Binding of Cbl to a Phospholipase Cγ1-docking Site on Platelet-derived Growth Factor Receptor β Provides a Dual Mechanism of Negative Regulation*

Ubiquitin conjugation to receptor tyrosine kinases is a critical biochemical step in attenuating their signaling through lysosomal degradation. Our previous studies have established Cbl as an E3 ubiquitin ligase for ubiquitinylation and degradation of platelet-derived growth factor receptor (PDGFR) α and PDGFRβ. However, the role of endogenous Cbl in PDGFR regulation and the molecular mechanisms of this regulation remain unclear. Here, we demonstrate that endogenous Cbl is essential for ligand-induced ubiquitinylation and degradation of PDGFRβ; this involves the Cbl TKB domain binding to PDGFRβ phosphotyrosine 1021, a known phospholipase C (PLC) γ1 SH2 domain-binding site. Lack of Cbl or ablation of the Cbl-binding site on PDGFRβ impedes receptor sorting to the lysosome. Cbl-deficient cells also show more PDGF-induced PLCγ1 association with PDGFRβ and enhanced PLC-mediated cell migration. Thus, Cbl-dependent negative regulation of PDGFRβ involves a dual mechanism that concurrently promotes ubiquitin-dependent lysosomal sorting of the receptor and competitively reduces the recruitment of a positive mediator of receptor signaling.

Receptor tyrosine kinases (RTKs) 2 control cellular behavior in response to extracellular peptide growth factors. Platelet-derived growth factor receptors (PDGFRs) are prototypical RTKs that play critical biological roles in many species including mammals and are implicated in disease states such as oncogenesis (1)(2)(3)(4)(5)(6). The two closely related PDGFRs (␣ and ␤) function as homo-or hetero-dimers with differential affinities for homodimers or heterodimers of four distinct ligands (PDGF A-D). Deficiencies of PDGF ligands or receptors produce unique and often severe developmental defects or other pathological manifestations indicating the important physiological roles of the PDGF-PDGFR ligand-receptor signaling networks (1,(7)(8)(9). Notably, the overexpression or activation of PDGFR␤ correlates with the development of human tumors, including gliomas (10,11), myelomonocytic leukemia (12), osteosarcoma (13), colon carcinoma (14,15), prostate cancer, and breast cancer (16,17). Given the critical physiological and pathogenic roles of PDGFRs, elucidating their regulatory mechanisms is of substantial interest.
We have previously demonstrated that the ubiquitin ligase Cbl (Casitas B-lineage lymphoma) functions as a negative regulator of PDGFRs (18 -20), and subsequent studies have extended the role of Cbl family proteins to a variety of other RTKs (21)(22)(23)(24)(25)(26)(27)(28). Cbl associates with activated RTKs via binding of its tyrosine kinase-binding (TKB) domain to phosphopeptide motifs generated by autophosphorylation of the receptor. Cbl TKB domain-binding sites were initially identified in ZAP70 and Syk through phospho-peptide library screening (29) and further characterized as a phosphotyrosine within a consensus sequence of (D/E)XpYXXX (where is a hydrophobic amino acid; usually Pro) (56,57). Additional distinct motifs, such as DpYR and RA(V/I)XNQpY(S/T), have been identified in the RTK c-MET and in the adaptor protein APS, respectively. Recruitment of Cbl results in RTK ubiquitinylation, which in turn serves as a signal for targeting the receptor to the lysosome for degradation via the evolutionarily conserved protein complexes called endosomal sorting complex required for transport 1-3 (30 -32). Enhanced lysosomal degradation is thought to limit the biological half-life of the active pool of RTKs, thus serving as a global attenuation mechanism; however, differential attenuation of signals originating in specific endosomal compartments remains to be investigated (19,20,33,34).
The Cbl RING finger domain is essential for the negative regulation of PDGFRs (18 -20), suggesting that ubiquitination provides one mechanism for Cbl-mediated negative regulation of PDGFRs. However, the functional role of endogenous Cbl proteins has not been addressed, and the mechanisms underlying the Cbl-PDGFR interaction remain unidentified. Several previous findings suggest that Cbl may interact with PDGFRs through its TKB domain because a point mutation (G306E) within the TKB domain that abolishes the binding of Cbl to its phospho-peptide targets reduces Cbl-PDGFR association (18). An intact TKB domain was also required for Cbl protein to promote the ubiquitinylation of PDGFR as well as for negative regulation of PDGFR-mediated cell proliferation and survival (20). Thus, identification of Cbl TKB domain-binding motif(s) on PDGFR is crucial to elucidate the mechanism of Cbl-mediated PDGFR regulation.
