Biochemical Journal

Research article

Metabolic inflexibility and protein lysine acetylation in heart mitochondria of a chronic model of Type 1 diabetes

Shraddha S. Vadvalkar , C. Nathan Baily , Satoshi Matsuzaki , Melinda West , Yasvir A. Tesiram , Kenneth M. Humphries

Abstract

Diabetic cardiomyopathy refers to the changes in contractility that occur to the diabetic heart that can arise in the absence of vascular disease. Mitochondrial bioenergetic deficits and increased free radical production are pathological hallmarks of diabetic cardiomyopathy, but the mechanisms and causal relationships between mitochondrial deficits and the progression of disease are not understood. We evaluated cardiac mitochondrial function in a rodent model of chronic Type 1 diabetes (OVE26 mice) before the onset of contractility deficits. We found that the most pronounced change in OVE26 heart mitochondria is severe metabolic inflexibility. This inflexibility is characterized by large deficits in mitochondrial respiration measured in the presence of non-fatty acid substrates. Metabolic inflexibility occurred concomitantly with decreased activities of PDH (pyruvate dehydrogenase) and complex II. Hyper-acetylation of protein lysine was also observed. Treatment of control heart mitochondria with acetic anhydride (Ac2O), an acetylating agent, preferentially inhibited respiration by non-fatty acid substrates and increased superoxide production. We have concluded that metabolic inflexibility, induced by discrete enzymatic and molecular changes, including hyper-acetylation of protein lysine residues, precedes mitochondrial defects in a chronic rodent model of Type 1 diabetes.

  • acetylation
  • cardiac mitochondrion
  • diabetes
  • free radical

INTRODUCTION

Diabetic cardiomyopathy is associated with contractility defects that may occur in the absence of other risk factors such as coronary artery disease. Mitochondrial bioenergetic deficits and increased free radical production are pathological hallmarks of the diabetic heart [14]. These mitochondrial alterations are thought to promote progression of diabetic cardiomyopathy [59]. However, the mechanisms underlying mitochondrial dysfunction and whether they precede or follow the development of disease is unknown. It is therefore critical to define the molecular changes that occur to mitochondria before cardiac deficits.

Deficits in mitochondrial function associated with diabetes have been linked to metabolic defects [4,6,10]. The heart's constant demand for energy comes primarily from mitochondrial oxidative phosphorylation, which accounts for 95% of the ATP produced [11]. To support oxidative phosphorylation, the heart derives 60–90% of its ATP from oxidation of fatty acids, with the remainder coming primarily from pyruvate and lactate [12]. However, under a chronic diabetic state, the ability of the heart to switch between available oxidizable substrates is impaired and, under these conditions, it relies almost exclusively on fatty acid metabolism [12]. This metabolic inflexibility may be expected to confer increased mitochondrial damage. This is, in part, because of the accumulation of potentially toxic incomplete products of fatty acid oxidation. Such products increase stress on cardiac mitochondria and increase insulin resistance [13,14]. In addition, in vitro studies show that mitochondrial respiration by fatty acid oxidation increases free radical production compared with that produced by other carbon sources [15]. This suggests that metabolic inflexibility itself may contribute to disrupted mitochondrial function and increased free radical production.

Lysine acetylation, post-translational modification of proteins, has emerged as an important metabolic regulatory mechanism in mitochondria. Electron transport chain, tricarboxylic acid cycle enzymes and fatty acid oxidation enzymes are all targets of this post-translational modification [1618]. However, the functional consequences on individual enzyme activities have not been fully resolved. Levels of protein lysine acetylation change between fasting and feeding [19,20]. In general, the change in acetylation and deacetylation of discrete targets appears to correspond to a co-ordinated shift in metabolic fuel selections. In addition, mitochondrial protein acetylation may also serve as an important means of regulating mitochondrial free radical production and metabolism. For example, acetylation of SOD (superoxide dismutase) 2 and ICDH2 (isocitrate dehydrogenase 2) decreases their activities, thereby decreasing dismutation of superoxide and decreasing the NADPH available for reducing glutathione respectively [21,22]. How protein lysine acetylation is affected in the chronic diabetic heart and how this relates to mitochondrial dysfunction and increased free radical production remain to be rigorously defined. We hypothesize that changes in protein lysine acetylation contribute to the development of diabetic cardiomyopathy.

