Characterization of the Receptor for Platelet-derived Growth Factor on Human Fibroblasts DEMONSTRATION OF AN INTIMATE RELATIONSHIP WITH A 185,000-DALTON SUBSTRATE FOR THE PLATELET-DERIVED GROWTH FACTOR-STIMULATED KINASE*

The receptor for platelet-derived growth factor (PDGF) on human foreskin fibroblasts has been char-acterized. The molecular weight of the PDGF-receptor complex was estimated by affinity labeling techniques to about 200,000, as determined by sodium dodecyl sulfate-gel electrophoresis performed under reducing conditions. Subtraction of the M, of reduced PDGF (18,000 to 15,000) gives a M, for the receptor proper of 185,000 (210,000). The mobility in sodium dodecyl sulfate-gel electrophoresis was similar whether or not reducing agents were present, suggesting that the receptor may be a single chain protein. The hydrody- namic size of the lZ5I-PDGF-receptor complex after solubilization with Triton X-100, corresponded to a M,. of -320,000, as determined by gel chromatography. Subtraction of the M, contributions from Triton X-100 and PDGF, respectively, gives a M, of -200,000 for the receptor itself, an estimate in good agreement with the value obtained from the affinity-labeling experi- ments. Several lectins were analyzed for their ability to inhibit binding of “‘1-PDGF to its receptor. It was found that wheat germ agglutinin and a lectin from Crotalaria juncea were effective inhibitors and that their inhibitory effects could be neutralized by N-ace-tylglucosamine and galactose, respectively, suggesting that the receptor contains these sugars. The properties of the receptor were compared with those of a 185,000-Da and incubation prolonged for another 10 min. Samples were then analyzed by SDS-gel electrophoresis. The pelleted beads were first washed 3 times with 0.2% Triton X-100. 20% glycerol, 50 mM NaCI, 20 mM Tris, pH 7.4, in the centrifuge and then subjected to phosphorylation conditions as described above. The material adsorbed to the beads was then desorbed by incubation for 3 min at 95 "C in SDS-sample buffer and analyzed hy SDS-gel electrophoresis. The amount of "P radioactivity associated with the 185-kDa component was quantified by densitometric scanning of the autoradiogram.

The receptor for platelet-derived growth factor (PDGF) on human foreskin fibroblasts has been characterized. The molecular weight of the PDGF-receptor complex was estimated by affinity labeling techniques to about 200,000, as determined by sodium dodecyl sulfate-gel electrophoresis performed under reducing conditions. Subtraction of the M , of reduced PDGF (18,000 to 15,000) gives a M , for the receptor proper of 185,000 (210,000). The mobility in sodium dodecyl sulfate-gel electrophoresis was similar whether or not reducing agents were present, suggesting that the receptor may be a single chain protein. The hydrodynamic size of the lZ5I-PDGF-receptor complex after solubilization with Triton X-100, corresponded to a M,. of -320,000, as determined by gel chromatography.
Subtraction of the M, contributions from Triton X-100 and PDGF, respectively, gives a M , of -200,000 for the receptor itself, an estimate in good agreement with the value obtained from the affinity-labeling experiments. Several lectins were analyzed for their ability to inhibit binding of "'1-PDGF to its receptor. It was found that wheat germ agglutinin and a lectin from Crotalaria juncea were effective inhibitors and that their inhibitory effects could be neutralized by N-acetylglucosamine and galactose, respectively, suggesting that the receptor contains these sugars.
The properties of the receptor were compared with those of a 185,000-Da component, being the major substrate for the membrane-bound PDGF-stimulated kinase. It was found that the 185,000-Da component behaved similar to the PDGF receptor in sodium dodecyl sulfate-gel electrophoresis, performed with or without reducing agents present. Further, the 185,000-Da component co-eluted with the PDGF receptor on a Sepharose 6B column, and had affinity for the same lectins that inhibited the binding of lZ5I-PDGF to its receptor. Finally, the 185,000-Dacomponent had affinity for PDGF immobilized on Sepharose beads, suggesting that it has PDGF-binding activity.