Here, we have identified the Cbl TKB domain-binding site on PDGFR␤ that overlaps with a previously known PLC␥1 SH2 domain-binding site. We provide evidence that endogenous Cbl is not only an important determinant of targeting activated PDGFR␤ for ubiquitin-dependent lysosomal degradation, but a negative regulator of PLC␥1-mediated chemotactic and migratory responses to PDGF. These findings reveal a novel dual mode of Cbl-mediated negative regulation of PDGFR via 1) ubiquitin-dependent degradation of the receptor and 2) competitive RTK binding with the positive signaling mediator PLC␥1.
Wound Healing and Transwell Migration Assays-Serumstarved confluent cell monolayers were scratched with a plastic pipette tip and washed twice with phosphate-buffered saline. The plates were refed with starvation medium with 20 ng/ml PDGF-BB and allowed to heal at 37°C. The cells were photographed at 0, 3, 6, 9, and 12 h with an inverted phase contrast microscope (Nikon, Eclipse TE2000-U) equipped with a CCD camera. For Boyden chamber Transwell migration assays, 10 5 cells in 400 l of serum-free medium were added to the upper chamber coated with fibronectin (10 g/ml), and the cells were allowed to adhere for 4 h. Subsequently, 400 l of starvation medium containing various stimulants was added to the lower chamber, and cell migration was allowed for 4 h. The cells on the upper surface were removed with a cotton swab, and the migrated cells on the lower surface were fixed and stained using the Diff-Quick stain kit (Dade Behring Inc., Newark, DE) as per the manufacturer's instructions. Each data point is a mean of six replicates, with average of three high power (400ϫ) fields/ filter. Statistical analysis used the Student's t test; p Ͻ 0.05 was considered significant. The experiments were repeated at least three times.
GST Fusion Protein Pulldown Assays-MCF-7 cells expressing PDGFR␤-GFP or its mutants were serum-starved and stimulated with PDGF BB (20 ng/ml) for 10 min prior to lysis. GST fusion protein pulldowns from 1-mg aliquots of lysate protein and 20 g of GST or GST-Cbl (Cbl-N, amino acids 1-357; and Cbl-N/G306E) fusion proteins were carried out as described (29), and bound proteins were visualized by Western blotting, as above.
Competition Between GST-PLC␥1 SH2-C and Cbl-N for PDGFR Binding-The GST fusion of the PLC␥1 C-terminal SH2 domain (residues 663-759) (PLC␥1 SH2-C) was generated by reverse transcription-PCR amplification from a human cDNA using primers (listed in supplemental Table S3) that incorporated BamHI and NotI sites for cloning into the pGEX-6P-1 bacterial expression vector (Pharmacia-LKB). The sequence-verified construct was used for expression and purification of the GST fusion protein as described (29). The GST tag was cleaved from fusion proteins using the PreScission protease, according to the manufacturer's protocols (GE Healthcare). For Cbl-N/ Cbl-N-G306E proteins, the GST tag was cleaved with human thrombin; contaminating thrombin was removed by incubating proteins with p-aminobenzamidine beads (Sigma) for 30 min at 4°C. Cleaved proteins were dialyzed against phosphate-buffered saline prior to binding assays.
For competitive binding assays, various concentrations of isolated PLC␥1SH2 domain protein were added to lysates of nonstimulated or PDGF BB-stimulated MEFs for 45 min at 4°C followed by pulldown using 20 g of GST or GST-Cbl-N (or Cbl-N/G306E) fusion proteins as above.
Confocal Immunofluorescence Microscopy-The cells were seeded on sterilized coverslips in regular growth medium for 24 h, serumstarved for 48 h, and stimulated with PDGF-BB as described above. The cells were fixed in 3.8% paraformaldehyde and immunostained as described (24) with anti-LAMP-1 antibody (1:1000) followed by Alexa Flour 594-conjugated goat antimouse antibody. Coverslips were mounted in Vectashield mounting medium containing 4Ј,6-diamidino-2-phenylindole for visualizing the nuclei (Vector Laboratories). The images were acquired with a Nikon C1 confocal microscope system under 600ϫ magnification.

