G Protein-coupled Receptors Desensitize and Down-regulate Epidermal Growth Factor Receptors in Renal Mesangial Cells*

Different types of plasma membrane receptors engage in various forms of cross-talk. We used cultures of rat renal mesangial cells to study the regulation of EGF receptors (EGFRs) by various endogenous G protein-cou-pled receptors (GPCRs). GPCRs (5-hydroxytryptamine 2A , lysophosphatidic acid, angiotensin AT 1 , bradykinin B 2 ) were shown to transactivate EGFRs through a protein kinase C-dependent pathway. This transactivation resulted in the initiation of multiple cellular signals (phos-phorylation of the EGFRs and ERK and activation of cAMP-responsive element-binding protein (CREB), NF- k B, and E2F), as well as subsequent rapid down-regu-lation of cell-surface EGFRs and internalization and desensitization of the EGFRs without change in the total cellular complement of EGFRs. Internalization of the EGFRs and the down-regulation of cell-surface receptors in mesangial cells were blocked by pharmacological inhibitors of clathrin-mediated endocytosis and in HEK293 cells by transfection of cDNA constructs that encode dominant negative b -arrestin-1 or dynamin. Whereas carried Cells serum-starved for h 6-well culture or dishes treated various conditions that endocytosis, potassium depletion hypertonic medium concanavalin A and monodansylcadaverine Cells were incu- bated with cycloheximide (5 m g/ml) for 30 min prior to treatment with mitogens to prevent the confounding effects of protein synthesis. Pro- cedures to inhibit endocytosis were also performed for 30 min prior to experimentation coincident with cycloheximide. The buffers used were as follow: potassium depletion , , and mg/liter dextrose); hypertonic medium (RPMI 1640, 0.5% bovine serum albumin, 4.5 g/liter dextrose, and m M sucrose); and medium for chemical inhibition by con- canavalin A (250 m g/ml) or 500 m M monodansylcadaverine (RPMI 1640, 0.5% bovine serum albumin, 4.5 g/liter dextrose). Cells were then shifted to 4 °C, and cell-surface EGFRs were measured by radioligand binding. M 5-HT in the presence and absence of various inhibitors of clathrin-mediated endocytosis, including potassium deple- tion buffer, ConA, and monodansylcadaverine ( ). Then cell-surface I-EGF binding was measured as described under “Experimental Procedures.” shows that these maneuvers attenuated the desensitization of the EGFR induced by 5-HT. shows that incu- with 30 did not impair the down-regulation of cell-surface functions through a PKC-dependent pathway.

receptors (GPCRs) are the two major families of receptors that convert extracellular signals into cellular physiological and mitogenic responses. Previously, the signals generated by RTKs and GPCRs were thought to be neatly compartmentalized, with very little cross-talk between or sharing of the signaling pathways. There is a new awareness that RTKs, such as the EGF receptor, and GPCRs possess the capacity for crosstalk during signal initiation and propagation. Cross-talk can take the form of using shared signaling pathways (1)(2)(3) or, for GPCRs, using RTKs themselves as signaling platforms (4 -12). Thus, contrary to relatively recent dogma, it is now abundantly clear that RTKs and GPCRs engage in extensive cross-talk with each other.
Just as there are similarities in the mechanisms that initiate the signaling pathways of GPCRs and RTKs, there might also be similarities in the mechanisms by which those signals are terminated or desensitized. Indeed, there is a growing body of evidence that GPCRs and RTKs share mechanisms that regulate signal desensitization. Desensitization is a group of processes through which receptors or components of their signaling pathways become less responsive after previous exposures to receptor ligands. Homologous desensitization occurs when cells become unresponsive only to subsequent activation of the receptor that was previously stimulated. This type of desensitization is usually mediated by receptor specific kinases (GRKs). Heterologous desensitization refers to attenuation of one receptor system by another and is usually mediated by broad spectrum serine/threonine kinase such as protein kinases C and A. A special form of heterologous desensitization may occur when RTKs desensitize GPCRs. RTKs can desensitize GPCRs by phosphorylating the GPCR (13), by phosphorylating heterotrimeric G proteins (14), or by other mechanisms (15,16). It is also possible that GPCRs could desensitize RTKs, but little is known about this phenomenon.
Renal mesangial cells possess many mitogenic GPCRs, including angiotensin II AT 1A (17), bradykinin B 2 (18,19), lysophosphatidic acid (20,21), and 5-hydroxytryptamine (5-HT 2A ) receptors (22). Mesangial cells also express RTKs, which may participate in the proliferative phase of chronic renal failure (23) or in the recovery from renal failure (24). Mesangial cells possess an epidermal growth factor (EGF) receptor (25) that stimulates proliferative cascades in those cells (26). It is somewhat paradoxical that mesangial cells should express so many mitogenic receptors in that, under normal circumstances, proliferation is highly restrained within the confines of the glomerulus. This suggests that the responsiveness of mitogenic receptors must be rigidly controlled in mesangial cells. One mechanism through which rigid control of mitogenic signaling in mesangial cells might be exercised is desensitization.
