EGF receptor ubiquitination is not necessary for its internalization

Abstract

Ubiquitination of the EGF receptor (EGFR) has been implicated in EGF-induced receptor internalization, lysosomal degradation, and down-regulation. Mutation of EGFR ubiquitination sites identified by mass spectrometry yielded receptor mutants that are weakly ubiquitinated and not down-regulated by EGF. However, these EGFR mutants were normally internalized. To examine whether this internalization was mediated by the residual ubiquitination, systematic mutagenesis of lysine residues in the kinase domain of the EGFR was performed to generate a receptor mutant that is not ubiquitinated. Mutations of a number of lysines inhibited kinase activity of the EGFR, thus leading to the inhibition of receptor internalization. However, a mutant lacking 15 lysine residues (15KR), which was negligibly ubiquitinated and normally phosphorylated, was internalized at a rate similar to that of the wild-type EGFR. As in the case of the wild-type EGFR, internalization of the 15KR mutant depended on the presence of clathrin, Grb2 adaptor, and Cbl ubiquitin ligase. These data imply that EGFR ubiquitination is not necessary for its internalization by clathrin-coated pits. Interestingly, the reconstitution of two major ubiquitination sites in the 16KR receptor mutant, which had impaired kinase activity and slow internalization kinetics, resulted in a partial rescue of ubiquitination and a complete rescue of receptor internalization. This result suggests that ubiquitination of the kinase-impaired receptor can mediate its internalization by the clathrin pathway. Altogether these data emphasize the robustness of the EGFR internalization process, which can be controlled by multiple kinase- and ubiquitination-dependent and -independent mechanisms.

Activation of the EGF receptor (EGFR) by EGF or other ligands at the cell surface triggers several signaling cascades leading to a cell-specific biological response (1). Activated EGFR also is rapidly endocytosed by clathrin-coated pits (2). After internalization, ligand–receptor complexes traffic through a series of endosomal compartments and are either returned back to the plasma membrane or degraded in lysosomes. The accelerated internalization and efficient sorting of activated EGFR to lysosomes result in the dramatic down-regulation of the receptor and serve as a negative-feedback regulation of receptor signaling (3). At the same time, internalized EGFR continues to signal from endosomes, and this endosomal signaling is thought to play an important role in determining the duration, intensity, and specificity of signaling processes (2).

Endocytosis and intracellular sorting of EGFR have been the most popular experimental models to study pathways and mechanisms of endocytic trafficking that are specific to signaling receptors. However, despite extensive research for more than two decades, the mechanisms of internalization and degradation of EGFR, as well as other receptor tyrosine kinases, are not well understood. Our understanding of the molecular interactions leading to the recruitment of activated EGFR into coated pits and receptor internalization is especially limited.

Recent studies reveal that EGFR and a number of other receptor tyrosine kinases are ubiquitinated in a ligand-dependent manner (4, 5). The E3 ubiquitin ligase responsible for ubiquitination of EGFR is Cbl, which is present as three isoforms in mammals (6, 7). Cbl proteins mediate EGF-induced EGFR down-regulation (6, 8). Thus, receptor ubiquitination has been proposed to serve as a sorting signal that targets activated receptor tyrosine kinases at the cell surface to clathrin-coated pits and in multivesicular endosomes to lysosomes (4, 9). Recently, ubiquitin conjugation sites in the kinase domain of the EGFR were identified (10). Mutational analysis of these lysine residues showed that EGFR ubiquitination is essential for the lysosomal targeting of the receptor, but appears to play no or a minimal role in the internalization step of endocytic trafficking (10). This conclusion was made on the basis of observing the normal internalization rates of EGFR mutants lacking major ubiquitination sites. However, these mutants preserved measurable residual ubiquitination because of the redundancy of the ubiquitin conjugation sites. Therefore, the possibility remained that this residual ubiquitination could be sufficient for the internalization of the EGFR mutants. To test this hypothesis, all lysine residues in the kinase domain of EGFR were mutated. These mutagenesis studies revealed that the receptor mutants that displayed negligible ubiquitination were still internalized at a rate comparable to that of the wild-type EGFR, suggesting that receptor ubiquitination is not necessary for internalization through clathrin-coated pits.

