PDH-E1α Dephosphorylation and Activation in Human Skeletal Muscle During Exercise: Effect of Intralipid Infusion

Seven healthy ([means ± SE] age 25.4 ± 1.3 years, height 181 ± 2 cm, and weight 74 ± 3 kg) normally physically active subjects participated in the present study. The subjects were given both written and oral information about the experimental protocol and procedures and were informed about any discomfort that might be associated with the experiment before they gave their written consent. The study was performed according to the Declaration of Helsinki and approved by the Copenhagen and Frederiksberg Ethics Committee, Denmark.

The day before each experimental trial, the subjects refrained from exercise and consumed a prepackaged dinner and evening snack, which was standardized based on the body weight of the subject (60 and 12 kJ/kg body wt, respectively). On the experimental day, the subjects arrived to the laboratory 2 h after consuming a standardized breakfast consisting of two slices of rye bread, butter, jelly, and water.

Each subject completed two identical trials in random order separated by ∼3 weeks with either intralipid/heparin or saline being infused in an antecubital vein. An arterial line was used for blood sampling. While the subjects remained resting in the supine position for 4 h followed by 3 h of two-legged knee extensor exercise at ∼50% of W_max (52 ± 12 Watts [mean ± SE]), a 20% intralipid emulsion (0.8 ml · kg body wt⁻¹ · h⁻¹; Fresinius Kabi, Uppsala, Sweden) and sodium heparin (24 units · kg body wt⁻¹ · h⁻¹, prime 3 units/kg body wt) were infused in one trial (intralipid) and saline in the other (control).

Blood samples were taken before infusion and at 60, 90, 120, 180, 210, and 240 min of rest during infusion and every 30 min during 3 h of exercise. Muscle biopsies were obtained from the vastus lateralis muscle using the percutaneous needle biopsy technique (21) with suction before infusion started, at 4 h of rest and after 45, 90, 135, and 180 min of exercise. Biopsies were taken from separate incisions using either leg and were rapidly frozen in liquid nitrogen.

Plasma FFA concentrations were determined using a Wako FA kit (Wako Chemical, Neuss, Germany) and an automatic spectrophotometer (Cobas FARA 2; Roche Diagnostic, Basel, Switzerland). Plasma insulin concentrations were determined using an insulin enzyme-linked immunosorbent assay kit (DakoCytomation, Glostrup, Denmark).

Muscle specimens were freeze dried and dissected free of blood, fat, and connective tissue under the microscope, and muscle glycogen content was determined as glycosyl units after acid hydrolysis using an automatic spectrophotometer (Cobas FARA 2; Roche Diagnostic) as previously described (22).

Total RNA was isolated from ∼20 mg of muscle tissue by a modified guanidinium thiocyanate–phenol-chloroform extraction method adapted from Chomczynski and Sacchi (23) as previously described (17). Reverse transcription was performed using the Superscript II RNase H⁻ system (Invitrogen, Carlsbad, CA) and Oligo dT as previously described (17). The amount of single-stranded DNA was determined in each reverse transcription sample using OliGreen reagent (Molecular Probes, Leiden, the Netherlands) as recently described (24).

The mRNA content of a given gene was determined by PCR using the fluorogenic 5′ nuclease assay with TaqMan probes (ABI PRISM 7900 Sequence Detection System; Applied Biosystems, Foster City, CA) as previously described (24) using the primer and probe sequences given in Table 1. The obtained critical threshold values reflecting the initial content of the specific transcript in the samples were converted to an arbitrary amount by using a standard curve obtained from a serial dilution of a representative pooled sample. For each sample, the amount of a given target cDNA was normalized to the total cDNA content in the sample.