In this study, we report that pretreatment of kidney mesangial cells with GPCR ligands (5-HT, bradykinin, lysophosphatidic acid) results in a PKC-dependent transactivation of EGFR followed by a profound decrease in the ability of EGF to initiate multiple signals including autophosphorylation of the EGF receptor (EGFR), phosphorylation of ERK, and regulation of transcription factor activities (NF-B, E2F, CREB). Furthermore, the desensitization pathway involves PKC and results in a dramatic internalization of native EGF receptors and transfected EGFR-GFP fusion proteins. Thus, preconditioning of cells by GPCR ligands may be a novel method to abrogate deleterious signals initiated by EGFR and other RTK.
Cell Culture and Transfection-Rat renal mesangial cells were obtained from cortical sections of kidneys from young 100 -150-gram Harlan Sprague-Dawley rats using standard sieving techniques (27). The kidneys were harvested in accordance with a protocol reviewed and approved by the Institutional Animal Care and Use Committee of the Medical University of South Carolina. Cells were incubated at 37°C in a humidified atmosphere of 95% air and 5% CO 2 and were subcultured every 1-2 weeks by trypsinization until a pure culture of mesangial cells was obtained. These cells were plated at a density of 2-5 ϫ 10 4 cells/ml in RPMI medium supplemented with 20% heat-inactivated fetal bovine serum and antibiotics (100 units/ml penicillin and 100 g/ml streptomycin). Cells used were from passages 6 -16. HEK293 cells were maintained in F12 medium (Life Technologies, Inc.) supplemented with 10% fetal bovine serum and 50 g/ml gentamicin (Life Technologies, Inc.) at 37°C in a humidified 5% CO 2 atmosphere.
Transfections were performed on 50 -70% confluent monolayers in 100-mm dishes, using LipofectAMINE, Lipofectin (Life Technologies), or FuGene™ 6 (Roche Molecular Biochemicals). Empty vectors were added to transfections to keep the total mass of DNA added per dish constant within experiments. 48 h prior to studies, cells were placed in serum-free medium supplemented with antibiotics and 0.1% bovine serum albumin.
Metabolic labeling of EGFR-Cells (ϳ1 ϫ 10 7 ) were grown in 100-mm culture dishes, washed twice with phosphate-free buffer (10 mM HEPES, pH 7.4, 137 mM NaCl, 3 mM KCl), and incubated in phosphate-free RPMI medium supplemented with 20 mM dextrose, 20 mM HEPES, pH 7.4, and 100 Ci of [ 32 P]phosphoric acid for 4 h at 37°C. Cells were then treated with mitogens for 3 min with or without pretreatment with inhibitors 30 min before stimulation. Cells were placed on ice, washed three times with ice-cold phosphate-buffered saline, and lysed in a modified radioimmune precipitation buffer (50 mM Tris-HCl, pH 7.4, 1% Nonidet P-40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM EDTA, 1 mM Na 3 VO 4 , 1 mM NaF, 500 M 4-(2-aminoethyl)benzenesulfonyl fluoride, 500 M EDTA, 1 M E-64, 1 M leupeptin, and 1 g/ml aprotinin). The cell lysates were rocked at 4°C for 1 h and then centrifuged to remove insoluble debris. The lysates were then diluted to a protein concentration of 1-2 mg/ml, and 500 l was used for immunoprecipitation using 5 g of anti-EGFR monoclonal antibody. The mixture was rocked for 1 h at 4°C, then 20 l of protein A-or protein G/A-agarose beads were added, and the mixture was incubated on a rocker for another 20 min at 4°C. The immune complexes were isolated by centrifugation, washed three times with radioimmune precipitation buffer, and then dissociated from the agarose beads by adding Laemmli buffer. The samples were heated to 90°C for 2 min and then loaded onto precast 4 -20% polyacrylamide gels (Novex, San Diego) and resolved under nonreducing conditions. The gels were dried and analyzed with a PhosphorImager.
Immunoblots-ERK immunoblots were performed essentially as described previously (28). The phospho-ERK antibody was used at 1:1000 dilution, whereas the control antibody, which recognizes equally well the phosphorylated and nonphosphorylated mitogen-activated protein kinase, was used at a 1:500 dilution as per the manufacturers recom-mendations. After treatment, cells were scraped into Laemmli buffer, boiled for 3 min, and subjected to SDS-polyacrylamide gel electrophoresis under reducing conditions with 4 -20% pre-cast gels (Novex). After semi-dry transfer to polyvinylidene difluoride membranes, the membranes were blocked with a BLOTTO buffer (5% defatted dried milk in 10 mM Tris, 150 mM NaCl, 1% Tween 20, pH 8.0). The membranes were incubated overnight with the BLOTTO containing the phospho-ERK antibody. The membranes were washed, then exposed to goat antirabbit alkaline phosphatase-conjugated IgG (1:1000) in BLOTTO for 1 h, and then washed again. Immunoreactive bands were visualized by a chemiluminescent method (CDP Star™, New England Biolabs) using pre-flashed Kodak X-AR film. For other immunoblots, cell extracts were incubated with 5 g/ml anti-EGFR or anti-phosphotyrosine monoclonal antibodies and visualized as described above, except the secondary antibody was a rabbit anti-mouse IgG alkaline phosphatase conjugate.