Results

Tyrosine Phosphorylation and Ubiquitination of Multilysine EGFR Mutants.

In our previous studies, EGFR mutagenesis guided by mass spectrometry resulted in the generation of EGFR mutants, which were partially ubiquitinated at a level of ≈15–20% of ubiquitination of the wild-type EGFR (10). These mutants contained a maximum of nine substitutions of lysines to arginines within the tyrosine kinase domain of the receptor. It was then hypothesized that, in the absence of the major ubiquitination sites identified by mass spectrometry, other lysines can be ubiquitinated. However, further mass spectrometry analysis did not detect these cryptic ubiquitination sites. At the same time, 5KR and 9KR mutants stably expressed in porcine aortic endothelial (PAE) cells were internalized at rates indistinguishable from those of wild-type EGFR. Therefore, to test whether residual ubiquitination of the 9KR mutant is responsible for its rapid internalization, other lysine residues in the EGFR were systematically mutated to arginines. Because all previously identified ubiquitination sites were located within the tyrosine kinase domain, we limited our mutagenesis experiments to 22 lysines located within this domain (Fig. 1A). Lys-721 was not mutated because it is a critical catalytic residue of the receptor kinase.

Fig. 1.

Mutations of lysine residues in the kinase domain of the EGFR and their effect on receptor phosphorylation. (A) Mutated lysine residues were mapped on a previously reported structure of the EGFR kinase domain (1M14). Lysines previously identified as ubiquitination sites by mass spectrometry are shown in red. Lysines mutations that did not affect receptor tyrosine phosphorylation are shown in brown. Lysines mutations, which resulted in a moderate or severe inhibition of receptor phosphorylation, are shown in yellow or green, respectively. (B) Wild-type (WT) and mutant receptors were transiently expressed in PAE cells. The cells were untreated or treated with 20 ng/ml EGF for 2 min at 37°C. EGFR immunoprecipitates were probed with phosphotyrosine (PY20) and EGFR (1005) antibodies. Examples of several representative experiments are shown. (C) The amounts of phosphotyrosine in the EGFR immunoprecipitates normalized to the total EGFR were averaged from two experiments performed as in B and expressed as a percentage of that phosphorylation of the WT EGFR.

A series of EGFR mutants, in which the original 9KR mutation was combined with various combinations of other lysine–arginine (KR) mutations, was generated. These mutants were initially transiently expressed in PAE cells that lack endogenous EGFR, and the extent of their EGF-induced tyrosine phosphorylation and ubiquitination was estimated. Mutations of several lysines resulted in a substantial reduction of the receptor tyrosine phosphorylation, indicative of the reduced kinase activity [see Fig. 1B and supporting information (SI) Table 1]. In some cases, the strong inhibitory effect of the mutations was difficult to predict because of the distant location of a surface-exposed lysine residue, such as Lys-782, from the ATP- and substrate-binding interface of the kinase (Fig. 1A).

In summary, transient expression experiments demonstrated that it is possible to generate a 15KR mutant (six additional lysines mutated in the template of the 9KR mutant) that did not display any detectable impairment of tyrosine phosphorylation (Fig. 1 B and C). Mutations of various additional lysines in the template of 14KR and 15KR mutants led to a different extent of inhibition of tyrosine phosphorylation of the receptor, with the K704R and K799R substitutions having the smallest effect (Fig. 1 and SI Table 1). Essentially similar results were obtained when EGFR mutants were transiently expressed in HEK293 cells (data not shown).

Western blot analysis of EGFR immunoprecipitates from transiently expressing cells showed that the extent of ubiquitination did not decrease in the 10KR, 11KR, and 12KR mutants, compared with the 9KR mutant. However, EGF-dependent receptor ubiquitination appeared to be reduced in the 13KR mutant and further reduced in the 14KR mutant (data not shown). Subsequent mutagenesis of the 14KR did not result in a detectable decrease in ubiquitination of the 15KR, 16KR, 18KR, and 20KR mutants, despite that the kinase activity of the latter three mutants was increasingly inhibited (data not shown).