Three polyclonal anti-PDH antibodies were made in sheep. One recognizing PDH independent of phosphorylation was made using a peptide corresponding to the COOH-terminal (296-309) part of the human PDH-E1α subunit (DPGVSYRTREEIQE). One recognizing PDH when phosphorylated on Ser²⁹³ (site 1) was made using a phospho-peptide corresponding to amino acid 287–299 in the COOH-terminal part of the human PDH-E1α subunit (YRYHGH(pS)MSDPGV). One recognizing PDH when phosphorylated on Ser³⁰⁰ (site 2) was made using a phospho-peptide corresponding to amino acid 294-306 in the COOH-terminal part of the human PDH-E1α subunit [MSDPGV(pS)YRTREE]. All peptides were synthesized with an additional cysteine at the NH₂-terminus, coupled to keyhole limpet hemocyanin and used to immunize sheep as previously described (25). Sera for the phosphospecific antibodies were first passed through columns containing immobilized dephosphopeptides, and the antibodies specific for the phosphorylation state were then bound to columns containing immobilized peptides of the same phosphorylation state as the antigen used for immunization and eluted as previously described (25). Using these antibodies, PDH-E1α subunit protein expression, as well as phosphorylation of PDH-E1α subunit on sites 1 and 2, were measured in different muscle preparations by SDS-PAGE (Tris-HCl 10% gel; BioRad, Copenhagen, Denmark) and Western blotting using polyvinylidene fluoride membrane and semidry transfer. After blocking (Tris-buffered saline with Tween 20 plus 2% skim milk), the membranes were incubated with the primary antibody (in Tris-buffered saline with Tween 20 plus 2% skim milk), followed by incubation with horseradish peroxidase–conjugated secondary antibody (in Tris-buffered saline with Tween 20 plus 2% skim milk). ECL-plus (Amersham, Birkerod, Denmark) was used as detection system.

Using a Kodak Image Station 440CF, one specific band at 40 kDa was detectable with all three antibodies, corresponding to the expected mass of the PDH-E1α subunit. When rat muscle was fractionated into mitochondrial and remnant fractions, the protein was exclusively found in the mitochondrial fraction (Fig. 1A). It was confirmed that no signal was lost to the pellet (17,500 g) when human muscle lysates were prepared from muscle homogenates made in a detergent containing buffer (Fig. 1B). The specificity of the antibodies was evaluated by peptide competition assays (Fig. 1C). Thus, the signal obtained in human muscle lysate, using the two phosphospecificspecific antibodies, was totally blocked by the corresponding phosphospecific peptide used for immunization but not by the corresponding dephosphopeptide or by the phospho/dephosphospecific peptides used for immunization/purification in the production of the other phosphospecific antibody. Similarly, the signal obtained with the nonphosphospecific antibody was blocked by the peptide used for immunization but not by the two phosphospecific peptides.

The two phosphospecific antibodies against PDH-E1α sites 1 and 2 were tested for phosphospecificity by phosphatase treatment of human lysates. A human muscle sample was cut in minute pieces and divided into two pools. Then, lysates were prepared as previously described (26) in a homogenisation buffer with or without phosphatase inhibitors (Na-pyrophosphate, β-glycerophosphate, NaF, and Na₃VO₄). Equal amounts of protein from the two lysates were then incubated in 10 mmol/l HEPES (pH 7.2), 5 mmol/l MgCl₂, 1 mmol/l EGTA, and 5 mmol/l MnCl₂ with or without 2 units/μg λ-phosphatase (New England Biolabs, Hertfordshire, U.K.) at 30°C for 2 h. The reaction was stopped by adding sample buffer, after which the samples were subjected to normal SDS-PAGE and Western blot analysis. The signal from the antibody against total PDH-E1α was not affected by the treatment, whereas the bands from the two phosphospecific antibodies were completely abrogated by the λ-phosphatase treatment (Fig. 1D).

