Protein Tyrosine Phosphatase PTPN3 Inhibits Lung Cancer Cell Proliferation and Migration by Promoting EGFR Endocytic Degradation

ORIGINAL ARTICLE Protein tyrosine phosphatase PTPN3 inhibits lung cancer cell proliferation and migration by promoting EGFR endocytic degradation

M-Y Li1,2, P-L Lai1, Y-T Chou3, A-P Chi1, Y-Z Mi1, K-H Khoo1,2, G-D Chang2, C-W Wu3, T-C Meng1,2 and G-C Chen1,2

Epidermal growth factor receptor (EGFR) regulates multiple signaling cascades essential for cell proliferation, growth and differentiation. Using a genetic approach, we found that Drosophila FERM and PDZ domain-containing protein tyrosine phosphatase, dPtpmeg, negatively regulates border cell migration and inhibits the EGFR/Ras/mitogen-activated protein kinase signaling pathway during wing morphogenesis. We further identified EGFR pathway substrate 15 (Eps15) as a target of dPtpmeg and its human homolog PTPN3. Eps15 is a scaffolding adaptor protein known to be involved in EGFR endocytosis and trafficking. Interestingly, PTPN3-mediated tyrosine dephosphorylation of Eps15 promotes EGFR for lipid raft-mediated endocytosis and lysosomal degradation. PTPN3 and the Eps15 tyrosine phosphorylation-deficient mutant suppress non-small-cell lung cancer cell growth and migration in vitro and reduce lung tumor xenograft growth in vivo. Moreover, depletion of PTPN3 impairs the degradation of EGFR and enhances proliferation and tumorigenicity of lung cancer cells. Taken together, these results indicate that PTPN3 may act as a tumor suppressor in lung cancer through its modulation of EGFR signaling.

Oncogene (2015) 34, 3791–3803; doi:10.1038/onc.2014.312; published online 29 September 2014

INTRODUCTION sorting EGFR to multivesicular bodies.15 Recently, Ali et al.16 Reversible tyrosine protein phosphorylation by protein tyrosine showed that the ESCRT accessory protein HD-PTP/PTPN23 kinases and protein tyrosine phosphatases (PTPs) acts as a coordinates with the ubiquitin-specific peptidase UBPY to drive molecular switch that regulates a variety of biological pro- EGFR sorting to the multivesicular bodies. A better understanding cesses.1,2 The receptor tyrosine kinase epidermal growth factor of the role of PTPs in regulating EGFR signaling will help to receptor (EGFR), the best characterized member of the ErbB family provide insights into the molecular mechanisms behind EGFR- receptors, acts as a critical regulator of numerous cellular mediated tumorigenesis. processes, including growth, proliferation and differentiation. PTPN3 (PTPH1) and the closely-related PTPN4 (PTPMEG) are Upon activation by its growth factor ligands, EGFR undergoes non-transmembrane PTPs that contain an N-terminal FERM dimerization and activation, leading to tyrosine phosphorylation (Band 4.1, Ezrin, Radixin, Moesin homology) domain followed by of the intracellular region of the receptor as well as many a single PDZ (PSD95, Dlg, ZO1) domain and the C-terminal PTP 3,4 cytoplasmic substrates. The activated EGFR is then internalized domain.1 They have been implicated in the regulation of cell by clathrin-mediated endocytosis and sorted into the endosomal growth and proliferation.17,18 However, their role in receptor compartments, through which it is either recycled back to the protein tyrosine kinase signaling is not clear. The dPtpmeg is the plasma membrane or transported to the lysosome for degrada- 5 Drosophila homolog of mammalian PTPN3 and PTPN4. Phenotypic tion. Because overexpression or constitutive activation of EGFR analyses have revealed that dptpmeg mutants exhibit aberrant has been implicated in the pathogenesis and progression of a mushroom body axon projection patterns in the brain.19 Besides variety of human malignancies,6 it is therefore crucial to under- its role in regulating neuronal wiring, the molecular function of stand how EGFR signaling is regulated. dPtpmeg has remained largely unknown. In this study, we Several PTPs have been implicated in the regulation of EGFR fi signaling. Among them, PTPRK, DEP-1 (PTPRJ), PTP1B (PTPN1), identi ed EGFR pathway substrate 15 (Eps15) as a substrate of SHP-1 (PTPN6), TCPTP (PTPN2), PTPN9 and PTPN12 have been dPtpmeg and PTPN3. Eps15 is known to be an endocytic adaptor fi 20 shown to downregulate EGFR signaling by dephosphorylating involved in the regulation of EGFR traf cking. PTPN3 depho- EGFR.7–11 The receptor-type PTP DEP-1 dephosphorylates EGFR on sphorylated Eps15 and promoted EGFR for lipid raft-mediated the cell surface and inhibits its internalization.7 On the other hand, endocytosis and lysosomal degradation. The ectopic expression of the endoplasmic reticulum-localized PTP1B has been reported to PTPN3 or Eps15-Y850F mutant in the non-small-cell lung cancer regulate EGFR signaling from endosomes.12,13 PTP1B promotes (NSCLC) cells inhibited cell proliferation, migration and tumor the sequestration of EGFR onto internal vesicles of multivesicular growth. Our findings uncover a novel role for PTPN3 in the bodies.14 The ESCRT (endosomal sorting complex required for regulation of EGFR endocytic trafficking, degradation and transport) complexes are known to play an important role in signaling.

1Institute of Biological Chemistry, Academia Sinica, Taipei, Taiwan; 2Institute of Biochemical Sciences, College of Life Science, National Taiwan University, Taipei, Taiwan and 3Institute of Biomedical Sciences, Academia Sinica, Taipei, Taiwan. Correspondence: Dr G-C Chen, Institute of Biological Chemistry, Academia Sinica, 128 Academia Road, Section 2, Taipei 115, Taiwan. E-mail: [email protected] Received 3 April 2014; revised 26 July 2014; accepted 16 August 2014; published online 29 September 2014 PTPN3 negatively regulates EGFR signaling M-Y Li et al 3792 RESULTS the activated form of Drosophila Ras (RasV12) (Figures 1k–m), but Drosophila Ptpmeg is involved in regulating the EGFR signaling not by coexpressing the constitutively active phosphoinositide 3- pathway kinase (Dp110-CAAX) or active Akt (myr-Akt) (Figures 1n–o). EGFR We previously performed genetic analyses to identify non- is known to induce extracellular signal-regulated kinase/MAPK transmembrane PTPs that could modulate border cell migration phosphorylation in future vein regions during wing develop- 21 ment.27 Strikingly, we found that clonal expression of dPtpmeg in during Drosophila oogenesis. Although RNA interference (RNAi)- 28 mediated downregulation of Drosophila non-transmembrane PTPs wing imaginal discs using the flipout/Gal4 system led to a did not have an obvious effect on border cell migration at stage marked reduction of MAPK phosphorylation in GFP-positive 10 egg chambers,21 we found that dptpmeg mutation caused dPtpmeg-expressing cells (Figures 1p, p’). Taken together, these accelerated migration of border cells and the clusters reached data indicate that dPtpmeg acts as a negative regulator of the oocyte prematurely at stage 9 (Figures 1a and b). The receptor EGFR/Ras/MAPK signal pathway. tyrosine kinases, platelet-derived growth factor (PDGF)- and vascular endothelial growth factor (VEGF)-related receptor (PVR) Identification of Eps15 as a substrate of dPtpmeg and EGFR, have been shown to play an important role in guiding 22,23 We have recently established a mass spectrometry (MS)-based migration of the border cells toward the oocyte. As shown in substrate trapping strategy to identify putative substrates of Figures 1c and d, ectopic expression of dPtpmeg with the border 26 fi PTP61F, the smallest non-transmembrane PTP in Drosophila. We cell-speci c Slbo-Gal4 driver impaired border cell motility. More- used the same approach to identify potential substrates of over, clonal analysis revealed elevated phosphotyrosine levels in dPtpmeg. The bacterially expressed wild-type (WT) form and the dptpmeg homozygous mutant border cell clones (marked by the substrate-trapping DA mutant form of dPtpmeg were incubated loss of green fluorescent protein (GFP); Figure 1e, e’), suggesting with Drosophila S2 cell lysates. After immunoprecipitation of that dPtpmeg may antagonize receptor tyrosine kinase-mediated dPtpmeg, the associated proteins were eluted, resolved by sodium tyrosine phosphorylation and border cell migration. dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) It has been reported that EGFR signaling regulates a variety of developmental processes in Drosophila, including wing vein and subjected to liquid chromatography tandem mass spectro- formation.24,25 To investigate the role of dPtpmeg in vein metry analysis. Several proteins, including Eps15, Pvr, Cortactin and Ter94 (mammalian VCP/p97), were identified to interact patterning, we used engrailed-Gal4 (en-Gal4) driver to restrict fi dPtpmeg expression in the posterior compartment of the wing; speci cally with the dPtpmeg-DA mutant (Figure 2a). Interestingly, fi therefore, the anterior part served as control. Interestingly, previous work has identi ed VCP/p97 as a substrate of PTPN3 in 17 fi expression of wild-type dPtpmeg, but not phosphatase-deficient mammalian cells. Moreover, the identi cation of Pvr as a fi mutant (dPtpmeg-CS), caused a wing vein missing phenotype putative substrate of dPtpmeg is consistent with the nding that (Figures 1f–h). The dPtpmeg-induced wing vein defects could be dPtpmeg plays a role in regulating border cell migration rescued by coexpressing dPtpmeg-RNAi (Figure 1i). Moreover, (Figures 1a–d). Here we have focused our study on Eps15. Eps15 ectopic expression of PTP61F,26 the Drosophila homolog of PTP1B, is a multidomain adaptor protein that contains EH domains at the in the developing wing did not affect the normal pattern of veins N-terminus and two ubiquitin-interacting motifs at the C-termi- 20 (Figure 1j). These results together suggest that the wing vein nus. Eps15 associates with components of the AP1 and AP2 defects are specifically caused by dPtpmeg in a phosphatase complexes and plays an important role in clathrin-mediated 20 activity-dependent manner. To determine the relationship endocytosis. between dPtpmeg and EGFR signaling, we examined whether To understand the relationships between dPtpmeg and Eps15, the dPtpmeg-induced wing vein defects could be modulated by we examined whether dPtpmeg genetically interacted with Eps15. components of the EGFR/Ras/mitogen-activated protein kinase Ectopic expression of Eps15 in the developing wing with en-Gal4 (MAPK) signaling pathway. Genetic analysis revealed that the wing caused a wing-notching phenotype. Notably, the Eps15-induced vein missing phenotype caused by dPtpmeg misexpression can be wing defects could be suppressed by coexpression of dPtpmeg suppressed by the gain-of-function mutation EgfrElp, by the (Supplementary Figure S1A), suggesting that dPtpmeg has an reduction of Ras inhibitor RasGAP1 or by the coexpression of antagonist effect on Eps15. The activation of EGFR has been

