Insulin Action Inhibits Insulin-like Growth Factor-I1 (IGF-11) Receptor Phosphorylation in H-35 Hepatoma Cells IGF-I1 RECEPTORS ISOLATED FROM INSULIN-TREATED CELLS EXHIBIT ENHANCED IN VITRO PHOSPHORYLATION BY CASEIN KINASE 11*

Insulin caused a rapid, dose-dependent increase in the binding of '261-insulin-like growth factor-I1 (IGF-11) to the surface of cultured H-36 hepatoma cells. The ["2P]phosphate content of the IGF-I1 receptors, immu- noprecipitated from extracts of H-36 cell monolayers previously incubated with [82P]phosphate for 24 h, was decreased after brief exposure of the cells to insulin. Analysis of tryptic digests of labeled IGF-11 receptors by bidimensional peptide mapping revealed that the decrease in the content of [s2P]phosphate occurred to varying degrees on three tryptic phosphopeptides. Thin layer electrophoresis of an acid hydrolysate of isolated IGF-I1 receptors revealed the presence of IS2PJ phosphoserine and [s2P]phosphothreonine. Insulin treatment of cells caused a decrease in the labeled phosphoserine and phosphothreonine content of IGF- I1 receptors. The ability of a number of highly purified protein kinases (CAMP-dependent protein kinase, protein kinase C, phosphorylase kinase, and casein kinase 11) to catalyze the phosphorylation of purified IGF-I1 receptors was examined. Casein

Insulin caused a rapid, dose-dependent increase in the binding of '261-insulin-like growth factor-I1 (IGF-11) to the surface of cultured H-36 hepatoma cells. The ["2P]phosphate content of the IGF-I1 receptors, immunoprecipitated from extracts of H-36 cell monolayers previously incubated with [82P]phosphate for 24 h, was decreased after brief exposure of the cells to insulin. Analysis of tryptic digests of labeled IGF-11 receptors by bidimensional peptide mapping revealed that the decrease in the content of [s2P]phosphate occurred to varying degrees on three tryptic phosphopeptides. Thin layer electrophoresis of an acid hydrolysate of isolated IGF-I1 receptors revealed the presence of IS2PJ phosphoserine and [s2P]phosphothreonine. Insulin treatment of cells caused a decrease in the labeled phosphoserine and phosphothreonine content of IGF-I1 receptors.
The ability of a number of highly purified protein kinases (CAMP-dependent protein kinase, protein kinase C, phosphorylase kinase, and casein kinase 11) to catalyze the phosphorylation of purified IGF-I1 receptors was examined. Casein kinase I1 was the only kinase capable of catalyzing the phosphorylation of the IGF-I1 receptor on serine and threonine residues under the conditions of our assay. Bidimensional peptide mapping revealed that the kinase catalyzed phosphorylation of the IGF-I1 receptor on a tryptic phosphopeptide which comigrated with the main tryptic phosphopeptide found in receptors obtained from cells labeled i n vivo with [s2P]phosphate. IGF-I1 receptors isolated by immunoadsorption from insulin-treated H-36 cells were phosphorylated i n vitro by casein kinase I1 to a greater extent than the receptors isolated from control cells. Similarly, IGF-I1 receptors from plasma membranes obtained from insulin-treated adipocytes were phosphorylated by casein kinase I1 to a greater extent than the receptors from control adipocyte plasma membranes. Thus, the insulin-regulated phosphorylation sites on the IGF-I1 receptor appear to serve as substrates in vivo for casein kinase I1 or an enzyme with similar substrate specificity. Previous studies have shown that the IGF-11' receptor is among a number of membrane proteins that are rapidly exposed on the cell surface membrane of target cells in response to insulin. This increase in the cell surface concentration of IGF-I1 receptors is caused by their redistribution from an intracellular membrane pool into the plasma membrane (1-3). The molecular mechanism which mediates this redistribution is unknown. However, it has been shown that IGF-11 receptors continually recycle in the absence of ligand (4). Thus, an increase in the exocytosis of the receptor, a decrease in its rate of internalization, or a combination of both, could cause an increase in the steady-state cell surface concentration of this molecule.
