Tyrosine Phosphorylation of the β2 Subunit of Clathrin Adaptor Complex AP-2 Reveals the Role of a Di-leucine Motif in the Epidermal Growth Factor Receptor Trafficking*

Tyrosine phosphorylation of the β2 subunit of clathrin adaptor complex AP-2 was detected in three types of cells treated with epidermal growth factor (EGF). The tyrosine phosphorylation was observed during recruitment of EGF receptors into coated pits at 4 °C and reached maximum at 37 °C at post-recruitment stages of endocytosis. An inhibitor of EGF receptor kinase completely abolished this phosphorylation in all cell types, whereas the inhibitor of Src family kinases partially inhibited β2 phosphorylation in A-431 cells but not in HeLa cells. By using β2 subunit tagged with yellow fluorescent protein that is effectively assembled into AP-2 complex, the major phosphorylation site of β2 was mapped to Tyr-6. Analysis of cells expressing dominant-interfering mutant μ2 subunit of AP-2 suggested that β2 phosphorylation is partially mediated by the receptor interaction with the μ2 subunit. Mutation of leucine residues 1010 and 1011 motif in the EGF receptor resulted in the severe inhibition of β2 tyrosine phosphorylation. From these data, we propose that interactions of the EGF receptor with AP-2 mediated by the receptor 974YRAL and di-leucine motifs may contribute to β2 tyrosine phosphorylation. Surprisingly, mutation of the Leu-1010/Leu-1011 motif resulted in impaired degradation of EGF receptors, suggesting the role of this motif in lysosomal targeting of the receptor.

Clathrin coats are specialized membrane structures that facilitate formation of endocytic vesicles and transport intermediates from the plasma membrane, Golgi complex, and possibly endosomes. Clathrin triskelions and clathrin adaptor complex AP-2 are the major components of the clathrin coats located at the plasma membrane, which are responsible for endocytosis of various proteins, lipids, and viral particles (1,2). AP-2 is a stable tetramer consisting of the following four subunits: 100-kDa proteins ␣ and ␤2, also called adaptins, 2 (50 kDa), and 2 (17 kDa) (3). Cryoelectron microscopy analysis suggested that AP-2 has so-called "trunk" or core domain that is composed of 2, 2, and amino-terminal parts of ␣ and ␤2. Flexible hinge domains connect trunk with ␣ and ␤2 "ear" domains (3). The crystal structures of the ears and the trunk domains of AP-2 have been solved recently (4).
AP-2 is engaged in interactions with multiple proteins and lipids. The main clathrin binding interface is located within the hinge domain of ␤2 (5), although ␣ subunit can also interact with clathrin (6). The amino terminus of ␣ subunit is capable of binding to phosphoinositides and inositol polyphosphates, and these interactions are thought to serve important roles in the regulation of membrane docking of AP-2 (7). The carboxylterminal domain of ␣ subunit (␣-ear) binds to several proteins containing DPF(W) motifs, such as Eps15, epsin, and CALM/ AP180 (8). Individual subunits are engaged into multiple intersubunit interactions within the trunk domain, which is necessary for the assembly of the AP-2 complex (4).
Besides the structural and scaffolding functions of AP-2 in coated pits, AP-2 is involved in the recruitment of endocytic cargo to clathrin coats by means of recognition of the specific internalization sequence motifs in the cytoplasmic tails of the cargo. The interaction of YXX⌰ motif with the 2 subunit of AP-2 is best understood at the functional and structural levels (9). Di-leucine-containing motifs may interact with AP-2 either through 2 (10) or ␤2 subunit (11). The interactions with AP-2 are critical for internalization of many types of cargo proteins because mutations in AP-2-binding motifs dramatically reduced their internalization rates (12). In contrast, despite that the interaction of the epidermal growth factor receptor (EGFR) 1 with 2/AP-2 was demonstrated by several techniques in vivo and in vitro, this interaction is not essential for EGFR internalization, and its involvement in EGFR endocytosis may be limited to certain experimental systems and cell types (13)(14)(15).