Mixed gastrocnemius muscle from Wistar rats (250g) were minced (4°C, 100 mmol/l KCl, 40 mmol/l Tris-HCl, 10 mmol/l Tris-base, 5 mmol/l MgSO₄, 5 mmol/l NaF, 1 mmol/l Na₃VO₄, 1 mmol/l Na₄P₂O₇, 5 mmol/l EDTA, 5 mmol/l EGTA, and 3 mmol/l Benzamidine, pH 7.5) and left to rotate end over end for 15 min. The suspension was homogenized with a Polytron (PT3100; Kinematica, Lucerne, Switzerland) before being centrifuged (750g for 10 min). The top layer of debris was removed. The homogenate was centrifuged again (15,000g for 10 min) to collect the mitochondria. A sample of the supernatant was taken as a cytosolic control (the remnant). The mitochondria rich pellet was resuspended in Tris-EDTA buffer (10 mmol/l Tris-HCl, 1 mmol/l EDTA) with protease inhibitor cocktail, pH 7.5. All samples were frozen in liquid nitrogen and stored at −80°C.

Muscle specimens were freeze dried and dissected free of blood, fat, and connective tissue under microscope and were homogenized (26). The supernatant (17,500g for 60 min, 4°C) was harvested, frozen in liquid nitrogen, and stored at −80°C. In a subset of samples, the homogenate, lysate, and pellet were saved for later analyses. Total protein content in the samples was determined using the bicinchoninic acid method (Pierce).

The activity of PDHa was determined in vitro as previously described (9,27,28) after homogenizing 8–15 mg of muscle tissue on ice for 50 s using a micro glass homogenizer (Kontes) and quick freezing the samples (<15 s) in liquid nitrogen.

Two-way ANOVA for repeated measures was applied to evaluate the effect of time and treatment using Student-Newman-Keuls’s post hoc test to locate differences. A one-way ANOVA for repeated measures was also applied to evaluate the effect of time separately for each trial. Differences were considered significant at P < 0.05, and a tendency is reported when 0.05 ≤ P < 0.1. The statistical analyses on mRNA data were conducted on logarithmic transformed ratios (target/cDNA content) to ensure a normal distribution. For presentation in figures, the mRNA data are given as fold changes relative to the preinfusion value in each trial, which was set to 1. Values presented are means ± SE.

In the control trial, the arterial concentration of plasma FFA remained similar to the prelevel (∼500 μmol/l) throughout the 4-h resting period and increased (P < 0.05) during exercise to ∼1,500 μmol/l at the end of exercise. At rest, intralipid infusion elevated (P < 0.05) the arterial plasma FFA concentration to ∼1,700 μmol/l. This increased further during exercise reaching 3,000 μmol/l after 3 h (Fig. 2).

Before infusion, arterial plasma triacylglyceride levels were ∼1,250 mmol/l and remained at ∼1,100 mmol/l in the control trial. Intralipid infusion resulted in plasma triacylglyceride levels between 1,700 and 2,000 mmol/l throughout the experiment, reflecting an efficient breakdown of the triacylglyceride provided in the intralipid emulsion.

To avoid effects of prolonged fasting, a light breakfast was provided to the subjects 2.5 h before the experiment was initiated, resulting in slight elevated plasma insulin levels at the beginning of infusion (27–40 pmol/l). The arterial plasma insulin concentration tended to decrease (P < 0.1) over the 4-h resting period in the control but with no change in the intralipid trial. The plasma insulin concentration decreased (P < 0.05) in the control trial relative to the preinfusion level to 15 pmol/l at 2 h of exercise, but this was not significant when related to 4 h rest. Intralipid infusion was, on the other hand, associated with lowered (P < 0.05) plasma insulin concentration relative to 4 h rest already at 1 h of exercise (10.9 pmol/l).

From similar muscle glycogen levels before infusion (588 ± 45 and 515 ± 37 mmol/kg dry wt in the control and intralipid trials, respectively), no changes were observed in muscle glycogen content over the 4-h resting period in either of the trials and no difference was present between the trials. Exercise induced in both trials a marked decrease (P < 0.05) in muscle glycogen concentration to 273 ± 38 and 160 ± 33 mmol/kg dry wt at the end of exercise in the control and intralipid trials, respectively. The rate of glycogen usage was more pronounced in both trials (P < 0.05) during the initial 45 min of exercise (control: 3.8 ± 0.7 and intralipid: 2.8 ± 0.5 mmol · min⁻¹ · kg⁻¹ dry wt) than the remaining part of exercise (control: 0.9 ± 0.4 and intralipid: 1.4 ± 0.2 mmol · min⁻¹ · kg⁻¹ dry wt). No significant differences were evident between trials, neither within specific time intervals nor over the total exercise period.