In an effort to gain insight into the mechanisms which maintain the basal steady-state distribution of IGF-I1 receptors, we have studied the phosphorylation state of these receptors in plasma membranes and low density microsomes from rat adipocytes treated with or without insulin. We reported that IGF-I1 receptors derived from the adipocyte plasma membrane were phosphorylated to a higher stoichiometry than the receptors derived from the low density microsomes ( 5 ) . Furthermore, insulin added to intact cells caused a marked decrease in the overall phosphorylation level of receptors in the plasma membrane with a time course that closely paralleled the increase in the cell surface receptor concentration. We have proposed that this insulin-mediated dephosphorylation of the IGF-I1 receptor may decrease its rate of internalization and thus leads to the increase in the steady-state cell surface number of receptors caused by this hormone (5). To test this hypothesis more detailed information on the characteristics of IGF-I1 receptor phosphorylation is required. In this paper we present data which indicate that in another cell type, the H-35 hepatoma cell, the effect of insulin to increase the number of IGF-I1 receptors on the cell surface was also accompanied by a decrease in the phosphorylation state of the receptor. Isotopic labeling experiments with [32P]phosphate revealed that both in freshly isolated adipocytes and in H-35 hepatoma cells, the IGF-I1 receptor was phosphorylated on serine and threonine residues. Furthermore, we report that purified casein kinase I1 catalyzed the phosphorylation of isolated IGF-I1 receptors in uitro. The phosphorylation of IGF-I1 receptors derived from plasma membranes obtained from insulin-treated adipocytes or from insulin-treated H-35 cells was increased compared to the phosphorylation of receptors obtained from control cells. 3117 These results suggest that the insulin-sensitive phosphorylated sites on the receptor may be phosphorylated in vivo by casein kinase 11 or an enzyme with similar substrate specificity.

MATERIALS AND METHODS
Cell Cultures-Rat H-35 hepatoma cells (gift from Dr. Gerald Litwack, Temple University) were grown as monolayers in Dulbecco's modified Eagle's medium containing 5% calf serum, 1000 units/ml of penicillin, and 100 pg/ml of streptomycin. Upon confluence, the cells were subcultured and seeded in 22.6-mm wells for binding studies or in 100-mm dishes for isotopic labeling with [32P]phosphate. After 72 h, when still subconfluent, the medium was replaced with serum-free Dulbecco's modified Eagle's medium, and the cells were grown for another 24 h. After this time, the medium was again replaced with phosphate-free, serum-free Dulbecco's modified Eagle's medium, supplemented with 25 mM Hepes, pH 7.4.
Measurement of IZI-IGF-II and Anti-IGF-XI Receptor IgG Binding-After 24 h in phosphate-free, serum-free medium, cells were stimulated by adding insulin (porcine, Lilly) at the concentrations indicated in the figures. After 10 min at 37 "C the medium was then removed and replaced by an ice-cold buffer (Buffer A) composed of 130 mM NaCl, 5 mM KCI, 1.2 mM CaCl2, 1.2 mM MgSO,, 10 mM Hepes, and 10 mg/ml bovine serum albumin, pH 7.4. The culture wells were placed on ice, and '261-IGF-II (60 Ci/g) was added at a final concentration of 5 nM, in the presence or absence of unlabeled IGF-I1 at a final concentration of 500 nM. After 2 h of incubation on ice, the monolayers were washed three times with 1 ml of cold Buffer A. The cells were then dissolved in 0.5 ml of 1 M NaOH, and the radioactivity waa counted in a Packard y counter. For the study of anti-IGF-I1 receptor I& binding, cell monolayers were incubated with 20 pg/ml of anti-IGF-I1 receptor IgG for 5 h at 4 "C. The monolayers were washed twice with 2 ml of Buffer A and then incubated with fresh Buffer A containing 2 pCi of '?-protein A (Du Pont-New England Nuclear). After 1 h at 4 "C, the monolayers were washed three times, dissolved in 1 N NaOH, and counted in a y counter.
