Stress granules and mTOR are regulated by membrane atg8ylation during lysosomal damage

Abstract

We report that lysosomal damage is a hitherto unknown inducer of stress granule (SG) formation and that the process termed membrane atg8ylation coordinates SG formation with mTOR inactivation during lysosomal stress. SGs were induced by lysosome-damaging agents including SARS-CoV-2ORF3a, Mycobacterium tuberculosis, and proteopathic tau. During damage, mammalian ATG8s directly interacted with the core SG proteins NUFIP2 and G3BP1. Atg8ylation was needed for their recruitment to damaged lysosomes independently of SG condensates whereupon NUFIP2 contributed to mTOR inactivation via the Ragulator-RagA/B complex. Thus, cells employ membrane atg8ylation to control and coordinate SG and mTOR responses to lysosomal damage.

Graphical abstract

Figure 1.

Lysosomal damage induces SG formation. (A) DIA LC/MS/MS quantitative analysis of proteins associated with lysosomes purified by LysoIP (anti-HA IP) from HEK293T cells expressing TMEM192-3xHA untreated or treated with 1 mM LLOMe for 30 min. Three groups of proteins are denoted: ESCRT components, green; autophagy factors, blue; SG components, purple. Scatter (volcano) plot shows log₂ fold changes and −Log₁₀ P values; n = 3 (see Table S1, Tab 1). Dashed line, significance cut-off (P < 0.05). (B) Protease accessibility analysis of proteins associated with purified lysosomes (LysoIP). Huh7 cells were treated with 2 mM LLOMe. LysoIP preparations (treated or not with detergent Triton X-100) were digested with 30 µg/ml proteinase K for 30 min and analyzed by immunoblotting. (C) Quantification by HCM of G3BP1 puncta. U2OS cells were treated with EBSS, 4 mM LOMe, 2 mM LLOMe, 200 µM GPN, or 400 µg/ml silica for 30 min. White masks, algorithm-defined cell boundaries (primary objects); red masks, computer-identified G3BP1 puncta (target objects). (D) Fluorescence confocal microscopy imaging of G3BP1. U2OS cells were treated with 2 mM LLOMe for 30 min and immunostained for endogenous G3BP1. Scale bar, 5 µm. (E) Quantification by HCM of G3BP1 puncta in BMM cells treated with 2 mM LLOMe or 100 µM NaAsO₂ for 2 h. Green masks, computer-identified G3BP1 puncta. (F) Quantification by HCM of G3BP1 puncta in U2OS cells treated with LLOMe at indicated doses or 100 µM NaAsO₂ in the presence or absence of 10 µg/ml cycloheximide (CHX) for 30 min. HCM images in Fig. S1 G. (G) Analysis of proteins associated with purified lysosomes (LysoIP; TMEM192-3xHA) from HEK293T cells treated with 2 mM LLOMe in the presence or absence of 10 µg/ml CHX for 30 min. TMEM192-2xFLAG, control. (H) Immunoblot analysis of eIF2α (S51) phosphorylation in BMM cells treated with 2 mM LLOMe or 100 µM NaAsO₂ for 2 h; eIF2α p-S51 quantification, n = 3. (I) Confocal microscopy analysis of G3BP1 (Alexa Fluor 488) and polyA RNA (Cy3-oligo[dT]) by FISH in U2OS cells treated with 2 mM LLOMe for 30 min. Scale bar, 5 µm. (J) HCM analysis of protein synthesis using Click-iT Plus OPP Alexa Fluor 488 Protein Synthesis Assay (Thermo Fisher Scientific) in U2OS cells treated with LLOMe at indicated doses or 100 µM NaAsO₂ or 10 µg/ml CHX for 30 min. (K) Immunoblot analysis of ATF4 and phosphorylation of 4EBP1 (S65) and eIF2α (S51) in U2OS cells treated with 2 mM LLOMe for indicated time points; quantification of ATF4 and phosphorylation of 4EBP1 (S65) and eIF2α (S51), n = 3. Ctrl, control (untreated cells). Data, means ± SEM; HCM: n ≥ 3 (each experiment: 500 valid primary objects/cells per well, ≥5 wells/sample). †, P ≥ 0.05 (not significant); *, P < 0.05; **, P < 0.01, ANOVA. See also Fig. S1. Source data are available for this figure: SourceData F1.

Figure 2.