The mechanisms of regulation of multiple interactions of AP-2 with cargo, membrane bilayer, clathrin, and accessory proteins are not fully understood. Serine/threonine phosphorylation of a ␤2 hinge appears to regulate ␤2 interaction with clathrin triskelions (16). Recently, phosphorylation of 2 was shown to play a role in the interaction of 2/AP-2 with cargo (17). Binding of phospholipids and phosphoinositides has also been proposed to modify the cargo binding ability of AP-2 (18).
In this work we found a new modification of AP-2, a tyrosine phosphorylation of ␤2 subunit, in cells stimulated with EGF. Mutagenesis of ectopically expressed ␤2 fused to yellow fluorescent protein (YFP) revealed one phosphorylation site in the amino terminus of ␤2 that is partially responsible for this phosphorylation. Analysis of EGFR mutants exposed the im-portance of a di-leucine motif of the receptor for ␤2 phosphorylation. This motif appears to be involved in the regulation of the turnover of EGFR protein.
Plasmid Constructs and Point Mutations-cDNA encoding rat ␤2 was provided by Dr. A. Nesterov (University of Colorado Health Sciences Center). Full-length ␤2 DNA was amplified using Pfu polymerase and inserted into pEYFP-N1 or pECFP-N1 vectors (Clontech, Inc., Palo Alto, CA) between XhoI and EcoRI restriction sites to generate ␤2-YFP and ␤2-CFP fusion proteins, respectively. All point mutations in ␤2-YFP were made using QuickChange mutagenesis kit according to the manufacturer's protocol (Stratagene). To generate YFP-tagged 2 subunit of AP-2, 2 internally tagged with HA epitope in pcDNA3.1 (15) was used as a template. Full-length YFP was amplified by PCR and cloned into an XhoI restriction site between an HA tag and the carboxylterminal sequence of 2 (corresponding to residues 237-435). All constructs and point mutations were verified by dideoxynucleotide sequencing.
Cell Culture and Transfections-HeLa cells were grown in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum, antibiotics, and glutamine. Porcine aortic endothelial (PAE) cells expressing wild-type or mutant EGFRs were described previously (20). PAE cells were grown in F12 medium containing 10% fetal bovine serum, antibiotics, and glutamine. Human epidermal carcinoma A-431 cells were grown in DMEM containing 10% calf serum, antibiotics, and glutamine. HeLa cells that inducibly express mutant 2 were described previously (15).
For transient expression, HeLa cells grown to 50 -80% confluency were transfected with pEYFP constructs using Effectene (Qiagen, Hilden, Germany). Transfected cells were split 1 day after transfection and used for experiments on the 2nd day. Cells were grown to about 90 or 50% confluency for immunoprecipitation or immunofluorescence experiments, respectively.
Immunoprecipitation of AP-2-A-431, PAE, or HeLa cells were treated with EGF at 37°C and washed with ice-cold Ca 2ϩ , Mg 2ϩ -free phosphate-buffered saline (CMF-PBS). The cells were solubilized by scraping with a rubber policeman in lysis buffer (50 mM HEPES, pH 7.3, 10% glycerol, 1% Triton X-100, 50 mM NaCl, 1 mM dithiothreitol, 1 mM EGTA, 5 mM EDTA, 10 mM sodium fluoride, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 544 M iodoacetamide, 10 g/ml aprotinin) containing 1 mM sodium orthovanadate, and by incubating further for 10 min at 4°C. The lysates were then cleared by centrifugation for 10 min at 14,000 ϫ g. The supernatants were recovered, and AP-2 complex was immunoprecipitated with saturating amount of AP.6 antibody specific to ␣-adaptin for 3 h at 4°C, followed by protein G-Sepharose beads for 1 h at 4°C. The precipitates were washed twice with lysis buffer supplemented with 100 mM NaCl and once without NaCl and then denatured by heating in the sample buffer. Immunoprecipitates and supernatants after immunoprecipitation were resolved on 7.5% SDS-PAGE followed by transfer to the nitrocellulose membrane and Western blotting with various antibodies to phosphotyrosine and AP-2 subunits. The primary antibodies were detected by species-specific secondary antibodies or protein A (Zymed Laboratories Inc., South San Francisco, CA) conjugated with horseradish peroxidase. The enhanced chemiluminescence kit was from Pierce.