EGFR-GFP Plasmid Construction-A bright green mutant of GFP, enhanced GFP (CLONTECH, Palo Alto, CA), was attached to the carboxyl terminus of human EGFR as previously described (29). This construct behaves like wild-type EGFR in assays of phosphorylation, protein-protein interactions, signal transduction, internalization, and degradation (29).
Down-regulation of EGFRs by GPCRs-Cells grown in 6-well plates were incubated with vehicle, EGF, or 5-HT prior to incubation with various concentrations of EGF (1-100 ng/ml) for various times at 37°C. Monolayers were then washed twice with ice-cold Hanks' balanced salt solution. Cells were then washed with cold acid wash buffer (50 mM glycine, 100 mM NaCl, pH 3.0) to dissociate bound EGF followed by three cold Hanks' balanced salt solution washings. Cells were then incubated with 50 pM 125 I-EGF for 90 min at 4°C in HEPES binding medium (RPMI 1640 with 40 mM HEPES, pH 7.4, 0.1% bovine serum albumin) in the continuing presence of vehicle or 5-HT. Cells were then washed three times with Hanks' balanced salt solution and dissolved in 1 ml of 1 M NaOH. The solubilized material was collected in scintillation vials and counted in a ␥-counter. Nonspecific binding was determined in quadruplicate wells containing 100 ng/ml unlabeled EGF and was subtracted from total binding to yield specific 125 I-EGF binding at each time point. Data were analyzed using Prism 2.0 software (GraphPad Software, San Diego, CA).
Electrophoretic Mobility Shift Assay-Oligonucleotides (E2F1, CREB, or NF-B transcription factor consensus binding sites) were end-labeled using T4 polynucleotide kinase and [␥ 32 P]CTP. Nuclear extracts were prepared exactly as described (30), and the electrophoretic mobility shift assay was modified from a previously published protocol (31). The reaction mixture comprised 10 g of nuclear extracts, 1-2 g of poly(dI-dC), 5 l of 5ϫ binding buffer (50 mM HEPES, pH 7.8, 5 mM spermidine, 15 mM MgCl 2 , 36% glycerol, 3 mg/ml bovine serum albumin, and 15 mM dithiothreitol). This mixture was incubated on ice for 15 min, and then 40,000 -70,000 cpm of 32 P-labeled oligonucleotide were added. The reaction mixture was incubated further for 15 min at room temperature. DNA-protein complexes were then resolved on 5% native polyacrylamide gels and quantified with a PhosphorImager after drying. For the CREB assays, cells were pretreated with Ro20 -1724 (4-(3-butoxy-4-methoxybenzyl)-2-imidazolidinone), a selective cAMPspecific phosphodiesterase inhibitor. For competition assays, nuclear extracts were preincubated with unlabeled oligonucleotides at room temperature for 15 min before adding 32 P to the labeled oligonucleotide. When possible, reactions were also carried out by using a 32 P-labeled oligonucleotide carrying mutations in the consensus regions to check the specificity of the binding reaction.
Inhibition of Endocytosis-Assays for the inhibition of endocytosis were carried out as described previously by Jockers et al. (32). Cells that were serum-starved for 24 h in 6-well culture plates or 100-mm dishes were treated with various conditions that inhibit endocytosis, including potassium depletion (33), hypertonic medium (34, 35), concanavalin A (36), and monodansylcadaverine (37,38). Cells were incubated with cycloheximide (5 g/ml) for 30 min prior to treatment with mitogens to prevent the confounding effects of protein synthesis. Procedures to inhibit endocytosis were also performed for 30 min prior to experimentation coincident with cycloheximide. The buffers used were as follow: potassium depletion buffer (20 mM HEPES, pH 7.4, 140 mM NaCl, 1 mM CaCl 2 , 1 mM MgCl 2 , and 4.5 mg/liter dextrose); hypertonic medium (RPMI 1640, 0.5% bovine serum albumin, 4.5 g/liter dextrose, and 500 mM sucrose); and medium for chemical inhibition by concanavalin A (250 g/ml) or 500 M monodansylcadaverine (RPMI 1640, 0.5% bovine serum albumin, 4.5 g/liter dextrose). Cells were then shifted to 4°C, and cell-surface EGFRs were measured by radioligand binding.