Cellular transcriptional response during lysosomal damage. (A) RNAseq analysis of the change in gene expression (HEK293T cells) in response to treatment with 1 mM LLOMe for 30 min. Scatter (volcano) plot shows log₂ fold change and −Log₁₀ P value for the genes identified in three independent experiments (see Table S1, Tab 3). Red dots indicate the genes downregulated; green dots indicate the genes upregulated. Dashed line, significance cut-off (P < 0.05). (B) Immunoblot analysis of DUSP1 expression level and ERK2 (T185/187) phosphorylation in HEK293T cells treated with 1 mM LLOMe for 30 min. (C) Immunoblot analysis of TFEB (S142) phosphorylation in U2OS cells treated with 2 mM LLOMe for 30 min. (D) Immunoblot analysis of ERK2 (T185/187) and TFEB (S142) phosphorylation in Huh7 cells transfected with scrambled siRNA as control (SCR) or DUSP1 siRNA treated with 2 mM LLOMe for 30 min. (E) Quantification by HCM of TFEB nuclear translocation in Huh7 cells treated with or without 530 nM ERK2 inhibitor AZD6244 for 2 h followed by 2 mM LLOMe for 30 min. Blue: nuclei, Hoechst 33342. Red: anti-TFEB antibody, Alexa Fluor 568. White masks, computer-algorithm-defined cell boundaries. Pink masks, computer-identified nuclear TFEB based on the average intensity of Alexa Fluor 568 fluorescence. (F) Immunoblot analysis of ERK2 (T185/187) and TFEB (S142) phosphorylation in Huh7 cells treated with or without 530 nM ERK2 inhibitor AZD6244 for 2 h followed by 2 mM LLOMe for 30 min. Ctrl, control (untreated cells). Data, means ± SEM; HCM: n ≥ 3 (each experiment: 500 valid primary objects/cells per well, ≥5 wells/sample). **, P < 0.01, ANOVA. Source data are available for this figure: SourceData F2.

Figure S1.

SG formation during lysosomal damage. (A) Immunoblot analysis of proteins associated with purified lysosomes (LysoIP; anti-HA immunoprecipitation TMEM192^3xHA) from HEK293T cells treated with 2 mM LLOMe or 100 µM NaAsO₂ for 30 min. TMEM192^2xFLAG, control. (B) Quantification by HCM of DCP1a and G3BP1 puncta in U2OS cells treated with 2 mM LLOMe for 30 min. PB, P-body. (C) Quantification by HCM of G3BP1 puncta in Huh7 cells treated with 2 mM LLOMe for 30 min. White masks, algorithm-defined cell boundaries (primary objects); Green masks, computer-identified G3BP1 puncta (target objects). (D) Quantification by HCM of Gal3 puncta in BMM cells treated with 2 mM LLOMe or 100 µM NaAsO₂ for 2 h. Red masks, computer-identified galectin-3 puncta. (E) Quantification by HCM of TIA1 puncta in U2OS cells treated with 2 mM LLOMe for 30 min. Red masks, computer-identified TIA1 puncta. (F) Quantification by HCM of TIA1 puncta in HeLa cells treated with 4 mM LLOMe for 30 min. Red masks, computer-identified TIA1 puncta. (G) Quantification by HCM of G3BP1 and Gal3 puncta in U2OS cells treated with increasing doses of LLOMe or 100 µM NaAsO₂ in the presence or absence of 10 µg/ml cycloheximide (CHX) for 30 min. (i) HCM sample images corresponding to Fig. 1 F. Red masks, computer-identified G3BP1 puncta. (ii and ⅲ) Green masks, computer-identified Gal3 puncta and corresponding quantification in iii. (H) Immunoblot analysis of eIF2α (S51) phosphorylation in U2OS cells treated with 2 mM LLOMe for 30 min and followed by 1 h washout. (I) Quantification by HCM of G3BP1 puncta in U2OS cells treated with 2 mM LLOMe for 30 min and followed by 1 h washout. Red masks, computer-identified G3BP1 puncta. (J) Immunoblot analysis of eIF2α (S51) phosphorylation in HEK293T cells treated with 2 mM LLOMe or 100 µM NaAsO₂ for 30 min. (K) Schematic summary of the findings in Fig. 1. (L) Quantification by HCM of Lysotracker Red (LTR) and G3BP1 puncta in parental HeLa WT and Gal3^KO cells treated with 4 mM LLOMe for 30 min. Red masks, computer-identified LTR puncta. Green masks, computer-identified G3BP1 puncta. (M) Quantification by HCM of poly(A) RNA (Cy3-oligo[dT]) in U2OS cells transfected with scrambled siRNA as control (SCR) or G3BP1/2 siRNA for single or double knockdown (DKD). Cells were treated with 2 mM LLOMe for 30 min. Red masks, computer-identified poly(A) RNA puncta. Ctrl, control (untreated cells). Data, means ± SEM; HCM: n ≥ 3 (each experiment: 500 valid primary objects/cells per well, ≥5 wells/sample). †, P ≥ 0.05 (not significant); **, P < 0.01, ANOVA. See also Fig. 1. Source data are available for this figure: SourceData FS1.