We conclude that the PDGF receptor and the 185,000-Da substrate for the PDGF-dependent kinase are intimately related and probably identical molecules.
In search for cellular events transmitting the mitogenic postreceptor signals of PDGF, we recently found that PDGF stimulates tyrosine-specific kinase activity in membrane preparations from human fibroblasts (17,18). A 185-kDa protein was identified as the major substrate for the PDGFdependent kinase in these membranes. A less pronounced phosphorylation of a 130-kDa protein was also noticed. This component has been shown to be a proteolytic cleavage product of the 185-kDa component.* Both the 185-and 130-kDa components are most likely integral membrane proteins (18). They also serve as substrates for the PDGF-dependent kinase after solubilization of the membranes with Triton X-100 (18). Stimulation of tyrosine phosphorylation has also been observed using membranes prepared from 3T3 cells (19). Further, Cooper et al. (20) have recently shown that PDGF stimulates tyrosine phosphorylation in intact cells. Exposure of 3T3 cells to PDGF resulted in stimulated phosphorylation of five components with M , values between 42,000 and 45,000; in human fibroblasts the phosphorylation of three analogous components were also stimulated.
Interestingly, two other growth factors, EGF and insulin, have also been shown to stimulate tyrosine-specific phosphorylation in membrane preparations (21,22) as well as in intact cells (23,24). Furthermore, kinase activities with the same unusual amino acid specificity have been observed in conjunction with cellular transformation with certain retroviruses, Rous sarcoma virus being the most studied example (25). An intriguing possibility is that phosphorylation of certain proteins on tyrosine residues is involved in the cellular mechanism leading to initiation of DNA synthesis.
Evidence has been presented that the receptor for EGF is associated with kinase activity which becomes activated after ligand binding, resulting in autophosphorylation of the recep-tor molecule (26). Furthermore, binding of insulin to its receptor results in phosphorylation of the 95-kDa chain of the receptor and possibly the insulin receptor molecule similarly contains kinase activity (22). Whether the PDGF receptor analogously becomes autophosphorylated after PDGF binding, or whether the receptor molecule is associated with kinase activity, is not known. In order to investigate this possibility, we have performed a comparative characterization of the PDGF receptor and the major substrate in the PDGFstimulated phosphorylation reaction (185 kDa). We show that the two components behave similarly in SDS-gel electrophoresis in the presence or absence of reducing agents, that they co-elute upon chromatography on Sepharose 6B after solubilization with Triton X-100, and that they have similar lectin binding properties. Further, it is demonstrated that the 185-kDa substrate has affinity for PDGF-Sepharose and can be phosphorylated after adsorbtion. We conclude that the PDGF receptor and the 185-kDa component are closely related and probably identical molecules. by chromat.ography on Sephadex G-25 in 1 M acetic acid containing 1 mg/ml bovine serum albumin. The resulting specific activities of the different '"1-HSAB-PDGF derivatives were -20,000 cpm/ng of protein. ' ' ' I radioactivity was determined in a gamma counter a t 70% efficiency.

Affinity Labeling Experiments with '251-HSAB-PDGF-Confluent
cultures of human fibroblasts in 50-mm Petri dishes (approximately lo6 cells) were washed once in binding medium (phosphate-buffered saline containing 0.9 mM CaCI2, 0.8 mM MgSO,, and 1 mg/ml bovine serum albumin). Separate cultures were then incubated in 2 ml of binding medium containing 20 ng/ml of radiolabeled PDGF derivatized with HSAB at the different cross-linker concentrations indicated above. As a control for specificity, '"I-HSAB-PDGF was also incubated with cells in the presence of unlabeled PDGF (0.5 pglml). Cultures were then washed 3 times in binding medium and given 2 ml of phosphate-buffered saline and subjected to UV irradiation for 10 min. During this time, cell cultures were kept on ice a t a 10-cm distance from the lamp (250 watts). In order to avoid unnecessary damage to proteins, the UV light was filtered through a 1-cm layer of 10% acetone in a glass vessel. Cells were then scraped from the culture dishes with a rubber policeman and collected by centrifugation a t 10,000 X g for 30 s. The cell pellets were then solubilized in 80 pl of 1% Triton X-100, 10% glycerol, 20 mM HEPES, pH 7.4, for 15 min a t 0 "C, and centrifuged a t 10,000 X p for 2 min: the suwxnatants Affinity Labeling Experiments with DSS-Human foreskin fibroblasts were grown to confluency on 50-mm Petri dishes. After washing once in binding medium, cells were incubated in binding medium a t 4 "C for 90 min together with 10 ng/ml of '"1-PDGF (radiolabeled as described (12) to a specific activity of 20,000 cpm/ng). Cells were then washed 3 times in binding medium and then incubated a t 20 "c in NaC1/Pi with various concentrations of DSS added from a stock solution in dimethyl sulfoxide (100 mM), prepared immediately before use. After 20 min of incubation, the reactions were quenched by the addition of methylamine to a final concentration of 50 mM. The supernatants were then removed, cells were scraped from the dishes as described above, and analyzed by SDS-gel electrophoresis.
