Mitochondrial complex I deficiency increases protein acetylation and accelerates heart failure

Abstract

Mitochondrial respiratory dysfunction is linked to the pathogenesis of multiple diseases, including heart failure, but the specific mechanisms for this link remain largely elusive. We modeled the impairment of mitochondrial respiration by the inactivation of the Ndufs4 gene, a protein critical for complex I assembly, in the mouse heart (cKO). Although complex I-supported respiration decreased by >40%, the cKO mice maintained normal cardiac function in vivo and high-energy phosphate content in isolated perfused hearts. However, the cKO mice developed accelerated heart failure after pressure overload or repeated pregnancy. Decreased NAD(+)/NADH ratio by complex I deficiency inhibited Sirt3 activity, leading to an increase in protein acetylation and sensitization of the permeability transition in mitochondria (mPTP). NAD(+) precursor supplementation to cKO mice partially normalized the NAD(+)/NADH ratio, protein acetylation, and mPTP sensitivity. These findings describe a mechanism connecting mitochondrial dysfunction to the susceptibility to diseases and propose a potential therapeutic target.

Figure 1. Mouse survival, cardiac function and mitochondrial assessment in Ndufs4 deficient mice

(A) Fold changes ±SEM of ETC gene expression in the cKO mice relative to CON (n=7). (B) Representative western blot for ETC proteins (n=4). (C) Mitochondrial state 3 respiration in permeabilized cardiac fibers. Pyruvate + malate was used as complex I substrate in the presence of ADP, then succinate was added to establish complex I + complex II (CI+CII) respiration and finally rotenone was added to block complex I and establish complex II (CII) respiration (mean ±SEM; n=3). (D) Kaplan-Meier survival curve of cKO and CON mice (n=19–25). Echocardiographic data depicting (E) fractional shortening (%), (F) LV end-diastolic dimension (mm) and (G) posterior wall thickness (mm) in CON (white) and cKO (black) mice over 30 months (n=6–13). (H) Cardiac tissue citrate synthase enzyme activity (n=13–15). Data are expressed as means ± SEM. *P<0.05 vs CON. (I) Electron microscopy images illustrating mitochondrial arrangement and morphological characteristic and (J) fold differences in total mitochondrial number. *P<0.05 vs CON. See also Figure S1 and Table S1.

Figure 2. Myocardial energetics and cardiac function in isolated perfused hearts

(A) Phosphocreatine (PCr), (B) ATP, and (C) inorganic phosphate measured by ³¹P NMR spectroscopy in isolated hearts perfused with mixed substrate buffer during normal workload and dobutamine challenge conditions. (D) Intracellular pH calculated from the chemical shift between PCr and Pi in isolated hearts perfused with mixed substrate buffer during normal workload and dobutamine challenge conditions. (E) Left ventricular developed pressure (LVDevP), (F) heart rate (HR), and (G) rate-pressure product (RPP), the product of LVDevP and HR, measured in isolated hearts perfused with mixed substrate buffer during normal workload and dobutamine challenge conditions. (H) Coronary flow, estimated by collecting the perfusate effluent over a 2-minute period, in Langendorff heart preparations during normal workload and dobutamine challenge conditions (n=5). Data are expressed as means ± SEM. *P<0.05 vs CON. See also Figure S2.

Figure 3. Cardiac function in response to cardiac stress

(A) Fractional shortening and (B) LV end-diastolic dimension before surgery (0wk) and at 2 or 4 weeks after TAC or sham surgery in male mice (^#P<0.05 vs the respective sham group and *P<0.05 vs CON-TAC group). (C) Heart weight normalized to tibia length and (D) lung edema index at 4 weeks after TAC or sham surgery (n=5–6 sham and n=8–13 TAC). (E) Average number of litters for female CON and cKO mice, (F) heart weight normalized to body weight and (H) fractional shortening of these mice (n=9 CON and n=13 cKO; *P<0.05 vs CON). (I) H&E stain of a typical CON and cKO heart after 6 gestational cycles. Data are expressed as means ± SEM. See also Figures S3–4.

Figure 4. Increased cell death in the cKO hearts

Representative image of (A) TUNEL and DAPI stain for apoptotic nuclei, (B) trichrome stain for fibrosis and (C) WGA stain to outlining the cell surface area in female CON and cKO mice after 6 gestational cycles and quantitation graphs of those (D–F; n=4 per group, *P<0.05 vs CON). (G) Mitochondrial swelling induced by Ca²⁺ pulsing (25 μM [Ca²⁺] increments in the reaction buffer) indicated by the black triangle measured as % decrease in the initial optical density (OD 540) in the presence or absence of 1 μM CsA (this experiment was repeated at least 5 times). (H) Representative Ca²⁺ uptake traces by mitochondria in the presence or absence of 1 μM CsA. Ca²⁺ was added to achieve 15 μM [Ca²⁺] increments in the reaction buffer (this experiment was repeated at least 3 times). Data are expressed as means ± SEM.