Figure 3.

PKR transmits lysosomal damage signals leading to SG formation. (A) Unique PKR peptides and intensity (DIA); LysoIP, n = 3 (see Table S1, Tab 1). Mann-Whitney U test (LLOMe treatment relative to Ctrl). (B) Immunoblot analysis of the phosphorylation of eIF2α (S51) in Huh7 cells transfected with scrambled siRNA as control (SCR) or HRI, PKR, PERK, and GCN2 siRNA for knockdown (KD). Cells were treated with 2 mM LLOMe for 30 min. The level of phosphorylation of eIF2α (S51) was quantified based on three independent experiments. (C) Immunoblot analysis of PKR (T446) and eIF2α (S51) phosphorylation in U2OS cells treated with or without PKR inhibitor 2-AP for 1 h followed by 2 mM LLOMe treatment for 30 min as indicated. (D) Quantification by HCM of G3BP1 puncta in Huh7 cells transfected with scrambled siRNA as control (SCR) or HRI, PKR, PERK, and GCN2 siRNA for knockdown (KD). Cells were treated with 2 mM LLOMe for 30 min. Red masks, computer-identified G3BP1 puncta. (E) Quantification by HCM of G3BP1 puncta in U2OS cells treated with or without PKR inhibitor 2-AP for 1 h followed by 2 mM LLOMe treatment for 30 min as indicated. (F) Quantification by HCM of G3BP1 puncta in U2OS cells treated with or without 210 nM imidazolo-oxindole C16 for 2 h followed by 2 mM LLOMe for 30 min. Red masks, computer-identified G3BP1 puncta. Ctrl, control (untreated cells). Data, means ± SEM; HCM: n ≥ 3 (each experiment: 500 valid primary objects/cells per well, ≥5 wells/sample). *, P < 0.05; **, P < 0.01, ANOVA. Source data are available for this figure: SourceData F3.

Figure S2.

SGs induced by lysosomal damage show limited and dynamic interactions with lysosomes. (A) Quantification by HCM of G3BP1 puncta in Huh7 cells transfected with scrambled siRNA as control (SCR) or RNASET2 siRNA treated with 2 mM LLOMe for 30 min. Red masks, computer-identified G3BP1 puncta. (B) Immunofluorescence confocal microscopy analysis of G3BP1 and LAMP2. U2OS cells were treated with 2 mM LLOMe for 30 min and stained for endogenous G3BP1 and LAMP2. Scale bar, 5 µm. (C) Quantification by HCM of overlaps between G3BP1 and LAMP2 in U2OS cells treated with 2 mM LLOMe for 30 min. White masks, algorithm-defined cell boundaries. Yellow masks, computer-identified overlap of G3BP1 and LAMP2. (D) Quantification by HCM of overlaps between FLAG-NUFIP2 and LAMP2 in U2OS cells expressing FLAG-NUFIP2 treated with 2 mM LLOMe for 30 min. Yellow masks, computer-identified overlap of FLAG-NUFIP2 and LAMP2. (E) Still frames from live-cell fluorescence imaging analysis of mCherry-G3BP1 and GFP-LAMP1. U2OS cells expressing mCherry-G3BP1 and GFP-LAMP1 were incubated with 2 mM LLOMe during live-cell fluorescence imaging. Arrows, the representative regions at indicated timepoint. (F) Zoom views of regions in E. Ctrl, control (untreated cells). Data, means ± SEM; HCM: n ≥ 3 (each experiment: 500 valid primary objects/cells per well, ≥5 wells/sample). †, P ≥ 0.05 (not significant); **, P < 0.01, ANOVA. Source data are available for this figure: SourceData FS2.

Figure S3.

