Raft-like lipid microdomains drive autophagy initiation via AMBRA1-ERLIN1 molecular association within MAMs

Abstract

Mitochondria-associated membranes (MAMs) are essential communication subdomains of the endoplasmic reticulum (ER) that interact with mitochondria. We previously demonstrated that, upon macroautophagy/autophagy induction, AMBRA1 is recruited to the BECN1 complex and relocalizes to MAMs, where it regulates autophagy by interacting with raft-like components. ERLIN1 is an endoplasmic reticulum lipid raft protein of the prohibitin family. However, little is known about its association with the MAM interface and its involvement in autophagic initiation. In this study, we investigated ERLIN1 association with MAM raft-like microdomains and its interaction with AMBRA1 in the regulation of the autophagic process. We show that ERLIN1 interacts with AMBRA1 at MAM raft-like microdomains, which represents an essential condition for autophagosome formation upon nutrient starvation, as demonstrated by knocking down ERLIN1 gene expression. Moreover, this interaction depends on the "integrity" of key molecules, such as ganglioside GD3 and MFN2. Indeed, knocking down ST8SIA1/GD3-synthase or MFN2 expression impairs AMBRA1-ERLIN1 interaction at the MAM level and hinders autophagy. In conclusion, AMBRA1-ERLIN1 interaction within MAM raft-like microdomains appears to be pivotal in promoting the formation of autophagosomes.Abbreviations: ACSL4/ACS4: acyl-CoA synthetase long chain family member 4; ACTB/β-actin: actin beta; AMBRA1: autophagy and beclin 1 regulator 1; ATG14: autophagy related 14; BECN1: beclin 1; CANX: calnexin; Cy5: cyanine 5; ECL: enhanced chemiluminescence; ER: endoplasmic reticulum; ERLIN1/KE04: ER lipid raft associated 1; FB1: fumonisin B1; FE: FRET efficiency; FRET: Förster/fluorescence resonance energy transfer; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; GD3: aNeu5Ac(2-8)aNeu5Ac(2-3)bDGalp(1-4)bDGlcp(1-1)ceramide; HBSS: Hanks' balanced salt solution; HRP: horseradish peroxidase; LMNB1: lamin B1; mAb: monoclonal antibody; MAMs: mitochondria-associated membranes; MAP1LC3B/LC3: microtubule associated protein 1 light chain 3 beta; MFN2: mitofusin 2; MTOR: mechanistic target of rapamycin kinase; MYC/cMyc: proto-oncogene, bHLH transcription factor; P4HB: prolyl 4-hydroxylase subunit beta; pAb: polyclonal antibody; PE: phycoerythrin; SCAP/SREBP: SREBF chaperone; SD: standard deviation; ST8SIA1: ST8 alpha-N-acetyl-neuraminide alpha-2,8 sialyltransferase 1; SQSTM1/p62: sequestosome 1; TOMM20: translocase of outer mitochondrial membrane 20; TUBB/beta-tubulin: tubulin beta class I; ULK1: unc-51 like autophagy activating kinase 1; VDAC1/porin: voltage dependent anion channel 1.

Keywords: AMBRA1; ERLIN1; autophagy; lipid rafts; mitochondria associated membranes.

Figure 1.

ERLIN1 distribution in MAM fractions from stimulated HBSS cells. (A) Protein components of subcellular fractions prepared from human 2FTGH (2F) fibroblasts cells, either untreated or treated with HBSS for 1 h, were detected by immunoblot analysis. MAM: mitochondria-associated membrane; Pmit: pure mitochondrial fraction; Cmit: crude mitochondrial fraction; ER; Cyt: cytosol; Nuc: nuclear fraction; Tot: total extracts. All fractions were analyzed by western blot using an anti-ERLIN1 mAb. The purity of all subcellular fractions was also tested by western blotting for the presence of specific markers: ACSL4 for MAMs, CANX for MAMs and ER, VDAC1 for mitochondria and MAMs, TOMM20 for mitochondria, TUBB/beta-tubulin for cytosol, LMNB1 for nuclei and P4HB for ER. Autophagy was checked by western blot analysis using rabbit anti-LC3 pAb or rabbit anti-SQSTM1 mAb. Loading control was evaluated using anti-ACTB mAb. A representative experiment among 3 is shown. Bar graph on the right shows densitometric analysis. Results represent the mean ± SD from 3 independent experiments. ***p ≤ 0.001, HBSS vs control cells. (B) Autophagy evaluation by western blot analysis using rabbit anti-LC3 pAb or rabbit anti-SQSTM1 mAb (first panel). Loading control was evaluated using anti-ACTB mAb. In the second panel, a bar graph showing densitometric analysis from 3 independent experiments and reported as the mean ± SD. ***p ≤ 0.001 HBSS vs control cells. Third and fourth panels show flow cytometry analysis of autophagy performed with a cyto-ID autophagy detection kit. Numbers represent the mean ± SD of the median fluorescence intensity obtained in three independent experiments. (C) Autophagy evaluation by flow cytometry (first to third panel) o by fluorescence microscopy (pictures on the right) after dual cell staining with anti-LC3 and anti-SQSTM1. For fluorescence microscopy, cells were also counterstained with Hoechst. All these different methods of detecting autophagy produced completely overlapping results

Figure 2.