Fluorescence Imaging-HeLa cells transiently expressing ␤2-YFP, ␤2-CFP, and/or 2-YFP protein for 2-3 days or PAE cell lines expressing wild-type or mutant EGFR were grown on glass coverslips. The cells were imaged live using CFP and YFP filter channels as described (21). In immunofluorescence experiments, the cells were washed with CMF-PBS and fixed with freshly prepared 4% paraformaldehyde (Electron Microscopy Sciences, Ft. Washington, PA) for 12 min at room temperature. The cells were mildly permeabilized using a 3-min incubation in CMF-PBS containing 0.1% Triton X-100 and 0.5% BSA at room temperature, further incubated in CMF-PBS containing 0.5% BSA at room temperature for 1 h with monoclonal AP.6 antibody to ␣-adaptin, X.22 antibody to clathrin heavy chain, or Ab528 antibody to EGFR, washed, and then incubated for 30 min with the secondary donkey anti-mouse IgG labeled with Cy3 (The Jackson Laboratories, West Grove, PA). Both primary and secondary antibody solutions were precleared by centrifugation at 100,000 ϫ g for 10 min. After staining the coverslips were mounted in Fluoromount-G (Southern Biotech Inc., Birmingham, AL) containing 1 mg/ml para-phenylenediamine. In experiments with PAE clones, the cells were incubated with 1 g/ml 4,6-diamidino-2-phenylindole (DAPI) for 5 min at room temperature to label cell nuclei. Threedimensional images of cells were obtained as described (22) using an epifluorescence imaging work station equipped with a ϫ100 oil immersion objective lens, cooled CCD, z-step motor, dual filter wheels, and a xenon 175-watt light source, all controlled by SlideBook 3.0 software (Intelligent Imaging Innovation, Denver, CO). Typically, 20 -30 serial two-dimensional images were recorded at 200-nm intervals. A Z-stack of images obtained was deconvoluted using a modification of the constrained iteration method. Final arrangement of all images was performed using Adobe Photoshop.
Degradation of [ 35 S]Methionine-labeled EGFR-PAE cell lines expressing wild-type or mutant EGFR were grown in 35-mm dishes and metabolically labeled overnight with [ 35 S]methionine (50 Ci per dish) in methionine-free DMEM containing 1% dialyzed fetal bovine serum. The cells were washed twice with binding medium (DMEM ϩ 0.1% BSA) and then incubated for 1 h in binding medium at 37°C. The cells were rinsed once more with binding medium and further incubated with or without 34 nM EGF in binding medium for the times indicated. At the end of the incubation, the cells were briefly rinsed with ice-cold CMF-PBS, and the plates were frozen at Ϫ80°C. After all incubations were completed, all cells were thawed and solubilized in a Triton X-100 lysis buffer supplemented with 1% deoxycholate. EGFR was immunoprecipitated with the antibody Ab528, and the immunoprecipitates were resolved on 7.5% SDS-PAGE. The gel was dried and exposed to x-ray film, and the amount of 35 S-labeled EGFR in immunoprecipitates was quantitated by densitometry.
EGFR Down-regulation-To monitor EGFR down-regulation, cells grown in 12-well dishes were incubated with EGF (34 nM) in binding medium at 37°C. After the indicated times, the cells were rinsed with ice-cold DMEM to remove unbound EGF and then incubated with ice-cold 0.2 M sodium acetate buffer, pH 4.5, containing 0.5 M NaCl for 2.5 min to remove surface-bound EGF followed by two rinses with cold DMEM to neutralize the acid. This acid wash procedure does not affect the binding properties of EGFR (23). The cells were then incubated with 4.25 nM 125 I-EGF at 4°C for 1 h followed by three rinses with cold DMEM to remove unbound 125 I-EGF. The cells were lysed in 1 M NaOH for 1 h at room temperature to determine cell-bound radioactivity. Nonspecific binding was measured for each time point in the presence of 100-fold molar excess of unlabeled EGF and was not more than 5-10% of the total counts.

RESULTS
␤2 Subunit of AP-2 Is Tyrosine-phosphorylated-In search of tyrosine-phosphorylated proteins that interact with the region of the EGFR important for receptor internalization (residues 945-1022), we performed affinity-binding experiments with the corresponding recombinant protein and lysates of EGFstimulated HeLa cells. Immunoblotting with phosphotyrosine antibodies detected only one protein band with an apparent molecular mass of ϳ100 kDa that was specifically bound to the 945-1022 EGFR fragment. Mass spectrometry analysis identified ␤2 subunit of AP-2 in this Coomassie-stained band. These experiments prompted us to investigate whether ␤2-adaptin is indeed tyrosine-phosphorylated in cells stimulated with EGF.
