Figures and data
Deletion of Ezrin increases the lysosomal pathway.
A, Gene Ontology of 530 genes differentially expressed at mRNA and protein levels (EZRKO versus WT). Bubble plot representing some of the most enriched GO terms regarding cellular components. Color and x axis represent minus logarithms of p Value. The size represents the numbers of genes enriched in a GO term. B, WT and EZRKO MEF cells were cultured in 6-cm cell plates for 24 hours, then fixed and immunostained with lysotracker and DAPI. Scale bar 10 µm. C, Data represent mean of lysotracker-positive cells ± SEM (n=3 experiments at least). Statistical test: unpaired t-test. D, MEF cells WT and EzrKO were cultured in 6-well plates for 24 hours, then fixed and immunostained with LAMP1 and LC3 antibodies and DAPI. Scale bar 10 µm (magnification 1 µm). E, Data represent mean of LAMP1-LC3 colocalization spots ± SEM (n=3 experiments at least). Statistical test: unpaired t-test. F, Model showing autophagic flux induction in EZR−/− cells. G, MEF EzrKO showed CTSB enzymatic activity increase compared to control cells. H, MEF cells WT and EzrKO were lysed and immunoblotted with NBR1, LAMP1, EZR, P62, LC3 antibodies or GAPDH antibodies as a loading control. The graphs show the mean NBR1, LAMP1, EZR, P62, LC3 levels relative to GAPDH ± SEM (n=3 experiments at least). Statistical test: unpaired t-test for NBR1, P62, LC3-I; Welch’s t-test for CTSD, LC3-II; Mann-Whitney test for LAMP1. Panel F created with BioRender.com.
Ezrin binds EGFR and regulates its activation.
A, Bubble plot representing the enrichment analysis of 530 DEGs performed in SILAC Phosphoproteomics data. Color and x axis represent minus logarithms of p Value. The size represents numbers of genes enriched in the indicated data. B, Physical interactions, obtained by GeneMANIA, highlight Ezrin and EGFR binding. C, Volcano plot of DEGs, with up-regulated EGFR and down-regulated MAP2 and ERBB2 (no threshold on Log2FC and 0.05 threshold on –Log10FDR). Legend: red dot, up-regulated gene; blue dot, down-regulated gene; grey dot, not significant gene. D,E Co-IP data for Ezrin-EGFR interaction. For the co-IP analyses, was used Ezrin antibody, conjugated with beads, and immunoblotted with EGFR antibody for WT and EZR−/− (D) and HeLa EZRT567D and EZRT567A (E) HeLa cells, respectively. Schematic representation of HeLa EZRT567D and EZRT567A co-IP (bottom). F, Confocal microscopy images showing EGFR (green) and EZR (red) co-localization on the membrane in HeLa WT cells (left) and magnified views of the regions are provided (right). Scale bar 10 µm (magnification 1 µm). Representative plots of co-localization profiles on the membrane between EGFR (green) and EZR (red). Data represent mean of EGFR-EZR colocalization spots ± SEM (n=3 experiments at least); G, Immunofluorescent labelling images of EGFR in HeLa WT and EZR−/− cells, observed by confocal microscopy. Scale bar 10 µm. H, Data represent fluorescence intensity ± SEM (n=3 experiments at least). Statistical test: unpaired t-test; I, Immunoblots and calculated levels (bottom) of HER3, pY845 EGFR, EGFR, pT222 MK2, MK2, pT180/pY182 p38 MAPK and P38 MAPK in HeLa WT and EZR−/− cells. Data are expressed as mean of pY845EGFR/EGFR, pT222 MK2/MK2 and pT180/pY182 p38 MAPK/P38 MAPK ratio ± SEM (n=3 experiments at least). GAPDH was used as loading control. Statistical test: unpaired t-test for pY845 EGFR; Mann-Whitney test for HER3, EGFR, pT222 MK2, pT180/pY182 p38 MAPK. Panel E created with BioRender.com.
Ezrin controls EGFR localization.
