PTEN modulates EGFR late endocytic trafficking and degradation by dephosphorylating Rab7

Abstract

Tumour suppressor phosphatase and tensin homologue deleted on chromosome 10 (PTEN) is a lipid phosphatase that negatively regulates growth factor-induced survival signalling. Here, we demonstrate that PTEN attenuates epidermal growth factor receptor (EGFR) signalling by promoting late endosome maturation by virtue of its protein phosphatase activity. Loss of PTEN impairs the transition of ligand-bound EGFR from early to late endosomes. We unveil Rab7, a critical GTPase for endosome maturation, as a functional PTEN interacting partner. PTEN dephosphorylates Rab7 on two conserved residues S72 and Y183, which are necessary for GDP dissociation inhibitor (GDI)-dependent recruitment of Rab7 on to late endosomes and subsequent maturation. Thus, our findings reveal PTEN-dependent endosome maturation through phosphoregulation of Rab7 as an important route of controlling EGFR signalling.

Figure 1. Loss of PTEN delays EGFR transport from early to late endosomes.

(a) Serum-starved control or PTEN shRNA-transduced HeLa cells were pulsed with alexa Flour 647-conjugated EGF (100 ng ml⁻¹) for 5 min. Cells fixed at different indicated times were imaged using confocal microscope after co-staining with antibodies against EEA1. Scale bar, 10 μm. (b) EGF-EEA1 co-localization in control and PTEN shRNA cells at different times after EGF addition was analysed by Manders method of pixel intensity correlation measurements using Image J/Fiji-Coloc2 plugin. Error bars indicate s.d. (n=100 cells for each time point from three independent experiments), *P<0.05 by Student's t-test. (c) The presence of EGF in late endosomes was analysed by co-staining with CD63 antibodies (scale bar, 10 μm) and (d) the extent of their co-localization in control and PTEN shRNA cells at 30 and 60 min post EGF addition was plotted (n=100 cells for each time point), Error bars indicate s.d. *P<0.05 by Student's t-test. (e) Overnight serum-starved control or PTEN shRNA transduced cells were treated with cycloheximide (50 μg ml⁻¹) for 1 h and stimulated with 100 ng ml⁻¹ EGF for indicated times. EGFR levels were tested by immunoblotting using specific antibody. (f) PTEN-deficient MDA-MB-468 breast cancer cells were transfected with PTEN WT (wild type), C124S (lipid and protein phosphatase dead), G129E (protein phosphatase active, lipid phosphatase dead) and Y138F (protein phosphatase dead, lipid phosphatase active) mutants. The EGFR levels at various times of post-EGF ligand addition is quantified using ImageJ and the levels of undegraded EGFR at each time point was calculated and plotted. Data plotted from three independent experiments; Error bars indicate s.d. *P<0.05, by Student's t-test.

Figure 2. Rab7 is PTEN-associated protein.

(a) HEK 293T cells transfected with the SFB-tagged Rab7 or (b) with SFB-tagged PTEN construct was subjected to immunoprecipitation (IP) with either control IgG or Flag antibody and the interaction of endogenous PTEN and Rab7 was determined by western blotting (WB) with their specific antibodies, respectively. (c) 293T cells were transfected with triple-tagged SFB-Rab5 or SFB-Rab7 and their interaction with PTEN was detected by immunoblotting with PTEN-specific antibody after immunoprecipitating with streptavidin (SBP) beads. (d) HEK293T cell lysates expressing SFB Rab7 WT or T22N (a dominant negative GDP bound) and Q67L (constitutively GTP-bound active) mutants were incubated with bacterially purified GST PTEN. The association of PTEN with Rab7 and its mutants was detected by immunoblotting with PTEN antibody. (e) SFB-Rab7 expressing HEK293T cell lysates pre-loaded either with GDPβS or GTPγS was incubated with glutathione–sepharose bound bacterially purified GST-PTEN and their binding was analysed by western blotting with Flag antibody. (f) GST Rab7 was loaded with either GTPγS or GDPβS in vitro. The association of PTEN with Rab7 was detected by immunoblotting with PTEN antibody after passing the 293T cell lysate through pre-loaded recombinant Rab7. (g) Agarose beads immobilized with bacterially expressed recombinant MBP-PTEN was incubated with either GST or GST-Rab7 proteins expressed in bacteria in the presence of GDP. The direct association of Rab7 with PTEN was detected by immunoblotting with GST antibody. Expression of the recombinant GST Rab7 and MBP-PTEN was shown by coomassie staining. (h) Schematic representation of N-terminal Myc-tagged PTEN (FL), along with its deletion mutants (D1–D4). (i) Myc-tagged PTEN constructs and SFB-Rab7 were co-expressed in HEK 293T cells, and the interaction of PTEN with Rab7 was detected by immunoblotting with anti-Myc antibodies after the cell lysates were pulled down with streptavidin beads.

