Suppression of EGFR endocytosis by dynamindepletion reveals that EGFR signaling occursprimarily at the plasma membraneLeiliane P. Sousaa, Irit Laxa, Hongying Shenb,c, Shawn M. Fergusonb, Pietro De Camillib,c, and Joseph Schlessingera,1

aDepartments of Pharmacology; bCell Biology; and cHoward Hughes Medical Institute, Yale University School of Medicine, New Haven, CT 06520

Contributed by Joseph Schlessinger, January 5, 2012 (sent for review December 13, 2011)

The role of endocytosis in the control of EGF receptor (EGFR) acti-vation and cell signaling was explored by using mouse fibroblastsin which dynaminwas conditionally depleted. Dynamin is a GTPaseshown to play an important role in the control clathrin mediatedendocytosis of EGFR and other cell surface receptors. In this report,we demonstrate that EGF binding activity and the display of highand low affinity EGFRs on the cell surface are not affected by dy-namin depletion. By contrast, dynamin depletion leads to a stronginhibition of EGFR endocytosis, robust enhancement of EGFR au-tophosphorylation and ubiquitination, and slower kinetics of EGFRdegradation. Surprisingly, MAPK stimulation induced by either lowor high EGF concentrations is not affected by dynamin depletion.While a similar initial Akt response is detected in control or dyna-min depleted fibroblasts, a somewhat more sustained Akt stimula-tion is detected in the dynamin depleted cells. These experimentsdemonstrate that dynamin-mediated endocytosis leads to attenua-tion of EGFR activation and degradation and that stimulation ofthe MAPK response and Akt activation are primarily mediated byactivated EGFR located in the plasma membrane.

membrane receptors ∣ tyrosine kinases

Epidermal growth factor receptor (EGFR) and other receptortyrosine kinases (RTK) undergo rapid internalization and de-gradation following ligand induced activation (1–3). At low phy-siological concentrations, EGF induced EGFR internalization isprimarily mediated by clathrin mediated endocytosis; a processblocked by siRNA silencing of clathrin heavy chain expressionor by overexpression of a dominant interfering dynamin mutant(K44) (4–6). By contrast, when high EGF concentrations are ap-plied, it is proposed that EGFR endocytosis is primarily mediatedby clathrin-independent mechanisms (3, 7, 8).

It was initially thought that the function of EGFR endocytosisand degradation was to terminate the signal initiated by EGFbinding to EGFRs located at the plasma membrane (1, 2). Sub-sequent studies demonstrated that EGFR and other RTKs, arealso capable of recruiting signaling molecules and transmittingsignals from endosomes (7, 9, 10). Moreover, it was proposed thatfollowing EGFR activation clathrin mediated endocytosis playsan important regulatory role in control of sustained activation ofthe MAPK/ERK signaling pathway and in Akt stimulation (4,11). The current prevailing view is that signals induced by acti-vated EGFR and other RTKs can be transmitted from the plasmamembrane as well as from endosomes and that the spatial loca-lizations of RTKs play an important role in the control of signalspecificity, duration, and robustness (4, 9–12).

Dynamin is large GTPase, that mediates the endocytic fissionof coated pits (13, 14). In this report we describe the analysis ofthe role played by dynamin in EGFR activation, endocytosis, andcell signaling. Using dynamin conditional knockout mouse fibro-blasts we demonstrate that the ligand binding characteristics andthe typical display of high and low affinity EGFR binding sites onthe cell surface are maintained and not affected by dynamin de-pletion. However, dynamin depletion leads to a strong inhibition

of EGFR endocytosis that is accompanied by enhanced tyrosineautophosphorylation, enhanced and prolonged EGFR ubiquiti-nation, and reduced EGFR degradation. We also demonstratethat depletion of dynamin expression results in selective enhance-ment in tyrosine phosphorylation of the p66 isoform of the adap-tor protein Shc. Moreover, while MAPK stimulation inducedby low or high concentrations of EGFare not affected by dynamindepletion, a somewhat more sustained Akt activation is observedin these cells. These experiments demonstrate that dynamin-mediated endocytosis results in attenuation of EGFR activation,autophosphorylation, and ubiquitination as well as in enhancedEGFR degradation. Furthermore, stimulation of the MAPK/ERK pathway and Akt activation can be effectively activated byEGFR located in the plasma membrane.

ResultsThe role of dynamin in the control of EGFR activation, endocy-tosis, and downstream signaling was explored by using fibroblastsderived from Dnm1 flox∕flox; Dnm2 flox∕flox; Cre-Esr1þ ∕ −mice(15, 16) following tamoxifen-induced gene recombination invitro. The Dnm1 and Dnm2 genes encode for dynamin 1 and 2,respectively, both of which are expressed in murine fibroblastswhile the third dynamin isoform, dynamin 3, is largely undetect-able in these cells. We therefore refer to these cells as dynamindepleted cells, or dynamin knockout cells (DKO).