Phosphorylation Assay-The phosphorylation assay was carried out as described (18), with modifications as detailed in individual experiments, using membranes prepared from human foreskin fibroblasts according to Thom et al. (27). Crude membranes, used in one experiment (see below), were prepared using the same method (27) up to, and including the first ultracentrifugation step.
SDS-Gel Electrophoresis-SDS-gel electrophoresis was performed according to Blobel and Dobherstein (28). Separating gels consisted of 7-12% polyacrylamide gradients and the stacking gels of 3% polyacrylamide. Gel dimensions of 300 x 200 X 1 mm were used and gels were run at room temperature overnight a t a constant current. Samples (40 pl) were prepared for electrophoresis by incubation together with 40 pl of SDS-sample buffer (80 mM Tris-HC1, pH 8.8, 3.6% SDS, 0.01% bromophenol blue with or without 10 mM dithiothreitol) for 3 min at 95 "C. Reduced samples were then alkylated with 50 mM iodoacetamide for 15 min or more a t 20 "C. After electrophoresis, gels were fixed in 10% trichloroacetic acid for 30 min, and then stained for protein with Coomassie brilliant blue R-250. Gels were dried under vacuum and heat, and subjected to autoradiography (18,29) using pre-exposed Kodak X-OMat AR or Fuji RX films and DuPont "Lightning Plus" intensifying screens. Exposure was a t -70 "C. The  &el Chromatography OR Sephnrose 6B-A Sepharose 6B column (0.7 X 28 cm) was prepared and equilibrated with elution buffer (0.5% Triton X-100, 0.15 M NaCI, 20 mM HEPES, pH 7.4, 10% glycerol, 1 mg of bovine serum albumin/ml). The column was run at 20 "C a t a flow rate of 2.4 ml/h and 0.4-ml fractions were collected. The samples that were analyzed on the Sepharose 6B column were obtained in the following way: (i) lz5I-PDGF (30 ng/ml) in 0.3 ml of binding medium (see above) was incubated with a confluent culture of human fibroblasts (about 1.5 X lo5 cells) for 90 min a t 4 "C. Cells were then washed 5 times in binding medium, scraped with a rubber policeman from the dish, and collected by centrifugation a t 10,000 X g for 30 s. The cell pellet was then solubilized in 100 p1 of elution buffer for 15 min at 20 "C and centrifuged a t 150,000 X g in an Airfuge (Beckman) for 10 min. The supernatant was then run on the column. (ii) A parallel culture was treated similarly but received only unlabeled PDGF (1 pg/ml) during the incubation with cells. Iz5I-PDGF (10,000 cpm) together with some additional unlabeled PDGF (250 ng) were added after cells had been washed. After solubilization, as described above, the sample was applied to the column. (iii) Crude membranes (see above) from human fibroblasts (100 pg of protein) were solubilized in 20 pl of 10% Triton X-100, 10% glycerol, 20 mM HEPES, pH 7.4, for 10 min a t 20 "C, and centrifuged a t 100,000 X g for 10 min. MnCh ( 3 mM) and PDGF (75 nM) were then added to the supernatant. After incubation for 10 min a t 0 "C, 10 p~ [32P]ATP (containing 9 pCi of radioactivity) was added, and incubation prolonged for another 10 min at 0 "C. After addition of a 1000-fold excess of unlabeled ATP to inhibit further incorporation of radioactivity, the sample was run on the column. Aliquots of individual fractions (75 p l ) were then analyzed by SDS-gel electrophoresis and autoradiography. 32P radioactivity associated with the 185-kDa component was then determined by scanning the autoradiograms in a spectrophotometer, and integrating the area corresponding to the 185-kDa band. Experiments with Lectins-The ability of various lectins to compete with '%PDGF for binding to its receptor was investigated using confluent cultures of human fibroblasts seeded on Linbro 24-well dishes (-1.5 X lo5 cells). Cells were first washed once with binding medium (see above) and then incubated at 4 "C for 90 min in 250 ~1 were then analyzed SDSrgel electrophorks. of this medium containing various concentrations of different lectins.

