Effect of Epidermal Growth Factor Receptor Internalization on Regulation of the Phospholipase C-γ1 Signaling Pathway*

The epidermal growth factor receptor (EGFR) ligands, epidermal growth factor (EGF), and transforming growth factor-α (TGFα) elicit differential postendocytic processing of ligand and receptor molecules, which impacts long-term cell signaling outcomes. These differences arise from the higher affinity of the EGF-EGFR interaction versus that of TGFα-EGFR in the acidic conditions of sorting endosomes. To determine whether EGFR occupancy in endosomes might also affect short-term signaling events, we examined activation of the phospholipase C-γ1 (PLC-γ1) pathway, an event shown to be essential for growth factor-induced cell motility. We found that EGF continues to stimulate maximal tyrosine phosphorylation of EGFR following internalization, while, as expected, TGFα stimulates markedly less. The resulting higher level of receptor activation by EGF, however, did not yield higher levels of phosphatidylinositol (4,5)-bisphosphate (PIP2) hydrolysis over those stimulated by TGFα. By altering the ratio of activated receptors between the cell surface and the internalized compartment, we found that only cell surface receptors effectively participate in PLC function. In contrast to PIP2 hydrolysis, PLC-γ1 tyrosine phosphorylation correlated linearly with the total level of Tyr(P)-EGFR stimulated by either ligand, indicating that the functional deficiency of internal EGFR cannot be attributed to an inability to interact with and phosphorylate signaling proteins. We conclude that EGFR signaling through the PLC pathway is spatially restricted at a point between PLC-γ1 phosphorylation and PIP2 hydrolysis, perhaps because of limited access of EGFR-bound PLC-γ1 to its substrate in endocytic trafficking organelles.

and differentiation of many cell types. At least five ligands are known to activate EGFR, including epidermal growth factor (EGF) and transforming growth factor ␣ (TGF␣). Progress has been made in the last two decades in elucidating structurefunction relationships for EGFR and other receptor tyrosine kinases, particularly in how signal transduction is modulated by self-phosphorylation of cytoplasmic tyrosine residues (1). This permits access to the kinase domain of EGFR (2) and allows the receptor to bind signaling proteins containing modular Src homology 2 (SH2) and phosphotyrosine-binding domains (3,4). Such interactions can affect the activity of the bound protein through transmission of conformational changes, enhancement of tyrosine phosphorylation, and/or localization in proximity to membrane-associated target molecules. One of the prominent signaling proteins activated by EGFR is the ␥1 isoform of phospholipase C (PLC) (5). This enzyme, which has two SH2 domains, catalyzes the hydrolysis of phosphatidylinositol (4,5)-bisphosphate (PIP 2 ), generating the second messengers diacylglycerol and inositol triphosphate and liberating PIP 2 -bound proteins (6). PLC-␥1 activity is positively modulated in vivo by association with EGFR and tyrosine phosphorylation by the receptor kinase, providing a link to ligand stimulation (7)(8)(9)(10).
Another consequence of EGFR activation is clustering of ligand-receptor complexes in clathrin-coated pits, which increases the rate of receptor internalization (11). Following endocytosis, receptor-ligand complexes and other components of the plasma membrane are delivered to early endosomes, where molecules are sorted for recycling back to the cell surface or degradation in lysosomes (12,13). Since the degradative route can yield down-regulation of total receptor mass and depletion of ligand from the extracellular milieu, endocytic trafficking has been recognized as an attenuation mechanism affecting long-term EGFR function (14,15). An unresolved question, however, is the contribution to signaling of the steady-state EGFR pool residing in pre-degradative internal compartments. It has been demonstrated that EGF remains predominantly associated with EGFR in sorting endosomes, and that internalized EGF-EGFR retain equal or greater tyrosine phosphorylation stoichiometry as well as competency in binding and phosphorylating signaling proteins (16 -21). This suggests that meaningful signal transduction might be extended after endocytosis of EGF (22,23). In contrast, the pH sensititivity of the TGF␣-EGFR interaction and differential trafficking of TGF␣ compared with EGF suggest that TGF␣ dissociates from EGFR under the acidic conditions of endosomes (24,25). At the pH found at the surface, EGF and TGF␣ exhibit indistinguishable * This work was supported in part by the National Science Foundation Biotechnology Program in the Division of Biological & Environmental Systems and the NIGMS, National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Supported by graduate fellowships from the National Science Foundation and the Merck/MIT Collaboration.
affinities for EGFR in an equilibrium competition assay (24). This disparity in ligand/receptor sorting could be responsible for differences in the cell responses to EGF and TGF␣. To evaluate such a possibility, it is first necessary to know whether internalized and surface complexes differ either qualitatively or quantitatively in signaling. We investigate here the effect of endocytosis and compartmentalization of EGFR on the magnitude of signaling through the PLC pathway. Because NR6 fibroblasts transfected with wild-type EGFR have been used extensively in previous studies of both PLC-␥1 activation (10, 26 -28) and endocytic trafficking of the EGFR (14, 29 -31), they were chosen as our model system. We employed a ligand-based approach to analyze the PLC pathway at three distinct points of regulation: tyrosine phosphorylation of EGFR, tyrosine phosphorylation of PLC-␥1, and hydrolysis of PIP 2 . We found that internalized EGFR are deficient in stimulating PLC function, and that the point of regulation lies downstream of PLC-␥1 tyrosine phosphorylation.