NUFIP2 exits nucleus and localizes to lysosomes upon damage and cooperates with Gal8 in mTORC1 response to lysosomal damage. (A) Immunofluorescence confocal microscopy analysis of G3BP1 and NUFIP2. Huh7 cells were treated with 2 mM LLOMe for 30 min and stained for endogenous G3BP1 and NUFIP2. Scale bar, 5 µm. (B) The NLS analysis of NUFIP2 by cNLS Mapper. The sequence in red, predicted NLS in NUFIP2, was deleted for generating NUFIP2ΔNLS. (C) Immunoblot analysis of NUFIP2 distribution in nuclear or postnuclear of Huh7 cells transfected with FLAG-NUFIP2 or NUFIP2ΔNLS after the treatment with 2 mM LLOMe for 30 min. (D and E) Confocal microscopy analysis (D) and quantification by HCM (E) of overlaps between mTOR and LAMP2 in U2OS transfected with scrambled siRNA as control (SCR) or NUFIP2 siRNA (NUFIP2KD) treated with 2 mM LLOMe for 30 min. Scale bar, 5 µm. (F) Immunoblot analysis of indicated proteins in U2OS cells transfected with scrambled siRNA as control (SCR) or NUFIP2 siRNA (NUFIP2KD) treated with 2 mM LLOMe for 30 min. The level of phosphorylation of ULK1 (S757), S6K (T389), and 4EBP1 (S65) was quantified based on three independent experiments. (G) Immunoblot validation of NUFIP2-knockout in Huh7 cells. #E7 was used in the following experiments, named as Huh7^NUFIP2-KO. (H) HCM images of Fig. 4C. Yellow masks, computer-identified overlap of mTOR and LAMP2. (I) Immunoblot analysis of indicated proteins in Huh7 cells transfected with scrambled siRNA as control (SCR) or TIA1 siRNA (TIA1^KD) treated with 2 mM LLOMe for 30 min. (J) Analysis of indicated proteins associated with lysosomes purified by anti-HA immunoprecipitation (LysoIP; TMEM192-3xHA) from HEK293T cells treated with 2 mM LLOMe for 30 min. TMEM192-2xFLAG, control. (K) Immunoblot analysis of the phosphorylation of ULK1 (S757), S6K1 (T389) and 4EBP1 (S65) in parental HeLa (WT) and TSC2-knockout HeLa cells (TSC2^KO) treated with 2 mM LLOMe for 30 min. (L) Immunoblot analysis of the phosphorylation of ULK1 (S757), S6K1 (T389), and 4EBP1 (S65) in HEK293T cells or HEK293T cells stably expressing constitutively active RagB GTPase (RagB^Q99L) treated with 2 mM LLOMe for 30 min. (M) GST pull-down assay of in vitro translated and radiolabeled Myc-tagged NUFIP2 or G3BP1 with GST or GST-tagged Gal8. AR, autoradiography. Ctrl, control (untreated cells). Data, means ± SEM; HCM: n ≥ 3 (each experiment: 500 valid primary objects/cells per well, ≥5 wells/sample). *, P < 0.05; **, P < 0.01, ANOVA. See also Figs. 4, 5, and 6. Source data are available for this figure: SourceData FS3.

Figure 4.

NUFIP2 contributes to mTOR inactivation during lysosomal damage. (A) Immunoblot analysis of FLAG-NUFIP2 or FLAG-NUFIP2ΔNLS associated with purified lysosomes (LysoIP; TMEM192-3xHA). Huh7 cells transfected with FLAG-NUFIP2 or FLAG-NUFIP2ΔNLS, treated or not with 2 mM LLOMe for 30 min. TMEM192-2xFLAG, control. (B) Quantification by HCM of G3BP1 puncta in parental Huh7 (WT) and NUFIP2-knockout Huh7 cells (NUFIP2^KO) treated with 2 mM LLOMe or 100 µM NaAsO₂ for 30 min. Red masks, computer-identified G3BP1 puncta. (C) Quantification by HCM of overlaps and confocal microscopy imaging of mTOR and LAMP2 in parental Huh7 (WT) and NUFIP2-knockout Huh7 cells (NUFIP2^KO) treated with 2 mM LLOMe for 30 min. HCM images in Fig. S3 H. Scale bar, 5 µm. (D) Immunoblot analysis of indicated proteins in parental Huh7 (WT) and NUFIP2-knockout Huh7 cells (NUFIP2^KO) treated with 2 mM LLOMe for 30 min. The level of phosphorylation of ULK1 (S757), S6K (T389), and 4EBP1 (S65) was quantified based on three independent experiments. (E) Immunoblot analysis of proteins associated with purified lysosomes (LysoIP) from HEK293T cells treated with 1 μM PP242 for 2 h or 2 mM LLOMe for 30 min or HEK293T cells stably expressing constitutively active RagB GTPase (RagB^Q99L) treated with 2 mM LLOMe for 30 min. Ctrl, control (untreated cells). Data, means ± SEM; HCM: n ≥ 3 (each experiment: 500 valid primary objects/cells per well, ≥5 wells/sample). †, P ≥ 0.05 (not significant); **, P < 0.01, ANOVA. See also Fig. S3. Source data are available for this figure: SourceData F4.

