Fig 2: Loss of VPS34 resulted in abnormal assembly of retromer complexes in oocytes. MII oocytes were collected from 5 W control and ZcKO mice after superovulation. a‐e) Immunofluorescent co‐staining of VPS26A (red) with LAMP1 (green) (a); RAB7 (red) with VPS26A (green) (b); RAB7 (red) with VPS35 (green) (c); RAB7 (red) with TBC1D5 (green) (d), M6PR (red) with LAMP1 (green) (e) in control and ZcKO oocytes. Scale bars = 50 µm. Blue, nuclear staining with Hoechst 33342; Merge, representing a superposition of fluorescence signals from three channels. f) Western blot of VPS34 and CTSD and M6PR in control and ZcKO oocytes. β‐tubulin was used as an internal control. g) Western blots of TBC1D5, VPS35, VPS26A and VPS26B in control and ZcKO oocytes of mice. β‐tubulin was used as a loading control. h) Immunofluorescent co‐staining of VPS26B (red) with LAMP1 (green) in control and ZcKO oocytes. i–m) Interactions of RAB7 with retromer components in control and Vps34 knockout (KO) 293T cells. Vps34 knockout 293T cell line was constructed using the CRISPR‐Cas9 system and co‐immunoprecipitation experiments were performed as follows: RAB7 (i), TBC1D5 (j), VPS26B (k), VPS26A (l) and VPS35 (m) were used to immunoprecipitate the proteins in control and KO 293T cells, and then anti‐RAB7, anti‐TBC1D5, anti‐VPS26A, anti‐VPS26B and anti‐VPS35 antibodies were used for Western blot detection. At least 3 independent replicates were performed for each experiment. n) Direct interaction between VPS34 and RAB7 by GST pull‐down assay.

Fig 3: VPS35 recycled EGFR to the cell surface and decreased EGFR degradation. (a) VPS35 knockdown enhanced EGF-induced EGFR degradation, whereas VPS35 overexpression inhibited EGFR degradation. Western blotting detected the alterations in the total EGFR degradation in indicated cells in response to EGF stimulation in the presence of cycloheximide (left panels). Densitometric analysis of EGFR blots from three independent experiments are shown as mean ± SD (right panels). Experiments were conducted three times independently. (b) Localisation analysis of EGFR after EGF and cycloheximide treatment at different time points. At 0 min, EGFR was located mainly on the cell surface in all indicated cells after overnight starvation. With the passing of time, AGS/shVPS35 cells showed weaker fluorescence intensity compared to AGS/Ctrl cells, while HS-746T/VPS35 cells retained stronger fluorescence intensity than cells in the control group did. RFP: EGFR. Experiments were conducted three times independently. (c) VPS35 knockdown inhibited EGFR recycling and VPS35 overexpression promoted EGFR recycling. Indicated cell surfaces were labelled with Sulfo–NHS–SS-biotin on ice and stimulated for 30 min with EGF at 37 °C. Recycling of EGFR was determined at different time points indicated in the graphs by western blotting (upper panels). The values in western blotting were mean ± SD from three independent experiments (lower panels). Experiments were conducted three times independently. (Scale bars: 100 μm).

Fig 4: VPS35 expression upregulated in tumour tissues and associated with poor prognosis of the patients with gastric cancer. (a) Upregulation of VPS35 in six digestive system cancer types in the GEPIA database. (b) Overall survival curves of the 384 patients in the gastric cancer cohort in GEPIA, with patients stratified based on VPS35 expression levels (Log-rank test). (c) Average immunohistochemistry (IHC) scores of VPS35 of patient samples in the Ruijin gastric cancer cohort (n = 96, Chi-square test). (d) Representative images of VPS35 IHC in gastric cancer and matched normal tissues from the Ruijin cohort. (e) Overall survival curves of our patient cohort, with patients stratified based on VPS35 expression levels (Log-rank test). (Scale bars: 100 μm, ∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001 and ∗∗∗∗, p < 0.0001.).

Fig 5: Identification of VPS35 binding to EGFR. (a) Gene Ontology pathway enrichment analysis of significantly differentially expressed genes (DEGs) between MKN28/Ctrl and MKN28/shVPS35 cells revealed by RNA-Seq. (three biological replicates, transcriptomic data were submitted to the GEO database (GSE210320). (b) KEGG pathway enrichment analysis of significantly DEGs (Fisher's exact tests). (c) A pairwise correlation of RNA expression of VPS35 with MAPK-associated genes of receptor components (i.e., EGFR, VEGFR2, MET, and FGFR3) from TCGA gastric cancer cohort. Left triangle: a colour plot for the expression of two molecules showing different correlation by colours; diagonal: the name for each molecule; and right triangle: correlation coefficient. (d) Correlation of RNA expression of VPS35 with several related genes from the GEPIA gastric cancer cohort (Spearman's correlation). (e) Flow cytometry analysis for AGS/Ctrl, AGS/shVPS35, MKN28/Ctrl, MKN28/shVPS35, HGC27/Ctrl, HGC27/VPS35, HS-746T/Ctrl, and HS-746T/VPS35 cells. Compared with cells in the control groups, AGS/shVPS35 and MKN28/shVPS35 cells had a significantly decreased EGFR fluorescence density (left two panels), while HGC27/VPS35 and HS-746T/VPS35 cells had a significantly increased EGFR fluorescence density (right two panels). Experiments were conducted three times independently. (f) Correlation of EGFR and VPS35 expression levels by linking IHC scores of EGFR to VPS35 in paired patient samples (Pearson's correlation). (g) Different binding proteins derived from immunoprecipitation and mass spectrometry (IP-MS) of MKN28 cells. (h) Interaction of VPS35 with EGFR determined by co-Immunoprecipitation (Co-IP) assays in MKN28 cells. (The student's t tests, ∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001 and ∗∗∗∗, p < 0.0001.).