Figure 1. Drosophila Ptpmeg negatively regulates EGFR signaling. (a) Fascin staining in a control and dPtpmeg mutant egg chamber at stage 9. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; blue) and actin was stained with phalloidin (red). Bar, 50 μm. (b) Quantification of the percentage of stage 9 border cell clusters that prematurely reach the oocyte in the indicated genotypes. Each stage 9 egg chamber was divided into three regions: 100% motility, 450% motility and o50% motility. Border cells of stage 9 egg chambers were scored and the percentage was presented in histogram form (N4100 for each genotype). (c) Control (Slbo-Gal44UAS-mCD8-GFP) and dPtpmeg (Slbo- Gal44UAS-mCD8-GFP/UAS-dPtpmeg) stage 10 egg chambers were stained with DAPI (blue) and phalloidin (red). (d) Quantification of border cell migration of the indicated genotypes in (c). To define defects in border cell migration, each stage 10 egg chamber was divided into three regions: 100% motility, 450% motility and o50% motility. Border cells of stage 10 egg chambers were scored and the percentage was presented in histogram form (N4100 for each genotype). BC, border cell. (e) A marked increase in phospho-Tyrosine (pTyr) levels (arrowhead) in dPtpmeg homozygous mutant border cell clones (GFP-negative cells). The dPtpmeg mutant clone was marked with white dashed lines and the surrounding GFP-positive cells were used as controls. Nuclei were stained with DAPI (blue). N410. Bar, 20 μm. (e’) Line scan across the dPtpmeg mutant clone to show the relative fluorescent intensities of phosphotyrosine (pTyr) in control (GFP-positive) and homozygous dPtpmeg mutant (GFP-negative) cells. (f–h) Compared with wild-type wing (f), ectopic expression of wild-type dPtpmeg (g), but not the phosphatase-deficient mutant dPtpmeg-CS (h), in the posterior compartment of the wing using engrailed-Gal4 (en-Gal4) caused a wing vein missing phenotype. The dashed line indicates the boundary between anterior (a) and posterior (p) compartments. (i) Coexpression of dPtpmeg-RNAi rescued dPtpmeg-induced wing vein defects. (j) Ectopic expression of PTP61F did not affect wing vein pattern. (k–o) The dPtpmeg-induced wing vein defects could be suppressed by EgfrElp (k) and gap1 (l) mutants, or by the coexpression of RasV12 (m), but not by coexpressing the constitutively active phosphoinositide 3-kinase (Dp110-CAAX) (n) or active Akt (myr-Akt) (o). Genetic analyses were performed for three times with 100% penetrance of the phenotype (N4100 for each genotype). (p) Clonal expression of dPtpmeg (GFP- positive cells) in the larval wing imaginal discs resulted in a marked decrease in phospho-MAPK staining (blue) in L3 and L4 proveins. The dPtpmeg-overexpressing clones were marked with white dashed lines and the surrounding GFP-negative cells were used as controls. Actin was stained with tetramethylrhodamine isothiocyanate (TRITC)-labeled phalloidin (red). N410. Bar, 50 μm. (p’) High-magnification view of the area indicated by a white square in (p).

Oncogene (2015) 3791 – 3803 © 2015 Macmillan Publishers Limited PTPN3 negatively regulates EGFR signaling M-Y Li et al 3793 shown to induce Eps15 tyrosine phosphorylation.29 To test coexpressed with Eps15 in S2 cells for substrate trapping. As whether dPtpmeg would recognize Eps15 as a substrate in vivo, shown in Figure 2b, we found a signiﬁcant amount of Eps15 to the full-length WT or DA mutant form of dPtpmeg was associate with the substrate-trapping DA mutant and substantially

