Physical and Functional Association of Lactate Dehydrogenase (LDH) with Skeletal Muscle Mitochondria

Background: The notion of mitochondrial lactate oxidation in skeletal muscle is controversial.

Results: Mitochondrial respiration increased in the presence of substrates and cofactors for the lactate dehydrogenase (LDH) reaction. Respiration was inhibited when mitochondrial pyruvate transport was blocked.

Conclusion: Extra-matrix LDH is associated with muscle mitochondria.

Significance: LDH is strategically positioned within skeletal muscle fibers to functionally interact with mitochondria.

Abstract

The intracellular lactate shuttle hypothesis posits that lactate generated in the cytosol is oxidized by mitochondrial lactate dehydrogenase (LDH) of the same cell. To examine whether skeletal muscle mitochondria oxidize lactate, mitochondrial respiratory oxygen flux (JO₂) was measured during the sequential addition of various substrates and cofactors onto permeabilized rat gastrocnemius muscle fibers, as well as isolated mitochondrial subpopulations. Addition of lactate did not alter JO₂. However, subsequent addition of NAD⁺ significantly increased JO₂, and was abolished by the inhibitor of mitochondrial pyruvate transport, α-cyano-4-hydroxycinnamate. In experiments with isolated subsarcolemmal and intermyofibrillar mitochondrial subpopulations, only subsarcolemmal exhibited NAD⁺-dependent lactate oxidation. To further investigate the details of the physical association of LDH with mitochondria in muscle, immunofluorescence/confocal microscopy and immunoblotting approaches were used. LDH clearly colocalized with mitochondria in intact, as well as permeabilized fibers. LDH is likely localized inside the outer mitochondrial membrane, but not in the mitochondrial matrix. Collectively, these results suggest that extra-matrix LDH is strategically positioned within skeletal muscle fibers to functionally interact with mitochondria.

Introduction

The intracellular lactate shuttle hypothesis posits that lactate formed during glycolysis can be continuously used as an energy source within the same cell (1). However, controversy continues to surround the notion of mitochondrial lactate oxidation in skeletal muscle (recently reviewed in Ref. 2). For example, Yoshida et al. (3) found negligible direct lactate oxidation in either subsarcolemmal or intermyofibrillar mitochondria isolated from rat red (oxidative) and white (glycolytic) muscles. Using the saponin-permeabilized fiber approach to study mitochondria in situ, Ponsot et al. (4) observed no increase in respiration after lactate was added to mitochondria supported by ADP and malate. In keeping with shifting paradigms (5) and the lactate dehydrogenase (LDH)² reaction (lactate + NAD⁺ ⇆ pyruvate + NADH + H⁺), we hypothesized that mitochondrial lactate oxidation would require exogenous NAD⁺, the LDH reaction cofactor. Herein, we demonstrate that skeletal muscle mitochondria are indeed capable of lactate oxidation. In skeletal muscle, mitochondrial-associated LDH is located outside of the mitochondrial matrix.

EXPERIMENTAL PROCEDURES

Animals and Reagents

Sprague-Dawley rats were purchased from Charles River and housed in a climate controlled environment with appropriate light:dark cycles. Rats were allowed to eat standard chow and drink water ad libitum. For experiments involving permeabilized fibers, rats were >110 days old and weighed 500–600 g at the time of sacrifice. All procedures were approved by the Animal Care Committees of St. Francis Xavier University and Dalhousie University and conformed to the standards of the Canadian Council on Animal Care. For experiments involving isolated mitochondria, rats were 6–10 weeks old at the time of sacrifice. Rats were sacrificed via an intraperitoneal injection of 250 mg/ml of sodium pentobarbital (Vetoquinol, QC; permeabilized fibers) or isofluorane/CO₂ chamber (isolated mitochondria). All other chemicals were purchased from Sigma. The stock concentration of l-lactate dehydrogenase (from rabbit muscle) used in the mitochondrial studies was 1850 units/ml in 75 mm Tris, pH 7.5.