The present study was undertaken to examine mitochondrial function and metabolic flexibility in OVE26 transgenic mice, an established model of chronic Type 1 diabetes. Our goal was to identify how mitochondrial fuel selection was affected in the heart before the onset of cardiac contractility deficits, and to determine underlying molecular changes that contribute to diabetes-induced pathology.

EXPERIMENTAL

Animals

OVE26 and age-matched wild-type (FVB) mice were obtained from Jackson Laboratories. Mice were fed on standard laboratory chow ad libitum. At 2–5 months, mice were killed by cervical dislocation. All animal procedures were in accordance with OMRF (Oklahoma Medical Research Foundation) Institutional Animal Care and Use Committee guidelines.

In vivo MRI (magnetic resonance imaging)

Cardiac function was measured by MRI as described previously [9]. Data were collected on an 11.74T wide-bore Bruker Spectrometer. Methods included with the Intragate package (Bruker Biospin) were used for collection of images. A total of 300 repetitions of a single slice placed axially at the mid-ventricular position were used for collection of images that were used to calculate fractional shortening. Black blood images were used for these measurements which allowed clear visualization of the left ventricle wall. The images representing three-quarters of a cardiac period were produced by retrospective means assuming a respiratory rate of 60 bpm (beats per minute) and a cardiac rate of 450 bpm using software provided by the Intragate method. The FOV (field of view) for these images was 2.56 cm×2.56 cm and the matrix size was 128×128, producing images with 200 μm in-plane resolution in a total acquisition time of 4.2 min. For ejection fraction calculations, images were collected with the same parameters as described above, but without presaturation and in multiple slices since only blood volume measurements are important in this case. The slice thickness was 1 mm and each slice was advanced manually to the next position and varied from six to seven slices depending on the size of the heart.

Image analysis was performed with Paravision software (version 5.0). For fractional shortening calculations, the black blood images were used and wall thickness from the reconstructed images were determined at ED (end-diastole) and ES (end-systole). The percentage of fractional shortening was calculated as %FS=[(ED−ES)/ED]×100. Conventionally (i.e. with echocardiograms),%FS is calculated by determining the diameter of the left ventricle wall at ED and ES. Raw wall thickness values are also included in Table 1. Percentage EF (ejection fraction percentage) was calculated by determining volumes of the left ventricle at ED and ES, so %EF=[(VD−VS)/VD]×100, where VD and VS are the diastolic and systolic volumes respectively. Sets of multislice images were used for calculation of volumes.

View this table:
Table 1 Comparison of MRI-derived left ventricle systolic function and wall thickness between FVB control and OVE26 mice

Results are means±S.D. from five experiments. *P<0.01 between the OVE26 and FVB controls as determined by Student's t test.

Isolation of cardiac mitochondria

Male OVE26 or age-matched FVB mice were killed by cervical dislocation. Their chest cavities were immediately opened and their hearts were perfused with 3–5 ml of ice-cold buffer containing 210 mM mannitol, 70 mM sucrose, 1.0 mM EDTA and 5.0 mM Mops, pH 7.4 (buffer A) via injection into the left ventricle. Hearts were excised and placed into 10 ml of buffer A and then minced with scissors. This was followed by Polytron homogenization. The homogenate was spun at 500 g for 5 min at 4°C, the supernatant was collected and passed through a cheesecloth, and spun again at 5000 g for 10 min. The resulting mitochondrial pellet was resuspended in approximately 60 μl of buffer A and protein concentration was determined by the BCA (bicinchoninic acid) method (Pierce) using BSA as a standard.

Mitochondrial respiration

Mitochondria were diluted to 0.25 mg/ml in 210 mM mannitol, 70 mM sucrose, 5.0 mM KH2PO4 and 10 mM Mops, pH 7.4 (buffer B) containing 10 mM glutamate/2.5 mM malate, 0.1 mM pyruvate/1.0 mM malate, 30 μM palmitoylcarnitine/0.1 mM malate or 10 mM succinate. For palmitoylcarnitine/malate experiments, 0.5 mg/ml BSA was added to buffer B. For chemically induced acetylation of FVB wild-type heart mitochondria, samples (0.25 mg/ml) were treated with the indicated amounts of Ac2O (acetic anhydride) in buffer B with 0.5 mg/ml BSA for 2.0 min at 20°C, and the reaction was then stopped by the addition of 3.0 mM lysine. Respiration was measured polarographically at 25°C with a Clark-style oxygen electrode (Instech). State 3 respiration [23] was initiated by the addition of ADP (0.5 mM). The starting amount of molecular oxygen in the 0.6 ml electrode chamber was based on the assumption that 265 nmol of molecular oxygen are dissolved per ml at atmospheric pressure and 20°C.