RESULTS
Endogenous Cbl Is Required for Efficient Ligand-induced Ubiquitinylation and Degradation of PDGFR␤-To address the role of endogenous Cbl in PDGFR regulation, we first assessed the extent of ligand-induced PDGFR␤ ubiquitinylation in Cbl ϩ/ϩ and Cbl Ϫ/Ϫ MEF cell lines. Ligand-dependent ubiquitinylation of PDGFR␤ was detected in both of the independently derived Cbl ϩ/ϩ MEF lines (HG and DB; Fig. 1, A and B); however, the extent of ligand-induced PDGFR␤ ubiquitinylation was greatly reduced in both Cbl Ϫ/Ϫ MEFs (HG and DB), indicating that endogenous Cbl is the major E3 mediating PDGFR␤ ubiquitinylation in MEFs. Anti-PDGFR␤ IP followed by anti-PDGFR␤ IB (Fig. 1, A and B) revealed that ligand-induced loss of the PDGFR␤ protein was substantially slower in both Cbl Ϫ/Ϫ MEFs; concomitantly, we observed a slower and less pronounced decrease in tyrosinephosphorylated PDGFR␤ (Fig. 1, A  and B). The pattern of PDGFR␤ phosphorylation indicated that impaired PDGFR␤ ubiquitinylation in Cbl Ϫ/Ϫ cells was not due to defective PDGFR␤ activation.
To exclude that defective PDGFR␤ ubiquitinylation and degradation in Cbl Ϫ/Ϫ MEFs was not an artifact of cell line derivation, we reconstituted the Cbl expression in both (HG and DB) Cbl Ϫ/Ϫ MEF lines (Fig. 1C). Comparison of the Cbl Ϫ/Ϫ /HA-Cbl versus Cbl Ϫ/Ϫ / vector cells demonstrated that Cbl reconstitution fully restored the PDGF-BB-induced ubiquitinylation and degradation of PDGFR␤ (Fig. 1, D and E). Correlating with a higher level of Cbl, PDGF-induced loss of phosphorylated and   . S1). Thus, the defect in PDGFR␤ ubiquitinylation and degradation in Cbl Ϫ/Ϫ MEFs is solely due to a lack of Cbl expression. The role of endogenous Cbl in ubiquitinylation and degradation of PDGFR␤ was independently verified in MCF-7 cells with stable knock-down of Cbl (supplemental Fig. S5).
Identification of the Cbl TKB Domain-binding Site on PDGFR␤-To identify the TKB domain-binding site(s) on PDGFR␤, we used GST fusions of Cbl TKB domain (GST-Cbl-N) or its nonbinding point mutant (GST-Cbl-N-G306E) to pull down wild type (WT) PDGFR␤ or its potential Cbl TKB domain-binding site Tyr 3 Phe mutants ectopically expressed in PDGFR␤-null MCF-7 cells (Fig. 2A). The previously described consensus sequence (N/D)XpYXXX (29) suggested the sequence around Tyr 1021 in PDGFR␤ (NDpYIIPL) as a potential Cbl TKB domain-binding motif (supplemental Fig. S6). We also included a PDGFR␤ mutant with a Tyr 3 Phe mutation in the Grb2 SH2 domain-binding site (Y716F) and the double mutants as controls. Analysis of cell lysates demonstrated that WT and mutant receptors were expressed at comparable levels and underwent similar levels of overall tyrosine phosphorylation upon PDGF BB stimulation ( Fig.  2A). The control GST-Grb2 SH2 domain pulled down WT or mutant PDGFRs as long as the Grb2-binding site Tyr 716 was intact (Fig. 2B). The GST-Cbl-TKB fusion protein could pull down the PDGF-stimulated WT PDGFR␤, whereas the G306E mutation abolished the TKB interaction with PDGFR␤ (Fig. 2C). Importantly, GST-Cbl-TKB failed to pull down the Y1021F mutant but could pull down the PDGFR␤ Y716F and Y1009F mutants (Fig. 2C). The inability to pull down Y1021F PDGFR␤ indicates that Tyr 1021 is the predominant Cbl TKB domain-binding site on PDGFR␤.
Competition between Cbl TKB and PLC␥1 SH2 Domain Binding to Phosphorylated PDGFR␤-Binding of Cbl TKB domain to a known PLC␥1 SH2 domain-binding site (42,43,50,51) raised the possibility that these proteins may compete for binding to PDGFR. We carried out in vitro competition studies with purified recombinant proteins to explore this possibility. As expected, the GST fusion of the C-terminal SH2 domain of PLC␥1 (GST-PLC␥1 SH2-C) pulled down the phosphorylated PDGFR␤ from lysates of PDGF-stimulated cells (Fig. 3A), and this interaction was dose-dependently reduced by incorporating increasing concentrations of purified Cbl-N in the pulldown assays, whereas the Cbl-N-G306E protein failed to do so (Fig. 3B). Conversely, GST-Cbl-N pull down of the phosphorylated PDGFR␤ from lysates of PDGF-stimulated cells was dose-dependently abrogated by incorporating increasing concentrations of recombinant PLC␥1 SH2-C. These results clearly demonstrate that Cbl TKB and PLC␥1 SH2 domains compete for binding to phosphorylated PDGFR␤.