Characterization
After washing 3 times in binding medium, incubation was cpntinued with "'I-PDGF (10 ng/ml) in 250 pl of this medium for 1 h. The cells were then washed another 3 times with the same medium. Cellassociated radioactivity was solubilized by incubation of the cell cultures with 0.5 ml of 1% Triton X-100, 109; glycerol, 20 mM HEPES, pH 7.4, 1 mg/ml bovine serum albumin for 30 min at 20 "C (12), and determined in a y counter. Using this protocol, about 9596 of the binding is specific in the sense that it can be inhibited by an excess of unlabeled PDGF (12).
The ability of certain monosaccharides to neutralize the lectininduced inhibition of "'I-PDGF binding was investigated in the following way. Cells were first incubated as described above with lectins a t 100 pg/ml (WGA, ConA, the lectin from Crotalaria juncea, or no addition). The cultures were then washed 3 times in binding medium and reincubated in 250 pl of this medium a t 4 "C for 60 min, together with different monosaccharides at 100 mM (n-methylmannoside, n-methylglucoside, N-acetylglucosamine, galactose, or no addition). After washing 3 times in binding medium, the '"I-PDGFbinding capacity of the cells was determined as described above.
To investigate the lectin-binding specificity of the substrates for the PDGF-dependent kinase, the following experiment was performed. Aliquots of human fibroblast membranes (10 pg) were incubated for 10 min at 0 "C in 0.5% Triton X-100, 20% glycerol, 20 mM HEPES, pH '7.4, 3 mM MnCI?, 1 mg/ml bovine serum albumin in the absence or presence of 75 nM PDGF in a total volume of 40 p l ; 15 pM ['"PIATP (containing 2.3 pCi of radioactivity) was then added, and incubation prolonged for another 10 min at 0 'C. After the addition of a 1000-fold excess of unlabeled ATP to inhibit further incorporation of ' "P radioactivity, samples were incubated for 30 min at 0 "C with immobilized lectins (40 pl of packed Sepharose heads containing WGA, ConA or the lectin from Crotalaria juncea, respectively). At this time, the beads were washed in the centrifuge, 3 times with 0.296 Triton X-100, 10% glycerol, 0.15 M NaCI, 20 mM Tris, pH 7.4, one time with the same buffer containing a higher ionic strength (0.5 M NaCI), and finally one additional time with 20 mM Tris, pH 7.4. The material adsorbed to the beads was then desorbed by incubation for 3 min at 95 "C in SDS-sample buffer and analyzed by SDS-gel electrophoresis.