EXPERIMENTAL PROCEDURES
Cell Culture and Quiescence Protocol-NR6 mouse fibroblasts transfected with wild-type human EGFR (NR6 WT) (14,32) were cultured in Corning tissue culture-treated dishes in a 5% CO 2 environment. All cell culture reagents were obtained from Life Technologies, Inc. The growth medium consisted of minimum essential medium (MEM) ␣, 26 mM sodium bicarbonate with 7.5% fetal bovine serum, 2 mM L-glutamine, 1 mM sodium pyruvate, 0.1 mM MEM nonessential amino acids, and the antibiotics penicillin, streptomycin, and G418 (350 g/ml). Cells were growth arrested at subconfluence using restricted serum conditions without G418 (MEM-␣, 26 mM sodium bicarbonate with 1% dialyzed fetal bovine serum, 2 mM L-glutamine, 1 mM sodium pyruvate, 0.1 mM MEM nonessential amino acids, and the antibiotics penicillin/streptomycin) for 18 -24 h prior to experiments. Experiments were carried out in an air environment using MEM-␣, 13 mM HEPES (pH 7.4 at 37°C) with 0.5% dialyzed fetal bovine serum, 2 mM L-glutamine, the antibiotics penicillin/streptomycin, and 1 mg/ml bovine serum albumin as the binding buffer.
Receptor Binding and Internalization Studies-Mouse EGF (Life Technologies, Inc.) or human TGF␣ (Peprotech) were iodinated with 125 I (NEN Life Science Products Inc.) using IODO-BEADS (Pierce), according to the manufacturer's protocol. The specific activities of labeled ligands were typically 150,000 -200,000 cpm/ng (Ϸ600 Ci/mmol). Quiescent cells in Corning 35-mm tissue culture dishes were equilibrated in binding buffer for 15 min, on a warm plate that maintains cells at 37°C, before challenge with 125 I-labeled ligand. Surface-bound and internalized ligand were discriminated essentially as described (11,33). Briefly, free ligand was removed by washing 6 times with ice-cold WHIPS buffer (20 mM HEPES, 130 mM NaCl, 5 mM KCl, 0.5 mM MgCl 2 , 1 mM CaCl 2 , 1 mg/ml polyvinylpyrrolidone, pH 7.4). Surface-bound ligand was then collected in ice-cold acid strip with urea (50 mM glycine-HCl, 100 mM NaCl, 1 mg/ml polyvinylpyrrolidone, 2 M urea, pH 3.0) for 5-8 min, and internalized ligand was released in 1 M NaOH overnight at room temperature. Nonspecific binding (Ͻ2%) was assessed in the presence of 2 M unlabeled human EGF (Peprotech) and subtracted from the total. Samples were quantified using a ␥-counter.
Removal of Surface-bound Ligand by Mild Acid Strip-At intermediate times during an experiment, surface-bound ligand was removed without compromising cell viability, using brief (1-2 min) treatments of ice-cold acid strip without urea (50 mM glycine-HCl, 100 mM NaCl, 1 mg/ml polyvinylpyrrolidone, pH 3.0) as indicated. By 1 min, this treatment is equally efficient in removing either EGF and TGF␣ (reproducibly 90 -93%) from the surface of NR6 cells.