Ligand Binding Activity of EGFR Is Not Influenced by Dynamin Deple-tion. In order to evaluate the effect of dynamin depletion on theligand binding characteristics of cell surface EGFRs, murine fi-broblasts were either pretreated with 4-hydroxy-tamoxifen or buf-fer alone and then subjected to quantitative binding experimentswith 125I-EGF. The effect of dynamin depletion on the ligandbinding characteristics of EGFR were determined by comparingdisplacement curves or Scatchard analyzes (17–19) of quantita-tive 125I-EGF binding experiments to live cells (Fig. 1). To con-firm that dynamin 1 and 2 were indeed depleted by tamoxifentreatment, samples of tamoxifen treated or untreated cells werelysed and subjected to immunoblotting with anti-dynamin antibo-dies (Fig. 1B). Control or dynamin depleted fibroblasts were in-cubated with 5 ng∕mL of 125I-EGF for 1 h at room temperaturein the presence of increasing concentrations of unlabeled EGF.The treated cells were lysed and the radioactive contents of thesamples were determined using a scintillation counter. The ex-periment presented in Fig. 1A depicts displacement curves of125I-EGF binding to control fibroblasts or to dynamin depletedfibroblasts. This experiment reveals similar 125I-EGF displace-

Author contributions: L.P.S., I.L., H.S., S.M.F., P.D.C., and J.S. designed research; L.P.S. andI.L. performed research; H.S., S.M.F., and P.D.C. contributed new reagents/analytic tools;L.P.S., I.L., H.S., S.M.F., P.D.C., and J.S. analyzed data; and L.P.S., I.L., H.S., S.M.F., P.D.C.,and J.S. wrote the paper.

The authors declare no conflict of interest.1To whom correspondence should be addressed. E-mail: Joseph.Schlessinger@Yale.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1200164109/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1200164109 PNAS ∣ March 20, 2012 ∣ vol. 109 ∣ no. 12 ∣ 4419–4424

BIOCH

EMISTR

Y

Dow

nloa

ded

by g

uest

on

June

16,

202

1

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1200164109/-/DCSupplementalhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1200164109/-/DCSupplementalhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1200164109/-/DCSupplementalhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1200164109/-/DCSupplementalhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1200164109/-/DCSupplementalhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1200164109/-/DCSupplemental

ment curves for control and dynamin depleted fibroblasts(Fig. 1B) with concentrations required for 50% inhibition, IC50of 0.62 and 0.40 nM, respectively. This experiment demonstratesthat the display of EGFR on the plasma membrane and the over-all binding affinity of EGFR towards EGF was not affected bydynamin loss. A similar conclusion was reached from Scatchardanalysis of quantitative binding experiments with concentrationsranging from 0.1 to 100 ng∕mL 125I-EGF to control or dynamindepleted fibroblasts. Similar numbers of EGFRs were present onthe cell surfaces of control and dynamin depleted cells and bothcells displayed a typical pattern of low and high affinity EGFRson their cell surfaces (Fig. 1C). It is noteworthy that under theconditions in which the 125I-EGF binding experiments were per-formed a substantial fraction of bound 125I-EGF molecules be-come internalized into dynamin depleted cells and even moreinto WTcells indicating that the profile of EGF binding charac-teristic is not sensitive to EGFR internalization. These experi-ments support the conclusion that the display of EGFR on thecell surface and the ligand binding characteristics of EGFRsare not affected by dynamin loss. However, these conclusionscontradict an earlier study demonstrating that overexpressionof a dominant interfering dynamin mutant (K44A) prevent highaffinity EGF binding and reduces EGF stimulation of EGFR au-tophosphorylation (20).

Endocytosis of EGFR Is Strongly Impaired in Dynamin Depleted Cells. Itwas previously reported that clathrin-mediated endocytosis isthe primary route for internalization of EGFR at low EGF con-centrations (1.5 ng∕mL), while at higher ligand concentrationsEGFR internalization may occur via both clathrin-dependent andclathrin-independent mechanisms (4, 8). To investigate whetherthe absence of dynamin affects EGFR internalization, we next ana-lyzed the kinetics of internalization of 125I-EGF using a previouslydescribed, well established quantitative procedure (21, 22). To thisend, control or dynamin depleted fibroblasts were incubated witheither low (1.5 ng∕mL) or high (100 ng∕mL) 125I-EGF concentra-tions for 90 min at 4 °C. At the end of the incubations, the cellswere washed and further incubated at 37 °C for different times. Toquantitatively determine the amount of EGFRs located at the cellmembrane, the cells were subjected to a mild acid wash at differenttime points to selectively release only cell surface bound 125I-EGF

molecules into the medium (21). The washed cells were then so-lubilized and the radioactive contents of the cell surface bound andthe internalized 125I-EGF molecules were separately determinedusing a scintillation counter (21).

The experiments presented in Fig. 2 A–H show that endocyto-sis of EGFR is strongly impaired in dynamin depleted fibro-blasts. The results show that after 15 min most cell surface EGFRin control fibroblasts became internalized in response to stimula-tion with either low or high EGF concentrations (Fig. 2 A–C andFig. 2 E–G). In contrast, a robust inhibition of EGFR internali-zation was detected in dynamin depleted fibroblasts stimulatedwith 1.5 ng∕mL (low dose) of EGF (Fig. 2B). Internalizationof EGFR was also compromised in dynamin depleted cells stimu-lated with 100 ng∕mL (high dose) of EGF (Fig. 2E). Under theseconditions approximately 70% inhibition of EGFR internaliza-tion was observed in dynamin depleted cells (Fig. 2G). These ex-periments also show that EGFR internalization is not completelyblocked in dynamin depleted cells. The partial internalizationthat takes place in dynamin depleted cells is probably not causedby the residual expression of dynamins in these cells, because re-

