The ubiquitin-specific protease USP2a prevents endocytosis-mediated EGFR degradation

Abstract

Ubiquitination of epidermal growth factor receptor (EGFR) is required for downregulation of the receptor by endocytosis. Impairment of this pathway results in constitutively active EGFR, which is associated with carcinogenesis, particularly in lung cancer. We previously demonstrated that the deubiquitinating enzyme ubiquitin-specific protease 2a (USP2a) has oncogenic properties. Here, we show a new role for USP2a as a regulator of EGFR endocytosis. USP2a localizes to early endosomes and associates with EGFR, stabilizing the receptor, which retains active downstream signaling. HeLa cells transiently expressing catalytically active, but not mutant (MUT), USP2a show increased plasma membrane-localized EGFR, as well as decreased internalized and ubiquitinated EGFR. Conversely, USP2a silencing reverses this phenotype. Importantly, USP2a prevents the degradation of MUT in addition to wild-type EGFR. Finally, we observed that USP2a and EGFR proteins are coordinately overexpressed in non-small cell lung cancers. Taken together, our data indicate that USP2a antagonizes EGFR endocytosis and thus amplifies signaling activity from the receptor. Our findings suggest that regulation of deubiquitination could be exploited therapeutically in cancers overexpressing EGFR.

Figure 1. USP2a localizes to early endosomes

GFP-USP2a co-localizes with EEA1 (a), while GFP does not (b). HeLa cells transiently expressing GFP-USP2a or GFP were used for immunofluorescence analysis. In panel (a) green indicates USP2a, red indicates EEA1. Right are merged pictures seen on left for both panels (a) and (b). White circles indicate co-localization in the magnified insets in (a). Yellow circles in (a) show lack of interference between red and green fluorescent channels. (c) Endogenous USP2a and early endosome markers (EEA1, TfR and Rab5) are enriched in heavier fractions from sucrose gradient centrifugation. The post-nuclear supernatant (PNS) from HeLa cells homogenate was separated by a discontinuous sucrose gradient ultracentrifugation. Fractions were collected from top to bottom and equal volumes from each fraction analyzed by western blot.

Figure 2. USP2a stabilizes EGFR and decreases ubiquitinated forms of the receptor after EGF stimulation

(a) Exogenous wild type USP2a, but not the catalytic USP2a mutant, stabilizes EGFR. HeLa cells transfected with empty vector, WT-USP2a or MUT-USP2a were serum starved for 12 h and then treated with100 ng/ml of EGF for the indicated times, with or without cycloheximide (CHX). Cells were lysed and analyzed by western blot. Quantification graphs of EGFR in the western blots are shown. Results are typical and representative of at least three experiments. (b) WT-USP2a but not MUT-USP2a stabilizes both p-EGFR and its downstream signaling molecule p-Stat3. Total Stat3 protein levels are not changed under USP2a or control conditions. Quantification graphs refer to CHX condition after 2 h of EGF. (c) Depletion of USP2a by siRNA #1 decrease retention of EGFR after 2h of EGF treatment as USP2a siRNA #2 (Supplementary Figure S5a). HeLa cells were transfected with anti-USP2a siRNA #1 or non-targeting siRNA. Left, knockdown of USP2a was detected by western blot; Right, cells were stimulated with EGF for 0 or 2 h, and then lysed for western blot. (d) Exogenous USP2a decreases EGFR ubiquitination. HeLa cells, EGF treated or not, were lysed with TGH lysis buffer and immunoprecipitated with anti-EGFR antibody. The precipitates and lysates were analyzed by western blot. The ubiquitinated EGFR in the precipitates were detected with anti-ubiquitin antibody (upper panels) or with anti-p-EGFR antibody (lower panel). (e) USP2a deubiquitinates both K48-Ub_n and K63-Ub_n. Flag-USP2a was purified from HeLa cell lysate by anti-Flag antibody conjugated agarose beads and eluted with 3xFlag protein. Purified USP2a was used in DUB assays with K48-Ub_n or K63-Ub_n as the substrate. Reactions were stopped by adding Laemmli buffer and analyzed by anti-ubiquitin antibody.