Confocal Laser Scanning Microscopy-HEK293 or mesangial cells were grown on round coverslips by placing coverslips at the bottom of the wells in 6-or 12-well culture plates. After rinsing in PBS, adherent cells transfected with the EGFR-GFP fusion protein were identified by incubating with rhodamine-concanavalin A (10 g/ml in PBS) for 2 min at 4°C. Cells were rinsed with PBS several times and then fixed with 4% paraformaldehyde in PBS for 15 min followed by quenching the fixative with three 5-min washes with 50 mM NH 4 Cl at room temperature. For single labeling of EGFR, cells on coverslips were fixed as described above and were then inverted onto 20 l of fluorescein isothiocyanate-conjugated anti-EGFR antibody raised against an extracellular epitope of the receptor (1:100) dilution in PBS with 1% goat serum and incubated in the dark for 2 h at room temperature. Cells were rinsed four times with PBS supplemented with 1% goat serum for 10 min. Coverslips were then mounted on a slide with Slow-Fade medium (Molecular Probes, Eugene, OR) and sealed with Cytoseal (Electron Microscopy Sciences, Fort Washington, PA) solution before scanning under a confocal microscope (Olympus Merlin™ IX70, Melville, NY). Analysis of Total cellular Complement of EGFRs-Cells in 100 mmculture dishes were treated with 30 g/ml cycloheximide or puromycin dihydrochloride for 1 h before treatment with 5-HT for different time periods, after which cells were washed and scraped into a modified radioimmune precipitation buffer as described above. Total EGFR protein was visualized by immunoprecipitation and immunoblotting as described above using anti-EGFR polyclonal antibody for immunoprecipitation and an anti-EGFR monoclonal antibody for immunoblotting. Fig. 1 shows that when rat renal mesangial cells were treated with 1 M 5-HT for 3 min, EGFRs became phosphorylated as detected by metabolic labeling and immunoprecipitation of EGFRs. The increase was dependent upon both the concentration of 5-HT (EC 50 ϭ 160 nM) and time of incubation, peaking at 3-10 min. EGFR phosphorylation was blocked almost completely when cells were pretreated with the specific EGFR tyrosine kinase inhibitor AG1478 for 30 min prior to exposure to 5-HT. Thus, the 5-HT 2A receptor transactivates the EGFR through the intrinsic kinase activity of the EGFR in a manner already shown to occur with other GPCRs such as those for angiotensin II (6), carbachol, lysophosphatidic acid, and thrombin (12,39). Because both ERK and PKC can induce phosphorylation of the EGFR (40), and because the 5-HT 2A receptor has been shown to activate both ERK and PKC (27,41), we used inhibitors of ERK kinase (MEK1) and PKC to determine which of those intermediates might be involved in 5-HT-induced phosphorylation of the EGFR. Fig. 1c shows that a PKC inhibitor (5 M GF109203X) greatly attenuated 5-HT-induced phosphorylation of the EGFR, whereas a MEK inhibitor (100 M PD98059) did not. This concentration of PD98059 nearly completely attenuates ERK activation by the 5-HT 2A receptor in these cells as previously determined by us (27). Thus, PKC (and not MEK/ ERK) seems to be involved in the transphosphorylation of the EGFR by the 5-HT 2A receptor in mesangial cells.

Transactivation of EGFRs by the 5-HT 2A Receptor-
The 5-HT 2A Receptor Stimulates Transcription Factors by Transactivation of EGFRs-Next, we examined the effects of 5-HT on the activation of three EGFR-stimulated transcription factors (E2F, CREB, and NF-B). Fig. 2 shows that acute treatment with either 5-HT or EGF induced activation of all three transcription factors as assessed by electrophoretic mobility shift assay. Moreover, the stimulation of all three transcription factors by 5-HT could be attenuated by preincubation with AG1478. Similarly, AG1478 blocked 5-HT-induced phosphorylation of ERK in mesangial cells (not shown). Those results suggest that the 5-HT 2A receptor in mesangial cells activates transcription factors through the intermediary actions of the EGFR.
Pretreatment of Mesangial Cells with 5-HT Attenuates Multiple Subsequent EGFR Downstream Signals-To further explore the similarities between the effects of 5-HT and EGF on EGFR function, we next studied the effects of prior treatment with 5-HT on the activation of downstream signals by EGF. Our rationale for those studies is that EGF treatment has been shown to desensitize the EGFR to subsequent activation by EGF. Thus, we hypothesized that 5-HT pretreatment might also desensitize the EGFR. Fig. 3 shows the results of studies in which EGF-induced ERK phosphorylation was assessed after pretreatment with vehicle or 5-HT. Those results clearly demonstrate that pretreatment of mesangial cells with 5-HT results in a marked attenuation of the ability of EGF to induce phosphorylation of ERK.