Antibodies

The following antibodies were used: anti-LDH (catalog number ab47010, Abcam); MitoProfile Total OXPHOS mixture (catalog number ab110411, Mitosciences); anti-Bcl-2 (catalog number 7382, Santa Cruz Biotechnology); Dylight549-conjugated donkey anti-mouse IgG (catalog number 711-505-152, Jackson ImmunoResearch Laboratories); Alexa 488-conjugated goat anti-mouse IgG (catalog number A11001, Invitrogen). Alexa Fluor® 680 goat anti-mouse IgG (catalog number A21057, Invitrogen); Alexa Fluor® 750 goat anti-rabbit IgG (catalog number A21039, Invitrogen).

Muscle Tissue

Permeabilized Fibers

Immediately following the confirmed death of the rat, the hindlimb was dissected and the gastrocnemius was removed and placed in ice-cold relaxing buffer X, consisting of (in mm): 50 MES, 7.23 K₂EGTA, 2.77 CaK₂EGTA, 20 imidazole, 0.5 DTT, 20 taurine, 5.7 ATP, 14.3 PCr, and 6.56 MgCl₂·6 H₂O (pH 7.1, 290 mOsmol). Two fiber bundles from each red and white portion of the gastrocnemius were gently separated along the longitudinal axis with needle tipped forceps (FST, Inc.) in ice-cold buffer X under magnification (Discovery V8, Carl Zeiss). Fiber bundles were then transferred to separate vials containing saponin (50 μg/ml) dissolved in 1.5 ml of buffer X for 30 min at 4 °C and frequently inverted. Following permeabilization, each fiber bundle was transferred to a separate vial containing a wash solution, buffer Z, consisting of (in mm): 105 K-MES, 30 KCl, 10 KH₂PO₄, 5 MgCl₂·6H₂O, 5 mg/ml BSA, pH 7.1. Fibers were kept in buffer Z for 20–40 min with frequent inversion until experimental testing.

Isolated Mitochondria

Similar procedures were performed for muscle extraction for isolated mitochondria, with the exception being that muscle was immediately placed in ice-cold isolation buffer (see “Mitochondrial Isolation” below). Additionally, liver was harvested in these rats, and placed directly in ice-cold mitochondrial isolation buffer.

Mitochondrial Isolation

Skeletal muscle mitochondria were isolated as described previously (3). Briefly, red gastrocnemius muscle was dissected away, and minced for 10 min on ice in isolation buffer consisting of (in mm): 50 Tris-HCl, 5 MgSO₄·7 H₂O, 5 EDTA, 100 KCl, 1.19 ATP, pH 7.4. Minced muscle was homogenized briefly with a Teflon pestle in a glass vial. Homogenate was centrifuged at 800 × g for 10 min. The supernatant, containing subsarcolemmal (SS) mitochondria, was removed and kept on ice. The pellet, containing intermyofibrillar mitochondria (IMF), was resuspended in isolation buffer without protease. We excluded protease in the isolation of IMF mitochondria, which has been the practice in previous studies, to address concerns about degradation of mitochondrial-bound LDH with protease treatment (6). The resuspension was centrifuged at 5000 × g for 5 min, after which, the supernatant was removed and discarded. The pellet was resuspended in isolation buffer and centrifuged at 800 × g for 10 min. The supernatant, containing IMF mitochondria was removed, and kept on ice. Both suspensions of SS and IMF mitochondria were centrifuged at 12,000 × g for 10 min, their pellets were resuspended in isolation buffer, and centrifuged once more at 12,000 × g for 10 min. The resulting SS and IMF pellets were resuspended in MiR05 respiration assay buffer, consisting of (in mm): 0.5 EGTA, 10 KH₂PO₄, 3 MgCl₂·6 H₂O, 60 K-lactobionate, 20 HEPES, 110 sucrose, and 1 mg/ml of fatty acid free BSA, pH 7.1. All isolated mitochondrial suspensions were kept on ice until respirometric analysis.

