Fig 2: Identification of potential FLOT1-/FLOT2-positive early endosome cargo.a Western blot analysis of FLAG-FLOT1/FLOT2 in early endosome-enriched fractions (SPIONs) and intact endosome-IP fractions (n = 1). b Data-independent acquisition (DIA)-based protein abundance fold-change ratios of SPIONs/SPIONs+IP fractions (n = 3, biologically independent samples over three independent experiments). Significant differentially regulated proteins are indicated (cut-offs: q-value: <0.05, fold change: >1.5). c STRING-based PPI analysis of selected protein categories overrepresented in the SPIONs+IP fraction. Node size corresponds to the protein abundance and line thickness to the PPI confidence score. SPIONs superparamagnetic iron oxide nanoparticles, IP immunoprecipitation, IN input, SN supernatant, FT flow through, WA wash, EL eluate, PPI protein–protein interaction. Source data are provided as a Source Data file.

Fig 3: Active RAB31 engages FLOTs to drive EGFR-containing ILV formation depending on cholesterol and ceramide in lipid raft microdomains.a Left, immunofluorescence of EGFR-HA (red) and Flag-RAB31Q65L (magenta) with CD63-GFP (green) in Flag-RAB31Q65L stable HeLa cells stably expressing shNC (negative control), shFLOT1, shFLOT2 or shFLOT1 and shFLOT2 and transiently expressing EGFR-HA and CD63-GFP under serum starvation (SS). Right up panel, the ratio of entry of EGFR-HA into CD63-GFP-positive LE/MVE in shNC (n = 9 fields), shFLOT1 and shFLOT2 (n = 9 fields), shFLOT1 (n = 12 fields), shFLOT2 (n = 12 fields). Right low panel, the ratio of entry of Flag-RAB31Q65L into CD63-GFP-positive LE/MVE in shNC (n = 9 fields), shFLOT1 and shFLOT2 (n = 9 fields), shFLOT1 (n = 12 fields), shFLOT2 (n = 12 fields). b Western blotting analyses of the concentrated conditional media from the indicated stable HeLa cells used in a. c Up panels, immunofluorescence of EGFR-HA (red) and Flag-RAB31 (magenta) with FLOT1-GFP (green) in the indicated stable HeLa cells transiently expressing EGFR-HA and FLOT1-GFP under SS. Low panel left, the ratio of co-localization of EGFR-HA with FLOT1-GFP-positive vesicle in Vector (n = 7 fields), RAB31WT (n = 8 fields) and RAB31Q65L (n = 9 fields). Low panel right, the ratio of co-localization of Flag-RAB31 with FLOT1-GFP-positive vesicle in RAB31WT (n = 8 fields) and RAB31Q65L (n = 9 fields). d Up panels, immunofluorescence of FLOT1-HA (red) and Flag-RAB31 (magenta) with CD63-GFP (green) in the indicated stable HeLa cells transiently expressing FLOT1-HA and CD63-GFP under SS. Low panel left, the ratio of co-localization of FLOT1-HA with CD63-GFP-positive LE/MVE in Vector (n = 7 fields), RAB31WT (n = 7 fields) and RAB31Q65L (n = 8 fields). Low panel right, the ratio of entry of FLOT1-HA into CD63-GFP-positive LE/MVE in Vector (n = 7 fields), RAB31WT (n = 7 fields) and RAB31Q65L (n = 8 fields). e Left, immunofluorescence of EGFR-HA (red) and Flag-RAB31Q65L (magenta) with CD63-GFP (green) in Flag-RAB31Q65L stable HeLa cells transiently expressing EGFR-HA and CD63-GFP and treated with DMSO, 5 µM GW4869, 5 µM simvastatin or 10 µM lovastatin under SS. Right up panel, the ratio of entry of EGFR-HA into CD63-GFP-positive LE/MVE in DMSO (n = 8 fields), GW4869 (n = 11 fields), simvastatin (n = 13 fields) and lovastatin (n = 12 fields). Right low panel, the ratio of entry of Flag-RAB31Q65L into CD63-GFP-positive LE/MVE in DMSO (n = 8 fields), GW4869 (n = 11 fields), simvastatin (n = 13 fields) and lovastatin (n = 12 fields). f Western blotting analyses of the concentrated conditional media from the indicated stable HeLa cells used in e. All data are means ± SD. Unpaired t-test was used to analyze the difference between the two groups. ****P < 0.0001, ***P < 0.001, NS, no statistical significance. Scale bars, 10 µm.