We used a similar paradigm to assess the effects of prior treatment with 5-HT on the ability of EGF to activate the three transcription factors shown in Fig. 2. Those results are shown in Fig. 4, a-c. Pretreatment with 5-HT greatly reduced tran- The inset shows representative autoradiographs with (ϩ) or without (Ϫ) AG1478. c, cells were preincubated with inhibitors (100 M GF109203X or 5 M PD98059) for 30 min prior to treatment with 1 M 5-HT for 3 min. The inset shows representative autoradiographs with and without inhibitors. Experiments were repeated at least three times. Error bars represent the means Ϯ standard errors. *, indicates p Ͻ 0.05 versus control; †, indicates p Ͻ 0.01 versus 5-HT without blocker as assessed using ANOVA and Fisher's protected least significant difference posthoc test for multiple comparisons. Similar results were found after immunoprecipitation of EGFRs followed by phosphotyrosine immunoblotting (not shown). scription factor activation as reflected by their binding with respective labeled consensus cis-elements. The specificity of the interactions of the transcription factors with their consensus oligonucleotides was confirmed by competition with unlabeled oligonucleotides and mutant (nonbinding) oligonucleotides. Figs. 3 and 4 show that multiple signals residing downstream from the EGFR can be attenuated by pretreatment with 5-HT, which suggested to us that desensitization of the EGFR most likely occurs at the level of the receptor itself. Therefore, we tested the effects of pretreatment with 5-HT on the ability of EGF to induce autophosphorylation of the EGFR. Fig. 5 shows that pretreatment of mesangial cells with 5-HT leads to a marked decrease in the ability of multiple concentrations of EGF to induce the phosphorylation of its receptor. The attenuation was consistent over a broad range of concentrations of EGF, suggesting that that this effect may be relevant under physiological conditions.

Pretreatment of Mesangial Cells with GPCR Ligands Attenuates EGFR Autophosphorylation-
If the effect of pretreatment of cells with 5-HT is truly important, we would expect that other GPCRs might also desensitize the EGFR. Indeed, attenuation of EGF-induced phosphorylation of the EGFR was observed when cells were pretreated with other mitogenic GPCR ligands such as bradykinin and lysophosphatidic acid (Fig. 6, a and b) as well as angiotensin (not shown). Thus, the ability of GPCR to desensitize EGFinduced phosphorylation of EGFR is not limited to the 5-HT 2A receptor. One of the major pathways that links G i and G qcoupled receptors to mitogenic signals in mesangial and other cells involves PKC (3,12,27,39). We therefore examined the effects of direct stimulation of PKC on the ability of EGF to induce phosphorylation of the EGFR. Fig. 6c shows that when cells were pretreated with 1 M phorbol 12-myristate 13-acetate (PMA), the ability of EGF to induce tyrosine phosphoryl- ation of the EGFR was reduced by at least 60%. These results suggest that PKC is involved in both transactivation and desensitization of the EGFR by GPCRs.
GPCR-induced EGFR Desensitization Is Associated with EGFR Internalization-One potential mechanism through which the EGFR could be desensitized is by internalization such that cell-surface EGFRs available for binding by EGF would be diminished. Thus, we measured cell-surface EGFRs by ligand binding and by confocal microscopy. For ligand bind-ing, cells were treated for 1 h with vehicle, EGF, or 5-HT and subjected to acid wash, and then cell-surface 125 I-EGF binding was measured. Fig. 7 shows that preincubation with either 5-HT (300 nM or 3 M) or EGF (20 ng/ml) resulted in a marked down-regulation of cell-surface 125 I-EGF binding, which was nearly complete after 10 min of pre-incubation. This downregulation of binding could be due either to decreased numbers of cell-surface receptors or to decreased affinity of the cellsurface receptors for EGF. Fig. 8 (panels A-D) shows the results of experiments in which cell-surface receptors were visualized in nonpermeabilized cells with a fluorescein isothiocyanate-conjugated anti-EGFR antibody (raised against an extracellular epitope of the EGFR). This method was used to visualize surface receptors on mesangial cells after incubation with vehicle (panel A), 300 nM 5-HT for 20 min (panel B) or 60 min (panel C), or EGF (20 ng/ml) for 60 min (panel D). The results show that there was a marked decrease in cell-surface EGFR after incubation with either 5-HT or EGF. Thus these two methods clearly demonstrate that preincubation with 5-HT or EGF reduces the number of cell-surface EGFRs in mesangial cells. Those studies cannot, however, distinguish between a redistribution of EGFRs to intracellular compartments and a loss of total EGFRs (from increased degradation or decreased synthesis). Thus, we used an EGFR-GFP fusion protein to assess whether 5-HT and EGF could induce