To more precisely characterize ubiquitination and endocytic parameters of the new multilysine EGFR mutants, we generated single-cell clones of PAE cells stably expressing 15KR, 16KR, and other receptor mutants. Immunoprecipitation of 15KR and 16KR from these cells revealed a low level of ubiquitination of these receptor mutants (Fig. 2). Quantifications suggested that the amount of ubiquitin associated with EGF-activated receptors in several clones of PAE/15KR and PAE/16KR cells was, on average, 1% of that detected in the wild-type EGFR (Fig. 2). At the same time, tyrosine phosphorylation and phosphorylation of Tyr-1068 (a major Grb2-binding site) was the same in the stably expressed 15KR mutant and wild-type EGFR (Fig. 2).

Fig. 2.

Ubiquitination and tyrosine phosphorylation of EGFR mutants stably expressed in PAE cells. (A) PAE cells stably expressing wild-type (WT) or mutant EGFR were untreated or treated with 20 ng/ml EGF for 2 min at 37°C. The EGFR immunoprecipitates were sequentially probed for ubiquitin, total phosphotyrosine (PY20), phosphotyrosine 1068 (pY1068), and total EGFR. Examples of three representative experiments are shown. Cell clone numbers are indicated after the names of the mutants. (B) The mean amounts of ubiquitin and phosphotyrosine in the EGFR immunoprecipitates were normalized to total EGFR (±SD) from three to four experiments performed as in A and expressed as a percentage of the ubiquitination and phosphorylation of the WT EGFR.

Internalization of Multilysine EGFR Mutants.

The data in Fig. 2 showed that PAE/15KR cells are a valuable model system to study receptor internalization of a mutant with normal kinase activity, but dramatically reduced ubiquitination, compared with 5KR, 9KR and Y1045F mutants, which were used in previous studies (10–13). Measurements of ¹²⁵I-EGF internalization rates were performed by using 1 ng/ml (low) physiological concentrations of EGF so that the rapid, clathrin-dependent pathway of endocytosis was not saturated (14). Under these conditions, ¹²⁵I-EGF was internalized in two independent clones of PAE/15KR cells at the rate comparable to the internalization rate of the wild-type EGFR (Fig. 3 A and B). This result strongly argues that receptor ubiquitination is not necessary for rapid internalization of EGFR. There also were no detectable differences in the EGF uptake rates when endocytosis of 100 ng/ml (high) EGF concentrations was compared (data not shown).

Fig. 3.

Internalization of ¹²⁵I-EGF by EGFR mutants. (A) WT and 15KR mutant-expressing PAE cells (clone 1) were incubated with 1 ng/ml ¹²⁵I-EGF for the indicated times at 37°C, and the ratio of internalized and surface ¹²⁵I-EGF was determined and plotted against time. (B) Rate constants (K_e) (±SD) of 1 ng/ml ¹²⁵I-EGF internalization were measured in cells stably expressing WT and 15KR receptors (clones 1 and 10). (C) PAE cells expressing 16KR (clone 2) and 16KR/2RK (clone 9) were incubated with 1 ng/ml ¹²⁵I-EGF for the indicated times, and the ratio of internalized and surface ¹²⁵I-EGF was determined and plotted against time. (D) Rate constants (K_e) (±SD) of 1 ng/ml ¹²⁵I-EGF internalization were measured in cells stably expressing WT EGFR and 16KR (clones 2 and 8) and 16KR/2RK (clones 9 and 11) mutants.

Unlike the 15KR mutant, 16KR displayed partially reduced (by ≈70%) tyrosine phosphorylation when stably expressed in PAE cells and stimulated with a subsaturating concentration of EGF (Fig. 2). Phosphorylation of Tyr-1068 in the 16KR mutant treated with 1 ng/ml EGF (concentration used to measure internalization rates) also was ≈30% of the phosphorylation of the wild-type EGFR mutant (data not shown). Analysis of EGF internalization in three independent clones of PAE/16KR cells revealed low rates of endocytosis of this mutant (Fig. 3 C and D). However, internalization of the 16KR mutant could be further reduced by 50% in cells depleted of clathrin-heavy chain (CHC) by siRNA (data not shown), thus indicating that the effect of lysine mutations in the 16KR mutant on clathrin-mediated internalization of EGFR was only partial. Because the level of ubiquitination of the 16KR mutant was essentially similar to that of the 15KR mutant, this reduced internalization of the 16KR mutant could not be attributed to a reduction of the receptor ubiquitination. However, because kinase activity is known to be required for clathrin-dependent internalization of EGFR (15–17), the low internalization rates of 16KR could be due to its partially inhibited tyrosine phosphorylation and/or kinase activity.