To demonstrate directly ␤2 tyrosine phosphorylation, AP-2 was immunoprecipitated from HeLa cells, and the presence of tyrosine-phosphorylated proteins in immunoprecipitates was tested by Western blotting (Fig. 1A). The tyrosine phosphorylation of a 100-kDa band that precisely overlapped with the ␤2 immunoreactivity was detected in cells treated with EGF. Because ␣C isoform of ␣ subunit of AP-2 runs on SDS gels similar to ␤2, the two subunits were separated using urea-SDS gels (24). Under these conditions ␤2 migrates significantly slower than ␣A and ␣C subunits (the latter is expressed at very low levels in HeLa cells). The tyrosine phosphorylation of a slowmigrating ␤2 subunit (ϳ118 kDa) was detected in urea gels, although the phosphotyrosine signal associated with ␤2 was weaker than in SDS gels (Fig. 1A). These results suggested that ␤2 adaptin is phosphorylated in an EGF-dependent manner.
The intensities of the phosphotyrosine antibody signal associated with ␤2 on Western blots were very low in HeLa cells expressing moderately low levels of EGFRs. The signal was detectable several minutes after continuous EGF treatment at 37°C and then decreased with time. In A-431 cells, expressing high levels of EGFRs, the tyrosine phosphorylation signal was much stronger and persistent during at least 30 min of EGF stimulation (Fig. 1B). ␤2 was not immunoprecipitated by several phosphotyrosine antibodies from cell lysates, presumably because the tyrosine-phosphorylated sequence of ␤2 was not accessible to antibodies and/or the amount of phosphorylated ␤2 molecules was very small compared with the total amount of tyrosine-phosphorylated proteins in lysates of cells treated with EGFR.
␤2 was only moderately phosphorylated when cells were stimulated with EGF at 4°C (Fig. 1C), conditions allowing EGFR activation, recruitment to coated pits but restricting endocytosis (22). For the maximum high level of phosphorylation, the incubation of cells with EGF at 37°C was required (Fig. 1C).
Inhibitors of EGFR Kinase Block ␤2 Phosphorylation-EGFR activation leads to activation of Src kinases that are responsible for tyrosine phosphorylation of clathrin heavy chain (25) and dynamin (26), the major functional components of coated pits. To determine whether EGFR kinase directly phosphorylates ␤2, or ␤2 phosphorylation is mediated through the activation of Src family kinases, the effect of EGFR kinase inhibitor PD168393 on ␤2 phosphorylation was compared with that of Src inhibitors. PD168393 completely blocked tyrosine phosphorylation of ␤2 in A-431 cells (Fig. 2). Src-family kinase inhibitor PP2 reduced ␤2 phosphorylation as compared with cells treated with the control compound, PP3. However, PP2 also partially inhibited EGFR kinase activity as evident from the decreased phosphorylation of the EGFR (also see (22)). A more specific Src family kinase inhibitor SU6566 (27) reduced ␤2 phosphorylation in A-431 cells by 25-30%. However, the same inhibitor did not affect ␤2 phosphorylation in HeLa cells (Fig. 2). These data suggested that ␤2 is phosphorylated either directly by the EGFR kinase or by another Src-unrelated tyrosine kinase that is activated by the EGFR. Src family kinases may contribute to ␤2 tyrosine phosphorylation in cells, such as A-431 cells, in which these kinases are effectively activated by EGFR.
In living cells ␤2-YFP (data not shown) and ␤2-CFP (Fig. 4A) were localized in small dots throughout the cell, a pattern of distribution typical of coated pits. When ␤2-CFP was co-expressed with the 2 subunit of AP-2 tagged with YFP, both subunits were highly co-localized (Fig. 4A). Immunoprecipitation experiments showed that 2-YFP was effectively incorporated into AP-2 complex (Fig. 3C), thus confirming that the localization of both ␤2-CFP and 2-YFP corresponds to localization of AP-2. Furthermore, ␤2-YFP was well co-localized in fixed cells with endogenous clathrin heavy chain and ␣-adaptin ( Fig. 4B), suggesting that ␤2-YFP is correctly targeted to coated pits. In some cells, a pool of ␤2-YFP was seen in the Golgi area. It is possible that overexpressed ␤2-YFP can incorporate into AP-1 complex located in the Golgi. Alternatively, overexpressed ␤2-YFP may form aggregates that concentrate in the Golgi area.