Figure 3. Rab7 is a substrate of PTEN protein phosphatase activity.

(a) In vitro phosphorylated Rab7 was incubated with wild type and different catalytically inactive mutants of PTEN and the released phosphate was assayed colorimetrically using the malachite green reagent (absorbance values measured at 620 nm). Data represent mean absorbance from three independent experiments; *P<0.05, by Student's t-test. (b) Alignment of partial Rab7 protein sequences was shown. Conserved phosphorylation residues at serine 72 and tyrosine 183 (amino-acid positions in human protein indicated on top) shown bold and underlined. (c) Bacterially purified GST Rab7 WT and other indicated mutants were subjected to in vitro kinase assay using total 293T cell lysate. The phosphorylated Rab7 was then used as a substrate for PTEN phosphatase activity and the released phosphate was assayed colorimetrically using the malachite green reagent (absorbance values at 620 nm). Data represent mean absorbance from three independent experiments; *P<0.05, by Student's t-test. (d) The custom made Rab7 phosphopeptides (67–77 aa; QERFQpSLGVAF or 178–188 aa, TEVELpYNEFPE) along with control non-phosphorylated peptides were incubated with PTEN WT or PTEN Y138F mutant for 60 min at 37 °C. The released phosphate was assayed colorimetrically using the malachite green reagent (absorbance values at 620 nm); n=3, *P<0.05 by Student's t-test. (e) Control and two individual PTEN shRNA expressing cells transfected with SFB-Rab7 were subjected to immunoprecipitation with streptavidin (SBP) beads and immunoblotted with either phosphoserine (p-ser) or phosphotyrosine (p-tyr) antibodies. (f) Control and PTEN shRNA transduced HeLa cells were transfected with SFB Rab7 WT or Rab7 2PD mutant. Rab7 and its phosphorylated species (indicated with asterisk) were detected by using Phos-tag PAGE after immunoprecipitating Rab7 from the cells.

Figure 4. Dephosphorylation of Rab7 is required for its endosomal membrane localization.

(a) HeLa cells expressing either control or PTEN shRNA were transfected with GFP-Rab7 and its localization to late endosomes/multivesicular bodies was determined by co-staining with CD63 antibody. Scale bar, 10 μm. (b) GFP-Rab7 wild type (WT), phospho-dead S72A, Y183F, S72A/Y183F (2PD) and phosphomimetic S72E, Y183D, S72E/Y183D (2PM) Rab7 mutants were transfected into HeLa cells and their endosomal localization was determined by confocal imaging after co-staining with CD63 marker. Scale bar, 10 μm. (c) 293T cells expressing either Rab7 WT, S72E, Y183D or Rab7 2PM mutant along with untransfected control cells were serum starved overnight. Cells were treated with cycloheximide (50 μg ml⁻¹) for 1 h and stimulated with 100 ng ml⁻¹ EGF for indicated times. EGFR levels were tested by immunoblotting using specific antibody.

Figure 5. PTEN-mediated Rab7 dephosphorylation is necessary for its interaction with GDI, GEF and effector proteins.

(a) HEK 293T cells transfected with either SFB-Rab7 WT or its various mutants were subjected to immunoprecipitation with streptavidin beads (SBP). The interaction of GDI with Rab7 and its mutants was analysed by immunoblotting with GDI-specific antibody. (b) Cells were transfected with SFB-Rab7 WT and its mutants along with Flag-tagged Ccz1 and the interaction of Rab7 with Ccz1 was determined by immunoblotting with Flag antibody after immunoprecipitating with SBP. (c) 293T cells expressing either control or two different PTEN shRNAs were co-transfected with SFB-Rab7 and Flag-Ccz1. The interaction of Ccz1 with Rab7 was analysed by immunoblotting with Flag antibody after immunoprecipitation with SBP beads. (d) Cells were transfected with SFB-Rab7 WT and its mutants along with GFP-tagged RILP and the interaction of Rab7 with RILP was determined by immunoblotting with GFP antibody after immunoprecipitating with SBP. (e) Cells were transfected with SFB-Rab7 WT and Rab7 S72A/Y183D mutant along with GFP-tagged RILP and the interaction of Rab7 with RILP was determined by immunoblotting with GFP antibody after immunoprecipitating with SBP. (f) Cells transfected with various indicated SFB-tagged Rab7 constructs were labelled with ³²P-orthophosphate. GDP and GTP levels were analysed by thin layer chromotagraphy after immunoprecipitating Rab7 from the cell lysates by using streptravidin sepharose. (g) The GTP/GDP-bound ratio of various Rab7 mutants quantified by using Phosphorimager was plotted.

Figure 6. A model for PTEN-Rab7 association in endosome maturation.

A proposed model to depict the role of PTEN in dephosphorylating inactive Rab7 in the cytosol, which facilitates its interaction with GDI for further presentation to Mon1-Ccz1 GEF activity at the endosomal membrane leading to active Rab7.