A number of studies demonstrated that the attachment of ubiquitin in frame to the cytoplasmic domains of various endocytic cargo can override the requirement for the regulated ubiquitination of the same cargo to promote its endocytosis and postendocytic sorting (18–20). To test whether restoration of ubiquitination can rescue the internalization of the nonubiquitinated and internalization-deficient 16KR mutant, two major ubiquitination sites, Lys-692 and Lys-713, were restored in the 16KR mutant by mutating arginines back to lysines. Stable PAE cell lines expressing the resulting 16KR/2RK mutant were generated. The reconstitution of two ubiquitin conjugation sites resulted in a significant ubiquitination of the 16KR/2RK mutant (Fig. 2A). The level of ubiquitination of the 16KR/2RK mutant was comparable with that of the 5KR, 9KR, and Y1045R mutants stably expressed in PAE cells (10). Tyrosine phosphorylation of the 16KR/2RK mutant was partially reduced, compared with the wild-type EGFR, although it was consistently higher than the phosphorylation of the 16KR mutant (Fig. 2). Interestingly, two independent clones of the PAE cell expressing the 16KR/2RK mutant internalized ¹²⁵I-EGF at a high rate similar to the internalization rate of the wild-type EGFR (Fig. 3 C and D). Thus, the partial reconstitution of ubiquitination was sufficient for a full rescue of internalization. This observation implies that receptor ubiquitination can mediate internalization of the receptor mutant that has impaired kinase activity.

Internalization of EGFR Lysine Mutants Requires Clathrin, Grb2, and Cbl.

To test whether ubiquitination-independent internalization of the 15KR mutant and ubiquitination-dependent internalization of the 16KR/2RK mutant are mediated by clathrin-coated pits, the clathrin-dependent pathway of internalization was blocked by CHC siRNA. This approach was established in our previous studies (21). Fig. 4 demonstrates that CHC siRNA strongly inhibited internalization of 1 ng/ml ¹²⁵I-EGF in cells expressing wild-type EGFR or both EGFR mutants.

Fig. 4.

Effect of siRNA knockdown of CHC, Grb2, and Cbl. Internalization rates of WT EGFR, 15KR (clone 1), and 16KR/2RK (clone 9) mutants were measured by using 1 ng/ml ¹²⁵I-EGF in cells transfected with control siRNA (mock), CHC siRNA, Grb2 siRNA, or a mixture of c-Cbl and Cbl-b siRNAs. The data are expressed as a percentage of the internalization rates in cells treated with nontargeting control siRNA.

Furthermore, the essential role of Grb2 and Cbl proteins in clathrin-mediated internalization of EGFR has been previously demonstrated (10, 11, 22–26). To test whether internalization of the 15KR and 16KR/2RK mutants used the same proteins, Grb2 and Cbl proteins (c-Cbl and Cbl-b) were knocked down by specific siRNA as described in refs. 10 and 11. Fig. 4 shows that depletion of Grb2 and Cbls had essentially similar effects on internalization of the wild-type and mutant EGFRs. The effects of Grb2 and Cbl knockdown on EGF internalization (reduction by 40–50%) were smaller than the effect of clathrin depletion (reduction by 60–70%). It is possible that Grb2-independent mechanisms of EGFR internalization through clathrin-coated pits exist. Alternatively, the incomplete effect of Grb2 and Cbl siRNAs can be due to the fact that residual Grb2 and Cbl remaining in the population of siRNA-treated cells are sufficient to support EGFR internalization. In contrast, even partial depletion of CHC had a dramatic effect on endocytosis in various cell types, including PAE cells. Regardless of the absolute effects of siRNAs, these experiments demonstrated that both 15KR and 16KR/2RK mutants are internalized through a clathrin-dependent pathway, and this internalization displays the same extent of the dependency on the presence of Grb2 and Cbl as does the internalization of the wild-type EGFR.