Figure 5. ROS production is not increased in the cKO hearts

(A) Measurement of H₂O₂ production using Amplex Red in isolated cardiac mitochondria from CON (white) and cKO (black) mice treated with pyruvate/malate (10/5 mM; P/M), succinate (10 mM; S), rotenone (10 μM; Rot), Antimicyn A (1 mg/ml; AntA), ADP (2.5 mM) or combinations of those (n=4) and (B) mitochondrial or cytosolic H₂O₂ production in isolated cardiomyocytes using a fluorescent probe targeted to mitochondria (Mito-Hyper) or to cytosol (Cyto-Hyper; n=27–43; *P<0.05 vs CON). (C) Data showing the change of MitoSOX fluorescence at 405 nm excitation during a 5 min period before (−) or after the addition of respiration substrates (P/M/ADP). Data are expressed as mean ± SEM (n=9–10 cells from 2–3 mice; ^#P<0.05 P/M/ADP vs. non-treated; *P< 0.05 cKO vs. CON). (D) Representative images of CON and cKO permeabilized cardiomyocytes showing the MitoSOX fluorescence at 5 min after the addition of respiration substrates (10 mM pyruvate, 5 mM malate and 2 mM ADP). (E) Cardiac tissue aconitase enzyme activity (n=3). Data are expressed as means ± SEM. (F) Representative western blot for mitochondrial catalase in hearts from CON, cKO and cKO/mCat mice. (G) Heart weight normalized to tibia length, (H) fractional shortening and (I) LV end-diastolic dimension at 4 weeks after TAC or sham surgery (n=11 for the cKO/mCat TAC group and n=5 for the rest of the groups). Data are expressed as means ± SEM. *P<0.05 vs the respective sham indicated by the dotted line and ^#P<0.05 vs CON-TAC.

Figure 6. NAD⁺/NADH regulates mitochondrial protein acetylation via SIRT3

(A) NADH and NAD⁺ concentrations in freeze-clamped mouse hearts from CON and cKO mice (n=5). (B) Sirt3 activity assay using a 3 mM of NAD⁺ and 0–9 mM of NADH. (C) Representative blot for acetylated lysine residues (Ac-K) and VDAC (loading control), and (D) a quantitation graph in cardiac mitochondrial extracts (n=4; vertical line in graph C indicates the area on the blot used for quantitation). (E) Western blot and (F) a quantitation graph of Ac-K residues in isolated cardiomyocytes from CON and cKO hearts overexpressing Sirt3. SIRT3 protein levels and SDHA (loading control) are also shown (n=3). (G) Representative linescan confocal images showing the laser induced mPTP opening in cultured adult cardiac myocytes as reflected by the sudden loss of membrane potential (TMRM, red color). (H) Summarized data showing the time from the start of scan to the mPTP opening (mPTP time), the shorter the time the more sensitive of the mPTP. n = 120–143 mitochondria from 21–25 cells isolated from 3 hearts in each group. All data are expressed as means ± SEM. *P<0.05 versus Con. ^#P<0.05 versus without AdSirt3. See also Figures S5.

Figure 7. Regulation of the mPTP by mitochondrial NADH and protein acetylation

(A) Representative NADH autofluorescence trace using isolated mitochondria stimulated with glutamate/malate (G/M; 10/5 mM), ADP (250 μM) or Rot (1 μM) added in this order consecutively in the same trace. (B) A representative western blot and (C) a quantitation graph for Ac-K residues in protein extracts of CON and cKO mitochondria pre-treated separately with G/M (10/5 mM), or ADP (250 μM), or Rot (1 μM) or nicotinamide (NAM; 10 mM) for 20 min at 30°C (vertical line in graph B indicates the area on the blot used for quantitation), SDHA was used as loading controls. *P<0.05 vs CON non-treated and ^#P<0.05 vs cKO non-treated. Representative Ca²⁺ uptake traces from (D) CON and (E) cKO mitochondria, isolated in the presence or absence of G/M (10/5 mM), ADP (250 μM), Rot (1 μM) or NAM (10 mM). Each experiment was repeated at least 3 times. (F) NAD⁺/NADH ratio in freeze-clamped mouse hearts from CON and cKO mice treated with vehicle or cKO treated with nicotinamide mononucleotide (NMN) as described in Methods (n=3). (G) Representative Western blots for Ac-K residues and VDAC (loading control) in CON and cKO cardiac mitochondria from mice treated with vehicle or NMN. (H) Ca²⁺ uptake traces and (I) quantitation graph from cardiac mitochondria isolated from CON and cKO mice treated with vehicle or NMN. Data in Panels F and I are expressed as means ± SEM. *P<0.05 vs CON-Veh and ^#P<0.05 vs cKO-Veh.