Inability to Interact with Cbl Alters the Lysosomal Trafficking of PDGFR␤-To assess whether Cbl regulates PDGFR through lysosomal degradation, a PDGFR␤-GFP chimera was generated FIGURE 2. Identification of phosphorylated Y1021 as the predominant Cbl TKB domain-binding site on PDGFR␤. A, MCF7 derivatives expressing C-terminally GFP-tagged chimeras of wild type PDGFR␤ (WT) or its mutants (Y716F, Y716F/Y21F, Y09F/Y21F, Y716F/Y09F, Y716F/Y21F, Y09F/Y21F, or Y716F/Y09F/Y21F) were serum-starved and stimulated with PDGF BB (20 ng/ml) prior to lysis. 100-g aliquots of lysate protein were subjected to anti-PDGFR␤ (upper panel) followed by anti-Tyr(P) IB (lower panel). B and C, 1-mg aliquots of protein lysate of MCF7 cells expressing the indicated PDGFR-GFP chimeras were used for binding to GST (control) or GST-Grb2-SH2 fusion protein and phosphorylated PDGFR chimeras visualized using anti-Tyr(P) IB. C, in vitro binding to GST, GST-Cbl-N (TKB domain) or GST-Cbl-N-G306E nonbinding mutant control fusion proteins was carried out as in B followed by anti-Tyr(P) immunoblotting. and stably expressed in non-PDGFR-expressing MCF-7 cells (as in Fig. 2A). We then examined the endocytic trafficking of the WT versus Y1021F PDGFR␤-GFP chimeras using confocal immunofluorescence microscopy. Upon ligand stimulation, WT PDGFR␤-GFP was internalized and initially colocalized with transferrin, indicating its entry into early and sorting endosomes (Fig. 4A, 10 min) and subsequently showed a lack of colocalization with transferrin (Fig. 4A, 30 min), as expected from trafficking of transferrin to the recycling endosomes and that of PDGFR␤-GFP chimera to the lysosomes (52). Indeed, PDGFR␤-GFP showed a nearly complete colocalization with the lysosomal marker LAMP-1 at 30 min, with only minor colocalization at 10 min (Fig. 4C). The signals of WT PDGFR␤-GFP chimera were essentially undetectable 90 min after PDGF stimulation (Fig. 4, A and C), consistent with its ligand-induced lysosomal degradation (19,20,33,34). In contrast, Y1021F PDGFR␤-GFP showed prolonged colocalization with transferrin (seen at 10 and 30 min) and only partial colocalization with LAMP-1 at 30 min (Fig. 4B). Furthermore, a substantial proportion of vesicles containing mutant PDGFR␤-GFP chimera were seen after 90 min, consistent with delayed degradation (Fig. 4, B and D). Confirming these results, biochem-ical analyses also showed impaired ligand-induced degradation of the non-Cbl-binding mutant Y1021F compared with WT PDGFR␤ (supplemental Fig. S2). Thus, the inability of Cbl recruitment via Tyr 1021 indeed leads to delayed sorting to the lysosome, whereas entry into early/sorting endosomes was unaltered.
Cells Expressing the Cbl TKB Domain Nonbinding Y1021F Mutant PDGFR␤ Display Reduced Cell Migration-Mutation of the Cbl TKB domain-binding site on Cbl targets has been shown to enhance downstream signaling (18 -20). However, because phosphorylated Tyr 1021 has been established as a PLC␥1 SH2 domain-binding site in PDGFR␤ (42,43,50,51), Tyr 1021 mutation in PDGFR␤ could impair PLC␥1-mediated functions, as indicated by a knock-in mutation in mice (53). To assess the functional consequences of the Y1021F mutation of PDGFR␤, we examined its impact on PLC␥1-mediated cell migratory response in our cell lines using a wound healing assay. PDGFR␤-null cells reconstituted with WT PDGFR␤ showed complete wound healing response within 12 h of culture in the presence of PDGF BB (Fig. 5). In contrast, the wound remained unclosed at 12 h in cells expressing PDGFR␤ Tyr 1021 (Fig. 5), or cells expressing the Y1009F/Y1021F or Y716F/Y1009F/Y1021F mutants (supplemental Fig. S3). Thus, mutation of the Cbl TKB domain-binding site on PDGFR␤, which coincides with a major PLC␥1-binding site, induces a loss of cell migration function. These results suggest that the major role of the Tyr 1021 phosphorylation site is to recruit PLC␥1, and direct binding of Cbl TKB domain and the ensuing Cbl recruitment could negate PLC␥1 signaling in addition to facilitating receptor ubiquitinylation and degradation. This possibility is further addressed below.