Enzyme activity assays

PDH (pyruvate dehydrogenase) activity was measured as follows. Mitochondria were diluted to 0.05 mg/ml in 25 mM Mops at pH 7.2 containing 0.025% Triton X-100. The reduction of 0.5 mM NAD+ was then followed spectrophotometrically at 340 nm (ϵ=6200 M−1·cm−1) upon the addition of 0.2 mM TPP (thiamine pyrophosphate), 0.04 mM CoASH and 10 mM pyruvate. NADH oxidase, an overall assessment of electron transport chain activity through complexes I–III–IV was measured as follows. Frozen and thawed mitochondria were diluted into 10 mM Mops at 0.025 mg/ml and the rotenone-sensitive rate of oxidation of 0.1 mM NADH was followed spectrophotometrically at 340 nm. Complex II activity was measured as follows. Mitochondria were diluted to 0.25 mg/ml in buffer B with 10 mM succinate for 2.0 min. This procedure reverses oxaloacetate-mediated inhibition and ensures that the enzyme is in the fully active state [24]. Samples were diluted 1:10 in a buffer containing 25 mM Mops, pH 7.2, and 0.025% Triton X-100. Complex II activity was measured spectrophotometrically by the TTFA (thenoyltrifluoroacetone)-sensitive reduction of ubiquinone 1 (50 μM) at 280 nm (ϵ=13700 M−1·cm−1) [24]. The presence of 0.025% Triton X-100 completely blocked NADH oxidation and ubiquinone reduction by complex I [25] and ubiquinol oxidation by complex III. The combined MDH (malate dehydrogenase) and glutamate/oxaloacetate transaminase activities were measured as follows. Mitochondria (0.025 mg/ml final concentration) were diluted in 25 mM Mops, pH 7.2, with 0.025% Triton X-100 and 5.0 mM malate. The reduction of 1.0 mM NAD+ was then followed at 340 nm. The forward MDH reaction quickly halts with the accumulation of oxaloacetate [24], but was found to rapidly proceed with the addition of 10 mM glutamate. Likewise, in the presence of glutamate and NAD+ alone, no production of NADH was detectable. This confirms previous reports of low glutamate dehydrogenase activity in the heart and that glutamate entry into tricarboxylic acid cycle proceeds primarily via transamination [26]. The MDH-mediated reduction of NAD+ in the presence of glutamate and malate was linear for >3.0 min.

SDS/PAGE and Western blot analysis

Mitochondrial proteins (amounts indicated below) were diluted into SDS/PAGE sample buffer containing 25 mM DTT (dithiothreitol). For chemically induced acetylation of FVB wild-type heart mitochondria, samples (0.25 mg/ml) were treated with indicated amounts of Ac2O in buffer B for 2.0 min at 20°C, and the reaction then stopped with the addition of 3.0 mM lysine. Following SDS/PAGE (4–12% polyacrylamide, NuPAGE Gel, Life Sciences), gels were either stained with Gel Code Blue (Life Sciences) according to the manufacturer's protocol, or proteins were transferred on to nitrocellulose membranes and blocked for 1 h at 23°C with 0.5% BSA in PBST (PBS with Tween 20). Blots were then incubated for overnight with a polyclonal primary antibody against acetyl-lysine (1:000 dilution, Cell Signaling Technology), SDHA [SDH (succinate dehydrogenase), subunit A] (1:100 dilution, Santa Cruz Biotechnology), the E2 subunit of KGDH [2-oxoglutarate dehydrogenase (α-ketoglutarate dehydrogenase)] [27], PDH (1:000 dilution, Abcam) or phospho-293 PDH E1 subunit (1:000 dilution, Abcam). Blots were then washed five times for 5 min with TBST (Tris-buffered saline with Tween 20) and incubated with goat anti-rabbit HRP (horseradish peroxidase)-conjugated antibody (1:20000 dilution, Thermo Scientific) for 1 h. Following three 5 min washes with TBST, blots were exposed to Super-Signal West Pico Chemiluminescent Substrate (Thermo Scientific) and imaged with an Alpha Innotech Fluorchem HD2 imaging system. Unless indicated, exposure times were chosen so that no band intensities exceeded saturation. Densitometry was performed by analysis with ImageJ software (NIH). Band intensities were compared with those of loading controls.

