Acyl-CoA Binding Protein Expression Is Fiber Type- Specific and Elevated in Muscles From the Obese Insulin-Resistant Zucker Rat

Female lean and obese Zucker rats (Harlan, Shaw’s Farm, Blacktorn, Bicestor, U.K.) were housed for at least 2 weeks in our animal facilities before experiments. The rats were kept on a 12 h–12 h light-dark cycle and fed rat food and water ad libitum as described previously (26). The experiments and procedures were conducted in conformity with laws and regulations controlling experiments and procedures for animals in Norway. The weights of the lean animals were 175–221 g, and the weights of the age-matched obese animals were 298–397 g (P < 0.0001).

The rats were anesthetized with an injection of ∼0.3 ml pentobarbital (50 mg/ml i.p.), and the muscles were dissected out. Epitrochlearis and split soleus were mounted at their resting length and incubated for measurement of glycogen synthesis as described previously (27). For measurement of ACBP, enzyme activities, and triglyceride concentration, extensor digitorum longus (EDL), the red medial part of the gastrocnemius, the white dorsal lateral part of the gastrocnemius, and some soleus muscles were frozen in liquid nitrogen and stored at −70°C until analysis.

The muscles were preincubated for 30 min in 3 ml of a modified Krebs-Henseleit buffer to which 0.1% BSA, 5 mmol/l HEPES, 5.5 mmol/l d-glucose, and 2 mmol/l pyruvate (pH 7.4) were added. For measurements of glycogen synthesis, the muscles were transferred to a fresh buffer containing 0.2 μCi/ml of d-[U-¹⁴C]glucose (298 mCi/mmol; NEN, Boston, MA) with and without 10 mU insulin/ml (Monotard; Novo Nordisk, Denmark). All incubations were performed at 30°C, and the buffers were continuously gassed with 95% O₂ and 5% CO₂. Immediately after incubation, the muscles were blotted on filter paper, frozen in liquid nitrogen, and stored at −80°C. The muscles were freeze-dried and weighed before analysis.

Glycogen synthesis was determined from incorporation of [¹⁴C]glucose into glycogen as described previously (27). In brief, freeze-dried muscle samples were dissolved in 500 μl of 1 mol/l KOH at 70°C for 20 min before 100 μl saturated Na₂SO₄ and 100 μl glycogen (25 mg/ml) were added and mixed. Ethanol (1.5 ml) was added, and the glycogen precipitated at −20°C overnight. After centrifugation (3,000g for 20 min at 4°C), the supernatant was discarded, the precipitate was dissolved in 500 μl distilled water for 20 min at 70°C, and the glycogen was reprecipitated with ethanol. The precipitate was dissolved in 300 μl distilled water, and 250 μl was added to 3 ml scintillation solution (Hionic-Fluor; Packard, Meriden, CT) and counted (TRI-CARB 460C; Packard).

Glycerol was extracted from freeze-dried muscles in 170 μl 3 mol/l perchloric acid for 20 min on ice. Of the extracts, 150 μl was neutralized with 250 μl 2 mol/l KHCO₃ and centrifuged (3,000g for 5 min at 4°C). Glycerol was determined fluorometrically as described by Chernick (28).

For hydrolysis of triglyceride to glycerol and free fatty acid (FFA), muscle samples were incubated overnight in 400 μl tetraethylammioniumhydroxide at 60°C in glass tubes with screw caps as described previously (29). After addition of 175 μl of 3 mol/l perchloric acid, the samples were centrifuged (3,000g for 10 min at room temperature), and 200 μl of the supernatant was neutralized with 250 μl 2 mol/l KHCO₃ and centrifuged again. Glycerol was determined fluorometrically as described by Chernick (28).

