Nucleocytosolic depletion of the energy metabolite acetyl-coenzyme a stimulates autophagy and prolongs lifespan

Abstract

Healthy aging depends on removal of damaged cellular material that is in part mediated by autophagy. The nutritional status of cells affects both aging and autophagy through as-yet-elusive metabolic circuitries. Here, we show that nucleocytosolic acetyl-coenzyme A (AcCoA) production is a metabolic repressor of autophagy during aging in yeast. Blocking the mitochondrial route to AcCoA by deletion of the CoA-transferase ACH1 caused cytosolic accumulation of the AcCoA precursor acetate. This led to hyperactivation of nucleocytosolic AcCoA-synthetase Acs2p, triggering histone acetylation, repression of autophagy genes, and an age-dependent defect in autophagic flux, culminating in a reduced lifespan. Inhibition of nutrient signaling failed to restore, while simultaneous knockdown of ACS2 reinstated, autophagy and survival of ach1 mutant. Brain-specific knockdown of Drosophila AcCoA synthetase was sufficient to enhance autophagic protein clearance and prolong lifespan. Since AcCoA integrates various nutrition pathways, our findings may explain diet-dependent lifespan and autophagy regulation.

Graphical abstract

Figure 1

ACS2 Depletion Ameliorates Age-Associated Autophagy in Yeast (A) Scheme of known major acetyl-CoA (AcCoA)-producing enzymes (bold characters) modulated in this study. (B) Representative immunoblot of GFP-Atg8p expressing yeast. Wild-type cells (tet-WT) were compared to strains carrying a doxycycline-repressible tet-O7 promoter controlling ACS2 transcription (tet-ACS2). Cells were grown in SC 2% glucose medium for 24 hr (day 1 of aging) in the presence (+Doxy) or absence (−Doxy) of 1 ng/ml doxycycline to induce knockdown of ACS2 (see also Figure S1). (C) Fluorescence microscopy of GFP-Atg8p expressing (under control of its natural pATG8 promoter) wild-type (tet-WT) and ACS2 knockdown (tet-ACS) cells grown in the presence of 1 ng/ml doxycycline (as shown in A) and chronologically aged for 3 days. Propidium iodide (PI) counterstaining served to visualize dead cells. Scale bars represent 5 μm. (D) Quantification of cells depicted in (C) (day 3) and of young (day 1) cells with 150–300 counts (blinded) for each replicate. Autophagic cells were defined as cells with clear vacuolar GFP fluorescence. Data represent means ± SEM (n = 4). ^∗∗p < 0.01. (E) Representative immunoblot analysis of day 1 (young controls) and day 3 (aged) cells shown in (C) and (D) using anti-GFP and anti-GAPDH (loading control) antibodies to detect “free-GFP” indicative of autophagic flux. (F–H) Representative micrographs (F), respective quantification (G), and immunoblot analysis (H) of wild-type (WT) and ACS1-deleted (Δacs1) yeast aged to day 3 compared to young (day 1) cells expressing GFP-Atg8p chimera as in (C)–(E). Data represent means ± SEM (n = 4).

Figure 2

Mitochondrial AcCoA Production by Ach1p- or Mpc1p-Associated Pathways Is Required for Autophagy during Aging (A) Fluorescence microscopy of GFP-Atg8p expressing wild-type (WT) and ACH1-deleted (Δach1) yeast grown to day 1 (young) and aged for 3 days in SC 2% glucose medium; propidium iodide (PI) counterstaining served to visualize dead cells. Scale bars represent 5 μm. (B) Quantification of cells depicted in (A) with 150–300 counts (blinded) for each replicate. Autophagic cells were defined as cells with clear vacuolar GFP fluorescence. Data represent means ± SEM (n = 4). ^∗∗∗p < 0.001. (C) Representative immunoblot analysis of cells shown in (A) and further aged to day 5 using anti-GFP and anti-GAPDH (loading control) antibodies to detect “free-GFP” indicative of autophagic flux (see also Figure S2). (D–F) Representative micrographs (D), respective quantification (E), and immunoblot analysis (F) of wild-type (WT) and MPC1 deleted (Δmpc1) yeast expressing GFP-Atg8p chimera as in (A)–(C) and aged to indicated time points. Data represent means ± SEM (n = 4). ^∗∗∗p < 0.001.