Superoxide anion (O2•−) production measurements

The NADH-supported rate of O2•− production was measured as described previously [28,29]. Briefly, the oxidation of hydroethidine by O2•− to the fluorescent product 2-hydroxyethidium was measured by fluorescence spectroscopy. The fluorescent signals were recorded (excitation, 480 nm; emission, 567 nm) over time utilizing 10 μM hydroethidine and 500 μM NADH. No change in fluorescence was seen with additional DNA, indicating endogenous nucleotides were sufficient for the formation of 2-hydroxyethidium. CuZn-SOD (8.0 units/ml) was used to test the specificity of the measurement for O2•−.

Statistical analysis

Data are presented as means±S.D. or S.E.M., as indicated. The data were evaluated utilizing a two-tailed Student's t test. The indicated statistical significance was assigned for P≤0.05.

RESULTS

Characterization of OVE26 mice

OVE26 mice are a well-characterized model for chronic Type 1 diabetes. These transgenic mice specifically overexpress calmodulin in the pancreas, impeding the release of insulin [30]. Blood glucose levels of OVE26 mice were routinely above 600 mg/dl (not shown). Mice that were used in the experiment were 2–5 months old. MRI analysis of cardiac contractility is shown in Table 1. Consistent with a previous study [31], the ejection fraction and percentage fractional shortening displayed an increased trend (20.6 and 11.8% respectively) in OVE26 relative to FVB wild-type, but failed to reach statistical significance. No differences were seen in ES and ED left ventricular wall thickness between OVE26 and FVB mice. Collectively, the data suggest subtle changes in cardiac function in 2–5-month-old OVE26 mice, without indications of overt cardiomyopathy.

Mitochondrial function

Mitochondrial yields from OVE26 mouse hearts were greater than those from FVB mice (0.94±0.06 and 0.82±0.06 mg of total protein respectively, n=7). This is consistent with a reported increase in mitochondrial biogenesis and content in this animal model [8]. It has been reported previously that OVE26 heart mitochondria from 5-month-old animals exhibit declines in respiration when using glutamate as an oxidizable substrate [8]. However, the capacity to use other carbon sources to support respiration had not been examined in this model before the present study. Typical oxygen trace measurements are shown in Figure 1(A) and ADP-dependent state 3 respiration rates are quantified in Figure 1(B). As shown in Figure 1, we observed severe deficits (52.1% decrease) in state 3 respiration when utilizing glutamate as a substrate. Respiration utilizing pyruvate as a substrate also showed a severe decline (45.5%). The amount of pyruvate used (0.1 mM) was chosen to minimize the confounding effects of high pyruvate concentrations on PDH and PDH kinase activities. This concentration is also well above the reported Km value of pyruvate in the PDH reaction (35 μM) [32]. Furthermore, similar declines were seen when 10 mM pyruvate was used (results not shown). Succinate-linked respiration, performed in the absence of rotenone, exhibited a 24.8% decline in state 3 respiration. Rotenone was omitted because of our previous finding that rotenone masks H2O2-mediated declines in succinate-linked respiration [24]. The deficits observed when using pyruvate or glutamate as substrates were to an extent that the transition from state 3 to state 4 was not readily discernible (Figure 1A). In contrast with other substrates, no differences between OVE26 and FVB wild-type respiration profiles were observed when using palmitoylcarnitine as a substrate. This result is significant because it suggests that the mitochondrial deficits demonstrated with other oxidizable substrates are not due to overt damage to mitochondrial integrity, electron transport chain activities, ADP/ATP transport or ATP synthase activity. We conclude that a major change to OVE26 heart mitochondria is a severe metabolic inflexibility.

Figure 1 OVE26 heart mitochondria have impaired mitochondrial respiration when using non-fatty-acid substrates

Mitochondrial respiration was measured, as described in the Experimental section, using 10 mM glutamate/2.5 mM malate, 0.1 mM pyruvate/1.0 mM malate, 10 mM succinate or 25 μM palmitoylcarnitine/0.1 mM malate. (A) Representative oxygen traces with each of the substrates used. (B) Results are mean±S.E.M. state 3 respiration measurements (n≥5). *P<0.05 between OVE26 and FVB controls as determined by Student's t test. Glu, glutamate; PC, palmitoylcarnitine; Pyr, pyruvate; Suc, succinate.