The ACBP content was determined using a multi-layer sandwich ELISA as described by Pawlak and Smith (30), with minor modifications. Muscles were homogenized (1 mg wet weight per ∼30 μl buffer) in a phosphate buffer (100 mmol/l KH₂PO₄, 100 mmol/l K₂HPO₄, 0.2 mmol/l EDTA, pH 7.0) for 3 × 30 s (Ultra-Turrax T8; IKA Labortechnik, Staufen, Germany); each burst was separated with 30 s. The samples were immediately sonicated for 3 × 30 s at 50 W (Vibra Cell; Sonics & Materials, Danbury CT). Samples were kept on ice during the whole procedure. Homogenate (50 μl) was stored at −80°C for protein analysis and the rest was centrifuged for 10 min (10,000g at 4°C). The supernatant was used for sandwich ELISA.

Recombinant rat ACBP was cloned and purified from Escherichia coli according to Mandrup et al. (31), and rabbit anti-rat ACBP IgG was produced and purified according to Knudsen et al. (32). Microtiter plates were coated with 0.5 μg polyclonal rabbit anti-rat ACBP IgG/100 μl Tris-buffered saline (TBS) buffer (8.7 mmol/l NaCO₃, 34.8 mmol/l NaHCO₃, 3.07 mmol/l NaN₃, pH 9.6) and incubated at 4°C overnight. The plates were washed once in PBS (137 mmol/l NaCl, 1.47 mmol/l KH₂PO₄, 7.45 mmol/l Na₂HPO₄, 2.68 mmol/l KCl, 3.07 mmol/l NaN₃, pH 7.4) containing 0.1% gelatin and blocked for 3 h at 37°C with PBS containing 1.0% gelatin. Plates were washed five times in PBS containing 0.1% gelatin and 0.1% Tween 20. The five-times washing step was carried out between all further incubation steps and is referred to as wash. Standards and the homogenates were diluted in TBS with Tween buffer (637 mmol/l NaCl, 2.68 mmol/l KCl, 8.91 mmol/l Tris [base + HCl], 3.07 mmol/l NaN₃, 0.1% Tween 20, 0.1% gelatin, pH 9.0), and 200 μl was added to the immobilized rabbit anti-rat ACBP IgG on the microtiter plates. After an overnight incubation at 4°C, the plates were washed, and further steps were performed at room temperature. Biotinylated rabbit anti-rat ACBP IgG was added (200 μl), and the plates were incubated for 2 h on an autoshaker. After washing, the plates were incubated for 50 min with avidin (200 μl) and washed. The plates were then incubated for 50 min with biotinylated alkaline phosphatase (200 μl) and washed. The chromogenic substrate p-nitrophenyl phosphate (PNPP) was added (200 μl) in a dilution of 1 mg PNPP per 1 ml buffer (0.5 mmol/l MgCl₂, 3.07 mmol/l NaN₃, 0.61% ethanolamine, pH 9.6), and the enzyme reaction was allowed to proceed for 20–40 min. The optical density was measured at 405 nm by an autoreader (Wallac, Victor², 1420 multilabel counter; Perkin Elmer, Turku, Finland).

For quantification, standards of rat ACBP ranging in concentrations from 0.1 to 6.0 ng/ml (11 concentrations) were prepared together with blanks and run in duplicate. Analyses were carried out using plates containing 96 wells, of which 24 were used for standard for calculation of concentration of the samples run on the plate. The linear part of the standard curve was used, normally between 1.5 and 4 ng/ml. Muscle samples were analyzed in triplicate at two different homogenate dilutions. Homogenates were typically diluted 200–1,000 times, depending on muscle type and protein concentration. Concentrations of ACBP in both dilutions should have been in the linear part of the standard curve and were reanalyzed if this was not the case. Homogenates were also reanalyzed if the ACBP concentration varied >10% in the triplicate determination at one of the dilutions. The ACBP concentration on each muscle is the mean of the concentration measured at the two dilutions. Interassay coefficient of variation (12 measurements from the same homogenate) was 2.9% (n = 12) in soleus from obese rats (high ACBP concentration) and 3.9% (n = 12) in white gastrocnemius (low ACBP concentration). Recovery of ACBP was 93.5% when added to homogenate of EDL from lean rats (n = 6) and 92.7% when added to homogenate of red gastrocnemius from obese rats (n = 6).