Figure 5.

Ragulator abundance and activity on damaged lysosomes is controlled by NUFIP2. (A) Summary of the quantitative changes in relevant proteins of mTORC1 signaling based on DIA LysoIP LC/MS/MS analysis. FC, fold change (see Table S1, Tab 1). (B) Immunoblot analysis of proteins associated with purified lysosomes (LysoIP; TMEM192-3xHA) from HEK293T cells treated with 2 mM LLOMe for 30 min. TMEM192-2xFLAG, control. (C ⅰ–iv) Immunoblot analysis of proteins associated with purified lysosomes (LysoIP) from parental Huh7 WT and NUFIP2-knockout Huh7 cells (NUFIP2^KO) treated with 2 mM LLOMe for 30 min (i); quantification (ii–iv), n = 3. (D) HEK293T cells stably expressing FLAG-Metap2 (control) or FLAG-LAMTOR2 transfected with scrambled siRNA (SCR) or NUFIP2 siRNA (NUFIP2^KD) were treated with 2 mM LLOMe for 30 min. Cell lysates were immunoprecipitated with anti-FLAG antibody and immunoblotted for indicated proteins. Quantification of interaction between RagA and LAMTOR2, n = 3. (E) HEK293T cells stably expressing FLAG-Metap2 (control) or FLAG-LAMTOR2 transfected with GFP or GFP-NUFIP2 were treated with 2 mM LLOMe for 30 min. Cell lysates were immunoprecipitated with anti-FLAG antibody and immunoblotted for indicated proteins. Quantification of interaction between RagA and LAMTOR2, n = 3. (F) HEK293T cells expressing FLAG (control) or FLAG-NUFIP2 were treated with 2 mM LLOMe for 30 min. Cell lysates were immunoprecipitated with anti-FLAG antibody and immunoblotted for indicated proteins. Data, means ± SEM; †, P ≥ 0.05 (not significant); *, P < 0.05; **, P < 0.01, ANOVA. See also Fig. S3. Source data are available for this figure: SourceData F5.

Figure 6.

NUFIP2 and Gal8 cooperate in mTOR response to lysosomal damage. (A) Quantification by HCM of overlaps between mTOR and LAMP2 in Gal8WT^HeLa (WT) or Gal8KO^HeLa (Gal8^KO) cells treated with 2 mM LLOMe for 30 min. Yellow masks, computer-identified overlap of mTOR and LAMP2. (B) Quantification by HCM of G3BP1 puncta in Gal8WT^HeLa (WT) or Gal8KO^HeLa (Gal8^KO) cells treated with 2 mM LLOMe for 30 min. Red masks, computer-identified G3BP1 puncta. (C) HEK293T cells stably expressing FLAG-LAMTOR2 with overexpression of GFP or GFP-NUFIP2 were transfected with scrambled siRNA as control (SCR) or Gal8 siRNA (Gal8^KD). Cells were treated with 200 µM GPN for 30 min. Cell lysates were immunoprecipitated with anti-FLAG antibody and immunoblotted for indicated proteins. (D) GST pulldown assay of in vitro translated and radiolabeled Myc-tagged NUFIP2 or G3BP1 with GST or GST-tagged Ragulator or Gal8. (E) GST pull-down assay of in vitro translated and radiolabeled Myc-tagged NUFIP2 with GST or GST-tagged Gal8. Quantification, n = 3. (F) Analysis of indicated proteins associated with lysosomes purified by anti-HA immunoprecipitation (LysoIP; TMEM192-3xHA) from HeLa WT, Gal8^KO, GABARAPs knockout (GBRP^TKO) or G3BP1 knockdown (G3BP1^KD) cells. Cells were treated with 200 µM GPN for 30 min. AR, autoradiography. Ctrl, control (untreated cells). Data, means ± SEM; HCM: n ≥ 3 (each experiment: 500 valid primary objects/cells per well, ≥5 wells/sample). †, P ≥ 0.05 (not significant); **, P < 0.01, ANOVA. See also Fig. S3. Source data are available for this figure: SourceData F6.