IGF-XI Receptor Phsphotylatwn in Intact Cells-For these experiments, the phosphate-free, serum-free Dubbecco's modified Eagle's medium was supplemented with [32P]phosphate (carrier free, Du Pont-New England Nuclear) at 1 mCi/ml final concentration. After 24 h of incubation, 1.5 X lo4 cells were treated with nothing or insulin at a final concentration of 10 nM for 10 min at 37 "C. The culture dishes (two 100-mm dishes) were placed on ice, the medium was quickly aspirated, and 1.5 ml/dish of an ice-cold lysis buffer containing 25 mM Hepes, pH 7.4, 1.5% Triton X-100, 1% sodium deoxycholate, 0.1% NaDodSO,, 1% bovine serum albumin, 0.5 M NaCI, 50 mM NaF, 100 p M NaVO6, 50 mM pyrophosphate, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride was added. The cell monolayers were scraped off the wells and tranferred into centrifuge tubes. The lysates were vortexed frequently over a 30-min period at 4 "C and then clarified by centrifugation at 4 ' C for 30 min at 100,000 Immunoadsorptwn and NaDodS04-Polyacrylamide Gel Electrophoresis-The H-35 clarified extracts from 1.5 X loT6 control or insulintreated cells were used for each immunoadsorption assay. These were performed using 10 pg of a rabbit polyclonal anti-IGF-I1 receptor immunoglobulin coupled to 5 pl of packed Affi-Gel 10 resin (Bio-Rad) prepared as previously described in detail (6). After 8-12 h at 4 'C, the resin was washed three times with 1 ml of lysis buffer and then once with 25 mM Hepes, 0.1% Triton X-100, pH 7.4. The samples were then boiled for 1 min in 50 mM Tris, 3% NaDodSO,, 0.005% bromphenol blue, 20% glycerol, pH 6.8, and separated on a 0.75-mmthick 6% polyacrylamide slab minigel, according to Laemmli (7). Gels were stained lightly with Coomassie Brilliant Blue, destained overnight, dried, and exposed to Kodak X-Omat AR film at -70 "C using Du Pont Cronex Lightning Plus intensifying screens. Phosphopeptide Mapping and Phosphoaminu Acid Analysis-Gel slices corresponding to the IGF-I1 receptor band were excised and digested with 5 pg of L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin (Worthington) in 80 p l of 0.1 M N-ethylmorpholine-acetate (Sigma), pH 8.3, for 24 h at 37 "C. The eluate was removed and the gel piece was then washed in 100 p1 of fresh Nethylmorpholine-acetate for another 2 h at 37 "C. The combined eluates contained more than 90% of the radioactivity. After drying, the phosphopeptides were resuspended in 7 pl of ice-cold performic acid, in order to convert cysteine residues into their stable cysteic acid derivatives (8). After 150 min at 0 "C, the acid was diluted with 250 pl of ice-cold water, frozen, and lyophilized to dryness. The phosphopeptides were then resuspended in 30% formic acid and spotted on a cellulose-coated thin layer plate (Machery-Nagel). The tryptic phosphopeptides were separated by electrophoresis in 1% ammonium carbonate at 450 V (approximately 13 mA) for 2.5 h, followed by ascending chromatography in 1-butanol/acetic acid/pyridine/water (15:3:1012, by volume). The plates were dried and exposed to Kodak X-Omat film at -70 "C for 3-4 days. Digestion with trypsin for up to 48 h did not increase the recovery of radioactive peptides from the polyacrylamide gel piece or the number of spots resolved by two-dimensional peptide mapping.
For phosphoamino acid analysis, the tryptic peptides eluted from the polyacrylamide gel pieces were lyophilized to dryness and resuspended in 500 pl of 6 N HC1. Hydrolysis was performed for 3 h at 110 "C, after which the hydrolysates were dried and resuspended in 1 mg/ml of phosphoserine, phosphothreonine, and phosphotyrosine (Sigma). Electrophoresis was then performed on cellulose-coated thin layer plates using acetic acid/pyridine/water (101:189 by volume). After drying, the phosphoamino acids were visualized by ninhydrin staining and autoradiography at -70 "C.
Phosphorylatwn of the IGF-II Receptor with Purified Protein Kinases-The IGF-I1 receptors from approximately 1 X lo-' H-35 cells or from rat adipocyte plasma membranes (see below) were solubilized in 1 ml of lysis buffer and immunoadsorbed onto 5 pl of packed Affi-Gel 10 resin. The resin was washed three t i e s with 1 ml of lysis buffer, three times with 1 ml of 25 mM Hepes, pH 7.4, 1.5% Triton X-100, 1% bovine serum albumin, 0.5% NaC1, 5 mM EDTA and then three times with 1 ml of 20 mM Tris, pH 7.4, 1 mM EDTA, 1 mM EGTA. The resin was then suspended in 20 pl of a phosphorylation buffer containing 50 mM Tris, pH 7.5,l mM dithiothreitol (or 0.1% mercaptoethanol for casein kinase 11) 1 mM EDTA, 1 mM EGTA, 6 mM MgC12 (or 3 mM CaC12 + 25 pg/ml phosphatidylserine for protein kinase C), and 0.1 M NaCl (for casein kinase 11).