Fig 4: Investigation of higher-order flotillin assemblies in lysosome- and early endosome-enriched fractions.a HeLa cells were stained with antibodies for FLOT1, FLOT2, and LAMP2 followed by microscopy imaging (n = 3). Lower panels display a zoom-in of the regions indicated in the full picture. Mander’s coefficients show the average degree of overlap of one protein population relative to another. Data are presented as mean values + SD (n = 3, cells examined over one experiment). Profile plots indicate the degree of colocalization for individual vesicles. b Analysis of FLOT1/FLOT2 in early endosomes (n = 1). Following pulsed SPIONs treatment, endosomes were enriched at the indicated chase times. Pre-CTSD serves as a marker for early endosomes. c Western blot analysis of early endosome-enriched fractions with marker proteins for different subcellular compartments (n = 1): Cytosol (ACTG2 and GAPDH), cytoskeleton (TUBA), endoplasmic reticulum (CANX), and mitochondria (SDHA). d Experimental workflow for early endosome enrichment and XL-LC-MS/MS analysis. Created with BioRender.com. e Overview of the cross-linking dataset obtained from early endosome-enriched fractions. f Overlap of unique FLOT1/FLOT2 cross-links for XL-LC-MS/MS analyses of early endosome- and lysosome-enriched fractions. g Western blot analysis of BN-PAGE-separated FLAG-IP eluates with/without cross-linking by DSSO (n = 5). IN/EL refers to individual FLAG-IP fractions. h Western blot analysis of SDS-PAGE separated FLAG-IP ELs with/without DSSO cross-linking for WT and FLOT1/FLOT2 residue 200–300 deletion mutants (Δ100) (n = 3). Scale bar = 5 µM; WCL whole cell lysate, IN input, WA wash, EL eluate, MCOE Mander’s coefficient, Dst. distance, LS lysosomes, EE early endosomes, XL cross-linking, SPIONs superparamagnetic iron oxide nanoparticles, M marker, WT wild type. Source data are provided as a Source Data file.

Fig 5: FLOT-1 is a predicted target of miR-485-5p. (A) Bioinformatics analysis predicted a miR-485-5p-binding site in the 3′UTR of FLOT1 mRNA. (B) The FLOT-1 plasmid transfection efficacy was verified in Eca 109, KYSE 30 and TE-1 cells transfected with FLOT-1 plasmid or pcDNA3.1 via western blotting. (C) FLOT-1 mRNA levels in KYSE 30, Eca 109 and TE-1 cells transfected with miR-485-5p mimics, miR-485-5p inhibitor or their corresponding controls (miR-NC and inhibitor-NC). (D) FLOT-1 protein levels of Eca 109 and KYSE 30 cells transfected with miR-485-5p mimics, miR-485-5p inhibitor or their corresponding controls (miR-NC and inhibitor-NC). (E) Effects of miR-485-5p on the translation of the reporter gene inserted downstream of the 3′UTR of FLOT-1 mRNA or the mutated 3′UTR of FLOT-1 mRNA in 293T cells. 293T cells were co-transfected as shown in the figure, and the luciferase activities were measured using a Dual-luciferase reporter assay. The luciferase activity was normalized and expressed as the ratio of firefly/Renilla luciferase activities. (F) IHC staining of FLOT-1 expression in esophageal squamous cell carcinoma tissue sections (scale bar, 200 µm). (G) FLOT-1 protein levels in Eca 109 and KYSE 30 cells co-transfected with mimics-NC and pcDNA3.1, miR-485-5p mimics and pcDNA3.1, miR-485-5p mimics and FLOT-1, mimics-NC and FLOT-1. The data are expressed as the mean ± SD. *P<0.05 and **P<0.01 vs. control group. FLOT-1, flotillin-1; miR, microRNA; UTR, untranslated region; NC, negative control; IHC, immunohistochemistry; WT, wild-type; MUT, mutant.