Tyrosine 6 Is the Major Phosphorylation Site in ␤2-To map phosphorylation sites in ␤2, we tested whether mutations of individual tyrosines in ␤2-YFP yield reduced phosphorylation of this fusion protein. Mutation of tyrosine 6 to phenylalanine decreased phosphotyrosine signal associated with ␤2-YFP by 50% (Fig. 5). Single mutations of all 23 other tyrosines in the trunk and hinge domains (residues 1-844) did not affect significantly tyrosine phosphorylation of ␤2-YFP. When Y6F mutation was combined with several other mutations (up to six Tyr-to-Phe mutations), no further decrease in phosphotyrosine signal was observed. Truncated version of ␤2-YFP-Y6F, in which the ear domain was removed (1-844), was phosphorylated to the same extent as the full-length Y6F mutant (Fig. 5). These data suggest that Tyr-6 is important for phosphorylation of ␤2 in cells stimulated with EGF, but there are other phosphorylation sites, which were not identified by mutagenesis. These additional sites are probably located within the aminoterminal trunk domain of ␤2. In support of this notion, tyrosine phosphorylation was detected in the trunk but not ear/hinge fragments of endogenous ␤2 protein that was proteolytically cleaved with elastase or trypsin (data not shown).
EGFR Phosphorylates ␤2 through Its Interaction with 2 and Possibly Direct Interaction with ␤2-We have shown previously that EGFR directly binds 2 subunit of AP-2 and that Tyr-974containing motif of the receptor is responsible for this interaction (14,15,28). Does this interaction with AP-2 mediate ␤2 phosphorylation? To test this possibility, EGFR mutants that were stably expressed in PAE cells were utilized (20). Stimulation of PAE cells expressing EGFR-Y974A mutant produced impaired tyrosine phosphorylation of ␤2 as compared with wild-type EGFR expressing cells (Fig. 6A). This result suggested that the interaction of the 974 YRAL motif of EGFR with AP-2 is partially responsible for ␤2 phosphorylation.
To confirm this conclusion directly, HeLa cells that inducibly express mutant 2 incapable of interaction with the EGFR were used (15). In these cells the induction of the expression of HA-tagged 2 D176A/W421A mutant upon removal of tetracycline results in efficient replacement of endogenous 2 from AP-2 complexes with the mutant 2 (Fig. 6B), thus preventing the interaction of EGFR with AP-2 through 2. Expression of mutant 2 (minus tetracycline) significantly reduced EGF-dependent tyrosine phosphorylation of ␤2 (Fig. 6B). These data suggested that the interaction of EGFR with 2 contributes to ␤2 phosphorylation. Together, the experiments with mutants of EGFR and 2 indicate that the direct interaction of 974 YRAL motif with 2 partially mediates ␤2 phosphorylation.
To define other mechanisms of ␤2 phosphorylation by the EGFR, several PAE cell lines expressing EGFR mutants in putative internalization and AP-2-binding motifs were used (20). Fig. 6A shows that the EGFR mutant, in which leucines 1010 and 1011 were replaced by alanines, weakly phosphorylated ␤2. These experiments suggested that Leu-1010/Leu-1011 motif is critical for EGF-dependent ␤2 phosphorylation in PAE cells.
