Ligand stimulation reduces platelet-derived growth factor beta-receptor susceptibility to tyrosine dephosphorylation.

Ligand binding to the platelet-derived growth factor (PDGF) beta-receptor leads to increased receptor tyrosine phosphorylation as a consequence of dimerization-induced activation of the intrinsic receptor tyrosine kinase activity. In this study we asked whether ligand-stimulated PDGF beta-receptor tyrosine phosphorylation, to some extent, also involved reduced susceptibility to tyrosine dephosphorylation. To investigate this possibility we compared the sensitivity of ligand-stimulated and non-stimulated forms of tyrosine-phosphorylated PDGF beta-receptors to dephosphorylation using various preparations containing protein-tyrosine phosphatase activity. Ligand-stimulated or unstimulated tyrosine-phosphorylated receptors were obtained after incubation of cells with pervanadate only or pervanadate, together with PDGF-BB, respectively. Dephosphorylation of receptors immobilized on wheat germ agglutinin-Sepharose, as well as of receptors in intact cell membranes, was investigated under conditions when rephosphorylation did not occur. As compared with unstimulated receptors the ligand-stimulated PDGF beta-receptors showed about 10-fold reduced sensitivity to dephosphorylation by cell membranes, a recombinant form of the catalytic domain of density-enhanced phosphatase-1, or recombinant protein-tyrosine phosphatase 1B. We conclude that ligand-stimulated forms of the PDGF beta-receptor display a reduced susceptibility to dephosphorylation. Our findings suggest a novel mechanism whereby ligand stimulation of PDGF beta-receptor, and possibly other tyrosine kinase receptors, leads to a net increase in receptor tyrosine phosphorylation.

Ligand binding to the platelet-derived growth factor (PDGF) ␤-receptor leads to increased receptor tyrosine phosphorylation as a consequence of dimerizationinduced activation of the intrinsic receptor tyrosine kinase activity. In this study we asked whether ligandstimulated PDGF ␤-receptor tyrosine phosphorylation, to some extent, also involved reduced susceptibility to tyrosine dephosphorylation. To investigate this possibility we compared the sensitivity of ligand-stimulated and nonstimulated forms of tyrosine-phosphorylated PDGF ␤-receptors to dephosphorylation using various preparations containing protein-tyrosine phosphatase activity. Ligand-stimulated or unstimulated tyrosine-phosphorylated receptors were obtained after incubation of cells with pervanadate only or pervanadate, together with PDGF-BB, respectively. Dephosphorylation of receptors immobilized on wheat germ agglutinin-Sepharose, as well as of receptors in intact cell membranes, was investigated under conditions when rephosphorylation did not occur. As compared with unstimulated receptors the ligandstimulated PDGF ␤-receptors showed about 10-fold reduced sensitivity to dephosphorylation by cell membranes, a recombinant form of the catalytic domain of density-enhanced phosphatase-1, or recombinant protein-tyrosine phosphatase 1B. We conclude that ligandstimulated forms of the PDGF ␤-receptor display a reduced susceptibility to dephosphorylation. Our findings suggest a novel mechanism whereby ligand stimulation of PDGF ␤-receptor, and possibly other tyrosine kinase receptors, leads to a net increase in receptor tyrosine phosphorylation.
Receptor tyrosine kinases (RTKs) 1 are critical components of signaling pathways that control cellular processes like proliferation, differentiation, migration, and metabolism. Ligand binding of RTKs often leads to dimerization and subsequent increases in autophosphorylation of tyrosine residues in the intracellular portion of the receptors (reviewed in Refs. 1 and 2). Autophosphorylation of intracellular receptor tyrosine residues controls the intrinsic tyrosine kinase activity and creates binding sites to recruit downstream signaling molecules (3,4). The mechanism whereby ligand-induced dimerization stimulates these phosphorylation events is incompletely understood but may involve a conformational change of the receptor or a proximity effect.