Experiments with PDCF-Sepharose-The interaction between PDGF-Sepharose and the substrates for the PDGF-dependent kinase was investigated in the following manner. Human fibroblast membranes (20 pg) were incubated for 30 min at 0 "C in 0.5% Triton X-100, 15% glycerol, 20 mM HEPES, pH 7.4, 75 mM NaCI, 1 mg/ml bovine serum albumin and PDGF-Sepharose (20 pl of packed beads) in a total volume of 75 pl. A parallel sample was treated similarly and incubated with an equal amount of a control gel consisting of an identical gel coupled with ethanolamine. After incubation, the samples were centrifugated a t 10,000 X g for 1 min; the supernatants and the pellets were then used separately in the phosphorylation reaction. Aliquots (25 pl) of the supernatants were incubated for 10 min a t 0 "C with 3 mM MnCI2 in the presence or absence of 75 nM PDGF; at this time ["PIATP was added to a final concentration of 15 p~ (4.5 pCi of radioactivity), and incubation prolonged for another 10 min. Samples were then analyzed by SDS-gel electrophoresis. The pelleted beads were first washed 3 times with 0.2% Triton X-100. 20% glycerol, 50 mM NaCI, 20 mM Tris, pH 7.4, in the centrifuge and then subjected to phosphorylation conditions as described above. The material adsorbed to the beads was then desorbed by incubation for 3 min at 95 "C in SDS-sample buffer and analyzed hy SDS-gel electrophoresis. The amount of "P radioactivity associated with the 185-kDa component was quantified by densitometric scanning of the autoradiogram.

RESULTS
Affinity Labeling of the PDGF Receptor with '"I-HSAB-PDGF-In order to obtain information on the molecular structure of the PDGF receptor on human foreskin fibroblasts, affinity labeling of the receptor was performed. In the first set of experiments, the photocatalyzable cross-linker HSAB was used. PDGF was derivatized with this cross-linker at three different molar ratios and then iodinated. After binding to cells at 4 "C for 90 min in the dark, the cell cultures were washed and exposed to UV light for 10 min. Cellassociated radioactivity was then analyzed by SDS-gel electrophoresis. As can be seen in Fig. 1, A- The total amount of radioactivity in this molecular weight region was, however, low ( 4 % ) ; the major part of radioactivity appeared at the position of reduced PDGF, indicating a low efficiency of the cross-linking reaction. In the control, where underivatized PDGF was used, faint bands with M , values 190,000 to 210,000 were occasionally also observed (Fig. 1, lane D). This suggests that a limited spontaneous covalent cross-linking of ""I-PDGF to its receptor may occur in conjunction with exposure to UV light. The specificity of the reaction was investigated by adding unlabeled PDGF (500 ng/ml) together with the 1251-laheled derivatives of PDGF during binding to cells. Under these conditions, the formation of the 190,000 to 210,000 bands was inhibited (Fig. 1). Addition of EGF, insulin, or fibroblast growth factor together with "'I-HSAB-PDGF did not inhibit the formation of the high M , bands (not shown). This supports the notion that the 190,000 to 210,000 bands represent covalent complexes between the PDGF receptor and "'I-HSAB-PDGF. Subtraction of the molecular weight of reduced PDGF __ 'I A higher amount of radioactivity appeared in the 200-kDa region in the unreduced gels compared to the reduced gels. The reason for this is not known, hut it may be related to the fact that PDGF consists of two disulphide-linked rhains; if only one of the chains is crosslinked to the receptor, radioactivity associated with the other is lost upon reduction. Analysis of the mobility of the 9-PDGF-receptor complex by SDS-gel electrophoresis in the absence of reducing agents, gave essentially the same result as when dithiothreitol was present (Fig. 1). This suggests that the receptor is a single chain protein.
Affinity Labeling of the PDGF Receptor by use of a Bifunctional Cross-linker-In another set of experiments, a homobifunctional cross-linker was used to covalently stabilize the binding of "'I-PDGF to its receptor. Cells were first given '"I-PDGF for 90 min at 4 "C, washed, and incubated together with various amounts of DSS for 15 min at room temperature. Cell-associated radioactivity was analyzed by SDS-gel electrophoresis. As can be seen in Fig. 2, a t low concentrations of DSS, l""IPDGF was cross-linked to form a complex of M, 200,000 to 220,000. When the DSS concentration was raised, additional complexes of M, -400,000 and higher were formed.
In analogy with the results of the photoaffinity labeling experiments, the electrophoretical pattern was similar whether or not reducing agents were present during analysis (Fig. 2).

Control experiments showed that the formation of high M,
"'I-labeled complexes was inhibited if unlabeled PDGF was included during the binding to the cells (not shown).