A

C

B

Fig. 1. Ligand binding characteristics of EGFR are not influenced by dynamindepletion. (A). Binding experiments were performed using a single concen-tration of 125I-labeled EGF in the presence of increasing concentration on un-labeled native EGF. The graphs depict competition experiments of 125I-labeled EGF binding to control (blue squares) or DKO (red circles) fibroblastsin the presence of increasing concentration of native EGF. Curve fitting tobinding data shown by lines and bars indicate standard deviation values.(B). Immunoblotting analysis of total cell lysates of control or DKO fibroblastswith anti-dynamin antibodies. Immunobloting with anti-AKT antibodieswas used as loading control. (C). Dissociation constants (Kd) and numbersof low and high affinity EGFR binding sites on control and DKO fibroblastswere determined using Scatchard analysis of quantitative 125I-EGF bindingexperiments to these cells. Similar results were obtained in three separate125I-EGF binding experiments.

A B

C D

E F

G H

Fig. 2. Endocytosis of EGFR is strongly impaired in dynamin depleted cells.Quantitative ligand internalization experiments using 1.5 ng∕mL (A–D) or20 ng∕mL (E–H) of 125I-EGF to control (A, E) and DKO (B, F) fibroblastsare shown. Surface bound 125I-EGF (blue), internalized 125I-EGF (red) anddegraded 125I-EGF (green) are shown. The TCA precipitatable radioactivity re-presenting intact released or recycled 125I-EGF molecules are not included inthe figures. The ratio of the amount of 125I EGF internalized vs. surface bound125I-EGF molecules are shown in (C) and (G) for control (solid line) and DKO(dashed line) fibroblasts. Each data point is the average value of duplicate re-sults. Data are presented as mean� SD, as indicated by the bars. Also shownimmunobloting analyzes with anti-dynamin antibodies or anti-EGFR antibo-dies to reveal dynamin or EGFR expression respectively, in control or DKO fi-broblasts. Similar results were obtained in three different experiments.

4420 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1200164109 Sousa et al.

Dow

nloa

ded

by g

uest

on

June

16,

202

1

sidual dynamin expression is extremely low (Fig. 2H). For thesame reason, it is also unlikely that the incomplete block ofEGFR endocytosis may reflect the presence of few cells whererecombination did not occur. A plausible explanation is that EGFinduced internalization of EGFR can follow both dynamin-dependent and dynamin-independent routes of internalization,especially when the cells are stimulated with high EGF concen-trations.

Ligand Induced Degradation of EGFR Is Compromised in DynaminDepleted Cells. Following EGF induced receptor activation andendocytosis, a fraction of the internalized EGFRs are degraded(3, 23, 24). The process of ligand induced EGFR eliminationfrom the cell surface and ensuing degradation that terminates sig-naling is known as “receptor down regulation” (25, 26). To ex-plore the effect of dynamin depletion on EGFR degradation,control or dynamin depleted fibroblasts were first incubated withcycloheximide for an hour to block protein synthesis and blockthe formation of new EGFRs. The cells were then treated with1, 10, or 100 ng∕mL of EGF, solubilized at different time pointsand subjected to immunoblotting analyses with anti-EGFR anti-bodies to reveal the amount of EGFR. Because minimal EGFRdegradation was detected in cells stimulated with 1 ng∕mL ofEGFafter 4 h, the experiment presented in Fig. 3 shows the effectof stimulation with 10 or 100 ng∕mL of EGF. As previously de-scribed (27), stimulation of control fibroblasts with 100 ng∕mL ofEGF led to robust degradation of EGFR with a half-time of ap-proximately 45 min (Fig. 3B). Slower kinetics of EGFR degrada-tion were detected in control fibroblasts treated with 10 ng∕mLof EGF with a half-time of 2–3 h (Fig. 3A). Previously describeddegradation products of EGFR were clearly detected after30 min of ligand stimulations in both experiments. The experi-ments presented in Fig. 3 A and B demonstrate that the stabilityof EGFR is increased in dynamin depleted fibroblasts that werestimulated with either 10 or 100 ng∕mL EGF. In dynamin de-pleted fibroblasts that were stimulated with 100 ng∕mL the half-time of EGFR degradation was extended to approximately 2 hwithout the appearance of EGFR degradation products seenin control stimulated cells. It is noteworthy that the typical EGFRdegradation products detected in control fibroblasts stimulated

with 100 ng∕mL EGF (21, 27) were not detected in dynamin de-pleted fibroblasts stimulated with the same EGF concentrationsuggesting that EGF induced degradation of EGFR may proceedvia a different mechanism in dynamin deficient cells.

