YXXM Motifs in the PDGF-β Receptor Serve Dual Roles as Phosphoinositide 3-Kinase Binding Motifs and Tyrosine-based Endocytic Sorting Signals*

Phosphoinositide 3-kinases (PI 3-kinases) are important regulators of endocytic trafficking. Previous studies have shown that mutant human platelet-derived growth factor-β receptors (PDGFR), which contain Phe in place of Tyr at the two p85/p110 PI 3-kinase binding sites (PDGFR-F/F), are defective for both p85 binding and ligand-stimulated degradation. This suggested that p85/p110 regulates PDGFR trafficking. However, more recent work has identified hVPS34, and not p85/p110, as the major PI 3-kinase regulating the movement of receptors through the endosomal system. To reconcile this discrepancy, we hypothesized that YXXM motifs in the PDGFR might play a second role as Tyr-based lysosomal sorting motifs (YXXΦ). To test this, we replaced both YXXM motifs with a motif from LAMP-1, YQTI. This mutant PDGFR (PDGFR-YQTI) still underwent PDGF-stimulated autophosphorylation but did not bind p85. In CHO cells, both wild-type and YQTI receptors showed PDGF-stimulated turnover, whereas F/F receptors did not. In addition, uptake and degradation of cell surface-labeled YXXM and YQTI receptors was fast relative to F/F receptors. We also constructed chimeras containing extracellular and membrane-spanning domains from CD25 (Tac) and cytoplasmic tails containing the YQTI motif, two YXXM motifs, or two mutant FXXM motifs. The YXXM and YQTI chimeras mediated lysosomal delivery of fluorescein isothiocyanate-labeled anti-CD25 antibodies, whereas the F/F chimera was defective. Thus, YQTI motifs can target PDGFR for degradation in the absence of p85/p110 binding, and the p85/p110 binding motifs from PDGFR are sufficient to target Tac chimeras to the lysosome. These data suggest that the YXXM motifs in the PDGFR serve two distinct functions: PI 3-kinase recruitment and lysosomal targeting.


EXPERIMENTAL PROCEDURES
Cell Culture and Transfections-Two CHO-derived lines, GRC LR-73 cells (22) and TRV cells (23), were cultured as described. Cells were transfected using LipofectAMINE-Plus (Invitrogen) and selected by culture in G418.
cDNA Constructs and Mutagenesis-The wild-type human PDGF-␤ receptor cDNA was obtained from Dr. D. Bottaro, EntreMed, Inc., Rockville, MD. The wild-type sequence was mutated to GF 740 MDMS-KAESVDFMPM (PDGFR-F/F) or GY 740 QTISKDESVGYQTI (PDGFR-YQTI) (bolded residues indicate mutations) using standard 4-primer PCR techniques. The Tac-YQTI construct (24), containing the extracellular and membrane-spanning domains from Tac and the C-terminal of LAMP-1 (RSHAGYQTI), was provided by Dr. M. Marcks, University of Pennsylvania. To make Tac-PDGFR chimeras, the residues DGGYM-DMSKDESVDYVPM or DGGFMDMSKDESVDFVPM were introduced in place of the LAMP-1 sequence using standard 4-primer PCR techniques. All constructs were confirmed by sequencing.
Western Blots-Samples were blotted with antiphosphotyrosine (PY-20, Transduction Laboratories) or rabbit polyclonal antibodies raised against the N-terminal SH2 domain of p85␣.
PDGF Uptake and Degradation-Quiescent TRV cells expressing WT or mutant PDGFRs were incubated with 50,000 cpm/ml 125 I-PDGF (PerkinElmer Life Sciences) in F12 medium containing 1 mg/ml BSA for 60 min at 4°C. Cells were washed with PBS at 4°C and warmed in F12, 1% BSA at 37°C. After various times the medium was removed, and proteins were precipitated with 10% trichloroacetic acid. Trichloroacetic acid-soluble counts in the supernatant were counted and expressed as a percentage of initial 125 I-PDGF bound. Untransfected control cells showed minimal PDGF degradation (data not shown).