Protein Assay

Isolated mitochondria were resuspended in 100 μl of lysis buffer, consisting of (in mm): 50 Tris, 150 NaCl, 50 Na₂HPO₄, 1 Na₃VO₄, 1 NaF, and 0.1% Nonidet P-40, and 0.25% sodium deoxycholate. The lysis buffer also contained 1 μl of a protease inhibitor mixture (Sigma, catalog number P8340). Samples were homogenized and incubated at 4 °C with constant agitation for 30 min, then centrifuged at 12,000 × g for 15 min. Protein concentration was determined in the supernatant of each sample with a modified Lowry method (Bio-Rad DC Protein Assay, catalog number 500-0116).

Respirometry

Mitochondrial respiratory oxygen flux (JO₂) was measured in high resolution using the Oxygraph-2k (OROBOROS Instruments, Innsbruck, AT). All mitochondrial samples were assessed in 2 ml of respiration assay buffer, consisting of buffer Z for the permeabilized fiber experiments, and MiR05 for experiments involving isolated mitochondria. Buffer Z was supplemented with 20 mm creatine to saturate the creatine kinase reaction (7) and 25 μm blebbistatin, an inhibitor of myosin ATPase, to prevent spontaneous, temperature-sensitive contractions (8). The concentration of O₂ in the experimental chambers was maintained between 100 and 300 μm for permeabilized fibers, and between 10 and 200 μm for experiments involving isolated mitochondria. All experiments were conducted at 37 °C. Instrumental background O₂ consumption was corrected using equations determined under the same parameters used for experimental data collection. Four respirometric protocols, involving sequential addition of substrates and inhibitors and/or titration of substrates, were used: 1) 5 mm ADP, 4 mm malate, 5 mm lactate, 250 μm NAD⁺ (permeabilized fibers), 500 μm NAD⁺, 10 mm lactate (isolated mitochondria), 10 μm cytochrome c, 250 μm α-cyano-4-hydroxycinnamate (CHC; permeabilized fibers only), 500 μm CHC, 1 mm CHC (isolated liver mitochondria), 10 mm glutamate (isolated mitochondria); 2) in permeabilized fibers only: 5 mm ADP, 4 mm malate, titration of pyruvate (50–1000 μm), 250 μm NAD⁺, 500 μm NAD⁺, 10 μm cytochrome c, 250 μm CHC, 500 μm CHC; 3) in permeabilized fibers only: 5 mm ADP, 4 mm malate, 500 μm NAD⁺, titration of lactate 0.5–15 mm, 500 μm CHC; 4) in permeabilized fibers: 5 mm ADP, 4 mm malate, 10 mm lactate, 1 mm NAD⁺, 9.25 units/ml of LDH, 10 μm cytochrome c, 5 μm CHC, 10 mm glutamate. For permeabilized fibers, JO₂ was expressed as picomol O₂ × s⁻¹ × mg⁻¹ tissue wet weight or as a percentage of glutamate-stimulated respiration. JO₂ in isolated mitochondria was expressed as picomol O₂ × s⁻¹ × mg⁻¹ of protein and presented as a percentage of glutamate-stimulated JO₂ to better compare the two mitochondrial subpopulations. For the lactate titrations (Fig. 2), basal JO₂ with ADP + malate + NAD⁺ was subtracted from each subsequent JO₂ elicited by addition of lactate (i.e., net JO₂). For tests involving protease treatment (Figs. 7 and 8), 10 μm trypsin was added to the oxygraph chambers containing buffer at 37 °C for 15 min prior to substrate (ADP + malate). 20 μm soybean trypsin inhibitor was added to both treatment and control chambers immediately after the 15-min typsin treatment.

FIGURE 2.