RTK net tyrosine phosphorylation is not only controlled by the receptor kinase activity but is also determined by the action of protein-tyrosine phosphatases (PTPs). Accumulating evidence suggest that PTPs are regulatory components of RTK signaling pathways. Antisense studies have demonstrated increased signaling via receptors for insulin, epidermal growth factor (EGF), and hepatocyte growth factor after attenuation of expression of the receptor-like PTP LAR (5-7), and disruption of PTP1B in mice results in enhanced insulin sensitivity (8,9). Furthermore, genetic studies in Caenorhabditis elegans have identified the receptor-like PTP CLR-1 as a negative regulator of signaling through the fibroblast growth factor receptor ortholog EGL-15 (10). Physical association between the insulin receptor and the receptor-like PTP LAR, as well as between PDGF ␤-receptor and the receptor-like PTP DEP-1, have also been demonstrated (11,12).
In this study we set out to investigate the possibility that reduced susceptibility to PTP action contributes to ligand-induced increases in net tyrosine phosphorylation of RTKs. The well characterized PDGF ␤-receptor was chosen as a prototype dimerization-activated receptor tyrosine kinase. The autophosphorylation sites of the PDGF ␤-receptor have been extensively studied and include a regulatory site, Tyr 857 , as well as numerous sites, which in their phosphorylated form act as binding sites for SH2 domain-containing proteins including c-Src, phospholipase C-␥, and phosphatidylinositol 3Ј-kinase (PI3-kinase) (reviewed in Ref. 13).
To study the effects of ligand stimulation on PTP sensitivity, preparations of tyrosine-phosphorylated ligand-stimulated and unstimulated PDGF ␤-receptors were obtained. Using these preparations we demonstrate that ligand-stimulated forms of the PDGF ␤-receptor display a reduced susceptibility to dephosphorylation, as compared with unstimulated forms.
Analysis of Receptor Dimerization-After overnight incubation in serum-free Ham's F-12, supplemented with 1 mg/ml BSA, PAE/ PDGF␤R cells were left unstimulated or treated with 100 ng/ml PDGF-BB for 60 min on ice, with 100 M pervanadate for 30 min at 37°C and 60 min on ice, or with pervanadate for 30 min at 37°C and then stimulated with 100 ng/ml PDGF-BB, in the presence of pervanadate, for 60 min on ice. After stimulation, cells were washed with ice-cold PBS. Ligand-receptor complexes were cross-linked by incuba-* 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. ‡ Present address: Dept. of Dermatology, Yamanashi Medical University, 1110 Shimokato, Tamaho, Nakakoma, Yamanashi 409-38, Japan.
Immunoprecipitation of PDGF ␤-Receptor and Analysis of Associating Proteins-For detection of PDGF ␤-receptor associating proteins, cells were lysed as described above and immunoprecipitated with the PDGF receptor antiserum CED raised against a peptide corresponding to the five carboxyl-terminal amino acids of the PDGF ␤-receptor. Immunoprecipitated PDGF ␤-receptor was detected with P-20 (Santa Cruz Biotechnology). Coprecipitating tyrosine-phosphorylated proteins were detected by immunoblotting with the phosphotyrosine antibody PY99 (Santa Cruz Biotechnology), and coprecipitating p85 was detected by a rabbit antiserum (Upstate Biotechnology).
Dephosphorylation Analysis of Immunoprecipitated PDGF ␤-Receptor-Cells were stimulated, and cell lysates were prepared as described above. Lysates were incubated with wheat germ agglutinin (WGA)-Sepharose (EC Diagnostics AB, Uppsala, Sweden) at 4°C overnight and then incubated in 15 mM iodoacetamide (Sigma) for 30 min at room temperature. The pellets were washed five times with lysis buffer without Na 3 VO 4 and once in a buffer containing 25 mM imidazol, 0.1 mg/ml BSA, and 10 mM dithiothreitol (DTT). The samples were incubated at 37°C with vigorous agitation for 15 min with a recombinant form of DEP-1, composed of an amino-terminal glutathione S-transferase domain, a DEP-1 portion encompassing human DEP-1 amino acids 997-1337, and a carboxyl-terminal hemagglutinin tag or with recombinant PTP1B. To stop the dephosphorylation reaction, the samples were washed once with lysis buffer containing 100 M Na 3 VO 4 and then eluted with overnight incubation at 4°C in 0.3 M N-acetylglucosamine (Sigma) in the presence of 3 M PDGF receptor-selective tyrosine kinase inhibitor AG1296 (16). After centrifugation, supernatants were incubated with a rabbit antiserum against the PDGF ␤-receptor for 2 h at 4°C and then with protein A-Sepharose (Amersham Pharmacia Biotech) for 1 h at 4°C. After washing three times with lysis buffer and once with 20 mM Tris, pH 7.5, the precipitates were heated for 3 min at 95°C in SDS sample buffer and subjected to SDS-PAGE, followed by immunoblotting with the rabbit antiserum PDGFR-3 or the phosphotyrosine antibody PY99.
