Fig 2: Rabaptin-5 CC2-2 binds directly to the CHMP4-binding and neighbouring ‘S’ pockets of HD-PTP. (A) Lysates from HEK293T cells transfected with GFP–Rabaptin-5 (RBPT5) truncation mutants and WT or L202D HA–HD-PTP-Bro1CC were immunoprecipitated (IP) and proteins detected by western blotting. Input is 2% of offered protein. Graph shows quantification of GFP-Rabaptin-5 binding to HA–HD-PTP. Values represent mean±s.d. (n=7). ***P<0.0001; ns, not significant (one-way ANOVA with Bonferroni's test for multiple comparisons). (B) Recombinant GST–Rabaptin-5 truncation mutants and WT or L202D His6–HD-PTP-Bro1CC were incubated with glutathione agarose beads and proteins detected by western blotting. Input is 3% of offered protein. Graph shows quantification of GST–Rabaptin-5 binding to His6–HD-PTP-Bro1CC. Values represent mean±s.d. (n=6). ***P<0.001 (one-way ANOVA with Bonferroni's test for multiple comparisons). (C) Recombinant GST–Rabaptin-5-CC2-2 and WT His6–HD-PTP-Bro1CC were incubated with WT or L217A CHMP4B peptide (aa 205–224), immobilised on glutathione agarose beads and proteins detected by western blotting. Input is 3% of offered protein. Graph shows quantification of GST–Rabaptin-5 binding to His6–HD-PTP-Bro1CC. Values represent mean±s.d. (n=4). **P<0.01; ns, not significant (one-way ANOVA with Bonferroni's test for multiple comparisons). (D) Residues in Bro1 domains involved in binding to effector peptides. Conserved residues are in black, residues that differ in Alix are in red. Lysates from HEK293T cells transfected with GFP–Rabaptin-5-CC2-2 and WT, F62E, R69L, H125A, T145K, R198M or L202D HA–HD-PTP-Bro1CC were immunoprecipitated and proteins detected by western blotting. Input is 2% of offered protein. Graph shows quantification of GFP–Rabaptin-5 binding to HA–HD-PTP. Values represent mean±s.d. (n=4). ***P<0.001; **P<0.01; ns, not significant (one-way ANOVA with Bonferroni's test for multiple comparisons).

Fig 3: AKTIP cooperates with TSG101 but not with ALIX for CHMP4B recruitment.(A-B) Representative images of shctr, shAKTIP, siTSG101 and shAKTIP/siTSG101 HeLa cells stained for CHMP4B (red) and a-tubulin (green) and relative quantification. (C-D) Representative images of shctr, shAKTIP, siALIX and shAKTIP/siALIX HeLa cells stained for CHMP4B (red) and a-tubulin (green) and relative quantification. For results in (B and D) at least 100 midbodies per condition were counted. Scale bars, 5µm.

Fig 4: ESCRT complex recruited to LLOMe-damaged lysosomes promotes cell survival.Human fibroblasts were exposed to 1 mM LLOMe and when indicated, the cell permeable calcium chelator BAPTA-AM (1 µM) was added 10 min prior to LLOMe exposure. Immunocytochemical staining with corresponding colocalization analysis of a CHMP4B (green) and LAMP2 (red), b ALIX (green) and NPC-1 (red), and c CHMP4B (green) and LAMP2 (red) in fibroblasts exposed to LLOMe for 15 min. Nuclei are stained with DAPI (blue) and merged images show colocalization in yellow (n = 5). d Caspase-3 like activity in cells exposed to LLOMe for 8 h (n = 3). e Quantification of apoptotic and necrotic cells using Annexin V/PI staining in fibroblasts exposed to LLOMe for 6 h (n = 3). f Immunoblot of LAMP2 in digitonin-extracted cytosolic fractions from cells exposed to LLOMe for 2 h (n = 8) with corresponding quantitative analysis. *p < 0.05, **p < 0.01 and ***p < 0.001 compared to control (a–b) or LLOMe only (c–f).

Fig 5: Rab8A is recruited to damaged lysosomes containing Galectin-3 and CHMP4B in response to calcium signalling ARAW264.7 macrophages were electroporated with EGFP-Rab8A and treated with 1 mM LLOMe for 30 min. Intracellular distribution of Galectin-3, CHMP4B or LC3B was visualised by immunofluorescence. Scale bar = 5 µm.BMacrophages were electroporated with EGFP-Rab8A and RFP-Galectin-3. Cells were then treated with 1 mM of LLOMe and monitored by live cell imaging. Snapshots at the indicated time points after LLOMe addition are shown. Scale bar = 5 µm.CRAW264.7 macrophages were treated with 1 mM of LLOMe, and LRRK2, Rab8A, CHMP4B, Galectin-3 and LC3B foci over time were analysed by high-content imaging. Data show the mean ± SEM of five biological replicates.D–HRAW264.7 macrophages were pre-treated with 10 µM BAPTA-AM for 1 h, and treated with 1 mM LLOMe for 30 min. (D) CHMP4B recruitment was monitored by immunofluorescence and high-content imaging. Scale bar = 10 µm. (E) Rab8A pT72 phosphorylation was analysed by Western blot. (F) LRRK2 and (G) Rab8A recruitment was monitored by immunofluorescence and high-content imaging. Scale bar = 10 µm. (H) Quantification of E and F. (D and H) Data represent the mean ± SEM of two to three independent biological experiments. One-way ANOVA followed by Sidak's multiple comparisons test. ns = non-significant; **P = 0.01, *P = 0.05. Source data are available online for this figure.