Degradation and Down-Regulation of EGFR Are Inhibited by Lysine Mutations.

Figs. 3 and 4 demonstrated rapid internalization of nonubiquitinated EGFR mutants such as 15KR. However, as shown in Fig. 5A, the 15KR mutant did not degrade in response to EGF stimulation. Furthermore, despite the rescue of internalization and the partial rescue of ubiquitination in the 16KR/2RK mutant, degradation of this mutant was severely impaired.

Fig. 5.

Mutations of ubiquitination sites slow degradation and down-regulation of EGFR. (A) Cells stably expressing WT EGFR, 15KR, 16KR, and 16KR/2RK mutants were incubated with 100 ng/ml EGF for the indicated times before lysis in the absence of orthovanadate and N-ethylmaleimide. EGFR was detected in cell lysates with antibodies 1005. (B) Cells stably expressing WT EGFR, 15KR, 16KR, and 16KR/2RK were incubated with 200 ng/ml EGF for the indicated times. The cells were then acid-washed to remove unlabeled EGF from surface receptors, and the number of surface EGF-binding sites was measured by using incubation with 50 ng/ml ¹²⁵I-EGF at 4°C for 1 h.

EGF-induced down-regulation of the surface EGFR was partially inhibited in 15KR and 16KR/2RK mutant-expressing cells (Fig. 5B). The partial effect can be explained by a combination of a slow lysosomal targeting leading to increased recycling and normal rates of internalization of these mutants. In cells expressing the 16KR mutant, the extent of inhibition of surface EGFR down-regulation was significantly higher presumably due to an additive effect of both reduced internalization (Fig. 3) and abolished degradation (Fig. 5A). Overall, these data are in agreement with the slow degradation and down-regulation rates of 5KR, 6KR, and 9KR mutants (10), suggesting that partial ubiquitination is not sufficient for effective lysosomal targeting of the EGFR.

Discussion

Modification of various endocytic cargo, including EGFR, by ubiquitination has been widely believed to be the mechanism responsible for endocytosis and postendocytic sorting of this cargo to lysosomes (27). However, in many cases, this theory has been difficult to test directly because ubiquitin conjugation sites have not been mapped in most of the ubiquitinated cargo. Our recent studies demonstrated that EGFR mutants that lack major ubiquitination sites and are poorly ubiquitinated (5KR and 9KR) were not efficiently sorted to lysosomes (10). Surprisingly, the same EGFR mutants were internalized at a normally high rate, which prompted the conclusion that receptor ubiquitination may not be necessary for endocytosis (10). At the same time, Cbl and its RING domain were shown to be necessary for internalization of these EGFR mutants, thus suggesting that a ubiquitination event mediated by Cbl is involved in EGFR internalization (10). We have suggested two explanations for this discrepancy. First, according to the threshold model, the residual ubiquitination of the cryptic sites in 5KR and 9KR mutants (10–20% of the maximal level of receptor ubiquitination) could be sufficient for receptor internalization, but not for the lysosomal targeting and degradation of these mutants. Second, Cbl activity is involved in endocytosis through ubiquitination of a yet unidentified protein, rather than ubiquitination of the receptor. In the present study, we tested the first hypothesis by systematically mutating multiple lysine residues in the EGFR and generating new EGFR mutants with negligible ubiquitination.

The critical result was obtained with the cells stably expressing the 15KR EGFR mutant. The extent of ubiquitination of this mutant was ≈1% of that of the wild-type EGFR. It is likely that both the amount of ubiquitin moieties per one EGFR monomer and a relative size of the pool of ubiquitinated 15KR receptors were dramatically reduced, compared with the wild-type EGFR. Given that EGFR internalization is a process controlled by the Michaelis–Menton kinetics, essentially the same rates of internalization of the 15KR mutant and the wild-type EGFR imply that the clathrin-mediated internalization of EGFR does not depend on receptor ubiquitination. This conclusion is in agreement with the data of an siRNA analysis, which showed that knockdown of the putative acceptors of the ubiquitinated cargo in clathrin-coated pits, such as epsin, Eps15, and Eps15R, did not have a specific effect on the clathrin-mediated internalization of EGFR (12, 21).