Cbl-deficient Cells Exhibit Increased PDGFR-PLC␥1 Association and Elevated Levels of Phosphorylated PLC␥1 in Response to PDGF-The overlapping Cbl TKB domain-and PLC␥1-SH2
domain-binding sites on PDGFR␤ suggested that Cbl deficiency might enhance PLC␥1 recruitment to PDGFR and its activation. To assess whether Cbl deficiency promotes increased PLC␥1 association with PDGFR␤, we coimmunoprecipitated PDGFR␤ with PLC␥1 in Cbl Ϫ/Ϫ /vector or Cbl Ϫ/Ϫ / HA-Cbl MEFs. Indeed, the level of association of PDGFR␤ with PLC␥1 was higher in Cbl Ϫ/Ϫ cells compared with the reconstituted cells (Fig. 6A). Anti-Tyr(P) immunoblotting also demon- FIGURE 3. Competition between Cbl TKB and PLC␥1 SH2 domains for binding to PDGFR␤. A, 1-mg aliquots of lysate proteins from unstimulated or PDGF BB-stimulated (20 ng/ml; 10 min) Cbl ϩ/ϩ MEF cells were incubated with 20 g of GST or GST-PLC␥1 SH2-C fusion proteins on beads, and bound phosphorylated PDGFR␤ was visualized using anti-Tyr(P) IB. 100-g aliquots of lysate protein were subjected to direct anti-Tyr(P) IB. The relative PDGFR␤ signals were quantified by densitometry using the Scion Image software (Scion Corp.) and are depicted as a proportion of signals observed with unstimulated cell lysates (set as 1). B, pulldown assay was carried out as in A except that indicated concentrations of purified Cbl-N or Cbl-N G306E proteins (cleaved from GST tags) were added for 45 min before the pulldown. C, purified recombinant PLC␥1 SH2-C at various concentrations was added to lysates prior to pulldowns with GST-Cbl-N (Cbl TKB domain) or GST-Cbl-N-G306E (nonbinding mutant) fusion proteins. Relative PDGFR␤ signals were determined by densitometry. OCTOBER 5, 2007 • VOLUME 282 • NUMBER 40 JOURNAL OF BIOLOGICAL CHEMISTRY 29341 strated a higher level of phosphorylated PDGFR in anti-PLC␥1 IPs in Cbl Ϫ/Ϫ MEFs. Use of phospho-specific antibodies indicated that PDGFR␤ Tyr 1021 was indeed hyperphosphorylated as were other sites (e.g. Tyr 1009 ) (supplemental Fig. S4).

Dual Mode of PDGFR␤ Regulation by Cbl
PDGF stimulation induced rapid PLC␥1 phosphorylation in Cbl ϩ/ϩ MEFs (seen at 3 and 10 min), which declined subsequently to reach nearly undetectable levels by 90 min (Fig. 6B,  lanes 13-18). In contrast, phospho-PLC␥1 levels were undi-  A and B, MCF-7 cells expressing PDGFR␤-GFP or PDGFR␤ Y1021F-GFP were serum-starved for 72 h and stimulated with 20 ng/ml PDGF BB for the indicated times. The cells were loaded with Alexa Fluor 546-conjugated transferrin for 10 min prior to fixation to visualize the early endosome/recycling endosomal compartment. The cells were imaged by confocal microscopy to visualize the colocalization of PDGFR␤-GFP (green) and Alexa 546-transferrin (red), seen as yellow in merged images. C and D, the cells were serum-starved and stimulated as in A and B and fixed, permeabilized, and immunostained for LAMP1 (H4A3 mouse monoclonal followed by Alexa Flour 594-conjugated goat anti-mouse antibody; red) followed by confocal microscopy. minished at 30 min in PDGF-stimulated Cbl Ϫ/Ϫ MEFs, started to decline by 90 min, and were still detectable at 270 min (Fig.  6B, lanes 1-6). The pattern of prolonged PLC␥1 phosphorylation was reversed by HA-Cbl reconstitution of Cbl Ϫ/Ϫ MEFs (Fig. 6B, lanes 7-12). Notably, total PLC␥1 levels were comparable in all cell lines (Fig. 6B, upper panel). To confirm this result, we further used MCF-7 cells with ectopically expressed PDGFR␤. As expected, PDGF BB induced PLC␥1 phosphorylation only in the cells with PDGFR␤ expression (Fig. 6D). When Cbl was knocked down in PDGFR␤-expressing MCF-7 cells using three distinct shRNAs (Fig. 6C, lanes 4 and 5), ligandinduced PDGFR␤ ubiquitinylation and degradation were impaired (supplemental Fig. S5). More importantly, Cbl knockdown prolonged the PLC␥1 phosphorylation induced by PDGF BB stimulation in PDGFR␤-expressing MCF-7 cells (Fig. 6D,  lanes 7-18), whereas the total PLC␥1 levels were unaltered in all cell lines (Fig. 6D, upper panel). Thus, in two distinct cellular systems, deficiency of Cbl protein prolonged the PLC␥1 phosphorylation upon PDGFR activation.