© 2015 Macmillan Publishers Limited Oncogene (2015) 3791 – 3803 PTPN3 negatively regulates EGFR signaling M-Y Li et al 3794 less with the WT form of dPtpmeg. Moreover, the tyrosine Figure S1D). As shown in Figures 2f and g, bacterially expressed phosphorylation level of Eps15 was markedly increased in S2 cells Flag-Eps15ΔN is tyrosine phosphorylated and interacts with both treated with double-stranded RNA knocking down dPtpmeg WT and DA forms of GST-PTPN3ΔN. We further confirmed the (Figure 2c). These results suggest that dPtpmeg plays a critical interaction using in vitro translated and phosphorylated Flag- role in regulating the function of Eps15 through direct tyrosine Eps15 and bacterially expressed GST-PTPN3ΔN (Supplementary dephosphorylation. Figures S1E and F). Moreover, the role of PTPN3 on dephosphorylation of Eps15 was examined by in vitro dephosphorylation Eps15 is a substrate of PTPN3 in mammalian cells assay with bacterially expressed Flag-Eps15ΔN (aa 590–897) and GST-PTPN3ΔN. We found that WT, but not the substrate-trapping dPtpmeg is the Drosophila homolog of two closely related Δ mammalian non-transmembrane protein tyrosine phosphatases: mutant (DA), form of GST-PTPN3 N promoted the dephospho- PTPN3 and PTPN4. It has been shown that PTPN4 is highly rylation of pEps15 (Tyr 850) in a dosage-dependent manner expressed in brain and plays an important role in motor learning (Figure 2h). These results together indicate that Eps15 is a and cerebellar plasticity.30,31 On the other hand, PTPN3 has been substrate of mammalian PTPN3. implicated in regulating cell growth and human tumorigenesis.17,32 Here we investigated whether Eps15 is a substrate of PTPN3-mediated Eps15 dephosphorylation promotes EGFR for PTPN3 in mammalian cells. HEK293T cells were transiently lysosomal degradation transfected with hemagglutinin (HA)-tagged wild-type PTPN3 Endocytosis plays an important role in the regulation of EGFR (PTPN3-WT) or substrate-trapping mutant PTPN3 (PTPN3-DA) or a activation and cell signaling.34 Upon EGF stimulation, activated PTP domain deleted PTPN3 (PTPN3-ΔPTP) with the Flag-tagged EGFR can be internalized to various subcellular compartments, Eps15. Immunoblotting of the anti-HA immunoprecipitates from recycling or to lysosome for degradation. Tyrosine phosphoryla- cell lysates revealed that Eps15 co-immunoprecipitated with the tion of Eps15 is known to play an important role in regulating WT and DA mutant forms of PTPN3, but not with PTPN3-ΔPTP EGFR endocytosis.33 To determine the role of PTPN3 in EGFR (Figure 2d). Eps15 has been shown to be phosphorylated at trafficking, we first checked endogenous PTPN3 localization by tyrosine residue 850 following EGFR activation.33 To determine immunofluorescence under basal and EGF-stimulated conditions. whether Y850 is the recognition site of PTPN3, we checked PTPN3 displayed a diffuse cytosolic distribution with scattered whether PTPN3 can interact with the Eps15 tyrosine small puncta at the periphery of H1975 and A549 lung cancer cells phosphorylation-defective mutant (Eps15-Y850F). Interestingly, not treated with EGF (Figure 3a and Supplementary Figure S2A). neither PTPN3-WT nor PTPN3-DA could co-immunoprecipitate When cells were stimulated with EGF for 5 min, PTPN3 became with Eps15-Y850F, suggesting that PTPN3 interacts with Eps15 in a distributed in a distinct punctate pattern that colocalized with phosphorylation-dependent mechanism. We next examined the EGFR. Next, we examined whether PTPN3 could interact and effect of WT and various mutant forms of PTPN3 on the tyrosine dephosphorylate EGFR upon EGF stimulation. As shown in phosphorylation levels of Eps15 with a pEps15 (Y850)-specific Supplementary Figure S1G, EGFR did not co-immunoprecipitate antibody. As shown in Figure 2e, EGF-induced tyrosine phosphor- with either WT or DA mutant form of PTPN3. Moreover, expression ylation of Eps15 was markedly diminished when it was of PTPN3-WT or PTPN3-CS did not cause any significant change of coexpressed with WT but not the catalytically inactive PTPN3 EGFR tyrosine phosphorylation at Y845, Y1045, Y1068, Y1148 (PTPN3-CS) or PTPN3-ΔPTP. We also found that, unlike PTPN3, and Y1173 tyrosine sites (Supplementary Figure S1H). Therefore, expression of WT or the CS mutant form of PTPN23 and Shp2 did PTPN3 is more than likely involved in EGFR signaling not affect EGF-induced Eps15 tyrosine phosphorylation specific through Eps15 in endosomal trafficking. To investigate (Supplementary Figures S1B and C). To determine whether PTPN3 whether PTPN3-mediated Eps15 dephosphorylation would affect can directly interact with Eps15, in vitro pull-down assay was EGFR internalization in response to EGF stimulation, lung cancer performed. As bacterial expression of the full-length Eps15 formed cells (H1975) stably expressing PTPN3 or Eps15 tyrosine insoluble aggregates, we performed glutathione S-transferase phosphorylation-deficient mutant (Y850F) were treated with (GST) pull-down assays with bacterially expressed Flag-Eps15ΔN EGF-Alexa 488 to stimulate EGFR endocytosis. Quantitative (amino acids (aa) 590–897) and GST-PTPN3ΔN (aa 507–909, internalization assay by flow cytometry showed that the rate of containing the PDZ and PTP domain) (Supplementary EGF-Alexa 488 internalization in PTPN3- and Eps15-Y850F-

Figure 2. Identification of Eps15 as a substrate for Drosophila Ptpmeg and human PTPN3. (a) Pervanadate-treated S2 cell lysates were incubated with either the WT or the DA trapping mutant form of HA-tagged dPtpmeg-PTP (catalytic domain of dPtpmeg). After immunoprecipitation using anti-HA antibody, dPtpmeg-PTP and its associated proteins were resolved by SDS–PAGE and stained with Sypro Ruby. The proteins associated with the DA mutant, but not the WT, form of dPtpmeg-PTP were excised for mass spectrometry analysis. The proteins identified in the dPtpmeg-trapping complexes are listed with protein score and numbers of peptides. (b) V5-tagged Eps15 was ectopically coexpressed with the HA-tagged WT form or the DA mutant form of dPtpmeg in S2 cells. After immunoprecipitation using anti-V5 antibody, the immunoprecipitates and total cell lysates (TCLs) were analyzed by immunoblotting with antibodies as indicated. (c) S2 cells treated with GFP double-stranded RNA (dsRNA; control) or dPtpmeg dsRNA were transfected with V5-tagged Eps15. Immunoprecipitated Eps15 was analyzed by immunoblotting with the indicated antibodies. The efficiency of dPtpmeg knockdown was verified by western blotting with dPtpmeg antibody (lower panel). The levels of phosphotyrosine (pTyr) were quantified with ImageJ (NIH, Bethesda, MD, USA) and normalized to the immunoprecipitated Eps15. Data are represented as mean ± s.d. from triplicate experiments. *P = 0.02. (d) HEK293T cells transfected with FLAG-tagged full-length Eps15 or Eps15-Y850F, together with HA-tagged PTPN3 WT, D811A (DA) or ΔPTP (PTP domain deletion mutant), were subjected to immunoprecipitations with anti-HA antibody. The immunoprecipitates and total cell lysates were analyzed by immunoblotting with antibodies as indicated. (e) Lysates of HEK293T cells transfected with FLAG-tagged full-length Eps15 WT or Y850F (YF), together with HA-tagged PTPN3 WT, C842S (CS) or ΔPTP, were immunoprecipitated with anti-FLAG antibody. The immunoprecipitates and cell lysates were analyzed by immunoblotting with antibodies as indicated. (f) Bacterially expressed Flag-Eps15ΔN is tyrosine phosphorylated at Tyr850. Equal amounts of bacterially expressed and purified Flag-Eps15ΔN WT and YF were analyzed by immunoblotting with antibodies as indicated. (g) Bacterially expressed Flag-Eps15ΔN was mixed with equal amounts of GST or GST-PTPN3ΔN WT or DA fusion proteins for in vitro pull-down assays. (h) PTPN3 dephosphorylates Eps15 in vitro. Purified GST (10 μg), GST-PTPN3ΔNDA (10 μg) or different amounts of GST-PTPN3ΔN (2, 5 and 10 μg) were incubated with FLAG-Eps15ΔN WT or YF in a PTPase buffer at 37 °C for 30 min. Numbers below lanes indicate the relative intensities of the pEps15 bands.

Oncogene (2015) 3791 – 3803 © 2015 Macmillan Publishers Limited PTPN3 negatively regulates EGFR signaling M-Y Li et al 3795 expressing cells was comparable to that of control cells (Figure 3b Y850F on the intracellular localization of EGF-Alexa 488 and and Supplementary Table S1). We further performed immuno- endosomal markers (EEA1 and Rab11) and the lysosomal marker ﬂuorescence assays to determine the effect of PTPN3 and Eps15- (LAMP-2). In control cells, we found that 5 min after internalization,

Figure 3. PTPN3-mediated Eps15 dephosphorylation accelerates sorting of EGFR for lysosomal degradation. (a) H1975 cells were treated without or with EGF (100 ng/ml) for 5 min and immunostained with anti-PTPN3 (red) and anti-EGFR (green) antibodies. Nuclei were stained with 4',6-diamidino-2-phenylindole (DAPI) in blue. The insets show a higher magnification of the area enclosed within the white box. Bar, 10 μm. (b) H1975 cells stably expressing HA-tagged PTPN3, FLAG-tagged Eps15-Y850F or an empty vector control were treated with EGF-Alexa 488 (100 ng/ml) for 1 h at 4 °C. Cells were then incubated at 37 °C for 5, 10 and 15 min for the internalization of EGF-Alexa 488. The rate of EGF internalization was determined by flow cytometry as described in the Materials and methods. Data represent the mean ± s.d. of three independent experiments. (c) H1975 cells stably expressing HA-tagged PTPN3, FLAG-tagged Eps15-Y850F or an empty vector control were treated with EGF-Alexa 488 (100 ng/ml) for 5 min. Ectopic expression of PTPN3 or Eps15-Y850F caused the colocalization of EGF-Alexa 488 and LAMP-2, compared with controls. Nuclei were stained with DAPI in blue. The insets show a higher magnification of the area enclosed within the white box. Bar, 10 μm. (d) Percentage colocalization of EGF-Alexa 488 with LAMP-2 in (c). Data are represented as mean ± s.d. of triplicates, with an average of 10 cells scored per experiment. ***Po0.001. (e, f) H1975 cells stably expressing HA-tagged PTPN3, FLAG-tagged Eps15- Y850F or an empty vector control were treated with 100 μM Bafilomycin A1 (Baf-A1) for 60 min, followed by incubation with 100 ng/ml EGF for the indicated times. Cell lysates were analyzed by immunoblotting with antibodies as indicated.