Leucines 1010 and 1011 Are Important for EGFR Turnover-Because Leu-1010/Leu-1011 motif is important for ␤2 phosphorylation, we speculated that this motif may regulate EGFR endocytosis. Previous studies showed that Leu-1010/Leu-1011 is not critical for internalization of EGFR through a rapid saturable pathway, although we observed some variation in the internalization rates among different clones of PAE cells ex- FIG. 3. Incorporation of ␤2-YFP and 2-YFP into AP-2 and tyrosine phosphorylation of ␤2-YFP. A, HeLa cells transfected with ␤2-YFP construct or mock-transfected for 2 days were lysed, and AP-2 complexes were immunoprecipitated (IP) with using AP.6. ␤2-YFP and endogenous ␤2 were detected by Western blotting of immunoprecipitates with Ab32. B, HeLa cells expressing ␤2-YFP for 2 days were treated with EGF (17 nM) for 5 min at 37°C and lysed, and AP-2 complexes were immunoprecipitated using antibody AP.6. Tyrosinephosphorylation of ␤2-YFP and endogenous ␤2 was detected by Western blotting using anti-phosphotyrosine antibodies RC20-HRP (pTyr) or 4G10. C, HeLa cells transfected with HA-2-YFP or mock-transfected were lysed, and AP-2 was immunoprecipitated using antibody AP.6. HA-2-YFP was detected in cell lysates and immunoprecipitates using anti-HA antibody, whereas ␣-adaptins were detected using antibody AC. 1. pressing this mutant (20). In particular, clones 1 and 16, used in this study, internalized 125 I-EGF about 20% slower than did cells expressing wild-type EGFR. Measurements of the degradation rates of metabolically labeled EGFR revealed significantly slower turnover of the L1010A/L1011A mutants compared with that of wild-type EGFR (t1 ⁄2 ϳ4 h for WT receptor versus Ն8 h for LL mutants; see Fig. 7A). EGF treatment reduced t1 ⁄2 of wild-type EGFR down to ϳ1.5 h. In contrast, EGF did not significantly accelerate the turnover of L1010A/L1011A mutants (Fig. 7A).
The slow degradation of mutant EGFR was also observed by fluorescence microscopy. After a 15-min stimulation of the cells with EGF, a similar pattern of endosomal localization of wildtype and L1010A/L1011A receptors was observed (Fig. 7B). After prolonged exposure of cells to EGF (2 h), mutant EGFR remained accumulated in the vesicular compartments, whereas a significant decrease in EGFR immunoreactivity was observed in cells expressing wild-type EGFR (Fig. 7B). In both cell lines, receptors were well co-localized with EEA.1 protein, a marker of early and intermediate endosomes during 2 h of continuous endocytosis (data not shown). The slow turnover of mutant EGFR resulted in the impaired EGF-dependent downregulation of the cell-surface receptors (Fig. 7C). Altogether, the data in Fig. 7 suggested that Leu-1010/Leu-1011 motif is involved in the regulation of the endosomal sorting of EGFR to the degradative pathway.

DISCUSSION
Protein phosphorylation is the major regulatory mechanism of clathrin-dependent endocytosis. In particular, serine and threonine phosphorylation of clathrin coat proteins plays an important role in the organization of macromolecular complexes during clathrin-mediated endocytosis (16,29). Tyrosine phosphorylation of coat proteins, such as clathrin heavy chain and dynamin, has been observed in cells stimulated with growth factors or hormones, although the precise role of these modifications is not fully understood (25,26). Hence we demonstrated tyrosine phosphorylation of the ␤2 subunit of AP-2 complex in cells stimulated with EGF. This is the first observation of tyrosine phosphorylation of AP complexes. Images of live cells were acquired through the YFP and CFP channels. B, HeLa cells expressing ␤2-YFP for 2 days were fixed with paraformaldehyde and stained with monoclonal antibodies to ␣-adaptin (AP.6) or clathrin heavy chain (X.22) followed by secondary donkey anti-mouse IgG conjugated with Cy3. The serial z-optical sections (thickness Ϫ0.2 m) were acquired through the Cy3 (red) and fluorescein isothiocyanate/green fluorescent protein (green) channels and deconvoluted as described under "Experimental Procedures." YFP and Cy3 images representing individual optical sections were merged after adjustment of both fluorescence signals to similar levels (Merge). Bars, 5 m.

FIG. 5.