a redistribution of EGFR within mesangial cells. This construct has already been used to demonstrate that EGF causes the EGFR-GFP fusion protein to internalize in a similar manner to wild-type EGFR in HEK293 and NIH 3T3 cells (29). We transiently transfected HEK293 cells with the EGFR-GFP construct and also with cDNA encoding the human 5-HT 2A receptor. Fig. 8, E-H, shows that both EGF and 5-HT induced redistribution of EGFR-GFP away from the cell surface and into a nuclear or perinuclear locale in HEK293 cells. Representative photomicrographs are shown for treatment with vehicle (panel E), with 300 nM 5-HT for 20 min (panel F) or 60 min (panel G), or with EGF (20 ng/ml) for 60 min (panel H). The red areas show the plasma membrane identified by rhodamine-concanavalin A after fixation of the cells, whereas the green areas represent the EGFR-GFP fusion protein (42). The yellow areas indicate superimposition of the red and green signals. We used a computer algorithm to provide a semiquantitative assessment (Lux units) of the subcellular localization of the EGFR-GFP fusion protein in HEK293 cells (Fig. 9). Those results showed that most of the fusion protein was located on or near the plasma membrane in quiescent cells, whereas the cell-surface receptors were reduced by Ϸ75% after stimulation with either EGF or 5-HT. (43), and it has been shown to mediate both PKC-dependent (27,44) and -independent effects (45) in those cells (see Fig. 1). We performed experiments using a specific PKC inhibitor (GF109203X) to establish a role for PKC in the down-regulation of cell-surface EGFR by 5-HT. Fig.  10 shows that in the absence of any inhibitor, both 5-HT and EGF resulted in a marked down-regulation of 125 I-EGF binding to intact mesangial cells after 60 min of incubation. Incubation with 3 M GF109203X nearly completely blocked the downregulation of the EGFR induced by 5-HT but was ineffective in blocking EGF-induced down-regulation. These data correlate well with those in Fig. 1c, which show that GF109203X blocks 5-HT-induced transphosphorylation of the EGFR. These data are also in keeping with a mechanism of action of PKC that occurs upstream of EGFR activation.

GPCR-induced EGFR Down-regulation Requires EGFR Internalization-Down-regulation of cell-surface receptors can
involve receptor internalization, degradation, or both. To study whether internalization of the EGFR is a component of the GPCR-induced down-regulation of cell-surface EGFRs, we transfected HEK293 cells with cDNAs encoding the EGFR-GFP fusion protein, the human 5-HT 2A receptor, and dominant negative forms of ␤-arrestin and dynamin GTPase. Dynamin is required for clathrin-mediated endocytosis, and a dominant negative version of dynamin (K44A dynamin) has been used previously to block internalization of both GPCRs and RTKs (46,47). A peptide fragment of ␤-arrestin 1, ␤-arrestin1-(319 -418), has been demonstrated previously to block GPCR-induced clathrin-mediated endocytosis (46,79) because it binds clathrin cages but not GPCRs. Fig. 11 shows that ␤-arrestin1-(319 -418) (panels C and D) and K44A dynamin (panels E and F) effectively prevent the 5-HT-induced endocytosis of EGFR in HEK293 cells as determined by confocal microscopy. Red indicates the decoration of the cell surface (post-fixation and treatment) by rhodamine-concanavalin A. The green signal is generated by the EGFR-GFP. Areas of overlap are indicated in yellow. In mock-transfected cells, most of the EGFR-GFP leaves a predominantly plasma membrane location and internalizes into intracellular compartments that seem to include the nucleus. In cells transfected with either of the dominant negative constructs, little 5-HT-induced internalization is seen after 60 min of treatment. Those results support a probable role for endocytosis in 5-HT-induced down-regulation of EGFRs in HEK293 cells. We also exposed rat mesangial cells to 5-HT in the presence and absence of various chemical inhibitors of endocytosis, including concanavalin A (ConA), monodansylcadaverine, and potassium depletion. Because no specific inhibitors of endocytosis are available, we had to use these multiple approaches to block endocytosis. Fig. 12a shows that these maneuvers attenuated the down-regulation of the EGFR induced by 5-HT (similar results of low temperature are not shown). We obtained similar results when endocytosis was blocked by low temperature or exposure to hypertonic sucrose (not shown). Fig. 12b shows that incubation with cycloheximide (30 g/ml) for 1 h did not impair the down-regulation of cell-surface EGFR by 5-HT, ruling out a significant contribution of EGFR protein synthesis in this effect. Thus, these studies demonstrate that internalization of the EGFR is a key component of its down-regulation by 5-HT. However, these studies do not demonstrate whether the functional desensitization of the EGFR induced by 5-HT also requires internalization.