EGF is dispersed in the cytosol and colocalized with the early hepatocyte growth factor stimulation (Supplementary Figure S4), endosomal marker EEA1 (Supplementary Figures S3A and B). suggesting that PTPN3-Eps15 plays a specific role in regulating However, EGF largely colocalized with LAMP-2 in PTPN3- or Eps15- EGFR signaling. We next investigated whether the PTPN3 and Y850F-expressing cells as compared with controls (Figures 3c and Eps15-Y850F-induced EGFR degradation could be blocked by d). Moreover, ectopic expression of PTPN3 in H1975 and A549 cells treating cells with a proteasomal inhibitor (MG132) or a lysosomal showed that WT PTPN3, but not the catalytically inactive mutant inhibitor (bafilomycin A1). Treatment of cells with bafilomycin A1 PTPN3-CS, resulted in a marked decrease in EGF-induced but not MG132 suppressed PTPN3- and Eps15-Y850F-induced phosphorylation and activation of MAPK (Figure 3e and degradation of EGFR (Figures 3e and f and Supplementary Figures Supplementary Figures S2B and S3C), suggesting that, like S5A and B). Together, these results indicate that PTPN3-mediated Drosophila Ptpmeg, PTPN3 negatively regulates the EGFR/MAPK Eps15 dephosphorylation promotes the ligand-bound EGFR for signaling pathway in lung cancer cells. Notably, PTPN3 over- lysosomal degradation and downregulates EGFR signaling. expression also caused a drastic reduction in the EGFR protein levels upon EGF stimulation (Figure 3e). In addition, like PTPN3, expression of Y850F mutation but not wild-type Eps15 caused a PTPN3 promotes EGFR degradation via non-clathrin-mediated similar reduction in the EGFR protein levels as well as EGF-induced endocytosis phosphorylation of MAPK (Figure 3f). Eps15 has also been shown In addition to the clathrin-mediated endocytosis, several studies to regulate the endocytic trafficking of the hepatocyte growth have reported that EGFR can also be internalized via caveolae/ factor receptor Met.35 The hepatocyte growth factor–Met signal- lipid raft-mediated endocytosis, a clathrin-independent path- ing plays a critical role in the growth, invasion and metastasis of way.38 Clathrin-mediated endocytosis has been shown to play a various human cancers.36,37 However, we noticed that PTPN3 and major role for sustained EGFR signaling, whereas non-clathrin- Eps15-Y850F mutant did not affect Met receptor levels after mediated endocytosis was shown to be preferentially for EGFR

Oncogene (2015) 3791 – 3803 © 2015 Macmillan Publishers Limited PTPN3 negatively regulates EGFR signaling M-Y Li et al 3797 degradation,39 although the molecular mechanisms in regulating Together, these results suggest that PTPN3-mediated Eps15 these two internalization processes remained unclear. Interest- dephosphorylation directs EGFR to lipid raft-mediated endocytosis ingly, Eps15 is reported to be involved in both clathrin-dependent for degradation. and -independent endocytosis of the EGFR.38,40 To investigate whether PTPN3-mediated Eps15 dephosphorylation plays a Ectopic expression of PTPN3 and Eps15-Y850F inhibits lung cancer role in partitioning EGFR to non-clathrin-mediated endocytosis growth for degradation, we used confocal microscopy to examine the Lung cancer is the leading cause of cancer-related deaths localization of EGFR, clathrin and caveolin-1 in control, PTPN3- worldwide. Accumulating evidence has shown that EGFR gene is or Eps15-Y850F-expressing H1975 cells. As shown in often amplified or mutated within the tyrosine kinase domain in Supplementary Figure S6, internalized EGF colocalized NSCLC.43,44 Because recent studies indicated an involvement of with clathrin in control but not in PTPN3- or Eps15-Y850F- 45 overexpressing cells (Supplementary Figures S6A and C). Notably, PTPN3 in lung tumorigenesis, we examined the role of PTPN3- internalized EGF colocalized with caveolin-1 in PTPN3- or Eps15- Eps15 in EGFR-dependent lung cancer growth. As shown in Y850F-overexpressing cells, but the controls did not Figures 5a and b, ectopic expression of PTPN3 or Eps15-Y850F (Supplementary Figures S6B and D). To further confirm our markedly inhibited cell proliferation in WST-1 conversion assay observation, we used the sucrose density gradient fractionation and colony formation assay compared with controls. We next approach to quantify the distribution of EGFR between clathrin- tested whether PTPN3 and Eps15-Y850F may have an effect on enriched compartments and caveolin-1 lipid rafts in control, lung cancer cell migration using the in vitro wound-healing and PTPN3- or Eps15-Y850F-expressing cells. The caveolin-enriched transwell migration assays. H1975 cells expressing PTPN3 or membranes (lipid raft) contained high levels of cholesterol and Eps15-Y850F migrated across the wound much slower than that of sphingolipids and migrated in the low-density region (fractions control cells over 24 h (Figure 5c). In the transwell assay, compared 3–5) of the sucrose gradient, whereas clathrin heavy chain was with control cells, there was a significant decrease in the number localized in the high-density non-lipid raft fractions (fractions of PTPN3- or Eps15-Y850F-expressing cells that migrated across 8–10) (Figure 4). In control H1975 cells, EGFR was enriched in the the transwell membrane to the underside of the inserts non-lipid raft fractions (Figure 4a). However, a substantial fraction (Figure 5d). These results indicate that PTPN3 and Eps15-Y850F of EGFR was found in the caveolin-enriched fractions in PTPN3- possess the potential for tumor suppression. To investigate and Eps15-Y850F-overexpressing cells (Figures 4b–d). Finally, we whether PTPN3 and Eps15-Y850F have a growth inhibitory effect examined whether PTPN3- or Eps15-Y850F-induced EGFR degra- on tumor formation and development in vivo, we subcutaneously dation could be affected by treating cells with methyl-β- injected H1975 cells stably expressing PTPN3, Eps15-Y850F or cyclodextrin (MβCD) and filipin that are known to disrupt the vector only controls into athymic nude mice and measured tumor integrity of membrane lipid rafts.41,42 As shown in Figures 4e and f volume over time. As shown in Figure 5e, overexpression of PTPN3 and Supplementary Figures S7A and B, the degradation of EGFR or Eps15-Y850F caused a marked decrease in the tumor growth was dramatically reduced in the presence of MβCD and filipin. rate (Figure 5e). We further examined the expression level of EGFR

Figure 4. PTPN3-mediated Eps15 dephosphorylation promotes EGFR endocytotic trafﬁcking through a non-clathrin pathway. (a–c) Lysates of H1975 cells expressing an empty vector control, PTPN3, or Eps15-Y850F were fractionated by sucrose density gradient centrifugation, and aliquots were immunoblotted with anti-EGFR, anti-clathrin and anti-caveolin-1 antibodies. Fractions 3–5 were enriched with caveolin-1 and denoted as lipid raft fractions, whereas fractions 8–10 were enriched with clathrin and denoted as non-lipid raft fractions. (d) Quantitative analysis for the relative distribution of EGFR in lipid raft fractions and non-lipid raft fractions. Data are represented as mean ± s.d. of triplicates. *P o 0.05. (e, f) H1975 cells stably expressing HA-tagged PTPN3, Eps15-Y850F or an empty vector control were treated with 100 μM MβCD for 60 min, followed by incubation with 100 ng/ml EGF for the indicated times. Cell lysates were analyzed by immunoblotting with antibodies as indicated.