Enhanced Ligand Induced Autophosphorylation and Ubiquitination ofEGFR in Dynamin Deficient Cells. We next examined the effect ofdynamin depletion on EGF stimulated EGFR tyrosine phosphor-ylation and ubiqutination (Fig. 4 A–C). Control or dynamin de-pleted fibroblasts were stimulated with 1.5, 5, or 100 ng∕mL ofEGF. Cell lysates of EGF stimulated or unstimulated cells weresubjected to immunoprecipitation with anti-EGFR antibodiesfollowed by immunoblotting with either anti-pTyr or anti-ubiqui-tin antibodies. The experiments presented in Fig. 4 A–C show anoverall strong increase in tyrosine phosphorylation and ubiquiti-nation of EGFR in dynamin depleted cells stimulated with1.5 ng∕mL of EGF. Interestingly bimodal activation of EGFRautophosphorylation was reproducibly observed in cells stimu-lated with 1.5 or 5 ng∕mL of EGF. The experiment presentedin Fig. 4A shows that in WTcells enhanced EGFR autophosphor-ylation occurs after 5 min of stimulation with 1.5 ng∕mL of EGFfollowed by reduced autophosphorylation at the 15 and 30 mintime points that is followed after 1 and 3 h by a second peak. Indynamin depleted fibroblasts stimulated with 1.5 ng∕mL EGF, astronger and earlier onset of the first peak was seen after 2 minfollowed by reduced autophosphorylation at the 5, 15, and 30 mintime points that is followed after 1 and 3 h by a strong secondpeak of autophosphorylation. These experiments also show thatthe levels of EGFR ubiquitination, which is weakly detected incontrol fibroblasts, is strongly stimulated in dynamin depleted fi-broblasts stimulated with 1.5 ng∕mL of EGF (Fig. 4A).

Upon stimulation with 5 ng∕mL EGF stronger enhancementof autophosphorylation of EGFR was detected in control anddynamin depleted fibroblasts (Fig. 4B) with a robust early onsetafter 2 min stimulation of the dynamin depleted cells. Similarly,enhanced ubiquitination of EGFR was also detected in cells sti-mulated with 5 ng∕mL of EGF (Fig. 4B).

While autophosphorylation and ubiquitination of EGFR weremuch more robust in both control and dynamin depleted fibro-blasts during the first thirty minutes of stimulation with 100 ng∕mL EGF (Fig. 4C) the initial decline and the second elevatedphase of autophosphorylation that was seen in 1.5 or 5 ng∕mLEGF stimulated cells at the 1 and 3 h time points were not detectedat this higher EGF concentration.

We also examined the effect of pretreatment with cycloheximideon ligand induced EGFR autophosphorylation and ubiquitination.These experiments showed minor effects of cycloheximide treat-ment on these posttranslational modifications (Fig. S1), indicatingthat preexisting pools of EGFR play a major role in the controlof EGFR autophosphorylation and ubiquitination during the first3 h of ligand stimulation.

Altered Pattern of Tyrosine Phosphorylation of Shc Isoforms in Dyna-min Depleted Fibroblasts. Shc is an adapter protein containingan SH2 and PTB domain that functions upon tyrosine phosphor-ylation by EGFR and other RTKs as an important link betweenEGFR and the RAS/MAP kinase signaling pathway by recruit-ment of Grb2/Sos complexes (28). Grb2 is an adaptor proteincomposed of one SH2 domain flanked by two SH3 domains whichplays an important role via direct or indirect interactions withEGFR and other RTKs to link between RTK stimulation andthe Ras/MAP kinase signaling pathway (2, 28). Most cells expressthree known Shc isoforms designated p42, p52, and p66 (29,30). The experiments presented in Fig. 4 D–F show that boththe p42 and p52 isoforms of Shc are similarly tyrosine phosphory-lated in control or dynamin depleted fibroblasts in response tolow (1.5 ng∕mL) or high (100 ng∕mL) EGF concentrations(Fig. 4 D–F, lower boxes). Interestingly, a more robust and sus-

A

B

Fig. 3. Ligand induced degradation of EGFR is compromised in dynamin de-pleted cells. Serum starved control or DKO fibroblasts were pretreated with10 μM of cycloheximide for 1 h followed by stimulation with 10 ng∕mL (A) or100 ng∕mL (B) of EGF for indicated times. Equal amounts of cell lysates weresubjected to immunobloting with anti-EGFR antibodies and as controls withanti-dynamin or anti-AKT antibodies. Arrow points to an EGFR degradationproduct that is detected in EGF stimulated WT cells and not in dynamin de-pleted cells. Similar results were obtained in three different experiments.

Sousa et al. PNAS ∣ March 20, 2012 ∣ vol. 109 ∣ no. 12 ∣ 4421

BIOCH

EMISTR

Y

Dow

nloa

ded

by g

uest

on

June

16,

202

1

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1200164109/-/DCSupplemental/pnas.1200164109_SI.pdf?targetid=SF1http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1200164109/-/DCSupplemental/pnas.1200164109_SI.pdf?targetid=SF1

tained tyrosine phosphorylation of the p66 isoform of Shc wasdetected in dynamin depleted fibroblasts stimulated by eitherlow or high EGF concentrations (Fig. 4 D–F, lower boxes). Theseexperiments showed similar Grb2 recruitment in Shc immunopre-cipitates in lysates from stimulated control or dynamin depletedfibroblasts suggesting that Grb2 becomes associated primarilywith the tyrosine phosphorylated p42 and/or p52 Shc isoformsand that the p66 isoform may be involved in mediating a Grb2independent process (31).