EGFR-Phosphotyrosine Sandwich ELISA-High-binding ELISA plates (Corning) were coated at room temperature overnight with 10 g/ml anti-EGFR monoclonal antibody 225 in PBS, then incubated at room temperature for 4 -18 h in blocking buffer (10% horse serum, 0.05% Triton X-100 in PBS). After various treatments in binding buffer as indicated, cells were washed once in ice-cold PBS supplemented with 1 mM sodium orthovanadate and 4 mM sodium iodoacetate, scraped into ice-cold lysis buffer (50 mM HEPES pH 7.0, 150 mM NaCl, 1% Triton X-100, 10% glycerol) supplemented with 1 mM sodium orthovanadate, 10 mM sodium pyrophosphate, 1 mM EGTA, 4 mM sodium iodoacetate, and 10 g/ml each of aprotinin, leupeptin, chymostatin, and pepstatin, and transferred to an Eppendorf tube. After 20 min of incubation on ice, cellular debris was pelleted for 10 min at 16,000 ϫ g, and the superna-tant of each sample was transferred to a new tube and kept on ice for analysis. Total protein in each sample was assessed using a Micro BCA kit (Pierce) according to the manufacturer's protocol. Each lysate was diluted to various extents in blocking buffer supplemented with 1 mM sodium orthovanadate and incubated in anti-EGFR-coated wells for 1 h at 37°C. The wells were then rinsed four times with wash buffer (10 mM Tris, pH 8.3, 300 mM NaCl, 0.1% SDS, 0.05% Nonidet P-40) and incubated with 0.5 g/ml alkaline phosphatase-conjugated RC20 anti-phosphotyrosine antibody (Transduction Laboratories) in blocking buffer for 1 h at 37°C. After four additional washes, the wells were reacted with 1 mg/ml p-nitrophenyl phosphate (Sigma) in 10 mM diethanolamine, 0.5 mM MgCl 2 , pH 9.5. The reaction rate was monitored by measuring absorbance at 405 nm in a 15-min kinetic assay, using a Molecular Devices microplate reader. The relative amount of EGFR-phosphotyrosine was determined from a binding plot of reaction rate versus micrograms of total lysate protein for each sample. Nonspecific control lanes in which the maximum lysate load was incubated in wells without 225 antibody yielded similar activities to 225 wells incubated without lysate.
Immunoprecipitation and Western Blotting-Cells were lysed in 1% Triton X-100, and total cell protein was determined as detailed above. Immunoprecipitations of equivalent total protein amounts were performed at 4°C for 90 min using 3-5 g of primary antibody precoupled to 10 l of protein G-Sepharose beads per sample. The beads were washed five times with ice-cold lysis buffer supplemented with 1 mM sodium orthovanadate, and the residual liquid was removed with a syringe. The beads in each tube were boiled for 5 min in 30 l of sample buffer (62.5 mM Tris, pH 6.8, 2% SDS, 100 mM dithiothreitol, 10% glycerol, 0.005% bromphenol blue), then clarified by centrifugation. Proteins were separated by SDS-PAGE (34) on 7.5% acrylamide gels and transferred to nitrocellulose membranes (35). Membranes were blotted for proteins as indicated and visualized using horseradish peroxidase-conjugated secondary antibodies and SuperSignal Ultra detection reagent (Pierce). Bands were detected and quantified using a Bio-Rad chemiluminescence screen and Molecular Imager. When reprobing of a blot was desired, bound antibodies were first removed for 1 h at 55°C in stripping buffer (62.5 mM Tris, pH 6.8, 2% SDS, 100 mM ␤-mercaptoethanol).
Determination of Internal EGFR-Phosphotyrosine-Internalized EGFR were isolated by labeling surface-accessible proteins for subsequent removal from cell lysates (36). Briefly, cells were washed 3 times with ice-cold PBS, pH 8.0, after specific treatments, and surface proteins were biotinylated at 4°C with 5 mg of sulfo-NHS-LC-biotin (Pierce) per 10-cm plate. Plates were washed once with PBS, once with PBS, 50 mM glycine, and once again with PBS. Cells were lysed in 1% Triton X-100 as described above, and EGFR were immunoprecipitated using 225 antibody precoupled to protein G-Sepharose. Proteins were eluted by boiling for 10 min in TNE buffer (50 mM Tris, pH 7.5, 140 mM NaCl, 5 mM EDTA) with 0.5% SDS. After adding 1 volume of lysis buffer supplemented with 1 mM sodium orthovanadate, biotinylated (surface) EGFR were removed using immobilized streptavidin (Pierce). Supernatants were subjected to SDS-PAGE and anti-phosphotyrosine immunoblotting.
PIP 2 Hydrolysis Assay-In vivo PLC activity was determined essentially as described (10). Briefly, cells were incubated with 5 Ci/ml myo-[2-3 H]inositol (American Radiolabeled Chemicals) during the growth arrest protocol. Unincorporated radioactivity was removed by two washes with PBS at 37°C just before the experiment. Following various treatments in binding buffer as indicated, cells were washed once with ice-cold WHIPS buffer, scraped into boiling dH 2 O, transferred to an Eppendorf tube, and kept on ice. Samples were boiled for 5 min, and cellular debris was pelleted for 5 min at 16,000 ϫ g. The concentration of cytosolic radioactivity in disintegrations/min/ml for each supernatant was determined by liquid scintillation counting of small aliquots, and equivalent volumes of samples were loaded onto minicolumns packed with 0.5 ml of anion exchange resin (AG 1-X8, formate, 100 -200 mesh; Bio-Rad) each. After washing each column with 20 ml of dH 2 O and 20 ml of 5 mM sodium borate, 60 mM sodium formate, inositol phosphate fractions were eluted with 200 mM ammonium formate, 100 mM formic acid. The disintegrations/min of inositol phosphate that accumulated during cell treatment was normalized to the total disintegrations/min applied to the anion exchange column for each sample.