Figure 3

Deletion of ACH1 or MPC1 Shortens Chronological Lifespan and Leads to Excess Release of Cellular Acetate (A–C) Chronological lifespan (CLS) analyses in SC 2% glucose medium of wild-type (WT) cells compared to Δacs1 (A), Δmpc1 (B), or Δach1 (C) cells (see also Figure S3). Survival was determined by colony-forming capacity (clonogenicity). Data represent means ± SEM (n = 4) of a representative aging experiment. (D–F) Propidium iodide (PI)-positive cells analyzed by flow cytometry to quantify age-induced cell death of experiments shown in (A)–(C). Data represent means ± SEM (n = 4) (see also Figure S4). (G–I) Extracellular acetate assessed from crude culture supernatants obtained at indicated time points of experiments shown in (A)–(C). Data represent means ± SEM (n = 4). ^∗p < 0.05 and ^∗∗∗p < 0.001.

Figure 4

Deletion of ACH1 or MPC1 Results in Upregulation of the Nucleocytosolic Acs2p Pathway, Causing Histone Hyperacetylation (A and B) Representative immunoblot (A) and densitometric quantification expressed as normalized Acs2p/GAPDH ratios (B) of protein extracts from wild-type (WT) and Δach1 yeast chronologically aged to indicated time points in SC 2% glucose medium (see also Figure S4). (C) Representative immunoblot analysis of 3-day-old WT and Δach1 cells similar to (A) using pan-acetyl-lysine antibodies. Crude protein extracts (Input) obtained in the presence of histone deacetylase and sirtuin inhibitors (see Supplemental Experimental Procedures) were subjected to immunoprecipitation (IP) at indicated protein concentrations using pan-acetyl-lysine antibodies to enrich acetylated proteins (Ac-Lys IP). GAPDH served as loading control. (D) Heatmap of protein ratios (Δach1 versus WT, log2 scale) corresponding to indicated genes. Data were obtained from mass spectrometric analyses of acetylated proteins enriched by IP similar to (C, 1 mg/ml protein). Extracts from stable-isotope (SILAC)-labeled cells were mixed prior to IP (prelysate), and mean SILAC-protein ratios after IP were normalized to ratios of prelysates to correct for changes in general protein abundance. Two biological replicates were performed (Rep. 1 and Rep. 2) with a regression coefficient of 0.76 of observed protein ratios. Heatmap represents identified proteins with reproducible 1.5-fold upregulation (red color) or downregulation (green color). (E and F) Representative immunoblots of whole cell acid extracts of wild-type (WT) and Δach1 (E) cells chronologically aged to designated time points. Blots were probed with antibodies against total histone H3 (loading control) or H3 acetylated lysines (K9Ac, K14Ac, K18Ac). Densitometric quantification (F) of relative acetylation was calculated as Ac-K/total H3 ratios normalized to ratios of WT at day 1. Data represent means ± SEM (n = 7–8) (see also Figure S5). (G and H) Survival analyses in SC 1.25% galactose/0.75% glucose medium of wild-type cells ectopically overexpressing ACS2 (ACS2 overexp.) compared to vector control cells (Vector control). Survival (I) was determined by colony-forming capacity (clonogenicity) and cell death (J) assessed by propidium iodide (PI)-positive cells analyzed by flow cytometry. Data represent means ± SEM (n = 4) of a representative aging experiment. (I and J) Autophagic flux determination by vacuolar protease-dependent GFP liberation after ACS2 overexpression compared to empty vector controls (Control) similar to (G) and (H). Representative immunoblots at indicated time points (G) and quantification by densitometry of free-GFP/GFP-Atg8p signal ratios (H) are shown. Data represent means ± SEM (n = 4). ^∗p < 0.05, ^∗∗p < 0.01, and ^∗∗∗p < 0.001.