Long-chain acyl-CoAs were extracted from muscle tissue as described previously (17). The amount of the six major species of acyl-CoA (palmitoyl, palmitoleoyl, stearoyl, oleoyl, linoleoyl, and linolenoyl) in muscle extracts were determined by high-performance liquid chromatography using a C18 reverse-phase column (Waters Corporation, Milford, MA) and appropriate standards (17). The sum of these six major species was used to reflect total long-chain acyl-CoA content.

The assay was performed as an isotope forward assay, described by Zierz and Engel (33). Muscles were homogenized (1 mg wet weight per 40 μl buffer) on ice for 3 min in Tris buffer (50 mmol/l Tris [base + HCl], 100 mmol/l KCl, 5 mmol/l MgSO₄, 1 mmol/l EDTA, 1 mmol/l ATP, pH 7.5 at 4°C) and sonicated as for ACBP analysis. All muscle samples were run as duplicates. To run one muscle sample, 900 μl reaction buffer was prepared (100 mmol/l Tris [base + HCl], 0.8 mmol/l palmitoyl-CoA, 2 mmol/l potassium cyanide, 1 mmol/l dithiothreitol, 0.5 mmol/l l-carnitine, 0.04 μCi l-[methyl-¹⁴C]carnitine [specific activity 57.0 mCi/mmol; Amersham, Amersham, U.K.]). The reaction was started by addition of 100 μl homogenate to the reaction buffer. After 20 min of reaction at 30°C, the reaction was stopped by addition of 1.5 ml ice-cold isobutanol (2-methyl-1-propanol) and 1 ml distilled H₂O, saturated with NH₃SO₄, and whirl-mixed. The formed carnitine palmitoyl-CoA produced occurs in the organic phase and was re-extracted after a second addition of 1 ml distilled water saturated with NH₃SO₄ and isobutanol. After the second extraction, the radioactivity of the carnitine-acyl-CoA complex was measured. A total of 500 μl of the organic phase was mixed with 10 ml scintillation fluid (Wallac, Optiphase Hisafe 3; Perkin Elmer) and counted for radioactivity (TRI-CARB 2700 TR; Packard).

Muscle samples were freeze-dried for 24 h, weighed, and homogenized (1 mg dry weight per 400 μl) in phosphate buffer (300 mmol/l K₂HPO₄, 300 mmol/l KH₂PO₄, pH 7.7). The maximal activity of citrate synthase and 3-hydroxy-acyl-CoA dehydrogenase (HAD) was determined by a fluorometer according to Lowry and Passonneau (34). The acetoacetyl-CoA used as substrate for HAD was prepared fresh immediately before analysis.

Protein concentration was analyzed using the commercially available BCA Protein assay kit (Pierce BCA protein reagent; Perstorp Life Sciences, Rockford, IL). The concentration was measured spectrophotometrically at 578 nm.

Results are presented as means ± SE. ANOVA was performed to establish differences between groups; Fisher’s protected least significant differences test was performed post hoc. P < 0.05 was considered as significant.

Insulin-stimulated glycogen synthesis was twice as high in soleus as in epitrochlearis muscle from lean rats (Table 1). Muscles from obese rats were severely insulin resistant, and the rate of insulin-stimulated glycogen synthesis was reduced by 66 and 85% in the slow-twitch oxidative soleus and the fast-twitch glycolytic epitrochlearis muscle, respectively (Table 1).

The concentration of ACBP varied between different muscles fiber types (Fig. 1). The highest concentration was observed in the slow-twitch soleus muscle and the lowest in white gastrocnemius. The mean concentration of ACBP was higher in muscles from the obese Zucker rat compared with muscles from lean Zucker rats (Fig. 1). The mean concentrations of ACBP in lean and obese soleus muscles were 0.46 ± 0.02 and 0.56 ± 0.03 μg/mg protein (P < 0.01), respectively (Fig. 1). The ACBP concentration in red gastrocnemius and EDL from obese Zucker rats was 25% higher than that in the muscles from lean rats. White gastrocnemius had the lowest ACBP concentration, 0.16 ± 0.005 μg/mg protein, and there was no significant increase in the obese rats (P = 0.25).