In conclusion, the experiments with the homobifunctional cross-linker, DSS support the results obtained with the pho- State-PDGF has been reported to stimulate phosphorylation in human fibroblast membranes (17). The molecular mass of the major substrates for the PDGF-stimulated kinase has been estimated to 185 and 130 kDa by SDS-gel electrophoresis in the presence of reducing agents (18). The 130-kDa component has been shown to be a proteolytic degradation product of the 185-kDa component.2 In order to allow a closer molecular comparison of the 185-kDa component and the PDGF receptor, we analyzed the mobility of the 185-kDa component also in SDS-gel electrophoresis without reducing agents present. Fig. 3 shows that the major substrate for the PDGF-stimulated kinase migrated as a 185,000 protein under nonreducing as well as reducing conditions. Thus, the behavior of the 185-kDa component in SDS-gel electrophoresis, with or without reducing agents present, was similar to that expected for the PDGF receptor, as judged from the affinitylabeling experiments (Figs. 1 and 2). Comparative Analysis of Hydrodynamic Size-Gel chromatography was used to investigate whether the binding of Iy51-PDGF to its receptor was stable after solubilization of the membranes with Triton X-100. '""I-PDGF was first bound to its receptor on intact human fibroblasts at 4 "C. Cells were then washed and cell-associated radioactivity solubilized with a buffer containing Triton X-100, and applied to a column of Sepharose 6B eluted with the same buffer. As can be seen in Fig. 4A, after this procedure, the major part of the radioactivity eluted early in the chromatogram in a broad peak at a Kay of about 0.3. In the control, where '2'I-PDGF was added together with unlabeled PDGF to Triton-solubilized material from the same cells, the major part of the radioactivity eluted later on the column, in two peaks co-eluting with '2sII-PDGF (KaV = 0.7) and free (Kav = l.O), respectively. This indicates that '"I-PDGF forms a high M , complex with its receptor which is stable in the presence of Triton X-100. Calibration of the column using water-soluble standard proteins of known M,, revealed that the '"1-PDGF-receptor complex eluted between ferritin ( M , = 440,000) and catalase (Mr = 232,000), a t a position corresponding to a M , of about 320,000. This figure includes, however, the weight of the detergent attached to the receptor. Triton X-100 form micelles of M , -90,000 (30). Thus, to obtain a M , for the receptor proper, 90,000 and in addition 30,000 for the PDGF molecule, has ta be subtracted from 320,000. This gives a M , around 200,000 for the receptor, which is in good agreement with the estimate derived from affinity labeling experiments. Fig. 4B shows an experiment where membranes from human fibroblasts were first solubilized with Triton X-100, and subsequently phosphorylated in the presence of PDGF. The sample was run on the same Sepharose 6B column and aliquots of the effluent fractions analyzed by SDS-gel electrophoresis and autoradiography. The amount of radioactivity associated with the 185-kDa component was then quantified. As can be seen, the elution position of this component was similar to the elution position of the IWI-PDGF-receptor complex.
Inhibition of Iz5I-PDGF Binding to Fibroblasts by Lectins-To further characterize the PDGF receptor on human fibroblasts, several lectins with different carbohydrate-binding specificities were tested for their ability to compete with '"I-PDGF for binding. Fig. 5 shows that WGA (which binds Nacetylglucosamine residues) was an effective competitor; at 20 Fg/ml, WGA inhibited 50% of the binding of 'T-PDGF. The lectin from Crotalaria juncea (which binds galactose and N-acetylgalactosamine residues) also had some inhibitory effect. Concanavalin A (which binds mannose and glucose residues) showed a limited inhibitory effect (Fig. 5). A small inhibitory effect was also observed with other lectins with similar specificity obtained from Lens culinaris, Pisum satiuum, Vicia sativa, and Vicia ervila (not shown). Lectins from the following sources, with the indicated specificities, were also tested and found not to interfere with binding of '*'I-PDGF: Vicia uilosa (N-acetylgalactosamine), peanut (galactose), Lotus tetragonolobus (fucose), and Helix pomatia ( Nacetylgalactosamine). To minimize the possibility of interactions between 'T-PDGF and the lectins, the binding was performed in two steps; cells were first exposed to lectins, then washed and incubated with lZ51-PDGF. Thus, the inhibition was likely to be due to an interaction between the lectin and the receptor. For experimental details see "Experimental Procedures." 100% corresponds to 3900 cpm bound/l.5 X lo5 cells.