Figure 5. PTPN3 and Eps15-Y850F negatively regulate cell proliferation of NSCLC in vitro and tumor growth in vivo.(a) A time course of cell proliferation by WST-1 assay of H1975 cells expressing PTPN3 (red line), Eps15-Y850F (green line) or vector only control (blue line). Data are represented as mean ± s.d. of triplicates. (b) Colony formation assay of H1975 cells stably expressing PTPN3, Eps15-Y850F or an empty vector control. The colonies were stained with crystal violet and counted. Data are represented as mean ± s.d. of triplicates. (c) H1975 cells stably expressing PTPN3, Eps15-Y850F or an empty vector control were cultured to confluent monolayer, wounded and incubated in growth medium containing 10 μg/ml mitomycin. Cell migration was observed with light microscope at indicated time points. (d) Cell migration was also assessed using in vitro transwell migration assay. Cells in (c) were plated onto the upper well of Transwell Boyden chamber and allowed to migrate for 24 h. Cells that migrated through the filter were stained with crystal violet and quantified in a microplate reader. Data are represented as mean ( ± s.d.) value of migrated cells (n = 3). (e) H1975 cells infected with PTPN3, Eps15-Y850F or vector only control were injected subcutaneously into the right and left sides, respectively, of the flank region of nude mice. Tumor volume was monitored for the indicated times (0, 7, 14, 21 and 28 days). Data are represented as mean ± s.d. of triplicates. (f) Immunohistochemical analysis of EGFR and pEps15 (Y850) in tumor xenografts from (e). Bar, 200 μm. *Po0.05; ***Po0.001.

Oncogene (2015) 3791 – 3803 © 2015 Macmillan Publishers Limited PTPN3 negatively regulates EGFR signaling M-Y Li et al 3799 and phospho-Eps15 in tumors harvested 4 weeks after injection. Eps15 is a multidomain adaptor protein that plays an important Immunohistochemical staining showed that PTPN3 and Eps15- role in regulating endocytic trafficking. In mammalian cells, Y850F drastically reduced the amount of EGFR and phospho- Eps15 can be phosphorylated by EGFR at tyrosine residue 850 Eps15 in excised tumors (Figure 5f). Together, these results upon EGF stimulation.29,33 Confalonieri et al.33 showed that suggest that PTPN3-mediated Eps15 dephosphorylation inhibits ectopic expression of Eps15-Y850F mutant impairs the internaliza- lung cancer formation and progression in vivo. tion of EGFR. On the contrary, using immunofluorescence and flow cytometry, we showed that ectopic expression of PTPN3 and Depletion of PTPN3 impairs EGFR degradation and enhances the Eps15-Y850F did not affect the internalization of EGFR. As we tumorigenic potential of lung cancer cells found a dramatic reduction of EGFR levels in cells expressing PTPN3 and Eps15-Y850F, one explanation for our results being We next examined whether ablation of PTPN3 would enhance the contradictory to previous findings33 might be because of the oncogenic properties of lung cancer cells. Quantitative real- difference in detection sensitivity. In addition to its role in – time PCR analyses revealed higher expression of PTPN3 in CL1-5 endocytic trafficking, Eps15 was reported to localize at the cells compared with other NSCLC cell lines (H520, H1975, A549, trans-Golgi network and regulate vesicle trafficking during the H1299 and H928) (Supplementary Figure S8A). Two different sets secretory process.47 However, immunofluorescence analysis of small hairpin RNAs (shRNAs) that can effectively downregulate of the intracellular distribution of endogenous PTPN3 or the expression of PTPN3 were stably expressed in CL1-5 human exogenously expressed HA-tagged PTPN3 indicated no signi- lung cancer cells (Supplementary Figure S8B). Intriguingly, ficant localization at the trans-Golgi network (data not shown), immunoblot analysis showed that depletion of PTPN3 caused an and this might suggest that PTPN3 is not involved in increase in the level of EGFR and EGFR-mediated MAPK Eps15-mediated protein sorting and vesicle trafficking at the phosphorylation (Figure 6a). After 24 h of EGF stimulation, the trans-Golgi network. level of EGFR expression decreased in scrambled shRNA knock- The ligand-activated EGFR and the transforming growth factor- down control cells (Figure 6b), but not in PTPN3 knockdown cells. β receptor have been reported to be endocytosed through a Consistent with our earlier observations, shRNA-mediated knock- clathrin-dependent as well as a clathrin-independent path- down of PTPN3 caused a marked increase of Eps15 p-Tyr 850 way.39,48 Segregation of these cell surface receptors through levels but there was no significant change of EGFR tyrosine distinct endocytic pathways is known to regulate downstream phosphorylation at Y845, Y1045, Y1068, Y1148 and Y1173 sites signal duration and receptor trafficking, although it is unclear how (Figures 6c and d). We further checked whether PTPN3 depletion it does this. Several lines of evidence indicate that PTPN3 and might affect EGFR endocytic trafficking by immunofluorescence Eps15-Y850F accelerate downregulation of EGFR via a clathrin- assay. At 5 min after stimulation, depletion of PTPN3 did not affect independent but lipid raft-dependent pathway. First, EGF-488 EGFR distribution or the endocytic pathway, compared with trafficking assay revealed that EGF-488 was largely colocalized controls (Figure 6e and Supplementary Figures S8C and D). After with lipid raft-associated protein caveolin-1 but not with clathrin 60 min of stimulation, EGF-Alexa 488 largely colocalized with in cells expressing PTPN3 or Eps15-Y850F (Supplementary LAMP-2 in control cells (Figure 6f). However, EGF was dispersed in Figure S6). Second, analysis of EGFR profile by sucrose the cytoplasm and did not significantly colocalize with LAMP-2 in gradient fractionation showed that EGFR was concentrated in PTPN3 knockdown cells. Moreover, we found that PTPN3 knock- clathrin-enriched non-lipid fractions in control cells (Figure 4a). down resulted in a substantial increase in CL1-5 lung cancer cell However, overexpression of PTPN3 and Eps15-Y850F led to a proliferation and cell growth, as observed by WST-1 and colony redistribution of EGFR to caveolin-1-enriched lipid raft fractions – β fi formation assay, respectively (Figures 7a and b). The in vivo effect (Figures 4b d). Third, disruption of lipid rafts with M CD and lipin of PTPN3 on tumor growth was examined in a xenograft model in suppressed PTPN3- or Eps15-Y850F-induced EGFR degradation nude mice. CL1-5 cells expressing shPTPN3 or a scrambled shRNA (Figures 4e and f). How does PTPN3-mediated tyrosine depho- control were subcutaneously injected into athymic nude mice and sphorylation of Eps15 regulate EGFR for lipid raft-dependent endocytosis and lysosomal degradation? Accumulating evidence tumor volume was measured over time. Compared with controls, has shown that EGFR ubiquitination is not essential for its depletion of PTPN3 significantly increased the tumor growth rate internalization, but appears to play an important role in (Figure 7c). Immunohistochemical staining showed that depletion endosomal sorting and lysosomal targeting of the receptor.49,50 of PTPN3 led to an increase in EGFR and phospho-Eps15 levels in One possibility is that tyrosine dephosphorylation of Eps15 by excised tumors (Figure 7d). Considered together, these results PTPN3 may affect the ubiquitination status of EGFR, accelerating demonstrate that PTPN3 indeed acts as a negative regulator of EGFR for lysosomal degradation. Recently, Roxrud et al.51 have EGFR signaling and EGFR-mediated lung cancer growth. identified an endosomally localized Eps15 isoform (Eps15b) that interacts with the Hrs (hepatocyte growth factor–regulated DISCUSSION tyrosine kinase substrate) complex to mediate EGFR degradation. We found that like Eps15, Eps15b is also a substrate of PTPN3 The EGFR belongs to the ErbB family of protein tyrosine kinases (M-Y Li and G-C Chen, unpublished data), indicating a dual role for and is a major regulator for both normal development and cancer fi 46 PTPN3 in the regulation of endocytic traf cking and endosomal progression. PTPs, which include receptor-like PTPs and sorting of EGFR. nonreceptor PTPs, are a group of tightly regulated enzymes Accumulating evidence has indicated that PTPs can function as thought to regulate tyrosine phosphorylation by antagonizing the 1,10 tumor suppressors or oncogenes depending on the substrate action protein tyrosine kinases. In this study, we provide the involved and the cellular context.32,52 It has been reported that first evidence that PTPN3 inhibits the EGFR signaling by targeting PTPN3 expresses in gastric cancer cells and may play a role in the receptor for lysosomal degradation. In Drosophila, dPtpmeg gastric cancer progression and differentiation53 PTPN3 has been antagonizes receptor tyrosine kinase activity and plays a role in found to coordinate with p38γ MAPK to promote Ras oncogenesis controlling border cell migration during oogenesis. Moreover, in colon cancer,54 and it has been found to stimulate breast cancer dPtpmeg negatively regulates the EGFR/Ras/MAPK pathway growth by inducing and stabilizing the protein expression of during wing morphogenesis. Substrate-trapping and biochemical vitamin D receptor.55 Interestingly, recent studies have also analysis further identified Eps15 as a substrate for dPTPmeg and indicated that PTPN3 plays a role in tumor suppression. Mutational PTPN3. Our results demonstrate that PTPN3 dephosphorylates analysis of the tyrosine phosphatome found PTPN3 along with five Eps15 and promotes EGFR for degradation in lung cancer cells. other PTPs (PTPRF, PTPRG, PTPRT, PTPN13 and PTPN14) to be