Figure 7.

Mammalian ATG8s participate in recruitment of NUFIP2 to damaged lysosomes. (A) Summary of the detected autophagy factors and their quantitative changes based on DIA LC/MS/MS analysis of LysoIP in HEK293T cells (see Table S1, Tab 5). FC, fold change. (B) DIA LC/MS/MS analysis of lysosomes purified by anti-HA immunoprecipitation (LysoIP; TMEM192-3xHA) from parental HeLa^WT or mATG8 knockout (hexa^KO) treated with or without 4 mM LLOMe for 30 min. Scatter (volcano; log₂ fold change and −Log₁₀ P values) plot of stress granule core proteins; n = 3 (see Table S1, Tab 9). Dashed line, significance cut-off (P < 0.05). (C) Analysis of indicated proteins associated with purified lysosomes. Lysosomes were purified by anti-HA immunoprecipitation (LysoIP; TMEM192-3xHA) from parental HeLa (WT), LC3^TKO, GBRP^TKO, and hexa^KO cells treated with 4 mM LLOMe for 30 min. (D) GST pulldown assay of in vitro translated and radiolabeled Myc-tagged NUFIP2 with GST or GST-tagged mATG8. GABARAP (GABA); GABARAPL1 (GABAL1); GABARAPL2 (GABAL2). (E) Quantification of D. Data (% binding) represents the percentage of the corresponding protein relative to its input. (F) GST pulldown assay of in vitro translated and radiolabeled Myc-tagged G3BP1 with GST or GST-tagged mATG8. GABARAP (GABA); GABARAPL1 (GABAL1); GABARAPL2 (GABAL2). (G) Quantification of F. Data (% binding) represents the percentage of the corresponding protein relative to its input. (H) Quantification of GST pulldown assay of in vitro translated and radiolabeled Myc-tagged NUFIP2 with GST or GST-tagged GABARAP deletions. Blots in Fig. S4 C. Data (% binding) represents the percentage of the corresponding protein relative to its input. (I) Quantification of GST pull-down assay of in vitro translated and radiolabeled Myc-tagged G3BP1 with GST or GST-tagged GABARAP deletions. Blots in Fig. S4 F. Data (% binding) represents the percentage of the corresponding protein relative to its input. AR, autoradiography. Data, means ± SEM; *, P < 0.05; **, P < 0.01, ANOVA. See also Fig. S4. Source data are available for this figure: SourceData F7.

Figure S4.

GABARAPs interact directly with NUFIP2 and G3BP1. (A) Immunoblot analysis of proteins associated with lysosomes purified by anti-HA immunoprecipitation (LysoIP; TMEM192-3xHA) from parental HeLa (WT), LC3^TKO, GBRP^TKO, and hexa^KO cells treated with 4 mM LLOMe for 30 min. (B) GST pulldown assay of in vitro translated and radiolabeled Myc-tagged NUFIP2 deletions with GST or GST-tagged GABARAP (GABA). (C) GST pull-down assay of in vitro translated and radiolabeled Myc-tagged NUFIP2 with GST or GST-tagged GABARAP deletions. (D) GST pulldown assay of in vitro translated and radiolabeled Myc-tagged NUFIP2/G3BP1 with GST or GST-tagged GABARAP mutants. (E) Immunoblot analysis of denaturing HA IP performed on untreated or 4 mM LLOMe treated (30 min) lysates from ATG8 knockout (hexa^KO) or ATG4/ATG8 knockout (deca^KO) HeLa cells expressing HA-GBRPL1-G. (F) GST pulldown assay of in vitro translated and radiolabeled Myc-tagged G3BP1 with GST or GST-tagged GABARAP deletions. (G) Summary of interactions between GABARAP and G3BP1. (H) GST pulldown assay of in vitro translated and radiolabeled Myc-tagged G3BP1 with GST or GST-tagged NUFIP2. (I) Quantification of Fig. S4 H. Data (% binding) represents the percentage of the corresponding protein relative to its input. (J) GST pulldown assay of in vitro translated and radiolabeled Myc-tagged NUFIP2 with GST or GST-G3BP1. (K) GST pull-down assay of in vitro translated and radiolabeled GFP-tagged NUFIP2 with GST or GST-tagged G3BP1 deletions. (L) Immunoblot analysis of interaction between NUFIP2 and G3BP1 in HEK293T cells transfected with FLAG or FLAG-NUFIP2 with 2 mM LLOMe treatment for 30 min. (M) HCM images corresponding to Fig. 8 C. White masks, algorithm-defined cell boundaries. Yellow masks, computer-identified overlap of mTOR and LAMP2. Ctrl, control (untreated cells). AR, autoradiography. Data, means ± SEM; HCM: n ≥ 3 (each experiment: 500 valid primary objects/cells per well, ≥5 wells/sample). **, P < 0.01, ANOVA. See also Figs. 7 and 8. Source data are available for this figure: SourceData FS4.