The enzymes were then added to the resin, and phosphorylation reactions were immediately started with the addition of 2 pl of a [y-32P]ATP stock (200 pM, 10 pCi/mmol, Amersham Corp.). After 60 min on ice, the reactions were stopped by the addition of Laemmli sample buffer. After boiling the samples for 1 min, the IGF-I1 receptor was separated on NaDodSO,, 6% polyacrylamide slab gels. Electrophoresis was monitored using prestained high molecular weight markers (Bethesda Research Laboratories) and allowed to continue until the 42-kDa marker (ovalbumin) reached the bottom of the gel. This allowed the electroelution of the unreacted [y-"PIATP and a better separation of the IGF-I1 receptor from the top of the separating gel. The kinases tested were: CAMP-dependent protein kinase catalytic subunits (Sigma), phosphorylase kinase (Sigma), protein kinase C from bovine retinae (9), and casein kinase I1 from rabbit skeletal muscle, which was purified as described previously (10).
In some experiments, IGF-I1 receptors from adipocyte plasma membranes were used as a substrate for casein kinase 11. For these studies, cells were isolated by collagenase digestion from rat epididymal fat pads as described by Rodbell (11). Cells were then treated with lo-* M insulin for 10 min at 37 "C, after which the plasma membranes were separated by differential centrifugation as previously described (12), but in a buffer containingphosphatase inhibitors (5). The membranes (500 pg) were solubilized and the IGF-I1 receptor was immunoadsorbed and used as a substrate for casein kinase I1 as described above. To quantitate the relative amounts of receptor immunoadsorbed from control or insulin-treated cell membranes, IGF-I1 receptor were transferred onto nitrocellulose paper (0.45 p~, Schleicher & Schuell) immediately after electrophoresis. The immunoblots were then treated sequentially with anti-IGF-I1 receptor I& and '261-protein A (Du Pont-New England Nuclear) as previously described (5).

RESULTS
It was important to assess whether or not the effect of insulin on IGF-I1 receptor phosphorylation occurred in cell types other than the rat adipocyte. The H-35 hepatoma cell has been shown to respond to insulin with increases in the uptake of IGF-I1 and incorporation of [3H]thymidine into DNA (14). Fig. 1 shows an experiment in which H-35 cells were treated with different concentrations of insulin for 10 min at 37 "C. Following this incubation, the cells were placed on ice, and the specific binding of '251-IGF-II was measured. Because at this low temperature endocytosis and exocytosis of receptors is blocked, the results reflect the number of IGF-I1 receptors present on the cell surface membrane. Insulin caused an approximately 2-fold increase in the binding of IGF-I1 to the surface of these cells (Fig. lA). The insulininduced increase in the binding of IGF-I1 could be due to an increase in the number of receptors on the cell surface or to an increase in the affinity of the receptor for IGF-11. To distinguish between these two possibilities, the effect of insulin on the binding of an anti-IGF-I1 receptor polyclonal antibody to the surface of the cells was investigated. Fig. 1B shows that the binding of the antibody to insulin-treated cells was increased approximately 2-fold compared to controls.
These results indicate that the increased binding of lZ5I-IGF-I1 is due to an increase in the number of IGF-I1 receptors on the cell surface in response to insulin.