No significant changes occurred in PDHa activity during the 4-h resting period in either trial. Exercise elevated the PDHa activity to 3.6- and 2.8-fold higher (P < 0.05) at 45–90 min of exercise than at 4 h rest in the control and intralipid trials, respectively. By the end of exercise, the PDHa activity had returned to the initial resting levels in the control trial but remained (2.6-fold, P < 0.05) elevated in the intralipid trial. Likely due to the multiple sampling, there was no significant overall difference in PDHa activity between the intralipid and control trials when analyzed using ANOVA (Fig. 3A). However, statistical evaluation of the data obtained at 3 h of exercise alone (paired t test) revealed a tendency for a difference between trials (P < 0.1).

Total PDH-E1α protein content was unchanged over the 4-h resting period and over the 3-h exercise period in both trials, and no difference was observed between trials (data not shown). Phosphorylation of the PDH-E1α subunit at site 1 (PDH-P1; P < 0.05) and at site 2 (PDH-P2; P < 0.1) was higher at rest with the intralipid infusion. During exercise, PDH-P1 and PDH-P2 decreased (P < 0.05) and the level reached was significantly lower in the control than in the intralipid trial. In accordance with the PDHa activity, the phosphorylation level was either not different (control trial) or still lowered (intralipid trial) at the end of exercise compared with 4 h rest (Fig. 3B–D).

Phosphorylation of both PDH-P1 and PDH-P2 was highly negatively associated with PDHa activity (Fig. 4A and B). The correlation coefficient (r²) obtained using a linear model was 0.34 and 0.40 for sites 1 and 2, respectively. However, visual inspection of the data (Fig. 4A and B) may suggest that a nonlinear relationship is present. In line, somewhat stronger correlations are evident if data are analyzed using an exponential model (r² = 0.43 and 0.46, for sites 1 and 2, respectively). The r² values predict that only ∼40% of the variation in PDHa activity can be explained by the changes in phosphorylation on one site indicating additional regulatory factors. In fact, a clear discrepancy between activity and phosphorylation is apparent at 4 h rest where activity was similar, but phosphorylation on both sites was different between the trials. It should be noticed, however, that the curve linear relationship at low PDHa activities is not solely due to data obtained at 4 h rest. A striking positive linear association (r² = 0.89) was also evident between the site 1 and the site 2 phosphorylation (Fig. 4C), indicating a similar and coupled regulation of both sites.

While PDK4 mRNA tended to increase (P < 0.1, 4.5-fold) over the 4-h resting period in the control trial, intralipid infusion resulted in a significant increase (6.5-fold) relative to before infusion. From this elevated level of PDK4 mRNA at 4 h rest, exercise induced a significant increase (five- to sixfold relative to 4 h rest) only in control (Fig. 5A). PDP1 mRNA tended to be lower (P < 0.1, 43%) at 4 h rest in the intralipid trial than before infusion. During exercise, PDP1 mRNA only decreased significantly (55%) in the control trial relative to 4 h rest. No differences were present between the two trials (Fig. 5B). No significant changes were apparent in PDK1, PDK2, PDK3, or PDP2 mRNA at rest or during exercise, and no differences were observed between trials.

The main findings of the present study are that exercise induces a decrease in phosphorylation of both site 1 and site 2 in the PDH-E1α catalytic subunit of the PDH complex in human skeletal muscle. This dephosphorylation is more pronounced when exercise is accompanied by high FFAs. Intralipid infusion also postpones or prevents the normal decrease in PDHa activity at the later stages of prolonged exercise. The phosphorylation state of PDH-E1α is inversely related to the PDHa activity during exercise. Interestingly, however, at rest, intralipid infusion for 4 h promotes PDH-E1α phosphorylation without a corresponding decreased PDHa activity.