Metabolic Labeling and Turnover Studies-TRV cells expressing the WT or mutant PDGFRs were labeled overnight in serum-free medium containing 200 Ci/ml [ 35 S]methionine/cysteine (PerkinElmer Life Sciences). Cells were washed in PBS and incubated at 37°C in the absence or presence of 20 ng/ml PDGF-BB (Calbiochem). After 2 h the cells were washed in cold PBS, lysed, and immunoprecipitated with anti-PDGFR␤ antibody (Calbiochem). Proteins were separated by SDS-PAGE, and visualized by autoradiography using Enhance (PerkinElmer Life Sciences). PDGF-stimulated degradation was determined as the difference between unstimulated and PDGF-stimulated degradation, expressed as a percentage of the initial labeled receptors.
Degradation of Receptors Labeled with Anti-PDGF Receptor Antibodies-Anti-PDGFR␤ antibodies (Calbiochem) were iodinated using Chloramine T and desalted by chromatography on Sepharose G25. Labeled antibodies were more than 98% trichloroacetic acid precipitable. Quiescent TRV cells expressing wild-type or mutant PDGFRs were incubated with labeled antibodies in medium containing 1% BSA for 90 min on ice, washed, and then warmed to 37°C in F12 medium, 1% BSA containing 20 ng/ml PDGF-BB (Calbiochem). At various times the medium was removed and precipitated with 10% trichloroacetic acid. Soluble radioactivity was counted and expressed as a percentage of initial cell-associated radioactivity.
Degradation of Biotinylated Receptors-TRV cells expressing wildtype or YQTI PDGFRs were labeled with sulfo-NHS-biotin (Pierce) on ice (25). The cells were either lysed directly or after warming to 37°C for 1 h in the absence or presence of 20 ng/ml PDGF. Receptors were immunoprecipitated and separated by SDS-PAGE, transferred to polyvinylidene difluoride membrane, and visualized with avidin-HRP and ECL chemiluminescence reagents (Amersham Biosciences). Blots were quantitated by densitometry using NIH Image software.
Immunofluorescence Studies-GRC LR-73 cells stably expressing Tac-YQTI, Tac-PDGFR, or Tac-PDGFR-F/F chimeras were incubated overnight in medium containing 100 g/ml leupeptin and pepstatin and then for an additional 4 h at 37°C with FITC-labeled anti-CD25 antibody (Ancel). Fixed cells were imaged with a Nikon E-400 upright fluorescence microscope equipped with a Nikon 60ϫ 1.4 N.A. plan-apo oil immersion, infinity-corrected objective. When indicated, cells were permeabilized with 0.1% Triton X-100 for 10 min and stained with anti-LAMP-1 or LAMP-2 antibodies (provided by Dr. Ana Maria Cuervo, Albert Einstein College of Medicine). Images were acquired with a Cohu 4910 camera and NIH Image 1.62 analysis software and assembled using Adobe PhotoShop.

Replacement of YXXM Motifs in the PDGF Receptor with GYQTI blocks p85
Binding without Affecting Autophosphorylation or Internalization-Mutation of Tyr 740 and Tyr 751 to phenylalanine abolishes p85 binding to activated PDGFRs and in-hibits post-endocytic sorting (19); autophosphorylation at other tyrosine residues is not affected (26). To test the hypothesis that YXXM motifs in the PDGFR act as tyrosine-based endocytic sorting motifs, we created mutant receptors in which the Y 740 MDM and Y 751 VPM motifs were both replaced with the sorting motif from LAMP-1, GYQTI. In TRV cells stably expressing the WT-PDGFR or the double YQTI mutant (PDGFR-YQTI), both WT and YQTI mutant receptors underwent PDGFstimulated autophosphorylation (Fig. 1A). However, only the WT receptor showed significant association with p85 (Fig. 1B). The preservation of tyrosine autophosphorylation in the PDGFR mutants is important, as defects in receptor tyrosine kinase autophosphorylation can affect endocytic behavior (27).