Figure 5

Acs2p Controls ATG7 Transcription through Epigenetic Histone Acetylation (A) ATG7 mRNA levels by quantitative reverse-transcriptase PCR (RT-qPCR) of wild-type and Δach1 cells aged to day 3 (see also Figure S5). Rel. mRNA levels are expressed as ratios of 18S rRNA normalized to wild-type cells by ΔΔCt-method. Data represent means ± SEM (n = 8). (B) Representative immunoblot analysis from wild-type (WT) and ACH1-deleted (Δach1) yeast expressing chromosomally tagged ATG7 by C-terminal 6HA fusion aged to indicated time points. Blots were probed with anti-HA and anti-GAPDH (loading control) antibodies. (C and D) Representative immunoblots (C) and densitometric quantification (D) of whole-cell acid extracts of wild-type and Δach1 cells combined with or without knockdown of ACS2 (tet-ACS2). Cells were chronologically aged to day 3 in the presence of 1 ng/ml doxycycline. Blots were probed with antibodies against total histone H3 (loading control) or H3 acetylated lysines (K9Ac, K14Ac, K18Ac). Data represent means ± SEM (n = 8). (E and F) ATG7 mRNA levels (E) by RT-qPCR and representative immunoblot analysis (F) of wild-type and Δach1 cells with or without knockdown of ACS2 (tet-ACS2) as in (C) and (E) aged to day 3. Rel. mRNA levels (E) are expressed as ratios to 18S rRNA normalized to wild-type cells by ΔΔCt method. Data represent means ± SEM (n = 7–8). (G and H) Representative immunoblot (H) and densitometric quantification (G) of histone H3 wild-type (H3-wt/wt) and H3 mutated (H3-K14,18Q/-K14,18R) strains carrying the GFP-Atg8p fusion to calculate “free-GFP/GFP-Atg8p Ratio” indicative of autophagic flux. Data represent means ± SEM (n = 4) (for strain details and supplemental data, see Figures S5D–S5F). ^∗p < 0.05, ^∗∗p < 0.01, and ^∗∗∗p < 0.001; n.s., not significant.

Figure 6

Knockdown of Acs2p Cures the Autophagy Defect in ach1 Mutants (A and B) Representative micrographs (A) and respective quantification (B) of wild-type and Δach1 cells expressing GFP-Atg8p chimera combined with or without knockdown of ACS2 (tet-ACS2). Cells were chronologically aged to day 3 in the presence of 1 ng/ml doxycycline and PI counterstained prior to epifluorescence microscopy (see also Figure S6). (C) Representative immunoblot analysis of cells shown in (A) aged until days 1, 2, and 3 to detect “free-GFP” indicative of autophagic flux (see Figure S6A for quantification). (D and E) Representative immunoblot analyses of GFP-Atg8p expressing wild-type (WT) and Δach1 cells either combined with deletion of SCH9 (D) or supplemented with or without 20 nM rapamycin (Rapa) (E) and aged until indicated time points. Blots were probed with anti-GFP and anti-GAPDH (loading control) antibodies to detect “free-GFP” indicative of autophagic flux (see Figures S6B and S6C for quantification). (F and G) Representative micrographs (F) and respective quantification (G) of GFP-Atg8p expressing yeast cells ectopically overexpressing ACS2 (ACS2 overexp.) or carrying the empty vector (wild-type) during nutrient depletion (starved) compared to SC 2% galactose control conditions (control). Counterstaining of cells with propidium iodide was used to visualize dead cells. (H) Representative immunoblot analyses of the experiment shown in (F) to detect “free-GFP” indicative of autophagic flux. ^∗∗∗p < 0.001.

Figure 7

Knockdown of Acs2p Partly Restores Survival of ach1 in an ATG7-Dependent Manner (A and B) Chronological aging in SC 2% glucose medium supplemented with 1 ng/ml doxycycline of wild-type and Δach1 cells combined with or without knockdown of ACS2 (tet-ACS2). Survival (A) was determined by colony-forming capacity (clonogenicity). Age-associated cell death (B) was assessed through propidium iodide (PI) staining analyzed by flow cytometry. Data represent day 1 normalized means ± SEM (n = 4). (C and D) Survival (C) and cell death (D) of chronological aging experiment similar to (A) and (B) but in the background of ATG7-deleted autophagy-incompetent cells (Δatg7). Data represent day 1 normalized means ± SEM (n = 4). (E and F) Drosophila lifespan analyses of male (E) and female (F) flies depleted for acetyl-CoA synthetase (AcCoAS) using RNAi-mediated knockdown (AcCoAS RNAi) compared to isogenized controls (Control (EGFP RNAi)). Log rank tests revealed p < 0.0001 for both sexes. (G) Immunofluorescence specific to p62-homolog ref(2)p of adult brain sections from 10-day-old flies compared to 30-day-old flies as depicted in (E). Representative confocal micrographs are shown with scale bars representing 25 μm. (H) Quantification of total ref(2)P intensity in the central brain region normalized to 10-day-old control flies (Control (EGFP RNAi)). Standard box plot represents data from seven to eight independent brains (whiskers indicate minimum and maximum values). ^∗p < 0.05, ^∗∗∗p < 0.001.