The experiments demonstrating slow endocytosis of the 16KR mutant and the restoration of endocytosis by adding back two ubiquitin conjugation sites in the 16KR/2RK mutant seem to contradict the results obtained with the 15KR mutant. We attributed the low internalization rates of the 16KR mutant to a reduced tyrosine phosphorylation of this mutant. First, the level of ubiquitination of stably or transiently expressed 16KR mutant of EGFR was not detectibly different from that of the 15KR mutant. Second, another multilysine mutant containing KR substitutions, which resulted in a severe inhibition of its kinase activity (18KR), displayed low internalization rates without a detectable reduction of receptor ubiquitination, compared with the 14KR, 15KR, and 16KR mutants (data not shown). Receptor kinase activity has been shown to be required for the recruitment of activated EGFR into clathrin-coated pits and receptor internalization (15–17). Clearly, kinase activity must be necessary for phosphorylation of tyrosines 1068, 1086, and 1045 involved in the recruitment of the Grb2–Cbl complex to activated EGFR, although receptor kinase may have other functions important for internalization.

The experiments with add-back mutants of the 16KR receptor, such as the 16KR/2RK mutant, demonstrated that ubiquitin moieties can mediate internalization in the absence of the receptor's full kinase activity. This experiment is analogous to the previously reported experiments in which ubiquitin polypeptide was attached to the C terminus of the EGFR mutant lacking the entire cytoplasmic domain (20). Such a truncated receptor was internalization-deficient, but became constitutively internalized upon attachment of the ubiquitin polypeptide (20). Unlike the latter experiments, ubiquitination of the 16KR/2RK mutant was EGF-dependent. Therefore, it can be suggested that, in principle, receptor ubiquitination can mediate clathrin-dependent endocytosis of the EGFR. However, ubiquitination of the EGFR is not necessary if another kinase-dependent pathway is operational. The mechanism of the latter pathway remains unknown, but it must involve Cbl-dependent ubiquitination of an unknown protein (protein X). Whereas the partial kinase activity of the 16KR/2RK mutant was evidently sufficient for recruitment of enough Grb2 and Cbl and ubiquitination of two available lysines, it appears insufficient for a Cbl-dependent ubiquitination of protein X, which is presumably necessary for internalization of the 15KR mutant. It is possible that impaired kinase activity results in insufficient phosphorylation of Tyr-371 in the Cbl linker domain. This phosphorylation event has been proposed to augment Cbl E3 ligase activity (8, 28). An alternative explanation of the high rate of 16KR/2RK internalization could be related to an apparent increase of the tyrosine phosphorylation of this receptor mutant, compared with the parental 16KR mutant (Fig. 2B). Although the increase was partial and did not convert the 16KR/2RK mutant into a fully phosphorylated EGFR, it is possible that such an increase was above the threshold in receptor phosphorylation, which is sufficient for rapid internalization by the mechanism used by the 15KR mutant.

In summary, our data suggest that EGFR is normally internalized by clathrin-coated pits by using a tyrosine kinase-, Grb2-, and Cbl-dependent mechanism. A ubiquitination event is involved in this pathway. However, ubiquitination of the receptor is not necessary. If kinase activity is partially inhibited, ubiquitination of the receptor can mediate its internalization. Thus, EGFR internalization can be mediated by multiple mechanisms. Such a robustness of this step of EGFR trafficking and the redundancy of the internalization mechanisms can explain why these mechanisms have been so difficult to elucidate.

Materials and Methods

Reagents.