Figure 8.

GABARAPs participate in mTOR inactivation but not in eIF2α phosphorylation in response to lysosomal damage. (A) Immunoblot analysis of the phosphorylation ULK1 (S757), S6K (T389), 4EBP1 (S65), and eIF2α (S51) in parental HeLa (WT), LC3^TKO, GBRP^TKO, and hexa^KO cells treated with 4 mM LLOMe for 30 min. (B ⅰ–ⅳ) Quantification of phosphorylation of ULK1 (S757; i), S6K (T389; ii), 4EBP1 (S65; iii), and eIF2α (S51; iv) in A; Quantification, n = 3. (C) Quantification by HCM and confocal microscopy analysis of overlaps between mTOR and LAMP2 in parental HeLa (WT), LC3^TKO, GBRP^TKO, and hexa^KO cells treated with 4 mM LLOMe for 30 min. HCM images in Fig. S4 M. Scale bar, 5 µm. (D) Quantification by HCM of G3BP1 puncta. Parental HeLa (WT), LC3^TKO, GBRP^TKO, and hexa^KO cells were treated with 4 mM LLOMe for 30 min. Red masks, computer-identified G3BP1 puncta. Data, means ± SEM; HCM: n ≥ 3 (each experiment: 500 valid primary objects/cells per well, ≥5 wells/sample). †, P ≥ 0.05 (not significant); **, P < 0.01, ANOVA. See also Fig. S4. Source data are available for this figure: SourceData F8.

Figure S5.

Atg8ylation participates in mTOR inactivation in response to lysosomal damage. (A) Quantification by HCM of overlaps between mTOR and LAMP2 in HeLa (WT), GBRP^TKO, and GBRP^TKO transfected with GFP-GABARAP/GABARAPL1/GABARAPL2. Cells treated with 4 mM LLOMe for 30 min. White masks, algorithm-defined cell boundaries. Yellow masks, computer-identified overlap of mTOR and LAMP2. (B) Quantification by HCM of G3BP1 puncta in HeLa (WT), GBRP^TKO, and GBRP^TKO transfected with GFP-GABARAP/GABARAPL1/GABARAPL2. Cells were treated with 4 mM LLOMe for 30 min. Red masks, computer-identified G3BP1 puncta. (C) Immunoblot analysis of GABARAP (GABA) lipidation in U2OS cells treated with 2 mM LLOMe for indicated time points. (D) Immunofluorescence confocal microscopy imaging of GFP-GABARAP and LAMP2. U2OS cells overexpressing GFP-GABARAP were treated with 2 mM LLOMe for 30 min and stained for endogenous LAMP2. Scale bar, 5 µm. (E) WB analysis of ATG9^KO, ATG3^KO in Huh7 cells. (F) WB analysis of FIP200^KO, ATG16L1^KO in Huh7 cells. (G) WB analysis of ATG3^KO, ATG13^KO in HeLa cells. (H) HCM images of Fig. 9 A. Yellow masks, computer-identified overlap of mTOR and LAMP2. (I) HCM images of Fig. 9 C. Yellow masks, computer-identified overlap of mTOR and LAMP2. (J) HCM images of Fig. 9 E. Yellow masks, computer-identified overlap of mTOR and LAMP2. (K) WB analysis of indicated proteins in ATG3^KO, ATG16L1^KO Huh7 cells. (L) Confocal microscopy imaging (i) and quantification by HCM (ii) of overlaps between mTOR and LAMP2 in parental Huh7 (WT) and ATG3 knockout Huh7 cells (ATG3^KO) transfected with scrambled siRNA as control (SCR) or NUFIP2 siRNA (NUFIP2^KD). Cells were treated with 2 mM LLOMe for 30 min. WB analysis of indicated protein in iii. (M) WB analysis of the expression of GFP-SARS-CoV-2^ORF3a in HeLa Flp-In^TetON GFP-SARS-CoV-2^ORF3a cells induced by tetracycline (Tet) for 16 h. (N) Immunoblot analysis of interaction between GCN1 and GFP-ORF3a in HEK293T Flp-In^TetON GFP-SARS-CoV-2^ORF3a cells induced by 1 µg/ml tetracycline (Tet) for 16 h. Cell lysates were immunoprecipitated with anti-GFP antibody and immunoblotted for endogenous GCN1. (O) Immunoblot analysis of interaction between GCN1 and ORF3a in HEK293T cells transfected with GFP or GFP-ORF3a. Cell lysates were immunoprecipitated with anti-GFP antibody and immunoblotted for endogenous GCN1. Ctrl, control (untreated cells). Data, means ± SEM; HCM: n ≥ 3 (each experiment: 500 valid primary objects/cells per well, ≥5 wells/sample). †, P ≥ 0.05 (not significant); **, P < 0.01, ANOVA. See also Figs. 9 and 10. Source data are available for this figure: SourceData FS5.