© 2015 Macmillan Publishers Limited Oncogene (2015) 3791 – 3803 PTPN3 negatively regulates EGFR signaling M-Y Li et al 3800 mutated in colorectal cancer.56 Moreover, Jung et al.45 analyzed data support a model in which PTPN3-mediated tyrosine depho- the transcriptome of two NSCLC cell lines and found that one sphorylation of Eps15 leads to EGFR degradation and tumor allele of PTPN3 is mutated in the NSCLC cell line H2228. They suppression in NSCLC cells. This study demonstrated that PTPN3 further showed that ectopic expression of PTPN3 inhibits and Eps15-Y850F overexpression reduced EGFR protein levels and the growth of NSCLC cells, although the molecular mecha- impeded the proliferation and migration of NSCLC cells. Moreover, nisms underlying the growth inhibition remain unknown. Our PTPN3 and Eps15-Y850F signiﬁcantly suppressed NSCLC tumor

Oncogene (2015) 3791 – 3803 © 2015 Macmillan Publishers Limited PTPN3 negatively regulates EGFR signaling M-Y Li et al 3801 growth in a subcutaneous xenograft model. Conversely, depletion help contribute to the development of a targeting intervention of PTPN3 enhanced EGFR stabilization and promoted NSCLC in NSCLC. tumorigenicity both in vitro and in vivo. Our findings are consistent with the idea that PTPN3 acts as a tumor suppressor in NSCLC. In conclusion, we have identified Eps15 as an evolutionarily MATERIALS AND METHODS conserved dPtpmeg/PTPN3 substrate that regulates EGFR signal- For additional information concerning Drosophila genetics, immunofluor- ing. The finding that PTPN3-mediated tyrosine dephosphorylation escence staining and cell migration assays, please refer to Supplementary of Eps15 modulates EGFR-dependent cancer progression may Information.

Figure 7. Depletion of PTPN3 promotes cell proliferation of NSCLC in vitro and tumor growth in vivo.(a) A time course of cell proliferation by WST-1 assay of CL1-5 cells infected with scramble (blue line), shPTPN3-A1 (green line) or shPTPN3-D1 (red line). (b) Colony formation assay of CL1-5 cells infected with scramble control or shPTPN3 (A1 and D1). The colonies were stained with crystal violet and counted. Data are represented as mean ± s.d. of triplicates. (c) CL1-5 cells infected with scramble control or shPTPN3-A1 were subcutaneously injected into the right and left ﬂank region of athymic nude mice. Tumor volume was monitored for the indicated times (0, 7, 14 and 21 days). Data are represented as mean ± s.d. of triplicates. (d) Immunohistochemical analysis of EGFR and phospho-Eps15 (Y850) in tumor xenografts from (c). Bar, 200 μm. *Po0.05; ***Po0.001.

Figure 6. Depletion of PTPN3 leads to an impairment of EGFR degradation. (a) CL1-5 cells infected with scramble control, shPTPN3-A1 or shPTPN3-D1 were treated with EGF (100 ng/ml) for the indicated times (0, 5, 15 and 30 min). Cell lysates were analyzed by immunoblotting with antibodies as indicated. (b) CL1-5 cells infected with scramble control, shPTPN3-A1 or shPTPN3-D1 were treated with cycloheximide and EGF (100 ng/ml) for the indicated times (0, 6, 12 and 24 h). Cell lysates were analyzed by immunoblotting with antibodies as indicated. (c) CL1- 5 cells infected with scramble control, shPTPN3-A1 or shPTPN3-D1 were stimulated with EGF (100 ng/ml) for 5 min and cell lysates were analyzed by immunoblotting with antibodies as indicated. (d) CL1-5 cells infected with scramble control, shPTPN3-A1 or shPTPN3-D1 were stimulated with EGF (100 ng/ml) for 5 min and cell lysates were immunoprecipitated with anti-EGFR antibody. The immunoprecipitates were analyzed by immunoblotting with pEGFR antibodies. (e, f) CL1-5 cells infected with scramble control, shPTPN3-A1 or shPTPN3-D1 were treated with EGF-Alexa 488 (100 ng/ml) for 5 min (e) or 60 min (f) and immunostained with anti-LAMP-2 antibody. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) in blue. The insets show a higher magniﬁcation of the area enclosed within the white box. Bar, 10 μm. Percentage colocalization of EGF-Alexa 488 with LAMP-2 is shown in the bottom panels. Data are represented as mean ± s.d. of triplicates, with an average of 10 cells scored per experiment. ***Po0.001.

Plasmids and antibodies 0.5% Triton X-100, 1 mM PMSF and protease inhibitor cocktail (Roche)) and Human PTPN3 complementary DNA was obtained from Open Biosystems homogenized by passing 10 times through a 27-gauge needle. The (Thermo, Waltham, MA, USA). HA-tagged PTPN3 WT, DA (D811A), CS supernatant was mixed with 1 ml of 80% sucrose in MES, and overlaid with (C842S) and ΔPTP (PTP domain deletion mutant, lacking residues 611 to 5 ml of 35% sucrose in MES followed by 3 ml of 5% sucrose in MES. The 913) were generated by PCR and subcloned into pcDNA and pSin-EF2-Pur samples were then centrifuged at 35 000 r.p.m. in a SW41Ti rotor lentiviral vector. Flag-Eps15-Wt and Flag-Eps15-Y850F were kind gifts from (Beckman, Brea, CA, USA) for 20 h at 4 °C. The sucrose gradients were Pier Paolo Di Fiore (IEO- European Institute of Oncology, Milan, Italy). The harvested in 1 ml fractions from the top of the gradient, and individual lentiviral shRNA clones used to knock down human PTPN3 were obtained fractions were analyzed by western blotting. from the National RNAi Core Facility of Academia Sinica (Taipei, Taiwan). The targeted sequences for these clones are PTPN3 shRNA #A: (clone ID: TRCN0000002788) 5′-CGTGTGTATGAAGAAGGTTTA-3′ and PTPN3 shRNA Xenograft tumorigenicity assay 59 #D: (clone ID: TRCN0000320903) 5′-CCAAGAGAGTTTATCCGAGAA-3′. Xenograft tumorigenicity assay was performed as previously described. Scramble shRNA was used as a control (Addgene, Cambridge, MA, USA). Briefly, virus-infected H1975 and CL1-5 cells were harvested, washed with Lentiviral production and infection were performed as previously phosphate-buffered saline and resuspended in RPMI-1640 medium. Cells described.57 Antibodies used for the study were: anti-pMAPK (Sigma, St (1 × 106) were then injected subcutaneously into the right and left side, Louis, MO, USA), anti-MAPK (Sigma), anti-LAMP-2 (Abcam, Cambridge, UK), respectively, of the flank region of 8-week-old male BALB/c nude mice anti-EEA1 (Cell Signaling, Danvers, MA, USA), anti-LAMP1 (Abcam), anti-Met (Rodent Model Resource Center, Taipei, Taiwan). Tumors were measured (Abcam), anti-EGFR and anti-phospho EGFR antibodies (Cell Signaling), with calipers every 7 days after injection. All mice were killed 28 days after anti-fascin (Developmental Studies Hybridoma Bank, Iowa City, IA, USA). injection, and tumors were surgically excised, weighed and photographed. Anti-PTPN3 antibody was a generous gift from Nicholas Tonks (Cold Spring Differences in tumor progression were statistically analyzed using Harbor Laboratory, Cold Spring Harbor, NY, USA). Anti-phospho Eps15 Student’s t-test. (Y850) antibody was generated by immunizing rabbits with a synthetic phosphopeptide (FANFSApYPSEEDMIC, bovine serum albumin coupled) corresponding to residues surrounding Y850 of mouse Eps15. Anti- Statistical analysis dPtpmeg antibody was generated by immunizing rabbits with a synthetic All experiments have been repeated for at least three times. Statistical peptide (RKPANAPKNRYRDISPYDC, ovalbumin coupled) corresponding to analysis was performed by Student’s t-test. Differences were considered residues of Drosophila dPtpmeg. significant if P-values were o0.05 (*), 0.01 (**) and 0.001 (***).