Similar Stimulation of MAP Kinase and Akt Responses in DynaminDepleted Cells.We next examined the effect of dynamin depletionon EGF stimulation of the MAP kinase (ERK) and Akt signalingpathways in response to either low (1.5 ng∕mL) or high(100 ng∕mL) EGF concentrations (Fig. 5). Control or dynamindepleted fibroblasts were stimulated with EGF. At different timepoints, the cells were solubilized and subjected to immunoblot-ting with antibodies that recognized either MAPK (anti-ERK) oractivated MAPK (anti-pERK). The samples were also subjectedto immunoblotting with antibodies that selectively recognize totalAkt and the activated form of Akt (anti-pAkt). The experimentspresented in Fig. 5 show a very similar profile of MAPK stimula-tion in control and dynamin depleted fibroblasts over a 3 h periodin response to 1.5, 5, or 100 ng∕mL of EGF stimulation. A similarAkt response was also observed in control or dynamin depletedfibroblasts stimulated with either low or high EGF concentra-tions. However, a somewhat more sustained Akt activation wasreproducibly detected in dynamin depleted cells stimulated withany of these EGF concentrations (Fig. 5).

DiscussionIt is well established that clathrin-mediated endocytosis of EGFRor other RTKs plays an important role in the control of receptordown regulation; a process mediated by intracellular degradationof both EGF and EGFR which results in signal termination (1, 2,21, 25, 26). Subsequent studies reporting experiments in whichclathrin mediated endocytosis of EGFR was blocked by eitherectopic overexpression of a dominant interfering dynamin mutant(5, 32) or by silencing the expression of clathrin heavy chain usingspecific siRNAs (4, 6) concluded that EGFR molecules interna-lized by means of clathrin mediated endocytosis are capable ofrecruitment and activation of critical intracellular signaling path-ways from endosomal compartments (5, 12, 33).

In this report we use dynamin depleted murine fibroblasts toexplore the role played by endocytosis in the control of EGFR

display on the cell surface, in regulation of EGFR activation,EGFR degradation, and in cell signaling via EGFR. Our experi-ments demonstrate that the expression and display of high andlow affinity EGFRs on the cell surface are not affected by dyna-min loss. These experiments contradict earlier studies demon-strating that overexpression of a dominant interfering dynaminmutant prevent high affinity EGF binding and reduces EGFRautophosphorylation (20). Ligand induced endocytosis of EGFRis strongly impaired in dynamin depleted fibroblasts stimulatedwith either low EGF concentrations [1-1.5 ng∕mL, conditionsunder which internalization of EGFR is primarily driven by cla-thrin mediated endocytosis (4, 8)], or high EGF concentrations(100 ng∕mL, a condition under which EGFR endocytosis isthought to be mediated by both clathrin-dependent and clathrin-independent mechanisms). The experiments performed withdynamin depleted fibroblasts stimulated with low EGF concen-tration provide an opportunity to address the role of clathrin-mediated endocytosis in the control of EGFR activation andsignaling via EGFR. These experiments clearly demonstrate thatduring the early phase of EGF stimulation under conditions inwhich the majority of EGFR are still located at the cell surface,autophosphorylation and ubiquitination of EGFR are stronglyenhanced indicating that autophosphorylation and ubiquitinationof EGFR are taking place primarily by activated EGFR locatedat the cell membrane. Autophosphorylation of EGFR leads torecruitment of Cbl which is responsible for the ubiquitinationand subsequent degradation of internalized EGFR molecules(34–36). The enhanced tyrosine autophosphorylation and ubi-quitination of EGFR observed under conditions in which EGFRendocytosis is strongly compromised suggests that tyrosine phos-phatases and deubiquitinating enzymes (DUB) may start to actshortly after activated EGFR become internalized and that Cblactivity is reduced prior to the onset of EGFR degradation bylysosomal enzymes. It is noteworthy that similar results wereobtained using HeLa cells in which endocytosis of EGFR wasimpaired by siRNA silencing of clathrin heavy chain. Namely,cells in which endocytosis is impaired by depletion of clathrinheavy chain expression also exhibit enhanced EGFR autopho-sphorylation and ubiquitination (Fig. S2). Moreover, similarMAPK stimulation was observed in control cells or in cells inwhich EGFR endocytosis was impaired by clathrin heavy chaindepletion (Fig. S3). Similar EGF stimulation of Map Kinaseof ERK Kinase (MEK) and ERK activation was previously de-scribed in HeLa cells in which clatharin expression was silenced

A B C

D E F

Fig. 4. Enhanced EGF induced autophosphorylation and ubiquitination of EGFR and altered pattern of phosphorylation of Shc isoforms in dynamin depletedcells. Serum starved control or DKO fibroblasts were stimulated with 1.5 ng∕mL (A, D), 5 ng∕mL (B, E) or 100 ng∕mL (C, F) of EGF for different times. Equalamounts of cell lysates were subjected to immunoprecipitation with either anti-EGFR or anti-Shc antibodies. (A–C) The anti-EGFR immunoprecipitates wereimmunobloted with anti-phosphotyrosine (pY) and reprobed with anti-ubiquitin (UB) antibodies. Enhanced tyrosine phosphorylation of EGFR is marked withasterisks. (D–E) Anti-Shc immunoprecipitates were subjected to immunobloting with anti-pY antibodies (asterisk marks the presence of an nonspecific proteinband observed in control cells) and reprobed using anti-Shc or anti-Grb2 antibodies. The efficiency of tamoxifen induced dynamin depletion as well as the levelsof EGFR expression in control and DKO fibroblasts were determined by immunobloting with anti-dynamin or anti-EGFR antibodies, respectively (not shown forclarity). Similar results were obtained in three different experiments.

4422 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1200164109 Sousa et al.