A, Western blot analysis of chemical crosslinked EGFR in HeLa WT and EZR−/− with (+) and without (-) EGF stimulation. Arrowheads indicate detected signals of dimeric and monomeric form of EGFR. B, Model showing the crosslinking effect of EGFR dimer formation in HeLa WT and EZR−/−. C, Immunofluorescence images of EGFR (green) and EEA1 (red) in HeLa WT (top) and EZR−/− (bottom) observed by confocal microscopy. Scale bar 10 µm (magnification 1 µm). Representative plots of co-localization profiles of EGFR in early endosome. HeLa EZR−/− cells do not show EGFR and EEA1 co-localization compared to control cells. D, Data represent mean of EGFR-EEA1 colocalization spots ± SEM (n=3 experiments at least). Statistical test: unpaired t-test. E, Representative immunoblots of EGFR in membrane (top) and endosomes (bottom) proteins in HeLa WT and EZR−/−. ZO-1 and EEA1 are used as membrane and endosomes extraction control, respectively. GAPDH are used as loading control. F, Schematic translocation of EGFR in the endosomes in HeLa WT compared to HeLa EZR−/−. G, HeLa cells were fixed and immunostained with endotracker and EEA1 (red) and DAPI (blue). Scale bar 10 µm (magnification 1 µm). H, Graph shows mean of endotracker-positive cells ± SEM (n=3 experiments at least). Statistical test: unpaired t-test. Panel B created with BioRender.com. Panel F created with BioRender.com.
EGFR migrates on the endosomes depending on Ezrin.
A, Immunofluorescence labelling images of EGFR-GFP (green), EEA1 (red) and DAPI (blue) after 3h of EGF stimulation (right) in HeLa WT (top) and EZR−/− (bottom). Magnified views of the regions in the boxes are provided in both Airyscan high-resolution microscopy and 3D-confocal microscopy. B, EGFR and EEA1 co-localization is expressed as a representative plot in HeLa WT (top) and EZR−/− (bottom). Scale bar 10 µm (magnification 1 µm). C, Data represent mean of EGFR-EEA1 colocalization spots ± SEM (n=3 experiments at least). Statistical test: One Way ANOVA.
EGF stimulation does not affect EGFR in absence of Ezrin.
A, Live cell imaging and model for EGFR (green) translocation from the membrane to the endosomes in HeLa WT (top) and EZR−/− (bottom) cells without EGF stimulation (T0) and with a progressive EGF stimulation (from T10’’ to T60’’). White boxes are magnifications that depict EGFR protein migration. Scale bar 1 µm). Please refer to Video 1. B, IEM (anti-GFP immunolabelling) of cycloheximide treated HeLa WT, WT + EGF, EZR−/− and EZR−/− + EGF cells expressing EGFR-GFP. Endosomes containing EGFR is shown in green. Scale bar 200 nm. Quantitative analysis (right) of EGFR positive endosomes expressed as mean ± SEM. Statistical test: generalized Linear Model with Likelihood Ratio (Poisson Regression). C, immunoblots and calculated levels (bottom) of HER2, pY845 EGFR, pY1068 EGFR, EGFR, pT202/Y204 p42/44 MAPK, pT180/pY182 p38 MAPK and P38 MAPK in HeLa WT and EZR−/− cells with (+) and without (-) EGF stimulation. Data are expressed as mean of pY845EGFR/EGFR and pT180/pY182 p38 MAPK/P38 MAPK ratio ± SEM (n=3 experiments at least). GAPDH was used as loading control. Statistical test: Unpaired t-test for HER2 WT, HER2 EZR−/−, pY845 EGFR EZR−/−, pY1068 EGFR WT, pY1068 EGFR EZR−/−, EGFR WT, EGFR EZR−/−, pT202/Y204 p44/42 MAPK WT, pT202/Y204 p44/42 MAPK EZR−/−, pT180/pY182 p38 MAPK WT, pT180/pY182 p38 MAPK EZR−/−; unpaired t-test with Welch’s correction for pY845 EGFR WT. D, Representative immunoblots of EGFR in membrane (top) and endosomes (bottom) proteins in HeLa WT and EZR−/− with (+) and without (-) EGF stimulation. ZO-1 and EEA1 are used as membrane and endosomes extraction control, respectively. GAPDH are used as loading control.
EGFR-Ezrin complex interacts with TSC1.