Mechanisms of mitochondrial respiratory deficiencies

OVE26 heart mitochondria exhibit severe deficiencies utilizing non-fatty-acid substrates. The inefficient utilization of pyruvate may be mediated, in part, by a decrease in PDH activity [33]. PDH activity decreased by 51.6% in OVE26 heart mitochondria compared with FVB controls (48.6±10.9 and 23.5±6.7 nmol of NADH·min−1·mg−1 for FVB and OVE26 respectively; means±S.D.). PDH regulation is multifaceted and includes inhibitory phosphorylation of the E1 subunit at multiple sites. Examination of the phosphorylation state of E1 subunits revealed a significant increase in OVE26 heart mitochondria compared with FVB controls (Figure 2). The ratio of band intensities of phosphorylated to non-phosphorylated E1 was 0.20±0.05 and 0.33±0.06 for FVB and OVE26 respectively (Figure 2). This result suggests that decreased PDH activity via hyper-phosphorylation may contribute to the inefficient utilization of pyruvate shown in Figure 1.

Figure 2 PDH activity is decreased in OVE26 heart mitochondria, and this corresponds to increased phosphorylation

Western blot (WB) analysis of total PDH E1 and phospho-293 E1 in heart mitochondria (20 μg of total mitochondrial protein per well) of three representative FVB wild-type and OVE26 mice. The E2 subunit of KGDH is shown as a loading control. The ratio of phospho-293 E1 to total E1, as determined by densitometry, is shown. Results are means±S.D. (n=6). *P<0.05 between OVE26 and FVB controls as determined by Student's t test.

Succinate-linked respiration is distinct from other respiratory substrates in that electrons are introduced at complex II instead of complex I. We therefore hypothesized that a decrease in complex II activity may contribute to decreased succinate-supported respiration in OVE26 heart mitochondria. Complex II activity is significantly impaired (28.3% decrease) in OVE26 heart mitochondria (SDH activities are 343.3±22.4 and 246±33.8 nmol of ubiquinone 1·min−1·mg−1 for FVB and OVE26 respectively; means±S.E.M.). Thus a decrease in complex II activity may contribute to a decreased rate of succinate-supported respiration by decreasing flux through complex II into the electron transport chain and by impeding the tricarboxylic acid cycle. In contrast, no decrease was observed in NADH oxidase activity, which is a combination of activities through complexes I–III–IV of the electron transport chain (633.5±56.8 compared with 662.6±70.3 nmol of NADH oxidized·min−1·mg−1 for FVB and OVE26 respectively). This supports further that deficiencies in NADH-supported respiration, observed with glutamate and pyruvate, are not due to deficiencies in the electron transport chain.

The oxidation of glutamate in heart mitochondria primarily proceeds via the transaminase reaction in which glutamate and malate are converted into 2-oxoglutarate and aspartate [26]. However, measurement of the combined MDH and glutamate/malate transaminase activity revealed no differences between OVE26 and FVB control heart mitochondria (results not shown). Thus the inefficient utilization of glutamate as an oxidizable substrate (Figure 1B) is not likely to be due to inhibition of this metabolic step.

OVE26 heart mitochondria exhibit hyper-acetylation of protein lysine residues

Recent studies have demonstrated that protein lysine acetylation is an important determinant in metabolic regulation. Specifically, protein lysine acetylation levels fluctuate in the fed compared with fasted state and this is likely to reflect changes in fuel selection and nutrients available for energy production. We therefore hypothesized that protein lysine acetylation may be altered in OVE26 heart mitochondria on the basis of their observed mitochondrial metabolic inflexibility. Protein lysine acetylation measured by Western blot analysis was significantly increased in OVE26 heart mitochondria. As shown in Figure 3(A), OVE26 heart mitochondria exhibit an approximate 2.5-fold increase in total acetylation levels compared with age-matched FVB controls. However, the total protein profiles looked similar between control and diabetic animals (Figure 3B). Comparing the mitochondrial fraction with the cytosolic fraction, we found that the majority of hyper-acetylated proteins in OVE26 hearts are mitochondrial (Figure 3B). Separation of isolated mitochondria into membrane and soluble fractions suggested that the majority of acetylated proteins are either membrane-bound or membrane-associated (results not shown).