Dow

nloa

ded

by g

uest

on

June

16,

202

1

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1200164109/-/DCSupplemental/pnas.1200164109_SI.pdf?targetid=SF2http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1200164109/-/DCSupplemental/pnas.1200164109_SI.pdf?targetid=SF2http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1200164109/-/DCSupplemental/pnas.1200164109_SI.pdf?targetid=SF3

by specific siRNA (37). These experiments provide further sup-port to the conclusion that MAP kinase stimulation is primarilymediated by activated EGFR located at the plasma membrane.

We have also shown that unaltered tyrosine phosphorylation ofthe p42 and p52 isoforms of the adapter protein Shc and complexformation between Shc and Grb2 are detected both in controland in dynamin depleted fibroblasts. By contrast, selective tyro-sine phosphorylation of the p66 isoform of Shc is detected indynamin depleted cells in which the majority of activated EGFRare located on the cell membrane.

Surprisingly and in contrast to earlier publications demonstrat-ing that MAPK signaling is compromised when endocytosis wasprevented (4, 5), MAP kinase stimulation by EGF was compar-able in control and in dynamin depleted cells. Thus, the experi-ments presented in this report demonstrate that MAP kinasestimulation is driven primarily by activated EGFR located atthe plasma membrane rather then by activated EGFR locatedinside the cell. Although MAP kinase activation is primarily

mediated by activated EGFR located on the cell surface, the factthat a similar MAP kinase response is observed in cells in whichthe majority of EGFR are internalized suggests that the fractionof activated EGFR that are located on the cell surface are cap-able stimulating the entire MAP kinase response.

Interestingly, EGF stimulation of Akt differs from the MAPkinase response, as we observed a somewhat more sustainedAkt stimulation in dynamin depleted fibroblasts stimulated witheither low or high EGF concentrations. This experiment suggeststhat Akt stimulation may be driven both by activated EGFRlocated at the cell membrane and by activated EGFR locatedin the membranes of intracellular organelles such as endosomes.These results also suggest, however, that termination of Akt sig-naling is enhanced by endocytosis of the EGFR.

All in all, while the experiments presented in this report donot rule out the possibility that activated EGFRs are capableof stimulating signaling pathways from within the cells; i.e., fromendosomes, it is clear that the MAP kinase response is primarilymediated by activated EGFR located at the cell membrane.Moreover, one of the functions of endocytosis is to suppressEGFR autophosphorylation and ubiquitination which are pri-marily taking place at the cell membrane likely by reduced actionof Cbl and possibly by the action of tyrosine phosphatases andDUB that operate during endocytosis.

Material and MethodsCells and Culture. Tamoxifen-inducible dynamin conditional knock-out mouse fibroblasts (dnm1 flox∕flox; dnm2 flox∕flox; Cre-Esr1þ∕0)were previously described (38). Mouse fibroblasts were grownin Dulbecco modified Eagle medium (DMEM) containing 10%serum and 1% streptomycin and penicillin mixture. Dynamin de-pleted fibroblasts were generated by treating the cells twice with2 μM 4-hydroxy-tamoxifen (Sigma) on sequential days. Six daysafter initiating tamoxifen treatment, cells were either plated intosix-well plates coated with collagen (BD Biosciences) or serumstarved overnight prior to EGF stimulation.

125I-Labeled EGF Binding and Internalization Experiments. MurineEGF (Biomedical Technologies) was labeled with 125I usingthe Iodogen method (Pierce) according to the manufacturer’s in-structions. Tamoxifen treated (DKO) and nontreated (control)cells were plated into six-well plates and allowed to grow over-night. 125I-EGF binding and internalization experiments wereperformed as previously described (19, 39, 40). For binding ex-periments, cells were incubated with the indicated concentrationof 125I-EGF at room temperature for 1 h in the presence of in-creasing concentration of nonlabeled EGF; conditions permittinginternalization of bound 125I-EGF molecules. Cells were lysed in0.5 M of NaOH overnight and their radioactive content quanti-fied in 10 mL Optifluor (Perkin Elmer) with a scintillation coun-ter (LS6500, Beckman Coulter). The half time of displacementcurves of 125I-EGF binding to control and DKO fibroblasts werecalculated by curve fitting using PRISM3 software (GraphPad).

Scatchard analysis of 125I-EGF binding experiments was car-ried out in triplicate using concentrations of 125I-EGF rangingfrom 0.1 to 100 ng∕mL. A 100-fold excess of nonlabeled EGFwas added for each assayed concentration to measure nonspecificbinding. The bound radioactivity was quantified as describedabove. The average values per well for control and DKO cellswere determined from two independent counts and used to cal-culate values of Bmax (number of receptors per cell) using non-linear curve fitting to saturation binding according to Scatchardanalysis as previously described (15, 17).

Quantitative analyzes of EGF internalization were performedby incubating cells with 1.5 or 20 ng∕mL of 125I-EGF for 90 minat 4 °C. The labeled cells were then washed twice with cold PBS toremove ligand excess followed by addition of prewarmed mediumand incubation at 37 °C for the indicated times. After an acid

A

B

C

Fig. 5. Similar stimulation of MAP kinase and AKT responses in dynamin de-pleted cells. Serum starved control or DKO fibroblasts were stimulated with1.5 ng∕mL (A), 5 ng∕mL (B), or 100 ng∕mL (C) of EGF. Cells were collected atthe indicated time points and equal amounts of lysates were subjected toimmunobloting with anti-AKT or anti-ERK antibodies. Membranes were re-probed with anti-pAKT or anti-pERK antibodies to monitor their enzymaticactivities. EGFR and dynamin levels were monitored by inmunoblotting withanti-EGFR or anti-dynamin antibodies respectively. Similar results were ob-tained in three different experiments.