Cbl Deficiency Promotes the PLC␥1-mediated Cell Migration Downstream of PDGFR-To assess whether Cbl deficiency promotes the PLC␥1-mediated biological responses, we compared Cbl ϩ/ϩ and Cbl Ϫ/Ϫ MEFs for PDGF-induced cell migration using a modified Boyden chamber assay. Indeed, the PDGF BB dose-dependent cell migration was significantly (p Ͻ 0.05) higher in Cbl Ϫ/Ϫ cells compared with that in Cbl ϩ/ϩ cells (Fig.  7A). To further assess whether the enhanced cell migration in Cbl Ϫ/Ϫ MEFs required PLC␥1 activity, we used chemical inhibitors of PLC. The specific PLC inhibitor U73122 (54) inhibited the PLC␥1 phosphorylation (Ն 90% at 1-3 M) as well as the phosphorylation of its downstream target MAPK; the inactive analog U73343 or Me 2 SO had no effect (Fig. 7B). U73122 but not U73343 blocked the PDGF-induced migration of MEFs (Fig. 7C), suggesting that a lack of Cbl results in hyperactivation of PLC␥1 by PDGFR, which in turn leads to enhanced downstream biological responses.

DISCUSSION
In this study, we establish that endogenous Cbl is critical for ligand-dependent ubiquitinylation and degradation of PDGFR␤. Importantly, our studies identify the Cbl TKB domain-binding site on PDGFR␤, which overlaps with the known PLC␥1 SH2 domain-binding site, thereby revealing a novel mechanism of negative regulation of a RTK by Cbl via competition for recruitment sites of a positive mediator of downstream signaling. Given the recent studies in lymphocytes that Cbl TKB domain-binding site on B-cell linker protein (BLNK) also corresponds to a PLC␥1 recruitment site (55), our analyses of a RTK (PDGFR) suggests that Cbl-mediated inhibition of PLC␥1 may provide a prevalent mechanism of reciprocal biological regulation of tyrosine kinase signaling.
Most of the analyses on Cbl-mediated ubiquitin modification and its role in lysosomal targeting of activated RTKs, including those of PDGFRs, have solely relied on the use of ectopically overexpressed Cbl proteins, thereby raising questions about the significance of endogenous Cbl in RTK regulation. Thus, our analyses of endogenous PDGFR␤ on two distinct MEF lines with or without Cbl expression, together with Cbl knock-down in a human cell line expressing ectopic PDGFR␤, provide the first direct evidence that Cbl indeed is the primary determinant of ligand-induced ubiquitinylation, lysosomal targeting, and degradation of PDGFR. These results further extend prior studies of human EGFR transfected into MEFs (24) and endogenous CSF-1 receptor in Cbl-null macrophages (25) and together provide a firm basis for a crucial regulatory role of Cbl in controlling activated RTK down-regulation via ubiquitin-dependent lysosomal targeting.
Prior studies of PDGFRs have demonstrated that an intact Cbl TKB domain is crucial for negative regulation. Yet, the Cbl TKB domain-binding site(s) on PDGFRs have not been identified. Previous work by us and others (29,56,57) has delineated a consensus Cbl TKB domain-binding motif (D/N)XpYXXX that is found in a number of known Cbl targets including EGFR, ZAP-70, Syk, and Sprouty 2 (29, 56, 58 -60). A close match of this motif corresponds to sequences surrounding Y1021 on human PDGFR␤ (DNDYIIPL), and our analyses ( Fig. 2 and supplemental Fig. S6) demonstrate that this motif indeed mediates Cbl TKB domain binding. The Y1021F mutation reduces the PDGFR␤ ubiquitinylation and lysosomal traffic (Fig. 4, B and  D), indicating that the ability of Cbl to interact with this motif is critical for Cbl-mediated regulation of PDGFR. Previous analysis of the PDGFR␤ Y1009F/Y1021F mutant in porcine aortic endothelial cells also showed diminution of PDGF-induced ubiquitinylation (61).