Figure 9.

Atg8ylation plays a role in mTOR inhibition and competes with SG formation during lysosomal damage. (A) Quantification by HCM and confocal microscopy imaging of overlaps between mTOR and LAMP2 in parental Huh7 (WT), ATG9^KO, and ATG3^KO treated with 2 mM LLOMe for 30 min. HCM images in Fig. S5 H. Scale bar, 5 µm. (B) Quantification by HCM of G3BP1 puncta. Parental Huh7 (WT), ATG9^KO, and ATG3^KO were treated with 2 mM LLOMe for 30 min. White masks, algorithm-defined cell boundaries. Red masks, computer-identified G3BP1 puncta. (C) Quantification by HCM and confocal microscopy imaging of overlaps between mTOR and LAMP2 in parental Huh7 (WT), FIP200^KO, and ATG16L1^KO treated with 2 mM LLOMe for 30 min. HCM images in Fig. S5 I. Scale bar, 5 µm. (D) Quantification by HCM of G3BP1 puncta. Parental Huh7 (WT), FIP200^KO, and ATG16L1^KO were treated with 2 mM LLOMe for 30 min. Red masks, computer-identified G3BP1 puncta. (E) Quantification by HCM and confocal microscopy imaging of overlaps between mTOR and LAMP2 in parental HeLa (WT), ATG13^KO, and ATG3^KO treated with 4 mM LLOMe for 30 min. HCM images in Fig. S5 J. Scale bar, 5 µm. (F) Quantification by HCM of G3BP1 puncta. Parental HeLa (WT), ATG13^KO, and ATG3^KO were treated with 4 mM LLOMe for 30 min. Red masks, computer-identified G3BP1 puncta. (G) Schematic summary of the findings in this study. Ctrl, control (untreated cells). Data, means ± SEM; HCM: n ≥ 3 (each experiment: 500 valid primary objects/cells per well, ≥5 wells/sample). †, P ≥ 0.05 (not significant); **, P < 0.01, ANOVA. See also Fig. S5. Source data are available for this figure: SourceData F9.

Figure 10.

Diverse pathological agents induce lysosomal damage and SG formation response. (A) Quantification by HCM of G3BP1 and ubiquitin (Ubi) puncta in BMM cells infected with Mtb strain Erdman or its ESX-1 mutant at MOI = 10 for 20 h. White masks, algorithm-defined cell boundaries. Red masks, computer-identified G3BP1 puncta. Green masks, computer-identified ubiquitin puncta. (B) Quantification by HCM of G3BP1 and ubiquitin puncta in U2OS cells treated with FuGENE HD-coated latex beads (Polybead Amino Microsphere) for 16 h. Green masks, computer-identified G3BP1 puncta. Red masks, computer-identified ubiquitin puncta. (C) Quantification by HCM of G3BP1 and Gal3 puncta in U2OS cells treated with 1 or 10 µg/ml Tau oligomer overnight. Red masks, computer-identified G3BP1 puncta. Green masks, computer-identified Gal3 puncta. (D) Quantification by HCM of G3BP1 and ubiquitin puncta in the constructed HeLa Flp-In^TetON GFP-SARS-CoV-2^ORF3a cells induced by 1 µg/ml tetracycline (Tet). Red masks, computer-identified G3BP1 puncta. Pink masks, computer-identified ubiquitin puncta. Data, means ± SEM; HCM: n ≥ 3 (each experiment: 500 valid primary objects/cells per well, ≥5 wells/sample). †, P ≥ 0.05 (not significant); *, P < 0.05; **, P < 0.01, ANOVA. See also Fig. S5. Source data are available for this figure: SourceData F10.