Sousa et al. PNAS ∣ March 20, 2012 ∣ vol. 109 ∣ no. 12 ∣ 4423

BIOCH

EMISTR

Y

Dow

nloa

ded

by g

uest

on

June

16,

202

1

wash to remove cell surface bound ligand molecules, the cell sur-face bound and intracellular 125I-EGF molecules were collectedfor each time point and quantified using a scintillation counter.The incubation medium was precipitated with 10% TCA (tri-chloroacetic acid) in order to quantify the amount of degraded125I-EGF molecules in the medium. All the results are presentedas a percentage of total cell associated 125I-EGF radioactivityafter 90 min incubation (time 0) at 4 °C (mean� SD). To revealthe efficiency of dynamin depletions, cell lysates were collectedfrom control and DKO cells plated in extra wells from each bind-ing and internalization experiment followed by immunoblottingwith anti-dynamin or anti-EGFR antibodies.

Immunoblotting and Immunoprecipitation Analysis.After serum star-vation, control or DKO fibroblasts were stimulated with 1.5, 5, 10,or 100 ng∕mL of EGF at 37 °C for the indicated time points andcollected in lysis buffer (50 mM Hepes, 150 mM NaCL, 1 mMEDTA, 1 mMEGTA, 10% glycerol, 1% TritonX-100, 25 mM NaF,10 μM ZnCl2, 1 mM NaVO4) that includes a protease inhibitor

cocktail (Roche). To carry out the degradation experiments,control and DKO fibroblasts were preincubated prior to ligandstimulation with 10 μMof cycloheximide for 1 h. Identical amountsof total cell lysates were subjected to inmunoprecipitation withanti-EGFR antibodies. The samples were also subjected toimmunoblotting analysis with anti-AKT, anti-pAKT (Cell SignalingTechnology), anti-ERK, anti-pERK, anti-ubiquitin (Santa CruzBiotechnology), anti-dynamin (clone 41, BD Biosciences), andanti-4G10, anti-phosphotyrosine antibodies. The anti-EGFR(ab328) antibodies used in the study were generated in our labora-tory. Primary antibodies were detected using anti-mouse HRP andProtein A-HRP (Santa Cruz Biotechnology), and visualized by achemiluminescence kit (Denville Scientific Inc.). Equal loadingof proteins analyzed by immunoblotting or immunoprecipitationanalyzes were confirmed by reprobing the stripped membranes(0.2 M NaOH, 5 min) with anti-AKT antibodies.

ACKNOWLEDGMENTS. The authors thank members of the Schlessinger groupfor helpful discussion.

1. Carpenter G, Cohen S (1979) Epidermal growth factor. Annu Rev Biochem 48:193–216.2. Lemmon MA, Schlessinger J (2010) Cell signaling by receptor tyrosine kinases. Cell

141:1117–1134.3. Sorkin A, Goh LK (2009) Endocytosis and intracellular trafficking of ErbBs. Exp Cell Res

315:683–696.4. Sigismund S, et al. (2008) Clathrin-mediated internalization is essential for sustained

EGFR signaling but dispensable for degradation. Dev Cell 15:209–219.5. Vieira AV, Lamaze C, Schmid SL (1996) Control of EGF receptor signaling by clathrin-

mediated endocytosis. Science 274:2086–2089.6. Huang F, Khvorova A, Marshall W, Sorkin A (2004) Analysis of clathrin-mediated

endocytosis of epidermal growth factor receptor by RNA interference. J Biol Chem279:16657–16661.

7. von Zastrow M, Sorkin A (2007) Signaling on the endocytic pathway. Curr Opin CellBiol 19:436–445.

8. Sigismund S, et al. (2005) Clathrin-independent endocytosis of ubiquitinated cargos.Proc Natl Acad Sci USA 102:2760–2765.

9. Murphy JE, Padilla BE, Hasdemir B, Cottrell GS, Bunnett NW (2009) Endosomes: alegitimate platform for the signaling train. Proc Natl Acad Sci USA 106:17615–17622.

10. Miaczynska M, Pelkmans L, Zerial M (2004) Not just a sink: endosomes in control ofsignal transduction. Curr Opin Cell Biol 16:400–406.

11. Taub N, Teis D, Ebner HL, Hess MW, Huber LA (2007) Late endosomal traffic ofthe epidermal growth factor receptor ensures spatial and temporal fidelity of mito-gen-activated protein kinase signaling. Mol Biol Cell 18:4698–4710.

12. Hupalowska A, Miaczynska M (2011) The new faces of endocytosis in signaling. Traffic3:9–18.

13. Ferguson S, De Camilli P (2012) Dynamin, a membrane remodeling GTPase. Nat RevMol Cell Biol , 13 pp:75–88.

14. Schmid SL, Frolov VA (2011) Dynamin: functional design of a membrane fissioncatalyst. Annu Rev Cell Dev Biol 27:79–105.