Although the previously identified Cbl TKB domain-binding motifs on other Cbl targets are apparently Cbl-selective, Tyr 1021 on PDGFR␤ is a known binding site for PLC␥1 SH2 domains (42,43,50). Potential interaction of Cbl, a negative regulator, and PLC␥1, a positive effector, with the same phosphotyrosine-containing motif predicted that (1), unlike other Cbl-regulated PTKs such as ZAP70 or Syk (29,58), Tyr 3 Phe mutation of Tyr 1021 may not render PDGFR hyperactive because of concurrent loss of interaction with PLC␥1, and (2) Cbl binding to a PLC␥1-binding motif on PDGFR␤ could provide an additional mechanism for negative regulation of receptor signaling distinct from Cbl-mediated PDGFR ubiquitinylation and lysosomal sorting. Indeed, Y1021F PDGFR␤, instead of being hyperactive, was defective in PDGF-induced cell migration, a function linked to PLC␥1 activation (36,41). Our results in this regard are consistent with published observations that PDGFR␤ Y1021F mutant is impaired in PDGF-induced migration in canine kidney epithelial cells (36), and a triple mutant (Tyr 3 Phe mutation of phosphatidylinositol 3-kinase binding sites at Tyr 739 and Tyr 750 together with Tyr 3 Ile mutation of Tyr 1020 ; equivalent to human Tyr 1021 ) was impaired in PDGF-induced cell proliferation and migration in mesangial cells (62). Strong evidence for functionally antagonistic effects of Cbl and PLC␥1 binding to the Tyr 1021 motif is provided by our analyses using Cbl-deficient cell lines and their parental and reconstituted controls, which demonstrated that lack of Cbl indeed promotes prolonged phosphorylation of PDGFR␤ (including that on Tyr 1021 ), prolonged PLC␥1 activation, and enhanced cell migration (Figs. 6, B and D, and 7A); the latter in our cell system was critically dependent on PLC␥1 activity as seen with the use of inhibitors (Fig. 7C).
Our findings that Tyr 1021 on PDGFR␤ is a shared binding site for Cbl and PLC␥1 may help explain, in part, the variability in the functional effects of Y1021F mutation when examined in different cell systems. For example, Y1021F mutant when expressed in canine kidney epithelial cells was shown to be defective in PLC␥1 activation and cell migration (36), whereas when expressed in porcine aortic endothelial cells, it did not affect cell migration in response to PDGF (63).  (bottom panel). B, enhanced PDGF-induced PLC␥1 phosphorylation in Cbl-deficient cells. MEFs (Cbl Ϫ/Ϫ , Cbl ϩ/ϩ , and Cbl Ϫ/Ϫ /HA-Cbl) were serum-starved for 48 h and stimulated with 20 ng/ml PDGF BB for the indicated time points. 100-g aliquots of cell lysate protein were subjected to anti-PLC␥1 IB and reprobed with anti-phospho-PLC␥1 (Tyr783) antibody. C and D, lentiviral infection was used to introduce Cbl shRNA or scrambled control shRNA into MCF-7-PDGFR-␤-GFP cells. 50-g aliquots of lysate protein were subjected to anti-Cbl (upper part of filter) and anti-␤-actin (lower part of filter; loading control) IB. Cbl Ϫ/Ϫ MEFs lysates (50 g) were served as a control (C). MCF-7-PDGFR␤-GFP cells with indicated shRNAs were serum-starved for 72 h and stimulated with 20 ng/ml PDGF BB for the indicated time points prior to lysis. 100-g aliquots of lysate protein were subjected to anti-PLC␥1 followed by anti-phospho-PLC␥1 IB (D).
The balance between the levels of Cbl and PLC␥1 and/or their relative recruitment to PDGFR could determine whether phosphorylation of Tyr 1021 functions as a positive or a negative regulatory mechanism. In addition, the effect of Tyr 1021 mutation could be influenced by how important a role other signaling pathways, such as the phosphatidylinositol 3-kinase activation, play in the cell migration response (63).