In order to investigate if this effect of insulin was accompanied by a change in the phosphorylation state of the IGF-I1 receptor, H-35 cells were labeled to constant specific activity with [32P]phosphate. Cells were then stimulated with insulin for 10 min and the IGF-I1 receptor from cell extracts was immunoprecipitated and separated on a 6% polyacrylamide gel. The culture dishes were then placed on ice, and '%IGF-II binding (A) or anti-IGF-I1 receptor IgG binding ( B ) was measured as described under "Materials and Methods." The nonspecific binding was defined as the radioactivity remaining when 500 nM of unlabeled IGF-I1 was included in the binding assay, and was typically 40% of the total radioactivity bound. Nonspecific antibody binding was assessed using IgG from a nonimmunized rabbit and represented -20% of the binding observed in the presence of anti-IGF-I1 receptor IgG. The nonspecific binding was subtracted from the total. The results presented are the mean of triplicates from one experiment, which was repeated twice, with similar results. In this specific experiment, 2003 and 1758 cpm were obtained in the receptors from control or insulin-treated cells, respectively. The bands were digested with trypsin, and 20% of the sample (400 and 350 cpm for control or insulin, respectively) was applied to the thin layer plates. The phosphopeptides were resolved by electrophoresis and chromatography as described under "Materials and Methods." The origin is indicated (0). weight bands that were detected in addition to the 250-kDa IGF-I1 receptor band showed variation from one experiment to another and were very prominent in one experiment in which phenylmethylsulfonyl fluoride was omitted. Therefore, they most probably represent proteolytic fragments of the receptor, which contain the phosphorylated sites within the molecule. The amount of 32P associated with the IGF-I1 receptor was quantitated by Cerenkov counting of the excised bands. Insulin caused a consistent decrease in the amount of [32P]phosphate present in the total cellular IGF-I1 receptor pool. The results from five different experiments, performed in succession, were analyzed individually, and the results were expressed as a percentage of the radioactivity present in the control samples. It was found that the radioactivity present in the receptor obtained from insulin-treated cells was 82 & 6% (mean +. S.E., r = 5) of that present in the receptors obtained from control cells which was 2527 & 676 cpm/band (mean k S.E.). Thus; these results indicate that the decrease in the phosphorylation state of the IGF-I1 receptor caused in response to insulin is small but statistically significant. Further analysis of the phosphorylation sites in the receptor was performed. For this purpose, IGF-I1 receptors isolated from 32P-labeled H-35 cells were digested with trypsin and analyzed by two-dimensional peptide mapping, using electrophoresis at pH 8.3 in one dimension and ascending thin layer chromatography in the second dimension. The most highly phosphorylated tryptic phosphopeptide (indicated by number 1) was characterized by a strong electrophoretic mobility and a weak mobility in ascending chromatography, suggesting a highly charged and polar species. A significant proportion of the radioactivity was found in additional tryptic peptides, labeled 2 and 3. Prolonged exposure of the thin layers revealed the presence of multiple tryptic fragments phosphorylated to a much lower stoichiometry (data not shown).
Insulin treatment caused a significant decrease in the phosphorylation of tryptic phosphopeptides 1 and 3, identified in control cells, as well as the virtual disappearance of phosphopeptide 2. These experiments indicate that in these cells the effect of insulin to increase the cell surface number of IGF-I1 receptors is accompanied by a decrease in the phosphorylation state of the receptor. This decrease is not observed exclusively on one tryptic peptide, indicating that insulin can decrease the 32P content of the IGF-I1 receptor on several of its phosphorylation sites.
The identity of the phosphorylated amino acids in 32Plabeled IGF-I1 receptors was studied by thin layer electrophoresis of an acid hydrolysate of the receptor. Both [32P]phosphoserine and [32P]phosphothreonine could be detected in receptors derived from control cells (Fig. 3). In contrast, the analysis of the labeled receptors from insulin-treated cells revealed only traces of phosphothreonine. It appears from these experiments that the insulin-mediated decrease in IGF-I1 receptor phosphorylation occurs on both serine and threonine residues.
As an approach toward identifying the enzymes that are involved in the phosphorylation and dephosphorylation of the IGF-I1 receptor, the ability of a number of highly purified protein kinases to catalyze phosphorylation of the purified receptor was tested. Among several enzymes employed (CAMP-dependent protein kinase, protein kinase C, phosphorylase kinase, and casein kinase 11), only casein kinase I1 was able to catalyze the phosphorylation of IGF-I1 receptors. Fig.  4 shows the result of an experiment in which casein kinase I1 was incubated for 30 min with an IGF-I1 receptor immunoprecipitate and [yS2P]ATP. The proteins were then separated on a 6% polyacrylamide gel. Autoradiography of such a gel revealed one phosphorylated band of 250 kDa, which corresponds to the molecular weight of the IGF-I1 receptor (A). As  It has been previously shown that the ability of the casein kinases to phosphorylate exogenous substrates is markedly dependent on the previous phosphorylation state of the substrate employed (15)(16)(17). Thus, we investigated whether the changes in the phosphorylation state of the IGF-I1 receptor produced by insulin could alter its ability to serve as a substrate for casein kinase 11. An experiment was performed in which the IGF-I1 receptor was immunoprecipitated from either control or insulin-treated H-35 cells and then incubated with casein kinase I1 and [y3'P]ATP for different times. The results of such an experiment are shown in Fig. 5. It can be seen that the rate of 32P incorporation into IGF-I1 receptors was significantly increased when the receptors were obtained from insulin-treated cells compared to control cells.