15. Ferguson SM, et al. (2009) Coordinated actions of actin and BAR proteins upstream ofdynamin at endocytic clathrin-coated pits. Dev Cell 17:811–822.

16. Shen H, et al. (2011) Constitutive activated Cdc42-associated kinase (Ack) phosphor-ylation at arrested endocytic clathrin-coated pits of cells that lack dynamin. Mol BiolCell 22:493–502.

17. Schlessinger J (1986) Allosteric regulation of the epidermal growth factor receptorkinase. J Cell Biol 103:2067–2072.

18. Mattoon D, Klein P, Lemmon MA, Lax I, Schlessinger J (2004) The tethered configura-tion of the EGF receptor extracellular domain exerts only a limited control of receptorfunction. Proc Natl Acad Sci USA 101:923–928.

19. Ozcan F, Klein P, Lemmon MA, Lax I, Schlessinger J (2006) On the nature of low- andhigh-affinity EGF receptors on living cells. Proc Natl Acad Sci USA 103:5735–5740.

20. Ringerike T, et al. (1998) High-affinity binding of epidermal growth factor (EGF) to EGFreceptor is disrupted by overexpression of mutant dynamin (K44A). J Biol Chem273:16639–16642.

21. Honegger AM, et al. (1987) Point mutation at the ATP binding site of EGF receptorabolishes protein-tyrosine kinase activity and alters cellular routing. Cell 51:199–209.

22. Sorkin A, Duex JE (2010) Quantitative analysis of endocytosis and turnover of epider-mal growth factor (EGF) and EGF receptor. Current Protocols in Cell Biology46:15.14.1–15.14.20 Chapter 15:Unit 15.14.

23. Kirisits A, Pils D, Krainer M (2007) Epidermal growth factor receptor degradation: analternative view of oncogenic pathways. Int J Biochem Cell Biol 39:2173–2182.

24. Madshus IH, Stang E (2009) Internalization and intracellular sorting of the EGF recep-tor: a model for understanding the mechanisms of receptor trafficking. J Cell Sci122:3433–3439.

25. Beguinot L, Lyall RM, Willingham MC, Pastan I (1984) Down-regulation of the epider-mal growth factor receptor in KB cells is due to receptor internalization and subse-quent degradation in lysosomes. Proc Natl Acad Sci USA 81:2384–2388.

26. Stoscheck CM, Carpenter G (1984) Down regulation of epidermal growth factor recep-tors: direct demonstration of receptor degradation in human fibroblasts. J Cell Biol98:1048–1053.

27. Honegger AM, Schmidt A, Ullrich A, Schlessinger J (1990) Separate endocytic pathwaysof kinase-defective and -active EGF receptor mutants expressed in same cells. J Cell Biol110:1541–1548.

28. Ravichandran KS (2001) Signaling via Shc family adapter proteins. Oncogene20:6322–6330.

29. Pelicci G, et al. (1996) A family of Shc related proteins with conserved PTB, CH1 and SH2regions. Oncogene 13:633–641.

30. Migliaccio E, et al. (1997) Opposite effects of the p52shc/p46shc and p66shc splicingisoforms on the EGF receptor-MAP kinase-fos signalling pathway. EMBO J 16:706–716.

31. Xi G, Shen X, Clemmons DR (2010) p66shc inhibits insulin-like growth factor-I signalingvia direct binding to Src through its polyproline and Src homology 2 domains, resultingin impairment of Src kinase activation. J Biol Chem 285:6937–6951.

32. Johannessen LE, Ringerike T, Molnes J, Madshus IH (2000) Epidermal growth factorreceptor efficiently activates mitogen-activated protein kinase in HeLa cells andHep2 cells conditionally defective in clathrin-dependent endocytosis. Exp Cell Res260:136–145.

33. Roepstorff K, Grøvdal L, Grandal M, Lerdrup M, van Deurs B (2008) Endocytic down-regulation of ErbB receptors: mechanisms and relevance in cancer. Histochem Cell Biol129:563–578.

34. Bertelsen V, Breen K, Sandvig K, Stang E, Madshus IH (2007) The Cbl-interactingprotein TULA inhibits dynamin-dependent endocytosis. Exp Cell Res 313:1696–1709.

35. Argenzio E, et al. (2011) Proteomic snapshot of the EGF-induced ubiquitin network.Mol Syst Biol 7:462.

36. Visser Smit GD, et al. (2009) Cbl controls EGFR fate by regulating early endosomefusion. Science Signaling 2:ra86.

37. Galperin E, Sorkin A (2008) Endosomal targeting of MEK2 requires RAF, MEK kinaseactivity and clatharin-dependent endocytosis. Traffic 9:1776–1790.

38. Ferguson SM, et al. (2007) A selective activity-dependent requirement for dynamin 1 insynaptic vesicle endocytosis. Science 316:570–574.

39. Yang Y, Yuzawa S, Schlessinger J (2008) Contacts betweenmembrane proximal regionsof the PDGF receptor ectodomain are required for receptor activation but not for re-ceptor dimerization. Proc Natl Acad Sci USA 105:7681–7686.

40. Mattoon DR, Lamothe B, Lax I, Schlessinger J (2004) The docking protein Gab1 is theprimary mediator of EGF-stimulated activation of the PI-3K/Akt cell survival pathway.BMC Biol 2:24.

4424 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1200164109 Sousa et al.

Dow

nloa

ded

by g

uest

on

June

16,

202

1