RESULTS
Differential Tyrosine Phosphorylation of Internalized EGFR by EGF versus TGF␣-Given the central role of EGFR autophosphorylation in initiating phospholipase C activity, we determined whether the tyrosine phosphorylation stoichiometry of EGFR (Tyr(P)/receptor) is altered upon internalization of EGF or TGF␣⅐EGFR complexes in NR6 WT cells. Based on the differential binding affinities of these ligands at endosomal pH, we expected that EGF would elicit a higher level of internal EGFR tyrosine phosphorylation than TGF␣. Saturating doses (20 nM) of radioiodinated EGF or TGF␣ were used to follow the levels of surface-bound and internalized ligand with time in NR6 WT cells (Fig. 1A). A decrease in surface complexes to a level of about 60% of the total was observed within 30 min, with a parallel increase in internalized ligand, in agreement with previously published results (14). The profiles of cell-associated EGF and TGF␣ in both compartments were indistinguishable in these experiments. This indicated that the initial trafficking of EGFR in these cells is similar following either EGF or TGF␣ treatment.
To distinguish between surface-associated and intracellular activated EGFR, we used a brief incubation with a mild acid wash. This treatment rapidly removes surface-bound ligand (both EGF and TGF␣ are dissociated equivalently). In addition, several studies have shown that it does not compromise cell viability (37)(38)(39). The kinetics of EGFR tyrosine phosphorylation were examined using two parallel stimulation protocols: a standard time course of stimulation with EGF or TGF␣ (20 nM) at 37°C, and a strip protocol (Fig. 1B). For the strip protocol, cells were stimulated with EGF or TGF␣ (20 nM) for 15 min at 37°C to allow internalization, treated with acid strip on ice for 1 min, and brought back to 37°C in the absence of ligand for 9 min. For EGF-treated cells, ligand was then added back at 37°C to determine whether receptor binding and signaling capacities were intact following the acid wash. As shown in Fig.  1B, EGF-treated NR6 WT cells displayed approximately 3 to 4 times higher EGFR-phosphotyrosine relative to TGF␣-treated cells following the surface strip, suggesting that the former ligand is more effective in maintaining activation of EGFR in internal compartments. Following readdition of EGF, Tyr(P)-EGFR returned to pre-strip levels, showing that the treatment does not compromise signaling in these cells.
To determine the stoichiometry of EGFR tyrosine phosphorylation, the levels of surface-bound and internalized 125 I-ligand were determined for the same time points and stimulation conditions shown in Fig. 1B. The ratio of Tyr(P)-EGFR/ total cell-associated ligand was then plotted versus the ratio of internalized ligand/total cell-associated ligand. If EGFR maintains a constant tyrosine phosphorylation stoichiometry, both with respect to time and cellular location, such a plot will have zero slope. This was indeed the case for EGF-treated NR6 WT cells ( Fig. 2A). EGFR maintained a nearly constant Tyr(P)-EGFR/cell-associated ligand before the strip, after the strip, and following readdition of EGF. This type of plot can also be used to determine whether EGFR is dephosphorylated following endocytosis, since the ratio of Tyr(P)/ligand would change from the surface to the internal value as the fraction of internalized ligand increased. TGF␣-treated cells displayed a decrease in Tyr(P)/ligand as the internal ligand fraction increased, and the extrapolated "surface" Tyr(P)/ligand value was very close to the mean phosphorylation stoichiometry observed for EGF (Fig. 2B). This suggests that in the case of cells treated with TGF␣, a significant fraction of internalized EGFR are dephosphorylated.
To examine the possibility that the receptor phosphorylation stoichiometries elicited by EGF and TGF␣ simply reflect differential activation of surface complexes, cell surface proteins were , with the relative level of Tyr(P)-EGFR thus defined as A 2 (r 2 typically Ͼ 0.99). biotinylated and cleared from EGFR immunoprecipitates. The remaining EGFR, presumably in intracellular compartments prior to cell lysis, were then subjected to anti-phosphotyrosine immunoblotting. As shown in Fig. 2C, EGF elicited significantly higher Tyr(P)-EGFR than TGF␣ in this assay, and phosphotyrosine levels were not altered by acid washing. This demonstrates that EGF induces a greater extent of internalized EGFR activation than TGF␣, although tyrosine phosphorylation of internal EGFR in TGF␣-treated cells is detectably higher than the unstimulated control. Taken together, our results indicate that tyrosine phosphorylation of internalized EGFR is strongly correlated with ligand occupancy in endosomes.