It was previously reported that insulin markedly decreased the phosphorylation of the IGF-I1 receptor derived from isolated adipocyte plasma membranes (5). The 32P content of the IGF-I1 receptors derived from insulin-treated adipocyte membranes was only 30-50% of that observed in receptors derived from control cells. This effect was larger than the one observed when studying the total cellular receptor pool in H-35 cells. For this reason, the ability of casein kinase I1 to catalyze phosphorylation of IGF-I1 receptors immunoprecipitated from control or insulin-treated adipocyte plasma membranes was investigated.
The amount of receptors present in the immunoprecipitate was quantitated by immunoblotting because plasma membranes from insulin-treated cells contain 2-3-fold more receptors than membranes from control cells. Fig. 6 shows the autoradiograph of an experiment in which immunoprecipitates from control or insulin-treated plasma membranes were phosphorylated in the presence of casein kinase 11. The immunoprecipitated IGF-I1 receptors were separated on a 6% polyacrylamide gel and then transferred onto nitrocellulose. Following autoradiography, the nitrocellulose was incubated with anti-IGF-I1 receptor IgG and l2'1-protein A. A %fold increase in the [32P]phosphate content of IGF-I1 receptors isolated from insulin-treated cells and incubated with [?-"PI ATP and casein kinase I1 was observed. The number of receptors estimated by the amount of 12' 1 associated with the receptor band was only 2-fold higher. These results indicate that insulin treatment of intact fat cells results in a 50% increase in the amount of [32P]phosphate that can be incorporated into plasma membrane IGF-I1 receptors in uitro by purified casein kinase 11. Figs. 5 and 6 show that IGF-I1 receptors from insulintreated cells were phosphorylated in uitro by casein kinase I1 to a higher stoichiometry than those receptors obtained from control cells. To further understand the nature of this effect, the tryptic phosphopeptide maps of IGF-I1 receptors obtained from control or insulin-treated cells, and subsequently phosphorylated in vitro by casein kinase 11, were compared. In Fig.  7, A and B,  or absence (C) of lo-' M insulin for 10 min at 37 "C. Cells were then homogenized, and plasma membranes were obtained. The membranes were solubilized, and the IGF-I1 receptor was immunoadsorbed and phosphorylated for 30 min with purified casein kinase I1 as described in the legend to Fig. 5. After electrophoresis the receptor was transferred onto nitrocellulose paper. The inset shows an autoradiograph of the nitrocellulose sheet containing the casein kinase I1 phosphorylated IGF-I1 receptor from control or insulin-treated cell membranes. After autoradiography, the nitrocellulose paper was incubated with 20 pg/ml anti-IGF-I1 receptor IgG, followed by '%I-protein A.
The IGF-I1 receptor bands were excised and '9 was determined in a Packard y counter and s2P by Cerenkov counting. The contribution of s2P radiation to the radioactivity determined by y counting was 7% of the Cerenkov counts and was subtracted from each sample.
Background radioactivity was determined for each lane by excising a similar-sized region of the paper below each band and was subtracted from each sample. The bar graphs represent the 32P cpm and the '%I cpm associated with the receptor from control (C) or insulin-treated cell membranes (n. This experiment was repeated twice with similar results. 2) are indicated. It can be observed that a significant proportion of the radioactivity incorporated into IGF-I1 receptors obtained from control cells by purified casein kinase I1 (Fig.  7A) was found in a tryptic phosphopeptide which had identical electrophoretic and chromatographic mobility as phosphopeptide 1 detected in the IGF-I1 receptor obtained from 32Plabeled cells (Fig. 2). Phosphopeptides 2 and 3 were not detectably phosphorylated by casein kinase I1 in receptors from control cells. In addition, casein kinase I1 phosphorylated a number of sites which were not detected in uiuo. The phosphopeptide map of the receptor obtained from insulintreated cells (Fig. 7B) was markedly different. In this case, two of the phosphopeptides (1 and 2) had identical mobilities to those found in IGF-I1 receptors from in uiuo labeled control cells (Fig. 2). In addition, several phosphopeptides, which were not detected in control cell-derived, casein kinase IIphosphorylated IGF-I1 receptors nor in receptors obtained from cells labeled in vivo with [32P]phosphate, could be observed.