Cell culture, immunoprecipitation and immunoblotting CONFLICT OF INTEREST Drosophila S2 cells were cultured at 25 °C in Schneider’s Drosophila fl medium (Invitrogen, Carlsbad, CA, USA) containing 10% fetal bovine serum The authors declare no con ict of interest. and 1 × penicillin/streptomycin antibiotics (Invitrogen). HEK293T cells were cultured at 37 °C in DMEM (Invitrogen) supplemented with 10% fetal ACKNOWLEDGEMENTS bovine serum and antibiotics. H1975 and CL1-5 lung cancer cells were grown in RPMI-1640 (Invitrogen) with 10% fetal bovine serum. For We thank Drs J Borst, PP Di Fiore, M McNiven, N Tonks, H Sun, KE Chen, the immunoprecipitations, cells transiently transfected with the indicated Bloomington Stock Center and Fly Core Taiwan for reagents. We are grateful to plasmids were scraped from dishes in lysis buffer (50 mM Tris–HCl, pH 7.4, Dr J Settleman for helpful comments on the manuscript, Dr C-C Hung for confocal 150 mM NaCl, 1% NP-40, 0.25% Na-deoxycholate, 1 mM EDTA, 10 mM NaF, microscopy assistance and Y-L Huang for peptide synthesis. This work was supported 2mM Na3VO4,1mM phenylmethylsulfonyl fluoride (PMSF) and protease by the National Science Council of Taiwan (NSC102-2311-B-001-027-MY3) and inhibitor cocktail (Roche, Indianapolis, IN, USA)). Cell lysates were Academia Sinica. immunoprecipitated with anti-FLAG or anti-V5 antibody and protein G-Sepharose beads (GE Healthcare, Piscataway, NJ, USA) at 4 °C for 3 h. These beads were washed three times with the lysis buffer. After resolution AUTHOR CONTRIBUTIONS by SDS–PAGE, the immunoprecipitates were subjected to western blot G-CC, M-YL and C-WW conceived and designed the experiments. M-YL, Y-TC, analysis. P-LL, A-PC, Y-ZM and G-DC conducted the experiments. G-CC, M-YL, K-HK and T-CM analyzed the data. G-CC and M-YL wrote the paper. In vitro pull-down assays GST-tagged WT or DA form of PTPN3ΔN (aa 507–909) (gift of Kai-En Chen, Academia Sinica, Taipei, Taiwan) and GST-tagged WT or YF form of FLAG- REFERENCES Δ – fi Eps15 N (aa 590 897) were puri ed from BL21 bacteria and eluted from 1 Tonks NK. Protein tyrosine phosphatases: from genes, to function, to disease. Nat Glutathione-Sepharose 4B beads (GE Healthcare) as suggested by the 7 – Δ Rev Mol Cell Biol 2006; :833 846. manufacturer. To remove the GST tag, GST-FLAG-Eps15 N was incubated 2 Hunter T. Tyrosine phosphorylation: thirty years and counting. Curr Opin Cell Biol with PreScission Protease (GE Healthcare) at 4 °C for 4 h. For binding 2009; 21:140–146. assays, FLAG-Eps15ΔN was incubated with equivalent amounts of GST or 3 Bogdan S, Klambt C. Epidermal growth factor receptor signaling. Curr Biol 2001; GST-PTPN3ΔN fusion proteins and immobilized on glutathione beads. 11: R292–R295. Following incubation, bead-bound proteins were separated with SDS– 4 Citri A, Yarden Y. EGF-ERBB signalling: towards the systems level. Nat Rev Mol Cell PAGE and analyzed by western blot analyses. Biol 2006; 7: 505–516. 5 Sorkin A, Goh LK. Endocytosis and intracellular trafficking of ErbBs. Exp Cell Res In vitro dephosphorylation assay 2009; 315:683–696. Bacterially expressed FLAG-Eps15ΔN was incubated with GST and GST- 6 Sebastian S, Settleman J, Reshkin SJ, Azzariti A, Bellizzi A, Paradiso A. The com- PTPN3ΔN WT or GST-PTPN3ΔN DA proteins at 37 °C for 30 min with in vitro plexity of targeting EGFR signalling in cancer: from expression to turnover. Bio- 1766 – PTPase buffer (25 mM Tris-HCl, pH 7.5, 2.5 mM EDTA, 1 mM EGTA, 5 mM chim Biophysic Acta 2006; :120 139. dithiothreitol, 1 mg/ml bovine serum albumin, 1 mM PMSF and protease 7 Tarcic G, Boguslavsky SK, Wakim J, Kiuchi T, Liu A, Reinitz F et al. An unbiased inhibitor cocktail (Roche)). Reactions were stopped by adding sample screen identifies DEP-1 tumor suppressor as a phosphatase controlling EGFR buffer and boiling for 5 min. Samples were resolved by SDS–PAGE and endocytosis. Curr Biol 2009; 19: 1788–1798. subjected to western blot analysis. 8 Yuan T, Wang Y, Zhao ZJ, Gu H. Protein-tyrosine phosphatase PTPN9 negatively regulates ErbB2 and epidermal growth factor receptor signaling in breast cancer cells. J Biol Chem 2010; 285: 14861–14870. Preparation of lipid raft membranes 9 Xu Y, Tan LJ, Grachtchouk V, Voorhees JJ, Fisher GJ. Receptor-type protein-tyr- Lipid raft isolation was carried out by sucrose density gradient osine phosphatase-kappa regulates epidermal growth factor receptor function. ultracentrifugation according to the method previously described.58 Briefly, J Biol Chem 2005; 280: 42694–42700. harvested cells were resuspended in 1 ml of ice-cold 2-(N-morpholino) 10 Tiganis T. Protein tyrosine phosphatases: dephosphorylating the epidermal ethanesulfonic acid (MES) buffer saline (25 mM MES, pH 6.0, 150 mM NaCl, growth factor receptor. IUBMB Life 2002; 53:3–14.