Effect of 5-HT Pretreatment on Total Immunoreactive EGFR Protein-The next question to ask was whether GPCR activation leads only to internalization of EGFR (removal from the cell surface), or whether a component of reduction of the total complement of receptors within the cell is involved. Fig. 13 illustrates experiments in which immunoblots were performed from whole cell lysates after incubation with 1 M 5-HT for up to 150 min in the presence of cycloheximide. Those experiments show that the total cellular complement of EGFR is markedly reduced by treatment with 5-HT despite the presence of cycloheximide. Cycloheximide alone had no effect on the amount of EGFR immunoreactivity in whole cell lysates (not shown). The initial decline in EGFR immunoreactivity is very gradual, diminishing only by 25% at 60 min. Thus, within the time frame of desensitization of the EGFR by 5-HT, there is only a small decline in the total number of EGFRs. After 60 min, the immunoreactivity drops off sharply. If the down-regulation of the total amount of cellular EGFR does not involve alterations in protein synthesis, then the degradation of EGFR is likely accelerated by incubation with GPCR ligands. Thus, the initial desensitization of the EGFR by the 5-HT 2A receptor appears to be related to internalization of the EGFR, whereas later effects may be due to degradation of the EGFR. DISCUSSION These studies demonstrate that GPCRs can transactivate EGFRs through a PKC-dependent pathway. This transactivation results in the initiation of multiple cellular signals, as well as subsequent internalization and desensitization of the EGFRs. What is new about this work is that we show that activation of GPCRs can profoundly desensitize a prototypical RTK, the EGFR. The effect is rapid, being manifested within minutes, and is associated with a rapid internalization of cellsurface EGFRs. Desensitization of the EGFR can be initiated by several GPCRs (5-HT 2A , lysophosphatidic acid, angiotensin AT 1 , bradykinin B 2 ) that classically couple to G q -type G proteins. Desensitization can also be mimicked by chemical activation of PKC by PMA. Activation of the 5-HT 2A receptor desensitizes a number of EGFR signals, including EGFR autophosphorylation, phosphorylation of ERK, and activation of transcription factors (CREB, NF-B, and E2-F). Internalization and down-regulation of cell-surface EGFRs induced by 5-HT (but not EGF) can also be blocked by pharmacological inhibition of PKC. Thus, GPCRs can induce desensitization of EGFRs by a process associated with the loss of cell-surface EGFRs through internalization. Our data also show that the 5-HT 2A receptor transactivates the EGFR in a manner already shown to occur with other GPCRs such as those for angiotensin II (6), carbachol, lysophosphatidic acid, thrombin (12,39) and for the ␤ 2 adrenergic receptor (48). The process of EGFR induced by EGF and GPCRs is distinct in that the GPCR signal is PKC-dependent whereas the EGF signal is not. These relationships are depicted in Fig. 14.
Although our data implicate PKC in the activation and desensitization of the EGFR induced by GPCRs, the mechanism of that process is undefined. Prenzel et al. (11) showed that release of heparin-bound EGF by a membrane-bound metalloproteinase-like enzyme mediated some of the effects of GPCRs to activate EGFRs. This enzyme resembled zinc-dependent proteases called ADAMs (cell-surface proteins that contain a disintegrin and metalloprotease domain), some of which can be activated by PKC (49,50). We tested this possibility by incubating cells with three different inhibitors of metalloproteinases, MMP-3 inhibitors I and II (Calbiochem) and BB94 (Bristol Biotech). None of those inhibitors had any effect on 5-HT 2A receptor-induced phosphorylation of EGFRs (not shown). Thus, those studies did not support a role for a PKC-activated membrane-bound metalloproteinase-like enzyme in GPCR-induced activation of EGFRs in rat renal mesangial cells.
Previous evidence that PKC is involved in transactivation of the EGFR is variable (10,51). Some studies have suggested potential roles for PKC in the negative regulation of EGFR signaling (52)(53)(54)(55)(56). In fact, Beguinot et al. (57) demonstrated that PMA could induce internalization of EGFRs and a transient decrease in cell-surface 125 I-EGF binding without inducing degradation of the EGFRs. Others, however, have suggested that PKC-dependent phosphorylation enhances and stabilizes EGFR levels and/or signaling (58 -62). PKC-␣ was shown to associate with EGFR and to increase its phosphorylation in transfected HEK293 and NIH3T3 cells (54). The authors hypothesized that PKC-mediated EGFR phosphorylation played a key role in EGFR internalization. In that regard, it is tempting to speculate that such a mechanism could link our finding that both PKC and EGFR internalization mediate the desensitization process. Another group showed that PKC-␣ reduces EGFR numbers without changing the affinity of EGF for the EGFR (63). What separates our current report from the previous work described in this paragraph is that we used endogenous prototypical G q -coupled GPCRs to activate PKC, whereas the other studies almost exclusively used chemical activation of PKC to study its effects on EGFR functions. Moreover, we have demonstrated a clear-cut desensitization of several signals that emanate from the EGFR by multiple different GPCRs.
Some have suggested that PKC-mediated effects on EGFRs include reductions of high affinity binding sites and tyrosine autophosphorylation (64), although others were not able to demonstrate that PKC was involved in down-regulating EGFR (65,66). Harada et al. (65) showed evidence that PKC could alter the affinity of EGF for the EGFR without down-regulating the receptor. Kaji et al. (67) showed that PKC decreased the affinity of the EGFRs for EGF without changing receptor number or by inducing internalization. In contrast, PKC-␣ was shown to reduce EGFR numbers without changing the affinity of EGF for the EGFR (63).
