receptor subunits, or play a role in coupling insulin binding to insulin bioactivity.