Preparation of Intact Cell Membranes and Dephosphorylation of PDGF ␤-Receptor in Vitro-After overnight incubation in serum-free Ham's F-12, supplemented with 1 mg/ml BSA, PAE/PDGF␤R cells were treated with pervanadate for 30 min at 37°C and 60 min on ice or with pervanadate for 30 min at 37°C and then stimulated with 100 ng/ml PDGF-BB, in the presence of pervanadate, for 60 min on ice and washed with ice-cold PBS and collected in 5 ml of ice-cold PBS. The cells were pelleted and incubated in hypotonic buffer (20 mM Tris, pH 7.5, 10 mM NaCl, 1% Trasylol) for 30 min at 4°C. The cells were homogenized with 80 strokes in a Dounce homogenizor, and the nuclei were pelleted at 3000 ϫ g for 5 min. The supernatants were then centrifuged at 100,000 ϫ g for 25 min. The membranous sediments were resolved in hypotonic buffer containing 10 mM DTT and 3 M AG1296 using a syringe with a 0.6 ϫ 26-mm needle and incubated with 15 mM iodoacetamide for 30 min at room temperature. Preparations were kept at Ϫ80°C until use. To obtain PTP activity containing membranes to be used in dephosphorylation reactions, membranous sediments from unstimulated parental PAE cells were prepared by hypotonic lysis and centrifugation, as described above. After centrifugation, membranes were resuspended in a buffer containing 25 mM imidazol, 0.1 mg/ml BSA, and 10 mM DTT. For the dephosphorylation assay, PAE/PDGF␤R membranes containing 10 mg of protein were mixed with membranes from parental PAE cells for 10 min at 37°C during vigorous agitation or incubated with recombinant DEP-1 for 15 min at 4°C. The dephosphorylation was stopped by addition of 100 M Na 3 VO 4 and an equal volume of 2ϫ lysis buffer for 10 min at 4°C. After centrifugation at 13,000 rpm for 15 min, the PDGF ␤-receptor was immunoprecipitated with PDGFR-3 antiserum, followed by immunoblotting with the PDGF ␤-receptor antibody P-20 or the phosphotyrosine antibody PY99.

Preparation and Characterization of Tyrosine-phosphorylated Ligand-stimulated and Unstimulated Forms of PDGF
␤-Receptors-To obtain tyrosine-phosphorylated forms of ligand-stimulated and unstimulated PDGF ␤-receptors, PAE cells stably transfected with the human PDGF ␤-receptor (PAE/PDGF␤R cells) were treated with pervanadate with or without PDGF-BB. To investigate whether receptors occurred as monomers or dimers after stimulation with PDGF-BB alone, pervanadate alone, or pervanadate, together with PDGF-BB, cross-linking of cell-surface proteins was performed, and WGA-Sepharose fractions were isolated. After elution with N-acetylglucosamine, PDGF ␤-receptors were immunoprecipitated and analyzed by SDS-PAGE and immunoblotting. As shown in Fig.  1, lower panel, dimeric receptors were recovered from cells stimulated with PDGF-BB, or pervanadate and PDGF-BB, but not from cells stimulated with pervanadate only. Furthermore, treatment with pervanadate only and pervanadate combined with PDGF-BB yielded receptors that were tyrosine-phosphorylated to the same extent, as determined by phosphotyrosine immunoblotting (Fig. 1, upper panel).
To determine the pattern of tyrosine phosphorylation of the ligand-stimulated and unstimulated receptor preparations, in vivo 32 P-labeled receptors were isolated by immunoprecipitations and subjected to two-dimensional phosphopeptide mapping ( Fig. 2A). No major differences were seen in the autophosphorylation patterns after the different stimulations.