Figure 3 Heart mitochondria from OVE26 mice exhibit hyper-acetylation, but no significant decrease in SIRT3 content

(A) Heart mitochondria from FVB wild-type or OVE26 mice were analysed by Western blotting using a commercially available (Cell Signaling Technology) anti-acetyl-lysine antibody. Comparison of two representative FVB wild-type and OVE26 mouse mitochondria are shown in the left-hand panel. A Western blot of mitochondrial protein, SDHA, is shown as a loading control. Densitometry of the entire lane was performed as described in the Experimental section (right-hand panel). Results are means±S.E.M. (n=5). *P<0.02 between OVE26 and FVB controls as determined by Student's t test. (B) Samples of 10 μg of either mitochondrial or cytosolic proteins from OVE26 or FVB mice were analysed by Western blotting using an anti-acetyl-lysine antibody. The blot was overexposed to demonstrate the large difference in protein lysine acetylation between cytosolic and mitochondrial proteins (left-hand panel). An SDS/PAGE gel with 20 μg of each sample per lane, and stained as described in the Experimental section, is shown in the right-hand panel. (C) Western blot analysis of SIRT3 in three representative mitochondrial samples from OVE26 and FVB heart mitochondria. A Western blot analysis against SDHA was performed as a loading control. The content of SIRT3 in FVB and OVE26 heart mitochondria was determined by densitometry as described in the Experimental section. Results are means±S.E.M. (n=5). There was a trend of decreased SIRT3 in OVE26 heart mitochondria that failed to reach statistical significance using Student's t test (P=0.11). AcK, acetyl-lysine; WB, Western blot. Molecular masses are indicated in kDa in (A) and (C).

SIRT3 (sirtuin 3) is the primary NAD-dependent deacetylase in mitochondria [35]. We therefore carried out Western blot analyses to determine whether hyper-acetylation is accompanied by a decrease in SIRT3 levels. SIRT3 is produced as a 49-kDa protein and processed when it enters the mitochondria. In mice, both the full-length (49 kDa) and processed versions of SIRT3 are found in the mitochondria [36]. It was found that the majority of SIRT3 detected in both the OVE26 and FVB heart mitochondria was in the form of the full-length enzyme (49 kDa). SIRT3 content trended lower in OVE26 heart mitochondria (20% decrease, n=5), but failed to reach statistical significance (Figure 3C).

Protein lysine acetylation may promote metabolic inflexibility

To evaluate the contribution of protein lysine acetylation to metabolic inflexibility we used an in vitro approach. Specifically, we hyper-acetylated proteins of isolated FVB control mitochondria by treatment with Ac2O. Ac2O preferentially acetylates primary amines at neutral pH and is used to selectively acetylate protein lysine residues [37]. We observed that treatment of FVB heart control mitochondria with increasing amounts of Ac2O led to a concentration-dependent increase in protein lysine acetylation, as measured by Western blot analysis (Figure 4A). Remarkably, treatment of FVB mitochondria with a low concentration of Ac2O (25 μM) induced the acetylation of proteins that were similar in profile to the banding pattern of OVE26 heart mitochondria (Figure 4B). This may reflect that these recognized proteins are highly abundant, most susceptible to acetylation or exceptionally recognized by the anti-acetyl-lysine antibody.

Figure 4 Acetylation profile and respiratory activities of control mitochondria treated with acetic anhydride is similar to OVE26

(A) FVB wild-type heart mitochondria (0.25 mg/ml) were treated with indicated amounts of Ac2O for 2.0 min at 20°C, and the reaction then stopped with the addition of 3.0 mM lysine. Samples were then analysed by Western blot analysis using an anti-acetyl-lysine antibody. (B) Control FVB heart mitochondria were treated with 25 μM Ac2O and analysed by Western blotting, as in (A). Acetylation is shown before and after treatment with Ac2O and compared with two unique OVE26 heart mitochondria samples. (C) The effect of Ac2O on state 3 respiration supported by three different oxidizable sources, measured as in Figure 1, is compared with FVB control and OVE26 heart mitochondria. Results are means±S.E.M. (n≥4). *P<0.05 between samples and FVB controls as determined by Student's t test. AcK, acetyl-lysine; Glu, glutamate; PC, palmitoylcarnitine; Pyr, pyruvate; WB, Western blot. Molecular masses are indicated in kDa in (A) and (B).