Studies on the roles of specific serine/threonine phosphorylation sites of the EGFR have had a similar lack of consensus. The EGFR can be phosphorylated on Thr 654 by PKC (66) and on Thr 669 by ERK (40). Phosphorylation of Thr 654 was shown to decrease high affinity EGF binding to the EGFR, but this residue was not involved in PKC-mediated down-regulation of the EGFR (64). Verheijden et al. (68) showed that PKC inhibits EGFR tyrosine kinase activity without changing receptor dimerization. One group suggested that PKC-and ERK-dependent phosphorylation of the EGFR receptor does not mediate desensitization of the EGFR (69). PKC can phosphorylate the EGFR at Thr 654 (66), but one group could not link phosphorylation of either Thr 654 or Thr 669 to down-regulation of the EGFR (69). On the other hand, Bowen et al. (70) showed that phosphorylation of Thr 654 blocked mitogenic stimulation by the EGFR. Another group showed that phosphorylation of Thr 654 inhibits ligand-induced internalization and down-regulation of the EGFR (58).
Internalization of the EGFR through clathrin-coated pits appears to be the major process through which desensitization of the EGFR by GPCRs occurs. The evidence for this is that pretreatment of mesangial cells with 5-HT results in 1) a decrease in cell-surface 125 I-EGF binding, 2) a translocation of an EGFR-GFP fusion protein away from the plasma membrane, and 3) multiple inhibitors of clathrin-mediated endocytosis preventing GPCR-driven internalization of the EGFRs. Moreover, the blockade of endocytosis prevents desensitization of the EGFR. Desensitization appears to be independent of protein synthesis, because the studies were performed in the presence of inhibitors of protein synthesis. Desensitization in the first 60 min of pretreatment with 5-HT appears to be largely independent of protein degradation because the amount of total cellular immunoreactive EGFRs decreases by only about 25%.
We used several distinct maneuvers to block GPCR-induced internalization of EGFRs including ConA, monodansylcadaverine, hypertonic medium, potassium depletion, low temperature, and dominant interfering constructs of ␤-arrestin and dynamin. Because no specific inhibitors of endocytosis are available, we had to use these multiple approaches to block endocytosis. ConA is a lectin that binds selectively to glycoprotein-associated terminal mannose residues and blocks their lateral movement in many cell types (36). Monodansylcadaverine blocks clathrin-mediated internalization proximal to endocytic vesicles (37,38). Hyperosmolarity interferes with clathrin-mediated endocytosis by preventing the formation of clathrin-coated pits (35,71). Potassium depletion interferes FIG. 14. Schematic depiction of GPCR regulation of EGFR. This scheme shows that GPCRs regulate various EGFR functions through a PKC-dependent pathway.
with clathrin-mediated endocytosis by preventing the formation of clathrin-coated pits (33,72). Incubation in hypertonic medium prevents formation of clathrin-coated pits (34). Dynamin is required for clathrin-mediated endocytosis, and a dominant negative version of dynamin (K44A dynamin) has previously been used to block internalization of both GPCRs and RTKs (46,47). A peptide fragment of ␤-arrestin1, ␤-arrestin1-(319 -418), has previously been demonstrated to block GPCR-induced clathrin-mediated endocytosis (46,79), presumably because it binds clathrin cages, but is unable to bind to receptors (73,74). Thus, the effectiveness of multiple strategies to block endocytosis supports a role for clathrinmediated endocytosis of the EGFR in its desensitization by GPCRs.
In our experiments, we observed that ConA significantly lowered the basal level of 125 I-EGF binding (Fig. 12). The explanation for this effect is most likely ConA-induced proteolytic cleavage of the EGFR, as recently described by Tang et al. (75) in vascular smooth muscle cells. That group also showed evidence that the cleavage event involved mainly the carboxyl terminus of the EGFR and did not interfere with 125 I-EGF binding. This portion of the EGFR contains three major (Tyr 1068 , Tyr 1148 , and Tyr 1183 ) and two minor (Tyr 992 and Tyr 1086 ) autophosphorylation sites (76,77) and binding/activation sites for phospholipase C-␥, adapter protein 2 (78), and Shc (79). The proteolytic effect of ConA on EGFR does not universally attenuate EGFR functions, however. In NIH3T3 cells, ConA does not affect EGFR functions such as EGF binding or proximal signals like the activation of phospholipase C-␥ or sodium-proton exchange activity (80,81).
Although our work has not yet delineated the precise molecular mechanisms through which PKC and clathrin-mediated endocytosis combine to rapidly desensitize EGFRs after activation of GPCRs, the existence of this process defines a novel form of cross-talk between GPCRs and EGFRs. Thus, GPCRs can both transactivate and desensitize EGFRs in kidney mesangial cells. The implication of these findings is that GPCR activation in some settings may precondition cells to become unresponsive to subsequent challenge with mitogens like EGF that bind to and signal through RTKs. Further work will be needed to: 1) define the molecular mechanisms of this process, 2) determine whether GPCRs linked to other G proteins (G s , G i ) can desensitize the EGFR, 3) determine whether GPCRs can desensitize other RTKs, and 4) establish whether this process occurs in other cell types.