To characterize the pattern of association of SH2 domaincontaining signaling molecules with the PDGF ␤-receptor after stimulation with pervanadate alone, or pervanadate and PDGF-BB, PDGF ␤-receptor was immunoprecipitated and subjected to immunoblotting with phosphotyrosine antibody, PDGF ␤-receptor antiserum, and p85 antiserum (Fig. 2B). With the exception of a component of 120 kDa, found exclusively coprecipitating with the ligand-stimulated receptor, both stimulations induced a similar pattern of PDGF ␤-receptorcoprecipitating tyrosine-phosphorylated proteins (Fig. 2B, top  panel). The tyrosine-phosphorylated component with an apparent molecular mass of about 200 kDa was recognized by an antibody to the PDGF ␤-receptor (Fig. 2B, second  Ligand Stimulation of PDGF␤R Protects from Dephosphorylation 27750 the top). Furthermore, equal amounts of p85 subunit of PI3kinase coprecipitated with the PDGF ␤-receptor after incubation with pervanadate alone and pervanadate and PDGF-BB (Fig. 2B, lower panel).
In conclusion, stimulation with pervanadate only, or pervanadate and PDGF-BB, yields PDGF receptors that display similar patterns of tyrosine phosphorylation and substrate association but differ with regard to occurrence as receptor dimers.
In Vitro Dephosphorylation of Unstimulated and Ligandstimulated PDGF ␤-Receptors Immobilized on WGA-Sepharose-To compare the susceptibility of unstimulated and ligand-stimulated PDGF ␤-receptors to dephosphorylation, WGA-Sepharose fractions from cells stimulated with pervanadate alone or pervanadate, together with PDGF-BB, were isolated. To avoid dephosphorylation mediated by PTPs present in the WGA-Sepharose fractions, these were inactivated by incubation with the alkylating agent iodoacetamide. After washes in vanadate-free buffers, the WGA-Sepharose fractions were subsequently mixed with the recombinant catalytic domain of DEP-1, a receptor-like PTP (Fig. 3A), or recombinant PTP1B (Fig. 3C). The dephosphorylation reaction was stopped by addition of vanadate. After elution of receptors with N-acetylglucosamine, receptors were immunoprecipitated and subjected to phosphotyrosine immunoblotting. As shown in Fig. 3, incubation of WGA-Sepharoseimmobilized PDGF ␤-receptors with either PTP1B or DEP-1 led to a dose-dependent dephosphorylation. However, both when PTP1B and when DEP-1 were used as the source of PTP activity, 10 -100-fold larger amounts were required for dephosphorylation of ligand-stimulated receptors as compared with unstimulated receptors. It is unlikely that the more abundant phosphorylation of ligand-stimulated receptors is because of rephosphorylation in trans by the kinases in the ligand-induced receptor dimer, because the analysis was done in the absence of ATP. Moreover, the same difference in susceptibility to dephosphorylation of ligand-stimulated receptors as compared with monomers was seen when the dephosphorylation was performed in the presence of the PDGF receptor kinase inhibitor AG1296 (data not shown).
Together, these experiments demonstrate that ligand-stimulated forms of PDGF ␤-receptors, when immobilized on WGA-Sepharose, display a reduced susceptibility to tyrosine dephosphorylation as compared with unstimulated forms.  Membranes containing unstimulated and ligand-stimulated tyrosinephosphorylated PDGF ␤-receptors were incubated with either membrane preparations from parental PAE cells (A) or recombinant DEP-1 (B). After dephosphorylation, the tyrosine phosphorylation status of PDGF ␤-receptors was determined by immunoprecipitation of receptors from membrane lysates, followed by SDS-PAGE and phosphotyrosine (P-Tyr) immunoblotting (IB). After stripping, membranes were reprobed with PDGF ␤-receptor antibody. The position of PDGF ␤-receptor is indicated by arrows to the left. The amount of phosphorylated receptor remaining, relative to the untreated control, is given below the phosphotyrosine immunoblotting panels.