by guest on March 23, 2020 http://www.jbc.org/ Downloaded from juncea, respectively. The small inhibition caused by ConA was neutralized by a-methylmannoside or a-methylglucoside. Thus, these findings were in concordance with the known specificities for the various lectins. The conclusion from these experiments is that the PDGF receptor contains carbohydrate, probably N-acetylglucosamine residues and galactose residues, in the vicinity of the PDGF-binding region.
Affinity of the Substrates for the PDGF-dependent Kinase for Various Lectins-To investigate whether the lectins which could inhibit "'I-PDGF-binding to its receptor, also bound the 185-and 130-kDa substrates for the PDGF-dependent membrane kinase, the following experiment was performed. Membranes from human fibroblasts were solubilized with Triton X-100 and subjected to the standard phosphorylation assay in the presence or absence of PDGF. Various lectins immobilized on Sepharose beads were then added to the incubations, and the adsorbed radioactivity analyzed by SDSgel electrophoresis. Fig. 7 shows that the 185-kDa as well as the 130-kDa components were effectively adsorbed to WGA-Sepharose. Immobilized Crotalaria juncea lectin and ConA Human fibroblast membranes were solubilized with Triton X-100 and incubated with PDGF-Sepharose or a control gel; the adsorbed (ads) and unadsorbed (unads) fractions were then separately subjected to the phosphorylation assay in the presence or absence of PDGF and analyzed by SDS-gel electrophoresis and autoradiography. See "Experimental Procedures" for further details.
also bound a small amount of these components. Thus, the lectins which inhibited the binding of "'I-PDGF to its receptor, also recognized the 185and 130-kDa substrates.
Affinity of the Substrates for the PDGF-dependent Kinase for PDGF-Sepharose-An important experiment in the evaluation of the hypothesis that the PDGF receptor and the 185-kDa component are the same protein, was to see whether the 185-kDa component had affinity for PDGF. For this purpose, fibroblast membranes were solubilized with Triton X-100 and exposed to PDGF-Sepharose and to a control gel consisting of ethanolamine-Sepharose, respectively. After incubation for 30 min at 0 "C, the unadsorbed fractions were incubated with ['"PIATP in the absence or presence of PDGF. Fig. 8 shows that the exposure of the solubilized membranes to PDGF-Sepharose resulted in a decrease in phosphorylation of the 185-and 130-kDa proteins in the unadsorbed fraction as compared to the solubilized membranes exposed to the control gel. When the corresponding PDGF-Sepharose beads were incubated with [''"PIATP after washing, a significant phosphorylation of the 185-and 130-kDa components was demonstrated (Fig. 8). The recovery of kinase activity on the beads, measured as stimulation of phosphorylation of the 185-kDa component, was estimated to be 10%. When the control beads were subjected to the same treatment, no phosphorylation was seen (Fig. 8). This indicates that the 185-kDa autoradiography. component and also the PDGF-dependent kinase activity have affinity for PDGF.

DISCUSSION
In this paper, we have shown that the PDGF receptor and the 185-kDa phosphoprotein, being the major substrates in the PDGF-stimulated phosphorylation reaction, have several characteristics in common. They have: 1) similar mobilities in SDS-gel electrophoresis with or without dithiothreitol present; 2) similar elution positions on Sepharose 6B after solubilization; and 3) similar lectin-binding specificities. Furthermore, the 185-kDa protein has affinity for PDGF-Sepharose. We conclude that they are closely related and probably identical.
The M , of the PDGF receptor was estimated by affinity- In the affinity labeling experiments using the photocatalyzable cross-linker, a small amount of 12'II-PDGF was bound to the receptor even in the absence of cross-linker (Fig. 1). Similarly, a spontaneous covalent binding of '"I-EGF to its receptor has been observed (32,33). This feature of "'I-EGF has been attributed to chemical modification of the EGF molecule in the course of iodination with the chloramine-T method (34). Since a similar procedure was used for radiolabeling of PDGF in the present studies, the spontaneous attachment of PDGF to its receptor, although less pronounced and also noticed only after UV irradiation, might be due to similar modifications of the PDGF molecule. The reason for the appearance of three distinct components in the M , region 190,000 to 210,000 after affinity labeling with the photocatalyzable cross-linker, is not known; possibly it might indicate proteolysis or damage to the receptor by UV irradiation.