Figure 3. USP2a enhances recycling of EGFR after EGF stimulation

(a) Exogenous WT-USP2a, but not the MUT-USP2a or AMSH, decreases internalized Rhod-EGF. HeLa cells expressing GFP-tagged WT- or MUT-USP2a were serum starved for 12 h and then changed to serum-free medium containing Rhod-EGF at 4°C. Subsequently, cells were chased for 15 min at 37°C. Any Rhod-EGF left on the plasma membrane was washed off with acid solution. Cells were fixed and photographed under fluorescence microscope. Quantification of internalized EGFR after 15 min of Rhod-EGF is shown at right. (b) Accumulation of internalized Rhod-EGF at different time points after EGF stimulation in the presence of exogenous WT-USP2a, MUT-USP2a or empty vector. (c) Effect of USP2a knockdown on EGFR internalization at different EGF time points treatments. HeLa cells were transfected with the siRNA to USP2a or the control siRNA, serum starved and treated with Alexa 488-EGF as with Rhod-EGF in (a–b). Internalization of EGFR was detected by measuring the signals from internalized Alexa488-EGF by FACS. (d) The amount of plasma membrane-localized EGFR was affected by exogenous WT-USP2a after 15 min of EGF treatment. HeLa cells are transiently expressing GFP tagged WT-USP2a, MUT-USP2a or empty vector. Experiments were done with an antibody against the N-terminus of EGFR and cells were analyzed by 2 channels flow cytometry. Non-transfected cells were used as internal controls. Results were shown as the amount of plasma membrane-localized EGFR (PM-EGFR, red florescence reading) in transfected cells (green) over to that in non-transfected cells (no green). (e) Amount of plasma membrane-localized EGFR decreases with USP2a siRNA treatment. Cells were treated with EGF 37°C, and then incubated with anti-N-EGFR antibody at 4°C, following incubation with PE labeled-secondary antibody at 4°C. Cells were subsequently analyzed by FACS. (f) More EGFR co-localized with recycling endosomal marker Rab11 in the presence of GFP-USP2a (left USP2a transfected cells) than the control (right non-transfected cells). Circles indicate co-localization of Rhod-EGF and Rab11 in lower panels.

Figure 4. USP2a stabilizes wild type and mutant EGFR in NSCLC cell lines and correlates with EGFR protein expression in human NSCLC

(a) USP2a siRNA #1 results in decreased amount of wild type and mutant EGFR in NSCLC cell lines after 2 hours of EGF treatment, as USP2a siRNA #2 (Supplementary Figure S5b). Experiments were done as in Fig. 2c. EGFR mutants (Del and L858R) were detected by specific antibodies to the mutant epitopes. (b) Transient expression of WT-USP2a results in accumulation of EGFR after 2h of EGF treatment in both NSCLC cells lines (H2009 and H358) and HeLa cells. Experiments were done as in Fig. 2b. (c) Expressions of EGFR and USP2a proteins in NSCLC are positively correlated. Protein expression of EGFR and USP2a in 16 cases of human NSCLC were determined by quantifying the protein bands in the western blot (Supplementary Figure S10), and normalized by β-actin. The plot shows weighted linear regression of every tumor’s EGFR expression (Y-coordinate) with its USP2a expression (X-coordinate). The higher the USP2a Tumor/Normal (T/N) expression ratio in a given case, the darker is the color of that point. The linear model fitted with weighted linear regression of the cases is shown as a dotted line (slope 0.8195, F-test p-value of association 0.0371).

Figure 5. Proposed model for the role of USP2a in the EGFR endocytic pathway

1. Plasma membrane-localized EGFR binds EGF and then dimerizes. 2. EGFR is activated by auto-phosphorylation, which in turn stimulates ubiquitination of the receptor. 3. Ubiquitinated EGFR is internalized into early endosomes, where USP2a is localized. 4. USP2a stimulates deubiquitination of EGFR. 5. Deubiquitinated EGFR is recycled back to the plasma membrane. As a result, USP2a prevents lysosomal degradation of EGFR.