Having determined that 25 μM Ac2O acetylates in a manner similar to that of OVE26 diabetic mice, we next examined the functional consequence of this treatment. It was found that 25 μM Ac2O preferentially inhibits glutamate- (30.1%) and pyruvate- (61.7%) supported respiration, but had less of an effect on palmitoylcarnitine-supported respiration (16.1% decrease). The inhibitory effect of Ac2O on oxidative phosphorylation is not surprising given the non-selective nature of this reagent and the high frequency of protein lysine residues. However, the differential sensitivity to Ac2O-mediated inhibition in the presence of various substrates may reflect the unique susceptibility of protein lysine residues of different metabolic pathways to modification.

Effect of acetylation on mitochondrial free radical production

Stimulation in mitochondria-derived free radical generation is implicated in the progression of diabetic cardiomyopathy. We next examined whether the metabolic inflexibility observed in OVE26 heart mitochondria was accompanied by an increase in free radical production. For these experiments, FVB wild-type and OVE26 mitochondria were disrupted by freeze–thaw and the production of O2•− at maximal respiratory rate was directly assessed by measuring the NADH-driven SOD-sensitive increase in 2-hydroxyethidium fluorescence. This assay has an advantage over fluorescence-based H2O2-detecting assays in intact mitochondria in that it is not dependent upon endogenous dismutation rates of superoxide. In addition, H2O2 detection rates in intact mitochondria may under-represent ETC-derived superoxide because it only determines the H2O2 that escaped via endogenous H2O2-scavenging pathways and diffused through both the inner and outer mitochondrial membranes. Representative kinetic traces of superoxide production are shown in Figure 5(A) and demonstrate measured superoxide rates are sensitive to addition of exogenous SOD1. As shown in Figures 5(A) and 5(B), there is a significant increase in superoxide production in OVE26 heart mitochondria (23.5%) compared with FVB controls. We next examined whether chemically induced metabolic inflexibility, using 25 μM Ac2O, was sufficient to increase mitochondrial superoxide production. As shown in Figure 5(B), treatment of FVB control mitochondria with the acetylating agent induced a 27.2% increase in superoxide production.

Figure 5 OVE26 heart mitochondria and FVB mitochondria acetylated with Ac2O produce more superoxide than FVB controls

(A) Superoxide production was measured fluorimetrically as the SOD-sensitive rate at which disrupted mitochondria oxidize hydroethidine in the presence of NADH. Traces are representative of 2-hydroxyethidium formation for FVB and OVE26 heart mitochondria in the absence or presence of CuZn-SOD (8.0 units/ml). (B) FVB control heart mitochondria were incubated in the presence of 25 μM Ac2O for 2.0 min, as indicated, and then the reaction stopped by the addition of 3.0 mM lysine. SOD-sensitive rates are compared with FVB control and OVE26 values. Results are means±S.E.M. (n=5). *P<0.05 between samples as determined by Student's t test.

DISCUSSION

In the present study, we demonstrate that heart mitochondria from a chronic model of Type 1 diabetes exhibit a significant decrease in maximal rates of respiration when using non-fatty-acid substrates. Surprisingly, the maximal rate of respiration as well as respiratory control ratio were unchanged between FVB and OVE26 heart mitochondria when palmitoylcarnitine was used as a substrate. This result demonstrates that mitochondria are not damaged to an extent that has affected mitochondrial integrity or capacity to perform maximal oxidative phosphorylation before the onset of diabetic cardiomyopathy. A similar observation was made to a lesser extent in Akita mice, another chronic model of Type 1 diabetes [6]. However, in that study, it was concluded that the decrease in non-fatty-acid-supported respiration was likely to be attributable to a deficit in complex I activity. However, in the present study, we observed no decrease in NADH oxidase activity in 2-month-old mice, suggesting that deficiencies in complex I activity are either not present or not to an extent that decreases overall electron transport chain activity.

The mechanisms responsible for the observed inhibition in respiration with non-fatty-acid substrates were shown to be multifaceted. They include a decrease in PDH and complex II activities. In the case of PDH, this decrease in activity is attributable in part to hyper-phosphorylation of the E1 subunit that is observed in OVE26 relative to FVB heart mitochondria. In addition, there may be other contributing factors, such as increased acetyl-CoA production [33,38], that contribute to feedback inhibition of PDH activity. Finally, it is possible that other post-translational modifications, such as oxidation or lysine acetylation, also contribute to decreased PDH activity. The mechanisms behind complex II inhibition are less clear. However, complex II has been described previously as a target of oxidative modification in an STZ (streptozotocin)-induced model of Type 1 diabetes [39]. In addition, it has been shown that complex II is negatively regulated by protein acetylation [40]. The relative contribution of oxidation and acetylation to decreased complex II activity in the diabetic heart is the subject of future studies.

