Restricted mitochondrial protein acetylation initiates mitochondrial autophagy

Abstract

Because nutrient-sensing nuclear and cytosolic acetylation mediates cellular autophagy, we investigated whether mitochondrial acetylation modulates mitochondrial autophagy (mitophagy). Knockdown of GCN5L1, a component of the mitochondrial acetyltransferase machinery, diminished mitochondrial protein acetylation and augmented mitochondrial enrichment of autophagy mediators. This program was disrupted by SIRT3 knockdown. Chronic GCN5L1 depletion increased mitochondrial turnover and reduced mitochondrial protein content and/or mass. In parallel, mitochondria showed blunted respiration and enhanced 'stress-resilience'. Genetic disruption of autophagy mediators Atg5 and p62 (also known as SQSTM1), as well as GCN5L1 reconstitution, abolished deacetylation-induced mitochondrial autophagy. Interestingly, this program is independent of the mitophagy E3-ligase Parkin (also known as PARK2). Taken together, these data suggest that deacetylation of mitochondrial proteins initiates mitochondrial autophagy in a canonical autophagy-mediator-dependent program and shows that modulation of this regulatory program has ameliorative mitochondrial homeostatic effects.

Keywords: Acetylation; GCN5L1; Mitochondrial autophagy; Parkin; SIRT3.

Fig. 1.

Depletion of GCN5L1 leads to mitochondrial accumulation of autophagy factors. (A) Western blots of isolated mitochondria from control (C), GCN5L1 (G) and SIRT3 (S) siRNA HepG2 cells with antibodies directed against p62, LC3, SIRT3, GCN5L1 and ubiquitylation (Ub) with VDAC as a loading control. (B) Relative mitochondrial protein ubiquitylation levels (representative blot shown in A) in the three groups. (C) Confocal microscopy in control or GCN5L1 siRNA cells showing colocalization of ds-Red-labeled mitochondria with GFP–LC3. Scale bars: 10 µm. (D) 2D fluorograms showing colocalization of ds-Red-Mito and GFP–LC3 as a distribution of pairs of pixel intensities (with greater diagonal alignment correlating to higher colocalization). (E) Quantification of the colocalization coefficient between ds-Red-Mito and LC3, p62, ubiquitin and Lamp1 displayed as Pearson coefficients in the colocalized volume (1, perfect correlation; 0, no correlation; –1, perfect inverse correlation). (F) Mitochondrial accumulation of autophagy mediators in response to the inhibition of lysosomal function by chloroquine. (G,H) Representative western blot analysis of concurrent SIRT3 and GCN5L1 siRNA on p62 and LC3-II mitochondrial accumulation in HepG2 cells. Protein levels are relative to VDAC. Control samples were normalized to 1, with KD levels determined relative to control values. Data are expressed as the mean±s.e.m. (n = 5 replicates per group). *P<0.05 versus scrambled siRNA control, ^#P<0.05 versus GCN5L1 siRNA levels.

Fig. 2.

Chronic GCN5L1 KD attenuated mitochondrial mass and protein levels via autophagic degradation. (A) Representative immunoblot of mitochondrial protein acetylation and the accumulation of autophagy mediators in isolated mitochondria following chronic lentiviral GCN5L1 (G) or control (C) shRNA expression in HepG2 cells. (B) Quantification of mitochondrial protein acetylated at lysine residues and ubiquitin (Ub) levels in control and GCN5L1 shRNA HepG2 cells. (C) Relative mitochondrial mass in lentiviral-infected HepG2 cells. (D) Representative immunoblot of whole-cell levels of the mitochondrial proteins glutamate dehydrogenase (GDH), ATP5a, NDUFA9 (complex I), VDAC and TOM20 following expression of GCN5L1 shRNA. (E) Quantification of mitochondrial proteins levels assayed in whole-cell preparations from control and GCN5L1 shRNA cells (F) Quantification of mKeima acidification as a measure of mitophagy in shRNA-infected cells (n = 3) relative to the untreated control (set at 1). (G) Histogram depicting the relative rate of incorporation of the heavy label into mitochondrial proteins in GCN5L1 shRNA cells to that in control shRNA HepG2 cells 48 hours after label introduction. (H) Representative immunoblot of whole-cell levels of mitochondrial proteins following lentiviral GCN5L1 shRNA expression in response to MG132 and chloroquine. (I) Quantification of mitochondrial protein levels relative to tubulin normalized to the control shRNA samples (set at 1). Data are expressed as the mean±s.e.m. with n≥4 experiments per study. *P<0.05 versus shRNA control group, **P<0.01 versus controls.

Fig. 3.

GCN5L1 depletion initiated mitophagy is dependent on Atg5 and p62 but not Parkin. (A,B) Representative mitochondrial protein immunoblots showing accumulation of autophagy mediators p62, LC3 and ubiquitylation in isolated mitochondria from wild-type (WT), Atg5 and p62 KO MEFs following GCN5L1 (G) shRNA and control (C) vector infection. (C,D) Representative whole-cell immunoblots showing mitochondrial proteins ATP5a and NDUFA9 in WT (C) and Atg5 KO (D) MEFs following shRNA infection. (E) Relative mitochondrial mass as measured by mitotracker green fluorescence in WT and Atg5 KO MEFs (n = 7). (F) Representative immunoblot of isolated mitochondria showing mitochondrial enrichment of p62, LC3-II and ubiquitin (Ub) after GCN5L1 shRNA infection in Parkin KO MEFs. (G) Quantification of mitochondrial protein ubiquitylation in Parkin KO MEFs in response to GCN5L1 KD. (H) Relative mitochondrial mass in control versus GCN5L1 shRNA in WT and Parkin KO MEFs (n = 3). Data are expressed as mean±s.e.m., **P<0.01.

Fig. 4.

GCN5L1 KD alters mitochondrial function and morphology. (A,B) Basal and maximal [dinitrophenol (100 µM)] oxygen consumption and glycolysis rates in response to chronic GCN5L1 KD in HepG2 cells (n = 3). (C) Cytometric analysis with TMRM to assess relative mitochondrial membrane potential in control and GCN5L1 shRNA HepG2 cells (n = 3). (D) Mitochondrial superoxide generation under basal conditions (control), GCN5L1 shRNA cells and in response to the inhibition of complex I of the electron transfer chain by rotenone (15 µM, 4 hour) relative to control levels (set at 1). (E) Assessment of the relative ionomycin-mediated mitochondrial permeability transition. The addition of cyclosporin A (2 µM, 30 minutes) attenuated this transition. (F) Paraquat-mediated (1.5 mM, 24 hours) cell death in control and GCN5L1 shRNA HepG2 cells. (G) Representative immunoblot of isolated mitochondrial proteins from wild-type (WT) and GCN5L1 KO MEFs to assess reversal of mitochondrial accumulation of autophagy proteins and restoration of acetylation by reconstitution of GCN5L1. (H) Quantification of mitochondrial protein acetylation in KO MEFs in response to transient transfect of GCN5L1. Data are expressed as mean±s.e.m. *P<0.05, **P<0.01 versus respective controls. (I) Representative electron micrograph of a GCN5L1 KO MEFs to visualize mitochondrial inclusions in autophagosome vacuoles and mitochondrial morphology. Scale bar: 200 nm.