The monoclonal antibody specific to phosphotyrosine (PY20) was obtained from BD Transduction Laboratories (San Diego, CA). The monoclonal antibody to EGFR (Ab528) was purchased from American Type Culture Collection (Manassas, VA). The monoclonal antibody to ubiquitin (P4D1) and the polyclonal antibody to EGFR (1005) were from Santa Cruz Biotechnology (Santa Cruz, CA). The monoclonal antibody to EGFR phosphotyrosine 1068 (pY1068) was from Cell Signaling Technology (Beverly, MA).

Plasmid Constructs and Point Mutations.

The EGFR construct was described in ref. 11. Point mutations in the EGFR 9KR construct (10) were generated by using the QuikChange site- and multisite-directed mutagenesis kits according to the manufacturer's protocols (Stratagene, La Jolla, CA). All point mutations were verified by automatic dideoxynucleotide sequencing.

Cell Culture and Transfections.

PAE cells were grown and transiently transfected as described (10). The cells were used for experiments 2 days after transfection. PAE cell lines stably expressing various EGFR mutants were generated as described in ref. 10.

In siRNA knockdown experiments, CHC siRNA duplex 2 (21), Grb2 siRNA duplex 3 (11), or a mixture of c-Cbl siRNA and Cbl-b siRNA duplexes (10) were resuspended in 1× siRNA universal buffer (Dharmacon, Lafayette, CO) to 20 μM. PAE/EGFR cell lines in six-well plates (50–60% confluency) were transfected with 10 μl of CHC siRNA2 duplex, the Grb2 siRNA3 duplex, or a mixture of 5 μl of c-Cbl siRNA and 5 μl of Cbl-b siRNA duplexes in 5 μl of DharmaFECT I reagent (Dharmacon) for 3 days. For mock transfections, the nontargeting siRNA duplex (Dharmacon) was used. The cells were replated to new 12-well plates 24 h before experiments.

Immunoprecipitation and Western Blotting.

To examine ubiquitination and tyrosine phosphorylation of EGFR by Western blotting, PAE cells transiently transfected with EGFR constructs or PAE/EGFR cell lines were grown in 60-mm dishes for 2 days, treated with 20 ng/ml EGF for 2 min at 37°C, and washed with Ca²⁺, Mg²⁺-free PBS. Cells were lysed by scraping with a rubber policeman in Triton X-100/glycerol solubilization buffer (21) supplemented with 10 mM N-ethylmaleimide and 1 mM sodium orthovanadate. The lysates were then cleared by centrifugation for 10 min at 14,000 × g. EGFR was immunoprecipitated with Ab528. The precipitates were washed twice with Triton X-100/glycerol solubilization buffer supplemented with 100 mM NaCl and once without NaCl and then denatured by heating in a sample buffer. The immunoprecipitates and aliquots of lysates were resolved on SDS/7.5% PAGE, transferred to the nitrocellulose membrane, and probed by Western blotting with various antibodies, followed by species-specific secondary antibodies or protein A conjugated with horseradish peroxidase (Zymed Laboratories, South San Francisco, CA). The enhanced chemiluminescence kit was from Pierce (Rockford, IL). Several x-ray films were analyzed to determine the linear range of the chemiluminescence signals, and the quantifications were performed by using densitometry.

In experiments examining EGFR degradation, PAE/EGFR cell lines in 12-well plates were incubated with 100 ng/ml EGF for indicated times and then lysed with Triton X-100/glycerol solubilization buffer as earlier, with the exception that orthovanadate and NEM were omitted from the lysis buffer to allow comparable efficiency of recognition of inactive and active EGFR by blotting with anti-EGFR antibody 1005.

Internalization and Down-Regulation of EGFR.

Mouse receptor-grade EGF (Collaborative Research, Bedford, MA) was iodinated by using a modified chloramine T method as described in ref. 11. ¹²⁵I-EGF internalization was measured by using 1 ng/ml ¹²⁵I-EGF, and the specific rate constant for internalization (K_e) was calculated as described in ref. 11. A low concentration of ¹²⁵I-EGF was used to avoid saturation of the clathrin-mediated internalization pathway.

The number of surface EGFR before and after EGF treatment was measured by using a ¹²⁵I-EGF binding assay as described in ref. 10.

Abbreviations

CHC

clathrin-heavy chain

EGFR

EGF receptor

PAE

porcine aortic endothelial.