Mutations of tyrosine residues in ␤2. HeLa cells were transfected with wild-type (WT) or mutant ␤2-YFP. After a 2-day expression, the cells were treated with 17 nM EGF for 5 min at 37°C and lysed, and AP-2 complexes were immunoprecipitated. Tyrosine phosphorylation of ␤2-YFP and endogenous ␤2 was detected in immunoprecipitates by Western blotting using RC20, and the amount of ␤2-YFP incorporated into AP-2 was determined using antibody Ab32. Y6F represents ␤2-YFP mutant in which Tyr-6 was replaced by Phe. Other mutants have additional point mutations or a carboxyl-terminal truncation made in the Y6F mutant (1-884/Y6F mutant, last residue is 884). Results are representative of 3-10 independent experiments with each mutant.
FIG. 6. Tyr-974 and Leu-1010/Leu-1011 of EGFR are important for ␤2 phosphorylation. A, PAE cell lines expressing wild-type EGFR (WT), Y974A or L1010A/L1011A (clone LL1) mutants were treated with EGF (17 nM) for 10 min at 37°C and lysed, and AP-2 was immunoprecipitated with antibody AP.6. Tyrosine phosphorylation of ␤2-adaptin was detected in immunoprecipitates using antibody RC20, and the amount of ␤2 protein was determined using antibody Ab32. EGFR was detected in supernatants of the immunoprecipitates using antibody 2913. B, HeLa cells expressing HA-2 W421A mutant under tet-off inducible promoter were grown in the presence or absence of tetracycline for 4 days, treated with EGF (17 nM) for 5 min at 37°C, and lysed. AP-2 was immunoprecipitated with antibody AP.6, and the immunoprecipitates were probed for tyrosine phosphorylation and ␤2 protein.
The replacement of endogenous 2 by the HA-2 mutant in AP-2 complexes was confirmed by blotting of AP-2 immunoprecipitates with polyclonal anti-2 antibody.
Because ␤2 phosphorylation was completely blocked by the EGFR kinase inhibitor and only partially inhibited by the Src kinase family inhibitor (Fig. 2), it is likely that ␤2 is a substrate of the EGFR kinase. It cannot be, however, ruled out, that ␤2 is phosphorylated by an unidentified tyrosine kinase activated by the EGFR. In A-431 cells, c-Src activated by EGFR may directly phosphorylate ␤2 or act by maximally activating the EGFR kinase through phosphorylation of tyrosine 845 in the kinase domain of the receptor. Interestingly, AP-2 was continuously phosphorylated (Fig. 1) and associated with EGFR after internalization in A-431 cells (28). It is therefore possible that ␤2 was tyrosine-phosphorylated after internalization in A-431 cells and that Src is involved in endosomal phosphorylation of AP-2 in these cells. A low level of ␤2 phosphorylation was observed at 4°C, conditions restricting endocytosis but allowing maximum recruitment of receptors into coated pits (22). The extent of tyrosine phosphorylation of ␤2 was maximally high after cell treatment with EGF at 37°C, suggesting that ␤2 phosphorylation occurs mainly during a post-recruitment step. This is consistent with the notion that the interaction with AP-2 is not necessary for the initial stages of EGFR endocytosis (15).
To examine the functional role of AP-2 tyrosine phosphorylation, we attempted to map phosphorylation sites, which would allow us to generate a dominant-interfering mutant of ␤2. However, we have not been able to map all phosphorylation sites. Mutation of Tyr-6 caused a 50% reduction of tyro-sine phosphorylation signal, suggesting that this residue could be the major phosphorylation site. This tyrosine is located in proximity to phosphatidylinositol-binding site on the surface of AP-2 core complex and is accessible for phosphorylation (4). Extensive mutagenesis of other tyrosines in ␤2-YFP did not reveal other phosphorylation sites. Furthermore, tandem nanospray mass spectrometry analysis did not reveal additional tyrosine phosphorylation sites. It is possible that multiple tyrosines are phosphorylated in addition to Tyr-6, and the mutations of these tyrosines are compensated by the phosphorylation of other tyrosines. It is also possible that phosphotyrosine antibodies cross-react with phosphoserine-or phosphothreonine-containing sequences in ␤2. In fact, mass spectrometry revealed phosphorylation of Ser-563 and Thr-647 of ␤2 (data not shown). Mutations of these residues slightly reduced the phosphotyrosine antibody signal associated with ␤2 (data not shown). However, this effect can be attributed to the conformational changes caused by mutations. Finally, we cannot formally rule out the possibility that the residual phosphotyrosine signal in Y6F-␤2-YFP is due to tyrosine phosphorylation