The dose responses (0 -20 nM) of EGF-and TGF␣-stimulated EGFR tyrosine phosphorylation were investigated as well, after 7.5 and 20 min of ligand challenge (Fig. 3). EGFR exhibited half-maximal tyrosine phosphorylation at 1-2 nM of either ligand, with TGF␣ values consistently and statistically lower than EGF values for the same dose (Fig. 3). This is also consistent with activation of surface EGFR to similar extents by the two ligands and a greater degree of internalized EGFR activation by EGF.
EGF and TGF␣ Are Equipotent in Stimulating PLC-mediated PIP 2 Hydrolysis-Having established that EGF yields higher levels of tyrosine phosphorylation of internalized EGFR than TGF␣, we next investigated whether these naturally occurring ligands could stimulate the PLC pathway to different extents. To this end, we employed a functional assay that assesses the hydrolysis of PIP 2 in intact cells. In vitro reactions using immunoisolated PLC-␥1 can be misleading, since the concentrations of PIP 2 and other membrane-associated signaling molecules in various compartments might be different. Following the liberation of soluble inositol triphosphate from PIP 2 , inositol phosphatases rapidly metabolize this intermediate to free inositol. Cell exposure to Li ϩ inhibits the breakdown of inositol phosphate (IP), potentiating its accumulation in the cytosol. Previous studies using NR6 WT and other NR6 transfectants in conjunction with the specific PLC inhibitor U73122 demonstrated that this assay is indeed a direct readout of PIP 2 hydrolysis (10).
PLC dose-response experiments were performed by incubating NR6 cells with 20 mM LiCl for 15 min, followed by stimulation in the continued presence of LiCl. Control experiments indicated that IP accumulation is roughly linear with time for at least 30 min of 20 nM EGF stimulation, that lithium is required for observable IP accumulation, that the basal level of IP in the absence of stimulation does not increase detectably with time, and that lithium treatment does not affect EGFR internalization (data not shown). The dose responses of EGFand TGF␣-stimulated PIP 2 hydrolysis were examined for stimulation times of 15 and 30 min (Fig. 4). These time scales allow for sufficient internalization of ligand to occur (Fig. 1A), and for stimulated IP accumulation to achieve adequate signal/noise ratios. As shown in Fig. 4, EGF did not gain any noticeable advantage over TGF␣ with respect to stimulation of the PLC pathway over the course of 30 min, despite higher levels of total cellular EGF-mediated EGFR activation at all doses (Fig. 3). This might be expected if the activation of PLC were saturable, i.e. if PLC-␥1 or PIP 2 were stoichiometrically limiting at submaximal Tyr(P)-EGFR (40). However, both ligand-induced PIP 2 hydrolysis and EGFR phosphotyrosine were half-maximal at similar EGF and TGF␣ concentrations (1-2 nM). Therefore, these results indirectly suggest that activated EGFR in internal compartments are deficient in stimulating PLC function. PIP 2 Hydrolysis Is Not Stimulated by Activated EGFR in the Endocytic Pathway-Although it seemed possible that active EGFR do not have access to PLC-␥1 and/or PIP 2 in intracellular trafficking compartments, our dose-response results did not address this point directly. The internal pool of EGFR does not constitute a large fraction of the total cellular EGFR in the NR6 cell line, obscuring its potential contribution to PLC activation. Thus, mild acid washing was again employed to test the relationship between cell surface and intracellular receptor pools with regard to signaling.
NR6 WT cells were pretreated for 15 min with 20 nM EGF or TGF␣ at 37°C, in the absence of lithium, to saturate and permit internalization of surface EGFR. This was followed by   FIG. 2. Analysis of EGFR tyrosine phosphorylation stoichiometry. Surface-bound and internalized 125 I-EGF were quantified for the same time points and conditions used in Fig. 1B (n Ն 3), and the ratio of Tyr(P)-EGFR/total cell-associated ligand is plotted versus the ratio of internal/total cell-associated ligand (mean Ϯ S.D. for both x and y axis values; y value S.D. determined by propagation of error). A, analysis of EGFR tyrosine phosphorylation stoichiometry in response to EGF: q, time course; f, strip protocol; solid line, mean of Tyr(P)-EGFR/ligand. B, analysis of EGFR tyrosine phosphorylation stoichiometry in response to TGF␣: E, time course; Ⅺ, strip protocol; solid line, mean of EGF-stimulated Tyr(P)-EGFR/ligand from A; dashed line, theoretical line describing complete dephosphorylation of EGFR upon internalization. C, tyrosine phosphorylation of internalized EGFR. Cells were stimulated with 20 nM TGF␣ (T) or EGF (E) for 15 min at 37°C. Where indicated, acid washing was carried out for 2 min on ice, followed by 5 min equilibration in binding buffer at 37°C. Surface biotinylation and clearance with immobilized streptavidin was employed to isolate internalized EGFR, as described under "Experimental Procedures," which was then subjected to SDS-PAGE and anti-phosphotyrosine immunoblotting.