The concentration of the three most abundant species of acyl-CoA (palmitoyl, linoleoyl, and oleoyl) was 6.5 ± 0.9 and 12.7 ± 1.5 nmol/g wet weight (n = 5; P < 0.01) in red quadriceps from lean and obese Zucker rats, respectively. These three species make up >90% of the measurable acyl-CoA (17), and the change in acyl-CoA in obese Zucker rats exceeds the increase observed in ACBP concentration.

CPT-1 activity showed similar fiber type differences as ACBP concentrations in lean rats. The CPT-1 activity was increased in the glycolytic muscles from obese rats (EDL and white gastrocnemius) but not in the oxidative muscles (soleus and red gastrocnemius) (Fig. 2). The mean CPT-1 activity in EDL from obese rats was 50% higher than that from lean rats (2.89 ± 0.19 vs. 1.93 ± 0.08 nmol · min⁻¹ · mg protein⁻¹; P < 0.0001) and similar to the activity in soleus from both lean and obese rats (Fig. 2). The CPT-1 activity in obese rats did not show the same pronounced fiber type differences as in muscles from lean rats. The correlation between the ACBP concentration and the CPT-1 activity was, therefore, stronger in muscles from lean animals (R = 0.70; P < 0.0001) than in muscles from obese animals (R = 0.36; P < 0.05) (Fig. 3).

HAD activity was 29% higher in red gastrocnemius muscles from obese rats than in red gastrocnemius muscles from lean animals (38.0 ± 2.0 μmol · g⁻¹ dry weight · min⁻¹ [n = 12] vs. 29.5 ± 1.5 μmol · g⁻¹ dry weight · min⁻¹ [n = 12]; P < 0.01, respectively) and 42% higher in EDL (23.8 ± 2.0 μmol · g⁻¹ dry weight · min⁻¹ [n = 12] vs. 16.8 ± 1.2 μmol · g⁻¹ dry weight · min⁻¹ [n = 10]; P < 0.001, respectively). Lean and obese rats had similar CS activity in both red gastrocnemius (66.6 ± 2.5 μmol · g⁻¹ dry weight · min⁻¹ [n = 12] vs. 62.0 ± 2.9 μmol · g⁻¹ dry weight · min⁻¹, respectively) and EDL (53.9 ± 3.3 μmol · g⁻¹ dry weight · min⁻¹ [n = 12] vs. 47.3 ± 2.7 μmol · g⁻¹ dry weight · min⁻¹ [n = 10], respectively). CS activity was higher in red gastrocnemius than in EDL in both lean and obese rats (P < 0.01).

Triglyceride concentration was increased in muscles from the obese Zucker rats (Fig. 4). Triglyceride concentration was 50% higher in soleus from obese rats compared with lean rats (P < 0.05). In all other muscles from obese rats, the triglyceride concentration was >100% higher than that in lean muscles (P ≤ 0.001). Glycerol concentration was also higher in muscles from obese rats than in muscles from lean rats (red gastrocnemius: 0.48 ± 0.04 vs. 0.16 ± 0.03 mmol/kg dry weight, P < 0.001; EDL: 0.87 ± 0.11 vs. 0.16 ± 0.03 mmol/kg dry weight, P < 0.001, respectively). In the obese Zucker rats, glycerol concentration was higher in EDL than in red gastrocnemius (0.87 ± 0.11 and 0.48 ± 0.04 mmol/kg dry weight, respectively; P < 0.001).

This is the first study of ACBP concentration in different muscle fiber types, and we report higher expression in oxidative muscles than in glycolytic muscles. The ACBP concentration is therefore found to correlate with CPT-1 activity. Another important finding is that the ACBP concentration is increased by 30% in muscles from the obese insulin-resistant Zucker rat. This increase is, however, much lower than the 90% elevation in acyl-CoA observed in red muscle from obese Zucker rats.