Mutation of the p85 binding sites in the PDGFR (PDGFR-F/F) affects post-endocytic trafficking but not internalization (19). We therefore compared the degradation of 125 I-labeled PDGF by the cells expressing wild-type, PDGFR-YQTI, and PDGFR-F/F receptors, a process that reflects both receptor internalization and post-endocytic sorting of ligands to the lysosome. As expected, release of degraded 125 I-PDGF was identical for all three lines (Fig. 2). Potential differences in post-endocytic processing of the receptors is not reflected in this measurement, since PDGF dissociates from its receptor in the acidic early endosome.
PDGF-YQTI Receptors Undergo Normal PDGF-stimulated Degradation-To evaluate the post-endocytic sorting of internalized wild-type and mutant receptors, the stable lines were labeled overnight with [ 35 S]methionine/cysteine and then chased for 2 h in the absence or presence of PDGF. As previ- ously shown (19), PDGF increases the degradation rate of WT receptors, whereas the PDGFR-F/F mutants show a loss of PDGF-stimulated degradation (Fig. 3A). However, PDGF-stimulated degradation in the PDGFR-YQTI cells was in fact greater than that in cells expressing wild-type receptors.
We also evaluated the kinetics of cell surface PDGFR turnover by labeling cells with 125 I-labeled anti-receptor antibodies at 4°C followed by incubation at 37°C in the presence of PDGF. Wild-type and YQTI receptors were degraded at similar rates, whereas degradation of PDGFR-F/F was significantly slower (Fig. 3B). It should be noted that the rates of degradation in this experiment reflect basal plus PDGF-stimulated turnover of cell surface receptors, as opposed to Fig. 3A, which measures PDGF-stimulated degradation of total cell receptors. To directly measure the effect of PDGF on the degradation of cell surface receptors, we biotinylated the cells at 4°C followed by incubation in the absence or presence of PDGF for 1 h and analysis by immunoprecipitation and avidin-HRP blotting. In cells expressing wild-type or YQTI receptors, PDGF-stimulated degradation was similar, with only 30 -40% of receptors remaining after 1 h of PDGF stimulation. In contrast, more than 85% of F/F receptors remained in cells stimulated with PDGF for 1 h (Fig. 3C). Thus, using three different approaches, we have shown that the PDGFR-YQTI receptors are degraded in a manner that is indistinguishable from wild-type receptors, despite their inability to bind p85/p110 PI 3-kinases.

YXXM Motifs Function as Endocytic Motifs in Chimeric
Receptors-The ability of the YQTI motif to supplant the YXXM motif in the PDGFR suggests that YXXM motifs may function as tyrosine-based sorting motifs. In this case, YXXM motifs might be able to function as endocytic motifs in another context. We therefore transfected cells with chimeric receptors consisting of the extracellular and membrane-spanning domains from CD25 (Tac) and an intracellular tail containing the 740 -754 domain on the PDGF receptor, which contains both YXXM motifs. Both wild-type (Tac-PDGFR) and mutant (Tac-PDGFR-F/F) chimeras were constructed. For comparison, we used a previously described Tac-YQTI chimera that is sorted from the cell surface to lysosomes (24). To measure delivery of the receptors to lysosomes, the cells were incubated with FITC-labeled anti-Tac antibodies for 4 h at 37°C in the presence of lysosomal protease inhibitors. The distribution of both Tac-YQTI and Tac-PDGFR chimeras were similar (Fig. 4A) and co-localized with a lysosomal marker (Fig. 4B). In contrast, the distribution of the Tac-PDGFR-F/F chimera showed minimal lysosomal accumulation and persistent cell surface labeling (Fig.  4A). These data demonstrate that the YXXM motifs can function as lysosomal sorting motifs in an exogenous context. Moreover, the Tyr-to-Phe mutation that blocks PDGF receptor degradation also disrupts lysosomal delivery of the Tac chimeras.
These data reconcile the apparent requirement for p85 binding during PDGFR sorting (19) with more recent studies showing a requirement for hVPS34 and its product, PI(3)P, during endosomal sorting (3). Our data show that post-endocytic sorting of the PDGFR does not require binding to p85/p110 PI 3-kinase but instead requires an appropriate tyrosine-based sorting motif. Thus, the LAMP-1 YQTI motif functions in the context of the PDGFR, and the PDGF YXXM motif functions in the context of the Tac chimera system. Our data suggest that the p85 binding sites in the PDGFRs serve a dual role in signaling and postendocytic processing.