We hypothesized that elevated plasma FFAs initiate a signal to spare carbohydrates by downregulating PDHa activity. However, in contrast to the changes in PDHa activity induced by a high-fat diet (5,6,9,29), intralipid infusion for 4 h was not associated with decreased PDHa activity in resting muscle. Interestingly, PDH-E1α phosphorylation was markedly elevated in the intralipid trial, and it could be speculated that more prolonged elevations in plasma FFA, by intralipid infusion, may lead to changes in PDHa activity similar to those seen by diet manipulations. FFAs have been suggested to play a role in the apparent dysregulation of PDH in a rodent model of type 2 diabetes (1), and intralipid infusion leads to muscle insulin resistance and impairs PDHa activation by insulin in human skeletal muscle (30). Thus, it may be speculated that type 2 diabetes is associated with elevated phosphorylation of PDH-E1α at rest and that such phosphorylation reduces the PDHa activation upon insulin stimulation.

The present study also demonstrates that intralipid infusion results in a sustained elevation rather than a decrease in the PDHa activity during prolonged exercise, suggesting that elevated plasma FFA does not elicit a signal to induce glucose sparing in skeletal muscle. In accordance with the exercise-induced changes and effect of intralipid infusion on PDHa activity, exercise elicited a pronounced drop in phosphorylation state of both PDH-1Eα sites 1 and 2 and intralipid infusion associated with more marked changes than in the control trial. Moreover, phosphorylation had returned to resting levels in the control at the end of exercise but remained reduced at both sites in the intralipid trial, relative to both preinfusion and preexercise levels. Although the decline in PDHa activity during later stages of prolonged exercise under normal conditions cannot be explained by changes in acute regulatory molecules like pyruvate and acetyl CoA (10), an increased PDK activity has been reported (31). Together, this may suggest that the sustained elevation in PDHa activity with intralipid infusion may be caused by an FFA-induced inhibition of PDK activity.

The elevated phosphorylation of PDH-E1α at rest during intralipid infusion did not compromise the regulation of PDHa activity during exercise, suggesting that regulatory mechanisms activated during exercise overrule the regulation exerted by elevated FFAs at rest. A generalization of these results obtained from healthy subjects to metabolic diseases may be difficult. However, to the extent that intralipid infusion can mimic a patophysiological state with elevated circulating FFA levels, the present results suggest that exercise may overcome a dysregulation of PDH activity at rest in type 2 diabetes. This could potentially explain the observation that lipid and glucose metabolism is normal during exercise in muscles of patients with type 2 diabetes.

The present study provides for the first time a direct measure of the phosphorylation state of PDH-E1α sites 1 and 2, which are thought to be the important phosphorylation sites for PDHa activity in skeletal muscle. Previous estimation of PDH phosphorylation has been based on the assumption that the in vitro activity assay measuring activity of PDH in the active form reflected the phosphorylation state and not other covalent modifications. By establishing these tools, we have taken one step toward elucidating the role of specific phosphorylation sites for PDHa activity and eventually to the regulatory role of upstream kinases/phosphatases. The necessity of differentiating PDH phosphorylation estimated from PDHa activity and that of a direct measurement of PDH-E1α site 1 and 2 phosphorylation is seen from the present results. Thus, after 4 h rest, a marked difference in PDH-E1α site 1 and 2 phosphorylation is evident between the control and intralipid trials, but this is not reflected in similar changes in PDHa activity. These findings may suggest that PDH, under some conditions, is subject to other covalent modifications, which are powerful enough to overrule at least to some extent the influence, which PDH-E1α site 1 and 2 phosphorylation exerts on PDHa activity. Similar conclusions may be drawn from the nonlinear relationship between either PDH-E1α site 1 or 2 phosphorylation and PDHa activity. However, it should be noted that phosphorylation of either one of the two sites is sufficient to render the PDC complex inactive (32). It might therefore be speculated that as the degree of phosphorylation of one site increases, the probability that the protein will already be phosphorylated at the other site will increase, causing the nonlinear relationship. In that aspect, it is interesting that changes in the phosphorylation of sites 1 and 2 were very similar in response to both exercise and intralipid infusion, and the phosphorylation state of sites 1 and 2 was highly correlated (r² = 0.89), which is in agreement with the findings that dephosphorylation of the sites occurs in a random order (33).