It has been hypothesized that accumulation of acyl-CoA in skeletal muscle causes insulin resistance (1–3,8) and that the concentration of acyl-CoA inversely correlates with insulin sensitivity (17,19). The effect of acyl-CoA on regulation of enzymes, however, depends on the free concentration, for which the concentration of ACBP is the most important determinant (10). The presented data on ACBP are therefore important for interpretation of acyl-CoA data in different muscle fiber types and when normal and insulin-resistant muscles are compared.

In the present study, a sandwich ELISA technique was used to quantify the concentration of ACBP. The concentration varied between 3 and 9 nmol ACBP/g wet weight (0.16–0.46 μg/mg protein) in the different muscle fiber types. In red gastrocnemius, we found a concentration of ACBP of ∼5 nmol/g wet weight, which is rather similar to the concentration of acyl-CoA (7.1 nmol/g wet weight) we found in red quadriceps. We previously found similar concentrations of acyl-CoA in red quadriceps and gastrocnemius (17), and our data suggest that the concentrations of acyl-CoA and ACBP are rather similar in skeletal muscle with normal insulin sensitivity.

The concentration of ACBP was ∼30% higher in muscles from obese Zucker rats. The increase in ACBP concentration was, however, much lower than the 90% increase in acyl-CoA we found in red muscles. This is the first comparison of ACBP concentration in normal and insulin-resistant muscles and shows a moderate elevation in the latter. On the other hand, acyl-CoA varies considerably. Elevations of ∼50% in concentrations of acyl-CoA have previously been reported in muscles from fat-fed insulin-resistant muscles (18). Furthermore, a threefold variation in acyl-CoA was observed in both humans (17) and rats (6,19) with varying insulin sensitivity. Our data show that not only is the increase in concentration of acyl-CoA much higher than the increase of ACBP, but that the absolute concentration of acyl-CoA may exceed the concentration of ACBP in muscles from obese rats. Because one molecule of ACBP binds to one molecule of acyl-CoA, the concentration of acyl-CoA may exceed ACBP binding capacity in muscles from obese rats. Although other proteins, such as sterol carrier protein 2 and FABP, can bind acyl-CoA, their affinity is much lower (10). Concentrations of acyl-CoA higher than ACBP will therefore cause a substantial increase in the free concentration of acyl-CoA. The larger increase in acyl-CoA than in ACBP in muscles from obese Zucker rats supports the hypothesis that an increased free concentration of acyl-CoA may contribute to insulin resistance.

ACBP is a cytosolic protein, whereas part of acyl-CoA occurs in mitochondria and peroxisomes, where acyl-CoA is metabolized, as well as in lipid membranes. It is therefore not possible to calculate the free concentration of acyl-CoA, even though we have quantitative data on both ACBP and acyl-CoA. Measurement of free concentration of acyl-CoA is an important physiological parameter, but this challenge has not yet been solved in any cell type. In addition to the low concentration, other methodological problems make these studies extremely difficult. Skeletal muscles have a substantial amount of mitochondria (35) and low amount of peroxisomes (36), but compartmentalization of acyl-CoA has not been studied in skeletal muscles. In the heart, the percentage of acyl-CoA in the mitochondria has been reported to be 23% (37), 60–70% (38), and 90% (39). These divergent results may reflect the difficulties in studying compartmentalization of acyl-CoA.

Regarding the increase in acyl-CoA in muscles from obese Zucker rats, it seems unlikely that a larger proportion of acyl-CoA is located in the mitochondria in muscles from obese rats because fat oxidation is decreased in these muscles (23,24), even though β-oxidative enzyme activity is increased (40; present study) (see also discussion about CPT-1 and malonyl-CoA below). The increase in acyl-CoA in obese Zucker rats may therefore be mainly cytosolic. The increase in ACBP and CPT-1 activity in muscle from obese rats may in fact be an attempt to protect the cell against too high a concentration of acyl-CoA in the cytosol.