DISCUSSION
The data presented in this paper indicate that the effect of insulin to cause an increase in the cell surface number of IGF-I1 receptors in H-35 hepatoma cells was accompanied by a decrease in the phosphorylation state of this receptor. This decrease had been previously observed to occur in isolated rat adipocytes, where IGF-I1 receptors continually recycle be- tween the plasma membrane and an intracellular microsomal membrane pool (4). A steady-state distribution of IGF-I1 receptors between these membrane fractions is maintained in which only 10-20% of the receptors are present on the plasma membrane. Upon addition of insulin, the amount of ["PI phosphate in plasma membrane receptors rapidly decreased to 30-50% of the control value, whereas the phosphorylation of the intracellular receptors was unaffected (5). The insulinmediated decrease in the [32P]phosphate content of the IGF-I1 receptor from H-35 cells was consistent but small. This small decrease may be due to the fact that the total IGF-I1 receptor pool from H-35 hepatoma cells was immunoprecipitated in these experiments. Thus, analogous to observations in adipocytes insulin may cause a selective decrease in the phosphorylation of receptors present exclusively in the H-35 plasma membrane, which may comprise a small proportion of the total receptor population. Methods to purify plasma membranes and intracellular membranes from H-35 cells are not yet available.
Both in H-35 cells (Fig. 3) and in freshly isolated adipocytes labeled with 32P (data not illustrated) the amino acid residues on the IGF-I1 receptor that were found to be phosphorylated were phosphoserine and phosphothreonine. Upon exposure of cells to insulin, a significant decrease in the [32P]phosphothreonine content of the IGF-I1 receptor was observed, together with a smaller decrease in ["P]phosphoserine. Previously it has been shown that, upon incubation of isolated adipocyte plasma membranes with [y3'P]ATP, the activity of a kinase which catalyzed the phosphorylation of IGF-I1 receptors on tyrosine residues (18) could be detected. Furthermore, this tyrosine phosphorylation reaction was decreased in membranes isolated from insulin-treated cells? For this reason, it was interesting to investigate if the tyrosine * Corvera, S., Whitehead, R., Yagaloff, K., and Czech, M. P. (1988) Biochem. J., in press. phosphorylation of the IGF-I1 receptor, which occurs in isolated membranes incubated with [-p92P]ATP, could also be detected in receptors obtained from intact cells labeled in vivo with [32P]phosphate. Under the experimental conditions employed, a significant amount of phosphotyrosine on the "Plabeled receptor could not be detected (Fig. 3). This negative result could be due to several technical reasons, such as the loss of receptor tyrosine phosphate during cell disruption or to the low recovery of tyrosine phosphate during acid hydrolysis. Alternatively, the tyrosine kinase activity of the adipocyte membrane may be under tight regulatory control in intact cells by a mechanism that is lost upon disruption of the cells. These possibilities are currently being explored in an effort to gain insight into the physiological significance of IGF-I1 receptor tyrosine phosphorylation. Exposure of 32P-labeled cells to insulin caused a marked decrease in the phosphothreonine and phosphoserine content of the IGF-I1 receptor. Thus, insulin may cause the inhibition of a serinelthreonine receptor kinase or the activation of a serinelthreonine phosphatase, or both. To gain insight into the enzymes that may catalyze phosphorylation of IGF-I1 receptors in uivo, the ability of a number of purified serine/ threonine kinases to phosphorylate the immunoaffinity purified IGF-I1 receptor was tested. Among several enzymes tested, only casein kinase I1 was able to phosphorylate the immunoprecipitated receptor in vitro. Analysis of a tryptic digest of the receptor phosphorylated by casein kinase I1 revealed that this enzyme phosphorylated the receptor on several sites. One of the peptides observed comigrated on electrophoresis and chromatography with a tryptic peptide derived from the receptor obtained from cells labeled in vivo with ["P]phosphate, suggesting that the receptor may be phosphorylated in vivo by casein kinase I1 or a similar enzyme. Phosphoamino acid analysis of the receptors obtained from intact cells revealed a predominance of phosphoserine over phosphothreonine, whereas the receptor labeled in vitro with casein kinase I1 contained predominantly phosphothreonine over phosphoserine. These results are consistent with the interpretation that the purified kinase cannot phosphorylate those sites that are already phosphorylated on the receptor when it is extracted from the cell (phosphoserine residues), but can readily phosphorylate those sites that are unoccupied with unlabeled phosphate (phosphothreonine residues). This interpretation is also consistent with the finding that the receptors obtained from insulin-treated cells, which contain less phosphate than controls, are phosphorylated to a greater extent in vitro by casein kinase 11. However, it is also possible that other kinases, in addition to casein kinase I1 or a casein kinase 11-like enzyme, may phosphorylate the receptor on serine residues in intact cells.