incubation with ice-cold acid wash for 2 min to remove surfacebound ligand. Cells were then returned to 37°C in the presence of 20 mM LiCl and various concentrations of TGF␣ (0 -20 nM), regardless of whether the cells were pretreated with EGF or TGF␣. This "surface titration" protocol allowed us to vary the level of surface-activated EGFR, relative to a constant level of internal-activated EGFR that depends on whether the cells were pretreated with EGF or TGF␣. This differs from the standard dose-response experiment, in which internal receptor activation is coupled to the level of surface receptor activation. PIP 2 hydrolysis was assayed 15 min following LiCl addition, and Tyr(P)-EGFR was assayed in a separate experiment 7.5 min following LiCl addition as an intermediate time point. Control experiments demonstrated that accumulation of inositol phosphate was evident within 5 min of lithium addition (data not shown). The control protocol was 15 min pretreatment with no ligand at 37°C, acid wash treatment, then a return to 37°C with 20 mM LiCl and no ligand.
EGF-pretreated cells yielded statistically higher levels of total EGFR phosphotyrosine than TGF␣-pretreated cells for each chase TGF␣ concentration (Fig. 5A), consistent with the continued EGF-stimulated phosphorylation of internalized EGFR. Despite this difference, PIP 2 hydrolysis activities were equivalent when EGF-and TGF␣-pretreated cells were stimulated with ligand in the chase (0.5-20 nM TGF␣; Fig. 5B). For the 20 nM chase concentration, using EGF instead of TGF␣ in the chase did not effect the level of PIP 2 hydrolysis observed in this assay (data not shown). Furthermore, in the absence of ligand in the chase, TGF␣-pretreated cells stimulated minimal PIP 2 hydrolysis, even though tyrosine phosphorylation of some internalized EGFR was detected under these conditions (see also Figs. 1C and 2C). These results are consistent with stim-ulation of PIP 2 hydrolysis by surface-localized EGFR only.
In apparent disagreement with this conclusion, however, EGF-pretreated cells exhibited statistically higher PIP 2 hydrolysis in the absence of ligand in the chase (Fig. 5B), in apparent disagreement with this conclusion. One possible explanation for this disparate result is that some internalized EGF, but not TGF␣, is recycled back to the surface complexed with EGFR (30), an effect that would be masked by exogenous ligand added to the medium. To determine whether recycled EGF/EGFR could account for the enhanced PIP 2 hydrolysis, a receptor-blocking antibody (10 g/ml 225 anti-EGFR) was included in the medium after acid wash treatment (Fig. 5C). This antibody causes accelerated dissociation of surface-bound EGF (data not shown) and therefore should reduce the level of surface signaling from recycled EGFR. Indeed, the presence of the antagonistic anti-EGFR antibody in the chase was able to inhibit PIP 2 hydrolysis in EGF-pretreated cells by 60%, but it had no effect on PIP 2 hydrolysis in TGF␣-pretreated cells (Fig.  5C). Taken together, these results indicate that active EGFR in internal compartments do not participate in PLC signaling.
To examine this hypothesis further, stimulated PIP 2 hydrolysis was plotted versus EGFR phosphotyrosine for both standard dose-response and surface titration experiments. Since the extent of EGFR tyrosine phosphorylation constitutes a readout and modulator of EGFR kinase activity, in addition to the defined role of EGFR phosphotyrosine in docking PLC-␥1 and other signaling proteins, it represents the input for cell signaling at the receptor level. If signaling downstream of EGFR autophosphorylation is not affected by internalization of the receptor, then the relationship between receptor phosphorylation and PIP 2 hydrolysis would be identical for both EGF and TGF␣. As shown in Fig. 6, this is clearly not the case for the PLC pathway in NR6 cells. For the surface titration experiments, the curves for TGF␣-and EGF-pretreated cells are parallel and shifted to the right by the constant level of internal Tyr(P)-EGFR, indicating that these receptors are not contributing to PIP 2 hydrolysis. In the case of the standard ligand treatment protocol, the curves of PIP 2 hydrolysis versus Tyr(P)-EGFR for both TGF␣ and EGF overlap until intracellular EGF concentrations become high enough to start occupying receptors. At this point the curves diverge, with EGF values shifted to the right of TGF␣ values since EGF is more effective than TGF␣ in stimulating internal Tyr(P)-EGFR (Fig. 6). Finally, for high ligand concentrations (and therefore high Tyr(P)-EGFR), the standard dose-response and surface titration curves for the same ligand converge, confirming that the nature of signaling is not affected by differences in the two experimental designs. These results are entirely consistent with the hypothesis that internalized EGFR, even when biochemically active, are far less effective than surface receptors in stimulating PLC-mediated PIP 2 hydrolysis.