Previously, an increased concentration of ACBP was reported in the heart and in liver of the fat-fed insulin-resistant rat (11), but this is the first study in skeletal muscles. Other proteins involved in transport and metabolism of FFA have, however, been studied in insulin-resistant muscles. Expression of fatty acid translocase (FAT) (CD36) is upregulated in skeletal muscle from streptozotocin-induced diabetic rats (41). Simoneau et al. (42) reported increased concentrations of plasma membrane FABP and FABP_c in muscles from insulin-resistant muscles, whereas Blaak et al. (43) found reduced concentrations of FABP_c in muscles from diabetic patients. In Zucker rats, studies so far have focused on liver and adipocytes. In adipocytes, fatty acid uptake and mRNA for fatty acid transporting protein, FAT, and mitochondrial aspartate aminotransferase were elevated in adipocytes (44,45). Our data show that ACBP is upregulated in insulin-resistant muscles, like many other proteins involved in metabolism of fat.

Diabetes is associated with abnormalities in fat metabolism, and insulin sensitivity inversely correlates with skeletal muscle concentration of triglycerides (46–48) and acyl-CoA (17). In the present study, we also found increased concentrations of triglycerides in insulin-resistant muscles, although the enzymes involved in fat metabolism were increased (Figs. 2 and 3). Despite higher concentrations of triglycerides and enzymes metabolizing FFA, glucose oxidation is increased in muscles with reduced insulin sensitivity (23,24,43,49,50). Inhibition of CPT-1 activity by an elevated concentration of malonyl-CoA in insulin-resistant muscles probably explains the reduced oxidation of fat and increased accumulation of triglycerides (1).

The rate of glycogen synthesis was reduced to low levels in muscles from the obese rats (Table 1). Although acyl-CoA inhibits glycogen synthase activity (51), this is probably not the reason for the reduction in glycogen synthesis because prior steps in glucose uptake are reduced. Recently, Thompson and Cooney (52) reported that acyl-CoA also inhibits hexokinase activity in skeletal muscles, but even glucose transport and translocation of GLUT4 is reduced in muscles from obese rats (15). The defect in insulin-stimulated glucose uptake in obese Zucker rats occurs at an early step, and insulin-stimulated activation of insulin receptor substrate (IRS)-1-associated phosphatidylinositol (PI) 3-kinase activity is reduced (53). The insulin resistance caused by prolonged infusion of FFA also reduces IRS-1-associated PI 3-kinase activity (54), and an interesting question is whether elevation of acyl-CoA may reduce IRS-1-associated PI 3-kinase activity.

The concentration of ACBP varied threefold between different muscle fiber types, with the highest concentration in the soleus and the lowest in the glycolytic white gastrocnemius. Because ACBP protects against a rise in free concentration of acyl-CoA, a high concentration of ACBP may have a favorable effect on insulin sensitivity. In fact, soleus muscles that have the highest concentration of ACBP also have higher insulin sensitivity than glycolytic muscles (16). Furthermore, the glycolytic epitrochlearis muscles become much more insulin resistant than the oxidative soleus muscle in the obese Zucker rats (Table 1) (15,55). An interesting idea is that the higher concentration of ACBP in soleus protects this muscle against the raised concentration of acyl-CoA and therefore against insulin resistance in the obese Zucker rat.

In conclusion, ACBP expression is fiber type-specific, with a higher concentration in oxidative muscles than in glycolytic muscles. Furthermore, the concentration of ACBP is 30% higher in muscles from insulin-resistant obese Zucker rats than in muscles from lean Zucker rats. The increase in ACBP was, however, much lower than the 90% increase in acyl-CoA concentration in insulin-resistant muscles, and our data support the hypothesis that increased free concentration of acyl-CoA may play an important role for the development of insulin resistance.

The study was supported by Team Denmark and the Research Council of Norway.

We thank Jorid T. Stuenæs for analysis of muscle triglycerides and glycerol and Bente Jørgensen for analysis of citrate synthase and HAD. We also thank Dr. Nick Oakes, AstraZeneca, R&D, Mölndal, Sweden, for tissue materials for analysis of acyl-CoA and Dr. Leonard Storlien and Nils J. Færgeman for comments on the manuscript.