Oncogene (2015) 3791 – 3803 © 2015 Macmillan Publishers Limited PTPN3 negatively regulates EGFR signaling M-Y Li et al 3803 11 Sun T, Aceto N, Meerbrey KL, Kessler JD, Zhou C, Migliaccio I et al. Activation of 36 Birchmeier C, Birchmeier W, Gherardi E, Vande Woude GF. Met, metastasis, multiple proto-oncogenic tyrosine kinases in breast cancer via loss of the PTPN12 motility and more. Nat Rev Mol Cell Biol 2003; 4:915–925. phosphatase. Cell 2011; 144:703–718. 37 Gherardi E, Birchmeier W, Birchmeier C, Vande Woude G. Targeting MET in cancer: 12 Eden ER, White IJ, Futter CE. Down-regulation of epidermal growth factor rationale and progress. Nat Rev Cancer 2012; 12:89–103. receptor signalling within multivesicular bodies. Biochem Soc Trans 2009; 37(Pt 1): 38 Sigismund S, Woelk T, Puri C, Maspero E, Tacchetti C, Transidico P et al. Clathrin- 173–177. independent endocytosis of ubiquitinated cargos. Proc Natl Acad Sci USA 2005; 13 Tonks NK. PTP1B: from the sidelines to the front lines! FEBS Lett 2003; 546: 102: 2760–2765. 140–148. 39 Sigismund S, Argenzio E, Tosoni D, Cavallaro E, Polo S, Di Fiore PP. Clathrin- 14 Eden ER, White IJ, Tsapara A, Futter CE. Membrane contacts between endosomes mediated internalization is essential for sustained EGFR signaling but dispensable and ER provide sites for PTP1B-epidermal growth factor receptor interaction. Nat for degradation. Dev Cell 2008; 15:209–219. Cell Biol 2010; 12: 267–272. 40 Benmerah A, Bayrou M, Cerf-Bensussan N, Dautry-Varsat A. Inhibition of clathrin- 15 Woodman P. ESCRT proteins, endosome organization and mitogenic receptor coated pit assembly by an Eps15 mutant. J Cell Sci 1999; 112(Pt 9): 1303–1311. down-regulation. Biochem Soc Trans 2009; 37(Pt 1): 146–150. 41 Orlandi PA, Fishman PH. Filipin-dependent inhibition of cholera toxin: evidence 16 Ali N, Zhang L, Taylor S, Mironov A, Urbe S, Woodman P. Recruitment of UBPY and for toxin internalization and activation through caveolae-like domains. J Cell Biol ESCRT exchange drive HD-PTP-dependent sorting of EGFR to the MVB. Curr Biol 1998; 141: 905–915. 2013; 23: 453–461. 42 Parpal S, Karlsson M, Thorn H, Stralfors P. Cholesterol depletion disrupts caveolae 17 Zhang SH, Liu J, Kobayashi R, Tonks NK. Identification of the cell cycle regulator and insulin receptor signaling for metabolic control via insulin receptor substrate- VCP (p97/CDC48) as a substrate of the band 4.1-related protein-tyrosine 1, but not for mitogen-activated protein kinase control. J Biol Chem 2001; 276: phosphatase PTPH1. J Biol Chem 1999; 274: 17806–17812. 9670–9678. 18 Gu M, Meng K, Majerus PW. The effect of overexpression of the protein tyrosine 43 Sharma SV, Bell DW, Settleman J, Haber DA. Epidermal growth factor receptor phosphatase PTPMEG on cell growth and on colony formation in soft agar in mutations in lung cancer. Nat Rev Cancer 2007; 7:169–181. COS-7 cells. Proc Natl Acad Sci USA 1996; 93: 12980–12985. 44 Nomura M, Shigematsu H, Li L, Suzuki M, Takahashi T, Estess P et al. Poly- 19 Whited JL, Robichaux MB, Yang JC, Garrity PA. Ptpmeg is required for the proper morphisms, mutations, and amplification of the EGFR gene in non-small cell lung establishment and maintenance of axon projections in the central brain of Dro- cancers. PLoS Med 2007; 4: e125. sophila. Development 2007; 134:43–53. 45 Jung Y, Kim P, Keum J, Kim SN, Choi YS, Do IG et al. Discovery of ALK-PTPN3 gene 20 van Bergen En Henegouwen PM. Eps15: a multifunctional adaptor protein reg- fusion from human non-small cell lung carcinoma cell line using next generation ulating intracellular trafficking. Cell Commun Signal 2009; 7: 24. RNA sequencing. Genes Chromosomes Cancer 2012; 51: 590–597. 21 Chen DY, Li MY, Wu SY, Lin YL, Tsai SP, Lai PL et al. The Bro1-domain-containing 46 Sibilia M, Kroismayr R, Lichtenberger BM, Natarajan A, Hecking M, Holcmann M. protein Myopic/HDPTP coordinates with Rab4 to regulate cell adhesion and The epidermal growth factor receptor: from development to tumorigenesis. Dif- migration. J Cell Sci 2012; 125(Pt 20): 4841–4852. ferentiation 2007; 75: 770–787. 22 Duchek P, Somogyi K, Jekely G, Beccari S, Rorth P. Guidance of cell migration by 47 Chi S, Cao H, Chen J, McNiven MA. Eps15 mediates vesicle trafficking from the the Drosophila PDGF/VEGF receptor. Cell 2001; 107:17–26. trans-Golgi network via an interaction with the clathrin adaptor AP-1. Mol Biol Cell 23 Duchek P, Rorth P. Guidance of cell migration by EGF receptor signaling during 2008; 19:3564–3575. Drosophila oogenesis. Science 2001; 291: 131–133. 48 Di Guglielmo GM, Le Roy C, Goodfellow AF, Wrana JL. Distinct endocytic pathways 24 Martin-Blanco E, Roch F, Noll E, Baonza A, Duffy JB, Perrimon N. A temporal switch regulate TGF-beta receptor signalling and turnover. Nat Cell Biol 2003; 5:410–421. in DER signaling controls the specification and differentiation of veins and 49 Bonifacino JS, Traub LM. Signals for sorting of transmembrane proteins to interveins in the Drosophila wing. Development 1999; 126: 5739–5747. endosomes and lysosomes. Annu Rev Biochem 2003; 72: 395–447. 25 Shilo BZ. Signaling by the Drosophila epidermal growth factor receptor pathway 50 Huang F, Goh LK, Sorkin A. EGF receptor ubiquitination is not necessary for its during development. Exp Cell Res 2003; 284: 140–149. internalization. Proc Natl Acad Sci USA 2007; 104: 16904–16909. 26 Ku HY, Wu CL, Rabinow L, Chen GC, Meng TC. Organization of F-actin via con- 51 Roxrud I, Raiborg C, Pedersen NM, Stang E, Stenmark H. An endosomally localized certed regulation of Kette by PTP61F and dAbl. Mol Cell Biol 2009; 29: 3623–3632. isoform of Eps15 interacts with Hrs to mediate degradation of epidermal growth 27 Gabay L, Seger R, Shilo BZ. In situ activation pattern of Drosophila EGF receptor factor receptor. J Cell Biol 2008; 180: 1205–1218. pathway during development. Science 1997; 277: 1103–1106. 52 Lessard L, Stuible M, Tremblay ML. The two faces of PTP1B in cancer. Biochim 28 Pignoni F, Zipursky SL. Induction of Drosophila eye development by dec- Biophysica Acta 2010; 1804:613–619. apentaplegic. Development 1997; 124:271–278. 53 Wu CW, Chen JH, Kao HL, Li AF, Lai CH, Chi CW et al. PTPN3 and PTPN4 tyrosine 29 Fazioli F, Minichiello L, Matoskova B, Wong WT, Di Fiore PP. eps15, a novel tyr- phosphatase expression in human gastric adenocarcinoma. Anticancer Res 2006; osine kinase substrate, exhibits transforming activity. Mol Cell Biol 1993; 13: 26: 1643–1649. 5814–5828. 54 Hou SW, Zhi HY, Pohl N, Loesch M, Qi XM, Li RS et al. PTPH1 dephosphorylates 30 Hironaka K, Umemori H, Tezuka T, Mishina M, Yamamoto T. The protein-tyrosine and cooperates with p38gamma MAPK to increase ras oncogenesis through PDZ- phosphatase PTPMEG interacts with glutamate receptor delta 2 and epsilon mediated interaction. Cancer Res 2010; 70: 2901–2910. subunits. J Biol Chem 2000; 275: 16167–16173. 55 Zhi HY, Hou SW, Li RS, Basir Z, Xiang Q, Szabo A et al. PTPH1 cooperates with 31 Kina S, Tezuka T, Kusakawa S, Kishimoto Y, Kakizawa S, Hashimoto K et al. vitamin D receptor to stimulate breast cancer growth through their mutual sta- Involvement of protein-tyrosine phosphatase PTPMEG in motor learning and bilization. Oncogene 2011; 30: 1706–1715. cerebellar long-term depression. Eur J Neurosci 2007; 26: 2269–2278. 56 Wang Z, Shen D, Parsons DW, Bardelli A, Sager J, Szabo S et al. Mutational analysis 32 Labbe DP, Hardy S, Tremblay ML. Protein tyrosine phosphatases in cancer: friends of the tyrosine phosphatome in colorectal cancers. Science 2004; 304: 1164–1166. and foes!. Prog Mol Biol Transl Sci 2012; 106: 253–306. 57 Hong CF, Chou YT, Lin YS, Wu CW. MAD2B, a novel TCF4-binding protein, mod- 33 Confalonieri S, Salcini AE, Puri C, Tacchetti C, Di Fiore PP. Tyrosine phosphoryla- ulates TCF4-mediated epithelial-mesenchymal transdifferentiation. J Biol Chem tion of Eps15 is required for ligand-regulated, but not constitutive, endocytosis. 2009; 284: 19613–19622. J Cell Biol 2000; 150:905–912. 58 Hawash IY, Kesavan KP, Magee AI, Geahlen RL, Harrison ML. The Lck SH3 domain 34 Grandal MV, Madshus IH. Epidermal growth factor receptor and cancer: control of negatively regulates localization to lipid rafts through an interaction with c-Cbl. oncogenic signalling by endocytosis. J Cell Mol Med 2008; 12: 1527–1534. J Biol Chem 2002; 277: 5683–5691. 35 Parachoniak CA, Park M. Distinct recruitment of Eps15 via Its coiled-coil domain is 59 Chou YT, Lin HH, Lien YC, Wang YH, Hong CF, Kao YR et al. EGFR promotes lung required for efficient down-regulation of the met receptor tyrosine kinase. J Biol tumorigenesis by activating miR-7 through a Ras/ERK/Myc pathway that targets Chem 2009; 284: 8382–8394. the Ets2 transcriptional repressor ERF. Cancer Res 2010; 70: 8822–8831.

Supplementary Information accompanies this paper on the Oncogene website (http://www.nature.com/onc)