The ability of Cbl to bind to PLC␥1-binding site on PDGFR␤ provides a biochemical basis for how Cbl could selectively dampen PLC␥1-mediated signaling output in cells far beyond its general negative regulatory effect as a result of targeting receptors to lysosomal degradation. Somatic knock-out of Cbl in a chicken B cell line was shown to dramatically increase the PLC␥2-dependent calcium flux and apoptosis (64). Biochemical analyses in this system suggested that the Cbl TKB domain can directly bind to phosphorylated BLNK, an essential adaptor to recruit PLC␥1 to the B cell receptor, and that Cbl competes for PLC␥1 binding (65). Because PLC␥ binds to multiple phosphotyrosine motifs on BLNK and the Cbl TKB domain-binding site on BLNK has not been directly identified, it is not known whether a shared binding site on BLNK may also account for Cbl-PLC␥ competition in lymphocytes. Consistent with this possibility, the PLC␥-binding sites on BLNK (65) correspond to potential Cbl TKB domain-binding motifs, with one motif (DDSY 115 ) a close match. Cbl was also found to attenuate PLC␥1 signaling downstream of EGFR, and it was suggested that Cbl may compete with PLC␥1 activation; however, no evidence for a shared binding site or another mechanism for direct competition was provided in this study (66). It is more likely that reduction of EGFR-mediated PLC␥1 activation by Cbl is a consequence of general negative regulation of EGFR rather than the competitive binding mechanisms deduced from our study.
The competition between Cbl and PLC␥1 can be considered an example of an increasing numbers of regulatory interactions that appear to fine tune the activity of Cbl as a negative regulator of RTK signaling. The interaction of Cbl TKB domain with Sprouty 2, itself a negative regulator of RTK signaling, can allow bimodal regulation with two negative regulators counteracting each other's action at certain times while allowing Cbl action at other times (67). Additional studies point to a potential reciprocal regulation between Cbl and c-Src in which Cbl-mediated ubiquitinylation facilitates Src degradation, whereas Srcdependent phosphorylation of Cbl targets it for auto-ubiquitinylation and degradation (68). Finally, activated Cdc42 appears to inhibit the EGFR degradation by sequestering Cbl in a Cbl-Cool-1/␤-Pix-CDC42 ternary complex (69). Thus, competition between Cbl and PLC␥ binding to RTKs or adaptor proteins may provide an analogous regulatory mechanism to more precisely regulate RTK signaling via specific pathways and control Cbl action temporally and/or spatially in an activated cell. It is notable, however, that in some cell types PLC␥ itself appears to be a target of Cbl-mediated negative regulation. For example, Cbl-b-mediated ubiquitinylation of PLC␥1 in T lymphocytes reduced its tyrosine phosphorylation and activation apparently in a proteolysis-independent manner (70). In another study, up-regulation in the levels of Cbl-b and other E3s in anergic T FIGURE 7. Enhanced PDGF-induced chemotaxis in Cbl-deficient MEFs is PLC-dependent. A, serum-starved Cbl Ϫ/Ϫ or Cbl ϩ/ϩ cells were allowed to attach to the top surface of Boyden chamber filters for 4 h and allowed to migrate toward medium containing the indicated concentrations of PDGF ␤ for 4 h. The migrated cells were counted under light microscopy. The data are presented as the means plus one S.D. of six replicates, and the asterisks mark statistically significant difference (p Ͻ 0.005). Similar results observed in both HG and DB MEFs. B, the indicated serum-starved MEFs were treated with various concentration of PLC inhibitor U73122, its inactive analog U73343, or Me 2 SO (vehicle control) for 10 min, followed by a wash with starvation medium. The cells were stimulated with PDGF BB (20 ng/ml) for 10 min prior to lysis. 50-g aliquots of lysate protein were subjected to anti-phospho-p44/42 MAPK, anti-phospho-PLC␥1 (Tyr 783 ), and anti-PLC␥1 IB. C, serumstarved MEFs (Cbl Ϫ/Ϫ , Cbl Ϫ/Ϫ /HA-Cbl, and Cbl ϩ/ϩ ) were incubated with PLC inhibitor U73122 or U73343 (1 M) for 10 min, washed with starvation medium, and trypsinized. The cells were analyzed for PDGF-induced chemotaxis as in A. OCTOBER 5, 2007 • VOLUME 282 • NUMBER 40 lymphocytes led to a ubiquitin-dependent loss of PLC␥1 protein, apparently because of lysosomal degradation (71).

Dual Mode of PDGFR␤ Regulation by Cbl
In summary, the results presented here demonstrate the essential role of endogenous Cbl in ligand-induced ubiquitinylation and lysosomal targeting of PDGFR␤. We identify the Cbl TKB domain-docking site on PDGFR␤ to be coincident with a known PLC␥1-binding site and demonstrate that loss of endogenous Cbl exaggerates the PLC␥1-mediated biochemical and cell biological responses. Our studies illustrate a biochemical mechanism of competitive binding of a negative regulator (Cbl) and a positive effector (PLC␥1) in fine-tuning signaling downstream of a prototypical RTK, PDGFR␤. Future studies using mutagenesis of Cbl, PDGFR␤, and PLC␥1 in the context of WT and Cbl-null cells should help assess the relative role of the competitive interaction of Cbl versus PLC␥1 in regulating RTK signaling.