Decreased respiration observed with succinate and pyruvate are largely explained by deficits in their respective enzymatic entry points into oxidative phosphorylation. However, it is less clear what is inducing the decrease in respiration with glutamate as a substrate. We found no changes in glutamate/oxaloacetate transaminase activity. We speculate that the observed hyper-acetylation may be contributing to a global change in mitochondrial metabolism that decreases anaplerotic reactions. This hypothesis is supported by our finding that in FVB control heart mitochondria, chemically induced acetylation preferentially inhibits respiration by non-fatty-acid substrates, mimicking the OVE26 characteristics.

Hyper-acetylation may be an important contributor to the increased oxidative stress associated with the progression of diabetes and diabetic cardiomyopathy. For example, acetylation of SOD2 is a negative regulator of SOD2 activity [21]. Likewise, acetylation of ICDH2 inhibits its activity [22], which may subsequently result in a decrease in the production of NADPH available for reduction of glutathione. Such changes in antioxidant capacity may occur concomitantly with acetylation-mediated inhibition of electron transport chain components such as complexes I and II. Acetylation has been shown to inhibit their activities [18,40], but its effects on subsequent increases in free radical production have not been determined.

The underlying cause of hyper-acetylation is not immediately apparent. The level of SIRT3 was decreased in OVE26 heart mitochondria, but not to a statistically significant extent. Decreases in SIRT3 activity, as mediated by post-translational modifications, may also be contributory to increased acetylation. For example, SIRT3 itself is a target of oxidative inactivation [41]. This may be significant given our findings of increased superoxide production in OVE26 heart mitochondria. It is also possible that the increase in acetylation may be promoted by the metabolic conditions of OVE26 mice. These mice exhibit high levels of circulating triacylglycerols [42], and, as we have demonstrated in the present study, are metabolically inflexible at the mitochondrial level of substrate selection. A constant use of fatty acids may be expected to increase acetyl-CoA content [38] and the ratio of NADH/NAD+. Increased acetyl-CoA, the most probable carbon donor for protein lysine acetylation, would promote acetylation, whereas the increased NADH/NAD+ ratio would impair the activity of SIRT3, which uses NAD+ as a cofactor.

As we have demonstrated, changes in protein lysine acetylation are likely to have significant effects on mitochondrial function. For example, non-specific chemical acetylation of control mitochondria preferentially inhibits mitochondrial respiration by non-fatty-acid substrates. Future studies identifying these preferentially acetylated proteins may offer insights into the mechanisms whereby metabolic dysfunction leads to irreversible loss of mitochondrial function and the progression of diabetic cardiomyopathy.

AUTHOR CONTRIBUTION

Shraddha Vadvalkar performed the majority of the biochemical experiments, including isolation of mitochondria, measurement of mitochondrial respiration and enzyme activities, and Western blot analysis. Nathan Baily carried out the experiments examining in vitro acetylation of mitochondrial proteins. Satoshi Matsuzaki measured superoxide production in isolated mitochondria and assisted with editing the paper before submission. Melinda West contributed to the rodent experiments, including animal husbandry and tissue collection. Yasvir Tesiram collected and analysed the functional MRI data and assisted with editing the paper before submission. Kenneth Humphries designed the experiments, analysed and interpreted the data, and wrote the paper.

FUNDING

This work was funded by the National Institutes of Health [grant number P20 RR 024215] from the Centers of Biomedical Research Excellence (COBRE) Program of the National Center for Research Resources.

Acknowledgments

We thank Dr John Knight for a review of the paper and helpful suggestions before submission.

Abbreviations: Ac2O, acetic anhydride; bpm, beats per minute; ED, end-diastole; ES, end-systole; ICDH2, isocitrate dehydrogenase 2; KGDH, 2-oxoglutarate dehydrogenase (α-ketoglutarate dehydrogenase); MDH, malate dehydrogenase; MRI, magnetic resonance imaging; PDH, pyruvate dehydrogenase; SDH, succinate dehydrogenase; SDHA, SDH, subunit A; SIRT3, sirtuin 3; SOD, superoxide dismutase; TBST, Tris-buffered saline with Tween 20

References

View Abstract