Ligand Stimulation of PDGF␤R Protects from Dephosphorylation 27751
Dephosphorylation of Unstimulated and Ligand-stimulated PDGF ␤-Receptors in Intact Cell Membranes-To investigate the sensitivity of unstimulated and ligand-stimulated PDGF ␤-receptors to the action of PTPs in a more physiological setting, their dephosphorylation in intact cell membranes was investigated. After stimulation of PAE/PDGF␤R cells with pervanadate only or pervanadate, together with PDGF-BB, membrane fractions were isolated and incubated with iodoacetamide to irreversibly block PTP activity in the membrane preparations. Membranes containing unstimulated and ligandstimulated tyrosine-phosphorylated PDGF ␤-receptors were then incubated with either membrane preparations from untransfected PAE cells or with recombinant DEP-1 (Fig. 4). Dephosphorylation was stopped by addition of vanadate, and the tyrosine phosphorylation status of PDGF ␤-receptors was determined by phosphotyrosine immunoblotting of receptors immunoprecipitated from membrane lysates (Fig. 4, A and B). Ligand-stimulated PDGF ␤-receptors displayed a reduced susceptibility to dephosphorylation as compared with the unstimulated form. Whereas 100 g of PAE membranes was required for 50% dephosphorylation of ligand-stimulated receptors, 10 g of PAE membranes led to more than 50% dephosphorylation of unstimulated receptors (Fig. 4A). A similar tendency was observed when recombinant DEP-1 was used as source of PTP-activity (Fig. 4B).
We thus conclude also that in intact cell membranes ligandstimulated PDGF ␤-receptors are more resistant to PTP-mediated dephosphorylation as compared with unstimulated PDGF ␤-receptors.
Our results clearly demonstrate that ligand-stimulated forms of the PDGF ␤-receptor display a reduced susceptibility to dephosphorylation by PTPs in vitro and suggest a novel mechanism contributing to the increased receptor tyrosine phosphorylation occurring after ligand stimulation.
The mechanism(s) underlying the difference between ligandstimulated and unstimulated receptors with regard to dephosphorylation susceptibility remains to be clarified. The characterization of the receptors with regard to tyrosine phosphorylation pattern did not reveal any major differences (Fig. 4A). Furthermore, analysis by two-dimensional phosphopeptide mapping of receptors after dephosphorylation indicated that dephosphorylation occurred of most sites (data not shown). This argues against the possibility of the difference being because of the presence of particular ligand-induced phosphatase-resistant sites in ligand-stimulated receptors. In contrast, a clear difference between the two receptor preparations was detected when they were analyzed with regard to occurrence as cross-linkable receptor dimers (Fig. 1). This suggests that dimerization might underlie differences in dephosphorylation sensitivity.
At least two principally different mechanisms whereby dimerization can affect sensitivity to PTPs can be envisioned; dimeric forms of the receptors can be intrinsically poorer substrates for PTPs, or alternatively, the difference between monomeric and dimeric forms could be caused by modulation of PTP activity. The latter possibility could involve an association with dimeric receptors of a PTP inhibitor or an association with monomeric receptors of a PTP activator. Earlier studies have demonstrated H 2 O 2 production and subsequent inhibition of PTP activity, following stimulation with PDGF or EGF (18,19). The PDGF-stimulated H 2 O 2 production was recently shown to be dependent upon PI3-kinase activation (20). Although simi-lar amounts of PI3-kinase were found associated with the monomeric and dimeric forms of the receptor in our study, the possibility that the difference in dephosphorylation between monomeric and dimeric receptors is caused by trans-inhibition of PTPs should be further studied. In this context, identification of the 120-kDa component, which was exclusively detected in immunoprecipitates of ligand-stimulated receptors (Fig. 2B), appears as an important goal for future studies.
Accumulating evidence suggests that PTPs act as negative regulators of ligand-activated tyrosine kinase receptors (15). There are also some indications suggesting that modulation of PTP activity is involved in ligand-independent activation of tyrosine kinase receptors. For example, UV-induced increase in EGF receptor tyrosine phosphorylation is caused by reduced receptor-directed PTP activity, rather than increased tyrosine kinase activity (21). In agreement with this notion, a direct inhibitory effect by UV light on the specific activities of PTP-␣, PTP-, DEP-1, and SHP-1 have recently been demonstrated (22). These observations, together with the findings of the present paper, thus suggest important functions for PTPs in controlling activation, as well as inactivation, of both unoccupied and ligand-occupied tyrosine kinase receptors.