AMBRA1-ERLIN1 association during autophagy induction in 2F cells. (A) 2F cells, untreated or treated with HBSS for 1 h, were lysed in lysis buffer, followed by immunoprecipitation with rabbit anti-AMBRA1 pAb. A rabbit IgG isotypic control (IpCtr) was employed. The immunoprecipitates were analyzed for the presence of ERLIN1 by western blot analysis, using anti-ERLIN1 mAb. A representative experiment among 3 is shown. As a control, the immunoprecipitates were assessed by immunoblot with anti-AMBRA1 mAb. Bar graph in the right panel shows densitometric analysis. Results represent the mean ± SD from 3 independent experiments. ***p ≤ 0.001 HBSS vs control cells. Autophagy was checked by western blot analysis using rabbit anti-LC3 pAb or rabbit anti-SQSTM1 mAb. Loading control was evaluated using anti-ACTB mAb. Results represent the mean ± SD from 3 independent experiments. ***p ≤ 0.001 HBSS vs control cells. (B) Quantitative evaluation of AMBRA1-ERLIN1 association by FRET technique as revealed by flow cytometry analysis. Numbers in the first and second panels indicate the percentage of FL3-positive events (FRET channel), obtained in one experiment representative of three. Third panels. Bar graphs showing the FRET efficiency, calculated according to the Riemann’s algorithm, of AMBRA1 and ERLIN1 molecular association. Data are reported as mean ± SD from three independent experiments. Fourth panels. Flow cytometry analysis of autophagy in Control (full gray curves) and HBSS treated cells (empty curves), with or without FB1, performed with a Cyto-ID Autophagy Detection kit. Numbers represent the mean ± SD of the median fluorescence intensity obtained in three independent experiments. A representative experiment among three is shown. In the fifth panel bar graphs showing the mean ± SD obtained from three independent experiments. **p ≤ 0.01 vs control

Figure 3.

AMBRA1-ERLIN1association during autophagy induction in 2F cells. (A) 2FTGH fibroblasts were treated with HBSS (Sigma, H9269) for 1, 2, 3 and 4 h. We considered the same cells treated with 50 nM rapamycin or with 100 nM torin 1 for 4 h as positive controls. At the end of treatment cells were harvested and analyzed by flow cytometry after single staining with Cyto-ID autophagy detection kit (left panel). In upper panels, the results obtained in a representative experiment are shown. Numbers represent the median fluorescence intensity values of the histograms. Bar graph in lower panel reports the mean ± SD of the results obtained in three independent experiments. (B) Increased colocalization between AMBRA1, ERLIN1 and a specific mitochondrial tracker upon autophagy induction. Untreated or HBSS-treated 2F cells were stained with MitoTracker (green), anti-AMBRA1 (red) and anti-ERLIN1 (blue) antibodies. Nuclei were stained with DAPI (Turquoise). To note, overlapping areas resulting from green, red and blue fluorescence in merge micrograph (see arrows in magnification of the boxed areas) indicate that colocalization of ERLIN1-AMBRA1-MitoTracker increases in cells treated with HBSS as compared to untreated cells. Images were acquired using a LSM 900, Airyscan SR Zeiss confocal microscopy and the co-localization between ERLIN1, AMBRA1 and MitoTracker was measured using the ZEN 3.0 Blue edition software and expressed as μm² per cell. A minimum of 30 cells/sample was analyzed, and the statistical analysis was performed using Student’s t-test, *p ≤ 0.05 (right panel). Scale bar: 10 μm

Figure 4.