Relatively little is known about the physiological role of casein kinase 11. This kinase has been identified in numerous mammalian and avian cells, in association with nuclei, membranes, mitochondria, and ribosomes (reviewed in Ref. 19). Casein kinase I1 has been shown to be involved in the phosphorylation of other insulin-sensitive phosphoproteins such as glycogen synthase (20) and phosphatase inhibitor 2 (21). Dephosphorylation of a peptide containing the sites for glycogen synthase kinase 3 on glycogen synthase is apparently responsible for the activation of this enzyme by insulin (22). Also, activation of Mg-ATP-dependent phosphoprotein-phosphatase involves a cyclic phosphorylation-dephosphorylation reaction on a site phosphorylated by glycogen synthase kinase 3 on inhibitor 2 (23). In both these systems, the phosphorylation of the glycogen synthase kinase 3 sites is critically dependent on the previous phosphorylation of the protein substrates by casein kinase I1 (21,24).
More recently casein kinase I1 has been shown to be associated with brain and liver coated vesicles, where it has been shown to phosphorylate stoichiometrically the @-light chain of clathrin (25). This finding raises the possibility that casein kinase I1 may play a role in regulating the interactions among coated vesicle proteins or between the proteins which undergo endocytosis through coated pita. The IGF-I1 receptor may be similar to other membrane receptors (reviewed in Ref. 26) in that ita internalization may involve its movement and concentration into coated pits. It is possible that phosphorylation of the receptor by the casein kinase I1 present in the coated pit might be important in anchoring the receptor to this structure, although no direct supporting data for this postulate is available.
An interesting finding presented in this report is that the amount of casein kinase 11-catalyzed phosphate incorporated into IGF-I1 receptors was significantly greater when receptors from insulin-treated cells were used as the substrate. It has been shown that the casein kinases recognize specific sequences in acidic protein substrates which are affected by the state of prior phosphorylation of the molecule. In the case of phosvitin, optimal incorporation of 3zP is achieved when the substrate is partially dephosphorylated (27). Studies with casein variants have shown that the ability of casein kinase to phosphorylate a specific site is greatly affected by the phosphorylation of the sequence surrounding the site (reviewed in Ref. 19). The tryptic phosphopeptide maps of IGF-I1 receptors obtained from insulin-treated cells and phosphorylated in vitro by casein kinase I1 revealed a number of phosphopeptides which were not detected in receptors derived from control cells and phosphorylated with this enzyme. Thus, it appears that the insulin-mediated decrease in the phosphorylation of IGF-XI receptors on specific sites enhances the ability of casein kinase I1 to catalyze phosphorylation of neighboring sites within the molecule. In addition, two of the sites which are dephosphorylated upon insulin treatment in uiuo (peptides 1 and 2, Fig. 2B) appear to be substrates for casein kinase 11. Therefore, the increased phosphorylation of the IGF-I1 receptor from insulin-treated cells by casein kinase I1 may also be partially due to the replacement of the phosphate groups on the dephosphorylated amino acids with ["PI phosphate in uitro. Studies to directly determine whether or not casein kinase I1 or a simiIar enzyme directly participates in IGF-I1 receptor phosphorylation and its modulation by insulin in uiuo are presently being conducted.