of the YFP moiety in the ␤2 fusion protein. The interaction of EGFR with AP-2 has been demonstrated by co-immunoprecipitation and in in vitro assays with purified proteins (30 -32). The observation of EGF-dependent AP-2 tyrosine phosphorylation provides yet another evidence for the interaction of EGFR and AP-2 in intact cells. However, the role of this interaction remains FIG. 7. Degradation and down-regulation of L1010A/L1011A EGFR mutant. A, PAE cell lines expressing wild-type EGFR (WT) or LL1010A/L1011A mutant (clones LL#1 and LL#16) were metabolically labeled with [ 35 S]methionine followed by incubation with or without EGF (34 nM) in binding medium at 37°C for the times indicated. The cells were then lysed, and EGFRs were immunoprecipitated with Ab528. The immunoprecipitates were separated by SDS-PAGE, and the amount of 35 S-labeled EGFR in immunoprecipitates was determined by autoradiography and quantitated by densitometry. The amount of labeled EGFR is expressed as percent of this amount at time point "0". The data represent mean values from three independent experiments. B, WT and LL1 cells grown on coverslips were incubated with 34 nM EGF at 37°C for 15 min or 2 h, fixed and stained with monoclonal antibody to EGFR (Ab528), followed by secondary donkey anti-mouse IgG conjugated with Cy3. Cellular nuclei were labeled by DAPI. The images were acquired through the Cy3 (red) and DAPI (blue) channels with the same integration time for each experimental condition. C, wild type (WT), LL1, and LL16 cells grown in 12-well plates were incubated with 34 nM EGF at 37°C for the indicated times. The residual number of surface EGFR was determined using 125 I-EGF binding assay as described under "Experimental Procedures" and expressed as percent to that number at time point 0.
unclear. It has been shown by several experimental approaches that the elimination of EGFR-2 interaction does not significantly affect EGFR internalization (14,15). In this study we show that leucines 1010 and 1011 in EGFR are important for ␤2 phosphorylation. This observation implies that the receptor LL motif may engage into interactions with AP-2. Such interactions must be, however, weak and unstable in detergent extracts, because they could not be detected by co-immunoprecipitation. It is also possible that the mutation of leucines 1010/1011 affects the conformation of the receptor carboxyl terminus, thus indirectly influencing the ability of EGFR to phosphorylate AP-2. As in the case of mutants of Tyr-974-containing motif, Leu-1010/Leu-1011 mutants of EGFR displayed high internalization rates (20). Therefore, the Leu-1010/Leu-1011 motif is unlikely to have a significant role in clathrin-dependent internalization of EGFR.
Surprisingly, L1010A/L1011A mutant of EGFR expressed in PAE cells displayed a substantially reduced rate of the turnover (Fig. 7). The data are consistent with the model by which this LL motif is involved in the regulation of lysosomal targeting of unoccupied and occupied EGFR. Such function may be related to a proposed interaction of this motif with AP-2 ( Fig.  6). Previous studies (28,33) suggested that AP-2 can be detected in endosomes and lysosomes and that EGFR/AP-2 interaction may persist after receptor internalization. Alternatively, L1010A/L1011A motif can function in intracellular sorting independently on its role in AP-2 phosphorylation. Interestingly, Leu-679/Leu-680 motif in the juxtamembrane domain of EGFR was implicated into lysosomal targeting of the EGFR (34). However, mutation of the juxtamembrane LL motif affected folding of the EGFR molecule, suggesting that these residues could have an indirect role in the lysosomal targeting of the receptor (35).
Recently, the role of EGFR ubiquitination in endosomal sorting has been demonstrated using several experimental approaches (36,37). In our experiments, the level of ubiquitination of wild-type and L1010A/L1011A mutant EGFR was not significantly different (data not shown), suggesting that the function of LL motif could be downstream of the ubiquitination step. It is possible that prolonged association of EGFRs with AP-2 after internalization interferes with recycling of EGFR from early endosomes, thus facilitating binding of EGFR to endosomal sorting complexes containing ubiquitin-binding proteins. Future studies should elucidate how di-leucine sequences and ubiquitin moieties are recognized by the endosomal machinery, and whether these two types of signals represent sequential or redundant steps of the EGFR endosomal sorting.