The appearance of components with M , 400,000 and higher after affinity-labeling with homobifunctional cross-linkers at high concentrations (Fig. 2), probably reflects the formation of receptor dimers and multimers, or the cross-linking of the PDGF receptor with other membrane proteins. These high molecular weight components were not observed when the more specific photocatalyzable cross-linker, HSAB, was used (Fig. 1).
The finding that the 185-kDa substrate for the PDGFdependent kinase has affinity for PDGF-Sepharose, suggests that the PDGF receptor and the 185-kDa component are the same protein. Other interpretations are, however, also possible; e.g. the 185-kDa component and the PDGF receptor may be distinct molecules, but linked to each other by noncovalent bonds, or, the 185-kDa component may have a nonspecific affinity for PDGF-Sepharose. However, in view of their similarities in M,, single chain state and lectin-binding specificity, we consider it more likely that they are indeed identical proteins. The 130-kDa fragment of the 185-kDa component also bound to PDGF-Sepharose (Fig. 8). This suggests that the PDGF-binding activity, as well as the lectin-binding activity (Fig. 7), is associated with a common part of the molecules. The conclusions above were drawn from experiments where Triton-solubilized membranes were exposed to PDGF-Sepharose, then the adsorbed and unadsorbed fractions were subjected to phosphorylation separately. When human fibroblasts were first phosphorylated in the presence of PDGF and then subjected to PDGF-Sepharose, only a minor part of the radioactivity associated with the 185-and 130-kDa components was adsorbed to the beads (not shown). This is probably due to the fact that the soluble PDGF binds very tightly to these components, and therefore blocks the binding to the immobilized PDGF.
It follows from the experiment with PDGF-Sepharose that the kinase activity also has affinity for this gel. Analogous to the discussion above, this may be taken to indicate that the 185-kDa PDGF receptor molecule itself has kinase activity, or that the kinase activity resides in a different molecule which is associated with the receptor by noncovalent bonds, or, which has some nonspecific affinity for PDGF-Sepharose. Since autophosphorylation is a common feature among kinases, we consider it likely that the PDGF-stimulated phosphorylation of the 185-kDa component represents autophosphorylation of the PDGF-dependent kinase, although definitive proof of this is still lacking. The recovery of activity that we obtained, measured as stimulation of phosphorylation of the 185-kDa component, was 10%. This low recovery might have several explanations. Possibly only a fraction of the PDGF-dependent kinase activity has affinity for PDGF-Sepharose, or, it may be due to inactivation of the PDGF kinase during the experiment. It might also be due to a decreased accessibility of the immobilized kinase for its substrates. The latter explanation is possible also if the phosphorylation of the 185-kDa component represents an autophosphorylation, as suggested above, provided that the autophosphorylation is intermolecular.
Thus, the receptors for three different growth factors, PDGF, EGF, and insulin are associated with tyrosine-specific kinase activity. Interestingly, PDGF (35), as well as EGF and insulin (36), has been shown to also stimulate serineor threonine-specific kinase activity in intact cells. In view of the different amino acid specificity, this is likely to be a secondary event. It is possible that some of the substrates for the receptor-associated tyrosine kinases are serine-or threonine-specific kinases, which may be activated by phosphorylation on tyrosine. Studies have been initiated to identify substrates in intact cells for the different tyrosine kinases (20,23,37). It is important for the understanding of the mechanism of action of growth factors, to elucidate the function of such substrates, and to investivate their possible role in growth control. It will be of further interest to compare the substrate specificities of these receptor-associated kinases with those of the tyrosine kinases associated with the oncogenes of certain retroviruses. Substrates common to all, or many of these kinases would be particularly interesting in relation to a possible role in control of cell proliferation.