Knocking down ERLIN1 expression hinders autophagy. (A) Cytofluorimetric and western blot evaluation of ERLIN1 expression level 48 h after specific siRNA transfection (#1 + #2). Representative experiments are shown. Numbers in the left panel represent the median fluorescence intensity. Bar graph showing the mean ± SD among 3 independent cytofluorimetric evaluations. (B) In the left panels, semiquantitative flow cytometry analysis performed with a Cyto-ID Autophagy detection kit of autophagy induced by HBSS in 2F fibroblasts knocked down for ERLIN1 or in cells transfected with control non-silencing siRNA. Numbers represent the median fluorescence intensity. Bar graph on the right shows the mean ± SD among 3 independent experiments. Pictures on the right show immunofluorescence analysis after cell staining with anti-LC3 (red) and anti-SQSTM1 (green) and Hoechst (blue). (C) Representative fluorescence images of Flag mATG16L1 expression in iERLIN1 2F cells subjected, or not, to nutrient deprivation, adding HBSS for 1 or 2 h. Scale bar: 10 μm. (D) Western blotting evaluation of ERLIN1 downregulation in Flag mATG16L1 expressing 2F cells. (E) Quantitative analysis of ATG16L1 dots area (μm²) per cell reported as the mean ± SD, *P ≤ 0.05, ****P ≤ 0.0001 using ANOVA 2-way test for repeated samples

Figure 5.

Autophagy flux evaluation in iERLIN1 cells. ERLIN1 expression was downregulated by siRNA using two different oligonucleotides in 2F cells and qRT-PCR (A) and western blotting (B) were performed to evaluate ERLIN1 mRNA levels from three independent experiments; *p ≤ 0.05 and ****p ≤ 0.0001 using Student’s t-test. (C) iERLIN1 cells were starved for the indicated time points and incubated with the lysosome inhibitor, bafilomycin A₁ or vehicle for 1 h before lysis. LC3 lipidation and SQSTM1 were detected by immunoblotting using specific antibodies; GAPDH was incubated as a loading control. The graph reports mean ± SD of LC3-II:GAPDH and SQSTM1:GAPDH from three independent experiments; *p ≤ 0.05, **p ≤ 0.01 using ANOVA 2-way test for repeated samples. (D) Western blotting evaluation of ERLIN1 downregulation in GFP-RFP-LC3-expressing 2F cells. (E) Representative fluorescence images of LC3 puncta in GFP-RFP-LC3-expressing iERLIN1 2F cells, subjected or not to autophagy induction by nutrient deprivation for 2 h (HBSS). The upper panel reports the amount of autophagosomes per cell measured as the mean area ± SD of RFP-GFP colocalizing puncta (μm²/cell); the lower panel reports the amount of autolysosomes reported as the mean area ± SD of RFP-only puncta (μm²/cell). **P ≤ 0.01, ****P ≤ 0.0001 using ANOVA 2-way test for repeated samples. Scale bar: 10 μm

Figure 6.

ERLIN1 downregulation sensitizes 2F cells to cisplatin-induced cell death. Analysis of cell death in 2F cells downregulated for ERLIN1 expression in response cisplatin (CDDP) treatment for 4 or 8 μM (48 h) monitored by (A) western blotting of cleaved PARP1, SQSTM1 and ERLIN1 and (B) densitometric analysis (bar graph in the right panel (n = 4, p-value **≤0.01, ***≤0.001 ****≤0.0001, *P ≤ 0.05, ****P ≤ 0.0001 using ANOVA 2-way test for repeated samples) or (C) by PI-ANXA5 incorporation by FACS; representative dot plots of ANXA5-propidium iodide by flow cytometry are showed; (D) Quantitative analysis of PI-ANXA5 incorporation. (n = 3, **p-value ≤ 0.01, ***p ≤ 0.001 ****p ≤ 0.0001 *P ≤ 0.05, ****P ≤ 0.0001 using ANOVA 2-way test for repeated samples)

Figure 7.

AMBRA1 interacts with ERLIN1 in AMBRA1 FL/F1-3 fragments overexpressed cells during autophagy. (A) 2F cells were coinfected with retroviral vectors encoding MYC-tagged AMBRA1 proteins or MYC-tagged beta-gal as a negative control; FL, full length; F1–3, fragments 1–3. (B) Cells, untreated or treated with HBSS for 1 h were lysed in lysis buffer. Protein extracts were subjected to IP using an anti-MYC antibody (IP MYC). Purified complexes were analyzed by WB using an anti-ERLIN1 (top) or anti-MYC (bottom) antibodies. A representative experiment among 3 is shown. Bar graph in the right panel shows densitometric analysis. Results represent the mean ± SD from 3 independent experiments. ***p ≤ 0.001 and ****p ≤ 0.0001. (C) A rabbit IgG isotypic control (IpCtr) was employed. WB was performed using an anti-ERLIN1 (top) or anti-MYC (bottom) antibodies; on right WB of the AMBRA1 mutants with the corresponding amino acid sequence boundary from total extracts is showed. A representative experiment among 3 is shown. (D) Modulation of autophagy by AMBRA1 FL/F1-3 fragments expressed cells. FL, or F1-3 untreated or treated cells with HBSS for 1 h were lysed in lysis buffer and the occurrence of autophagy was analyzed by LC3-I to LC3-II conversion using rabbit anti-LC3 pAb or by degradation of SQSTM1 using rabbit anti-SQSTM1 mAb. Loading control was evaluated using anti-ACTB mAb. Densitometric analysis of the band density ratio of LC3-II relative to ACTB is reported. A representative experiment among 3 is shown. Bar graph in the right panel shows densitometric analysis. Results represent the mean ± SD from 3 independent experiments. ***p ≤ 0.001 and ****p ≤ 0.0001, HBSS vs control cells. (E) shBECN1 cells and shRNA control cells were transfected with an expression vector encoding MYC-tagged AMBRA1 and cultured in nutrient-rich conditions or nutrient starvation for 1 h. AMBRA1 was immunoprecipitated using an anti-MYC antibody. Immunocomplexes and total extracts (input) were analyzed by western blot using anti AMBRA1 and ERLIN1 antibodies. HSP90AA1 levels were analyzed in total extracts as loading control