Compartmentalization of PLC Activity Is Not Due to Differences in PLC-␥1 Tyrosine Phosphorylation-Our dose-response and surface titration experiments indicated that PLC activity is inhibited following EGFR endocytosis, and that loss of signaling occurs between EGFR activation and PIP 2 hydrolysis. To determine if this was due to an inability of the EGFR to induce tyrosine phosphorylation of PLC-␥1, we used the same surface titration protocol used above (EGF or TGF␣ pretreatment, surface strip, and TGF␣ chase). The same conditions used to measure EGFR phosphorylation in Fig. 5A were used to examine PLC-␥1 phosphorylation.
After electrophoresis and membrane transfer, the blots were probed for the presence of PLC-␥1 or pY. Shown in Fig. 7A is a typical experiment in which PLC-␥1 was visualized following immunoprecipitation with anti-Tyr(P) antibodies. Qualitatively, PLC-␥1 tyrosine phosphorylation mirrored EGFR tyrosine phosphorylation in NR6 WT cells, in that EGF-pretreated cells always exhibited higher Tyr(P)-PLC-␥1 than TGF␣-pre- treated cells for the same chase stimulation. Essentially identical results were obtained when PLC-␥1 was immunoprecipitated and then visualized with anti-Tyr(P) antibodies.
We quantified the results of these experiments using a Bio-Rad Molecular Imager. Control experiments verified that there was a linear relationship between micrograms of total protein from the same lysate subjected to immunoprecipitation and the detected band intensity (data not shown). To ascertain quantitatively whether tyrosine phosphorylation of PLC-␥1 is affected by EGFR endocytosis, Tyr(P)-PLC-␥1 (averaged over three experiments) was analyzed as a function of Tyr(P)-EGFR for each condition (Fig. 7B). The relationship between tyrosine phosphorylation of the EGFR and PLC-␥1 was the same in the case of both EGF-and TGF␣-pretreated cells, showing that tyrosine phosphorylation of PLC-␥1 is not affected by the localization of active receptors. As EGFR-mediated PIP 2 hydrolysis is dependent on a surface localization, this suggests a signaling restriction downstream of PLC-␥1 phosphorylation.

DISCUSSION
While receptor down-regulation and ligand depletion via the endocytic pathway are known to negatively modulate signal transduction mediated by EGFR and other receptor tyrosine kinases (14,15), there is no a priori reason to suspect that internalized receptors in sorting endosomes cannot participate in signaling. Because its kinase and substrate binding activities continue to face the cytosol in early endosomes, internalized EGFR have the potential to signal as long as they remain ligated (20). Indeed, internalized EGFR in rat liver are competent in both binding and phosphorylating the adaptor protein Shc, which helps localize the exchange factor Sos for potential interactions with the Ras GTPase (21). In this cellular context, surface complexes are rapidly desensitized, while internal complexes somehow escape this regulation, implying that compartmentalized feedback mechanisms exist (16,18). Another possibility is that endocytosis might affect signaling specificity. Cells overexpressing a dominant-negative mutant of the dynamin GTPase are defective in both inducible endocytosis of EGFR and EGF-responsive tyrosine phosphorylation of the cytoplasmic signaling protein phosphatidylinositol 3-kinase.
This suggested that receptor internalization is required for full manifestation of some signals but not others (15). However, it has been reported recently that the mutant cells are also defective in high affinity EGFR/ligand binding, suggesting that dynamin might modulate aspects of signal transduction at the surface (41). The concept of compartmentalized "separation" of signaling pathways has also been implicated in the activation of sphingomyelinases mediated by tumor necrosis factor (23,42). Thus, it is possible that signaling from the endosomal compartment versus the surface plays an important role in dictating the outcome of receptor tyrosine kinase stimulation.
It is now appreciated that many intracellular reactions, especially those at or just beyond the receptor level, are regulated by subcellular localization (43,44). Interestingly, post-receptor targets such as PIP 2 and Ras are membrane-associated, implying that they are readily compartmentalized. Since previous studies on EGFR signaling in internal compartments have not probed beyond tyrosine phosphorylation of cytoplasmic proteins, we examined activation of the PLC pathway by the EGFR to the level of PIP 2 hydrolysis. We used a physiologically relevant ligand-based approach, rather than comparing results from variant cell lines, based on previous studies suggesting that TGF␣ dissociates from EGFR in endosomes to a much greater extent than EGF (24,25). Our study is the first to show that the magnitude of signal transduction through a specific pathway, at the level of proximal target modification, is affected by EGFR internalization.