In agreement with fasting- and high-fat diet–induced upregulation of PDK4 transcription, mRNA and/or protein content in both rodent and human skeletal muscle (7,18,20,34), resting in the fasted state for 4 h resulted in a small elevation in PDK4 mRNA in the control trial and a more clear increase in PDK4 mRNA content, when the circulating FFA concentration was elevated by intralipid infusion. A recent study (35) reported no effect of intralipid infusion on PDK4 mRNA in resting skeletal muscle, but it may be noted that a nonsignificant change in PDK4 mRNA seemed apparent in that study. Taken together, FFAs do appear to have a regulatory effect on the PDK4 gene expression in resting human skeletal muscle.

The present results also demonstrate that although the plasma FFA concentration at rest in the intralipid trial was at a level similar to that observed during exercise in the control trial, the PDK4 mRNA content was far less than the level observed during exercise. This shows that FFAs cannot be the sole regulator of PDK4 gene expression in human skeletal muscle. This is in accordance with previous studies (12,19,36) suggesting that other factors are also potentially important in regulating PDK4 gene expression. Insulin infusion has been shown to reduce PDK4 mRNA expression in both rat (36) and human (4) skeletal muscle. Decreased plasma insulin concentration is therefore also a possible factor eliciting the increases in PDK4 transcription in human skeletal muscle during fasting and prolonged exercise. Thus, the different plasma FFA and plasma insulin levels may jointly have contributed to the difference in PDK4 mRNA expression in the intralipid and control trial in the present study. However, except for a possible role of FFAs at rest, no simple explanation is apparent during exercise.

The observed changes in PDK4 mRNA are in agreement with previous studies in humans (20,37) and rats (36) and reflect the distinct gene regulation of the PDK isoforms. We recently have shown that PDP1 mRNA is downregulated in human skeletal muscle in recovery from a 75-min exercise bout (20), and the present results reveal that the PDP1 mRNA content can already be reduced during prolonged exercise, while PDP2 mRNA is unaffected. Together, these findings underline the isoform-specific nature of the regulation of PDKs and PDPs in human skeletal muscle. Because the PDK and PDP isoforms not only have a unique tissue distribution but also isoform-specific sensitivity to activators and inhibitors (38,39), isoform-specific regulation of PDK and PDP expression may have an important impact on PDH-E1α phosphorylation and PDH activity and may be targets in dysregulation of PDH in type 2 diabetes, as previously indicated (1).

In conclusion, exercise is associated with decreased phosphorylation of PDH-E1α at both site 1 and 2 in human skeletal muscle, and intralipid infusion modifies this response with even lower phosphorylation of both sites. These changes in PDH-E1α phosphorylation are associated with corresponding changes in PDHa activity, including a more sustained elevation in the PDHa activity during the later stages of exercise with intralipid. The present findings reveal an inverse association between PDH-E1α phosphorylation and PDHa activity in human skeletal muscle. However, short-term elevation in plasma FFAs does not change PDHa activity, although phosphorylation of PDH-E1α is increased.

The study was supported by grants from the Danish National Research Foundation, the Danish Medical Research Council, the Danish Natural Science Research Council, the Novo Nordisk Foundation, and an Integrated Project (LSHM-CT-2004-005272) from the European Commission. The Centre of Inflammation and Metabolism is funded by the Danish National Research Foundation (no. 02-512-55). The Copenhagen Muscle Research Centre is supported by grants from the Copenhagen Hospital Corporation and the University of Copenhagen. J.W. was supported by a Hallas Møller Stipendium from the Novo Nordisk Foundation.

The authors thank the subjects who participated in the study for their extraordinary effort. The technical assistance of Kristina Møller Kristensen, Carsten Nielsen, Birgitte Jessen, and the medical assistance of José Calbet are gratefully acknowledged.