A, co-IP analysis for Ezr-TSC1 (left) and EGFR-TSC1 (right) interaction. For co-IP analyses, Ezrin (left) and EGFR (right) antibodies were used. The proteins immunoprecipitated were blotted for TSC1 and AKT antibodies in HeLa WT and EZR−/−. B, HeLa WT and EZR−/− cells were lysed and immunoblotted with pS939 TSC2, TSC2, PT389 P70 S6 Kinase, P70 S6 Kinase, pS473 AKT, AKT, pS65 4E-BP1, 4E-BP1 and GAPDH as a loading control. Data represent the mean of pS939 TSC2/TSC2, T389 P70 S6 Kinase/P70 S6 Kinase, pS473 AKT/AKT and pS65 4E-BP1/4E-BP1 ratio ± SEM (n=3 experiments at least). Statistical test: unpaired t-test for pT389 P70 S6 Kinase, pS473 AKT; Unpaired t-test with Welch’s correction for pS939 TSC2; Mann-Whitney test for pS939 TSC2. C, pP70 S6 Kinase western blotting with insulin time course in HeLa WT (up) and EZR−/− (bottom) cells. Graph shows the mean of pP70 S6/P70 S6 ratio ± SEM (n=3 experiments at least). Statistical test: One Way ANOVA for WT and KO curve (pairwise comparisons with reference T0). D, Representative confocal images of LAMP1 and TSC1 (left) and LAMP1 and TSC2 (right) immunofluorescence in HeLa WT and EZR−/− cells. Magnified insets of TSC1/2 localization are shown. Scale bar 10 µm (magnification 1 µm). E, Data represent mean of LAMP1-TSC1 (left) and LAMP1-TSC2 (right) colocalization spots ± SEM (n=3 experiments at least). Statistical test: unpaired t-test for LAMP1-TSC1; Unpaired t-test with Welch’s correction for LAMP1-TSC2. F, pS473 AKT western blotting with EGF time course in HeLa WT (left) and EZR−/− (right) cells. Graph shows the mean of pS473 AKT/AKT ratio ± SEM (n=3 experiments at least). Statistical test: One Way ANOVA with Dunnett’s post-hoc test for WT curve; Kruskal-Wallis test with Dunn’s post-hoc test for KO curve (pairwise comparisons with reference T0). NS: not significant. G, HeLa WT, WT + EGF, EZR−/− and EZR−/− + EGF cells were immunostained with EGFR (green), TSC1 (red) and EEA1 (gray). Representative magnifications are shown. Scale bar 10 µm (magnification 1 µm). H, Data represent mean of TSC1-EEA1 colocalization spots ± SEM (n=3 experiments at least). Statistical test: One Way ANOVA.
Ezrin depletion induces EGFR-mediated retinal degeneration.
A, Schematic representation of used CRISPR/Cas9 strategy to generate Ezrin−/− medaka lines. The red box highlighted the deleted nucleotides in the Ezrin exon 1 gene. B, WT and Ezrin−/− medaka proteins were immunoblotted with Ezrin antibody and Actin as a loading control. C, Stereo-microscopic representative images of WT and Ezrin−/− medaka at stage 40. Scale bar 1 mm. D, Immunoblots and calculated levels (right) of pT1462 TSC2, pS473 Akt, LC3-I, LC3-II, pY845 Egfr and Egfr, pS65 4E-BP1 in WT and Ezrin−/− medaka fish. Data are expressed as mean of pT1462 TSC2/TSC2, pS473 Akt/Akt, pS65 4E-BP1/4EB-P1 and pY845EGFR/EGFR ratio ± SEM (n=3 experiments at least). Actin was used as loading control. Statistical test: Unpaired t-test. E, Representative confocal images of LAMP1 immunofluorescence in WT and Ezrin−/− medaka fish. Magnified insets of RPE LAMP1 localization are shown. Scale bar 10 µm. RPE: retinal pigment epithelium; ONL: outer nuclear layer; INL: inner nuclear layer. F, Medaka WT and Ezrin−/− fish were immunostained with EGFR. Scale bar 10 µm. G, Immunofluorescence labelling images of RHO (left) and ZPR1 (right) in WT and Ezrin−/− fish. Magnified views of the regions in the boxes are provided at the bottom. Scale bar 10 µm. H, Confocal images showing representative TUNEL positive cells on cryosection from WT and Ezrin−/− medaka lines. Scale bar 10 µm. Graph shows the mean of number of TUNEL positive cells for retina ± SEM (n=3 experiments at least). Statistical test: unpaired t-test.
EGFR/Ezrin/TSC complex molecular pathway
Diurnal inactivation of Ezrin leads to incapacity of EGFR to dimerize. The absence of active EGFR on the endosome causes the migration of TSC complex on the lysosome, where it inhibits mTORC1C1. On the contrary, the nocturnal activation of Ezrin favorites the phosphorylation and dimerization of EGFR, that translocate from the plasma membrane on the early endosome. Ezrin localized with EGFR, on the endosomal membrane, binds TSC complex, preventing mTORC1C1 inactivation on the lysosome. Created with BioRender.com.