We found that internalized EGF⅐EGFR complexes retain a maximal tyrosine phosphorylation stoichiometry, whereas EGFR internalized in response to TGF␣ binding are dephosphorylated to a significant extent. However, the two ligands are equipotent in stimulating PIP 2 hydrolysis, the functional outcome of PLC-␥1 activation. By manipulating the relative levels of surface and internal receptor activation independently, we showed that active EGFR in internal compartments stimulate little if any hydrolysis of PIP 2 . This deficiency was not due to a location-specific difference in PLC-␥1 tyrosine phosphorylation, since this event correlates with receptor phosphorylation in either compartment. Therefore, we concluded that the spa-FIG. 7. Analysis of PLC-␥1 tyrosine phosphorylation. NR6 WT extracts were prepared as for Fig. 5A, using the surface titration protocol with TGF␣ (T) or EGF (E) pretreatments. Tyrosine-phosphorylated PLC-␥1 was immunoprecipitated from equal levels of total cellular protein using PY20 anti-phosphotyrosine antibody (Transduction Laboratories) and detected by immunoblotting with an anti-PLC-␥1 mixed monoclonal (Upstate Biotechnology) (n ϭ 2; representative data shown, A). As a check, Tyr(P)-PLC-␥1 was also detected once by anti-PLC-␥1 immunoprecipitation/anti-phosphotyrosine immunoblotting with similar results; this blot was reprobed with anti-PLC-␥1 to confirm roughly equal total levels of PLC-␥1 (not shown). B, plot of Tyr(P)-PLC-␥1 versus Tyr(P)-EGFR. For each of the three Tyr(P)-PLC-␥1 experiments, the data was expressed relative to the maximum band intensity, and the mean for each condition is plotted versus Tyr(P)-EGFR from Fig. 5A to compare TGF␣-(E) and EGF (q)-pretreated cells. The unstimulated point (᭛) is also shown. The dotted line is the least squares linear fit of all the data points (R 2 Ͼ 0.99). tial requirements for PIP 2 hydrolysis are defined at a step subsequent to PLC phosphorylation by EGFR.
The simplest interpretation of our data is that PLC-␥1 associated with EGFR in pre-degradative trafficking organelles (endocytic vesicles, early endosomes, and recycling endosomes) do not have access to PIP 2 . Studies in numerous cell types have indicated that active maintenance of PIP 2 levels is required for meaningful PLC signaling. This function is carried out by phosphatidylinositol transfer protein, which directs transport of phosphatidylinositol between cellular membranes (45). Therefore, exchange among lipid pools by membrane-phase sorting must be much slower than enzymatic turnover by PLC and other enzymes (46), indicating that the concentrations of PIP 2 and other lipids present in low amounts are not likely to be homogeneous among distinct cellular membranes. Interestingly, phosphatidylinositol transfer protein forms a signaling complex with EGFR, phosphatidylinositol 4-kinase, and PLC-␥1 in response to EGF, suggesting that PIP 2 supply and hydrolysis are coupled. When phosphatidylinositol transfer protein and PLC-␥1 are depleted from the cytosol of A-431 cells, both proteins must be added exogenously to reconstitute EGFR-mediated PLC activity (47). The requirement of cofactors for maintenance of PLC function could restrict PIP 2 hydrolysis to the plasma membrane, especially if one or more of these proteins are preferentially recruited by surface EGFR. Furthermore, PIP 2 concentrated in caveolae microdomains may comprise the EGFR-responsive substrate pool (48), which is probably segregrated from the bulk membrane delivered to endosomes via clathrin-coated pits.
Our results also have implications regarding the functional difference between EGF and TGF␣ as naturally occurring agonists for the same receptor. Physiologically, EGF and TGF␣ may have evolved as ligands that vary in their ability to mediate long-term receptor/ligand processing (24,30,49) but not in their compartmentalization of signal transduction. Although EGFR-transfected NR6 fibroblasts do not maintain a high percentage of internalized receptors at steady state, other cells display a rapid redistribution of receptors to internal pools (50). In these cases, internalization could act as a potent shut-off mechanism in response to chronic stimulation of EGFR. While the above situation likely applies to regulation of PLC-␥1 signaling, it is unclear whether other EGFR-mediated pathways are activated or even augmented in endosomes as has been suggested. If internalized receptors can signal through other intermediates, an important question is whether endocytosis is a requirement for signaling or if internal receptors simply continue activities initiated at the cell surface. In either case, compartmentalization of membrane-associated molecules would provide an additional level of signaling control, by affecting the spatiotemporal selectivity of enzymes that coordinate different cell functions.