Figure 8.

Involvement of lipid raft-associated to MAMs on AMBRA1-ERLIN1 interaction. (A) Upper panel, Crude mitochondrial fractions obtained from F2 cells, either untreated or treated with HBSS for 1 h, were subjected to Percoll gradient fractionation. After centrifugation, high-purity MAM fractions were obtained. Isolated MAMs and pure mitochondrial fractions (Pmit) were subjected to IP using an anti-ERLIN1 pAb. A rabbit IgG isotypic control was employed (IpCtr). Immunoprecipitates were analyzed by western blot analysis, using anti-MFN2, or with anti AMBRA1. As a control, the immunoprecipitates were assessed by immunoblot with anti-ERLIN1 mAb. A representative experiment among 3 is shown. Bar graph shows densitometric analysis. ***p ≤ 0.001 HBSS vs control cells. Lower panel. Modulation of autophagy was analyzed by LC3-I to LC3-II conversion using rabbit anti-LC3 pAb or by degradation of SQSTM1 using rabbit anti-SQSTM1 mAb. Loading control was evaluated using anti-ACTB mAb. A representative experiment among 3 is shown. A representative experiment among 3 is shown. Bar graph shows densitometric analysis. ****p ≤ 0.0001 HBSS vs control cells. (B) In parallel, the immunoprecipitates were spotted onto nitrocellulose strips and incubated with anti-GD3 R24 mAb, anti-MFN2 or with anti-AMBRA as described in Materials and Methods. A rabbit IgG isotypic control was employed. As a control, the immunoprecipitates were checked using anti-ERLIN1 mAb. A representative experiment among 3 is shown. A positive control was obtained using pure standard GD3 (STD). A representative experiment among 3 is shown. Bar graph shows densitometric analysis. **p ≤ 0.01 HBSS vs control cells

Figure 9.

Knocking down MFN2, ST8SIA1 or ERLIN1 expression impairs the interaction AMBRA1-ERLIN1 and AMBRA1-MFN2, hindering autophagy. (A) Cytofluorimetric and western blot evaluation of MFN2 (left panel) and ST8SIA1 (right panel) expression level 48 h after specific siRNA transfection. Numbers in the upper panels represent the median fluorescence intensity. Bar graph on the right shows the mean ± SD among 3 independent cytofluorimetric experiments. (B) FRET analysis of AMBRA1-ERLIN1 association in cells knocked down for MFN2 (middle) or ST8SIA1 (right), or in cells transfected with control siRNA (left panel), treated or not with HBSS for 1 h. Numbers indicate the percentage of FL3 (FRET channel)-positive events obtained in one experiment representative of 3. Bar graphs on the right show the evaluation of FE, according to the Riemann algorithm. Results represent the mean ± SD from 3 independent experiments. (C) Semiquantitative analysis of autophagy, performed by flow cytometry with a Cyto-ID Autophagy detection kit and by western blot evaluating LC3-I to LC3-II conversion using rabbit anti-LC3 pAb, in cells knocked down for MFN2 and ST8SIA1 or in cells transfected with scrambled siRNA. Numbers in the upper panels represent the median fluorescence intensity. Upper bar graph shows the mean ± SD among 3 independent cytofluorimetric experiments. Bottom bar graph shows densitometric analysis considering 3 different experiments. **p ≤ 0.01 between indicated samples. (D) FRET analysis of AMBRA1-MFN2 association in AMBRA1 FL knocked down for ERLIN1 and treated with HBSS for 1 h or in cells transfected with control non-silencing siRNA and treated with HBSS for 1 h. Numbers indicate the percentage of FL3-positive events obtained in one experiment representative of 3. Bar graph on the right shows the evaluation of FE, according to the Riemann algorithm. Results represent the mean ± SD from 3 independent experiments. **p ≤ 0.01 between indicated samples