We do not yet know the intracellular receptor for the PDGFR YXXM motif. We could detect binding between the intracellular tail of the PDGFR and the -subunits from AP-2 and, to a lesser extent, AP-3 adapters in two-hybrid studies (data not shown). However, this binding was not blocked by mutation of Tyr 740 and Tyr 751 to phenylalanine. Thus, additional sorting motifs in the receptor may also be involved in post-endocytic sorting. This is consistent with the finding that Tyr 579 in the juxtamembrane region of the receptor is important for ligandinduced internalization (29). With regard to the sorting of the PDGFR, a Tyr-to-Phe substitution within the YXXM motif is not tolerated. Aromatic substitutions are tolerated in some endocytic motifs although not in lysosomal acid phosphatase and a truncated mannose 6-phosphate receptor (30,31), or at Tyr 579 in the PDGFR itself (29).
Ubiquitination of cell surface receptors and receptor tyrosine kinases is important for their internalization and subsequent degradation (32,33). For the EGF and PDGF receptors, recruit- ment of Cbl, which has ubiquitin-ligase activity, leads to receptor ubiquitination and endocytosis (34 -36). Monoubiquitination of an internalization-defective EGF receptor or a truncated PDGFR lacking the intracellular domain is sufficient to drive endocytosis and degradation (37,38). In contrast, in cblϪ/Ϫ fibroblasts, EGFR internalization is normal whereas degradation is impaired (39). Although these studies point to Cblmediated ubiquitination as an important regulator of endocytic sorting, it is unlikely that this accounts for the phenotype of the PDGFR Phe 740 /Phe 751 mutant. Cbl binds to the PDGFR via an N-terminal kinase-binding domain, which contains an atypical SH2 domain (40). Although the preferred phosphotyrosine motif for the N terminus of Cbl was predicted to be D(N/D)X(pY) (41), the crystal structure of the N-terminal domain complexed with the Zap-70 binding peptide suggests that Cbl binds to the motif NX(pY)XXE⌽ (40). Neither sequence is consistent with the DGGYMDMSKDESVDYVPML sequence in the PDGFR. Although recruitment of Cbl to the PDGFR can also occur via the APS adapter protein, APS binds to Y1021 in the PDGFR but not to Tyr 740 /Tyr 751 (43).
Although the YXXM motif in the PDGFR appears to serve two distinct functions during receptor signaling versus receptor targeting, it most likely does not do both at the same time. Structural studies on -subunit/YXXM motif interactions suggest that a phosphotyrosine residue should not be tolerated (44). Thus, during the initial phase of receptor activation, binding of p85 SH2 domains to phosphorylated YXXM motifs could preclude interactions with the cellular sorting machinery. Once PDGF dissociates in the acidic environment of the endosome, the balance between ligand-stimulated autophosphorylation and phosphatase-mediated dephosphorylation would shift, leading to a loss of p85 binding. Alternatively, recent data suggest that receptor dephosphorylation occurs when the receptor reaches membrane-associated PTP2A in the endoplasmic reticulum (45). In either case, this would facilitate the sorting of the internalized receptor toward the degradative pathway. This is consistent the finding that the epidermal growth factor receptor undergoes endosomal dephosphorylation prior to degradation (28). This model is related to one proposed for the CTLA-4 receptor in T-cells, where the phosphorylation of a YVKM motif determines whether the receptor binds to the Syp tyrosine phosphatase or to AP-2 adapters (42). However, in the case of CTLA-4, tyrosine phosphorylation of Tyr 165 prevents endocytosis of the receptor. This is not the case for the PDGFR.
Our data shows that p85 binding motifs in the PDGFR act as tyrosine-based sorting motifs for the lysosomal delivery of internalized receptors. Dephosphorylation of tyrosine-based motifs may serve as a switch between receptor signaling and degradation, thereby providing a synergistic mechanism for the termination of receptor-mediated signals.