Respiratory kinetics for lactate in permeabilized red (RG) and white (WG) saponin-permeabilized rat gastrocnemius muscle fibers. A, net mitochondrial respiratory O₂ flux (JO₂) in red (closed symbols) and white fibers (open symbols) supported by increasing concentrations of lactate (0.5–15 mm) titrated atop 5 mm ADP + 4 mm malate and 500 μm NAD⁺. Michaelis-Menten kinetic curves: WG, Y = 4.681 · X/(12.08 + X); RG, Y = 10.03 · X/(4.615 + X). B, apparent K_m (K_m,app) of RG and WG mitochondria for lactate, as computed from kinetic curves. C, maximal net JO₂ (V_max) of RG and WG mitochondria supported by lactate, with 5 mm ADP + 4 mm malate and 500 μm NAD⁺, as computed form kinetic curves. n = 3. **, main effect for muscle fiber type, p < 0.01; *, significantly different from WG, p < 0.05. Data are mean ± S.D.

FIGURE 7.

Effect of protease treatment (10 μm trypsin for 15 min at 37 °C) on mitochondrial respiratory O₂ flux (JO₂) in red saponin-permeabilized (30 μg/ml) rat gastrocnemius muscle fibers supported by components of the LDH reaction. No significant difference was observed between trypsin-treated mitochondria versus control. As with untreated fibers, trypsin-treated fibers responded to lactate + NAD⁺. n = 3. Control, significantly greater than ADP + malate, and lactate JO₂, aa, p < 0.01; aaa, p < 0.001, significantly less than cytochrome c JO₂; b, p < 0.05, significantly greater than all other JO₂; ccc, p < 0.001. Protease, significantly greater than ADP + malate, and lactate JO₂; d, p < 0.05; dd, p < 0.01; ddd, p < 0.001, less than cytochrome c JO₂; ee, p < 0.01, significantly greater than all other JO₂; fff, p < 0.001. Data are mean ± S.D.

FIGURE 8.

Immunoblot for LDH and IMM protein complex II in permeabilized muscle fibers. Red gastrocnemius fibers permeabilized with 30 μg/ml of saponin and treated with 10 μm trypsin for 15 min revealed that LDH remains present even after trypsin treatment (lane 2). Commercially available purified LDH is shown without (lane 3) or with (lane 4) trypsin treatment.

Immunofluorescence

Dissected muscle fibers were fixed in 4% paraformaldehyde for 1 h, treated with 100 mm glycine for 30 min, and permeabilized with Tris-buffered saline containing 0.1% Tween 20 (w/v) for 20 min. The samples were blocked for 1 h with PBS containing 1% gelatin. Fibers were incubated with anti-LDH (1 μg/ml) for 1 h at room temperature, washed three times with PBS, and incubated with the secondary antibody goat anti-rabbit conjugated with Dylight 549 for another hour. Following washes, the same samples were incubated with either anti-Bcl-2 (1/500) or anti-Mito mixture (1/1000) for 2 h, washed, and mounted with DAKO mounting media. Images were collected using a fluorescent confocal microscope (Carl Zeiss). For the different panels shown in Figs. 4–6, identical settings were used among panels. Experiments were repeated three times in triplicate. Colocalization/correlation analyses of data presented in Fig. 4 were performed using ImageJ software (9).

FIGURE 4.

High-resolution confocal laser scanning microscopic imaging of immunolabeled LDH (red in A and B), IMM proteins (green in A) and Bcl-2 (green in B). Antibodies against Bcl-2 were used as a label of the outer mitochondrial membrane. Antibodies against the respiratory chain proteins were used to label the IMM. Pearson correlation coefficients for colocalization were 0.88 for IMM with LDH, and 0.73 for Bcl-2 with LDH. The scatter plots show the intensity of the red (Ch-1) and green (Ch-2) pixels.

FIGURE 5.

Confocal laser scanning microscopic imaging of immunolabeled IMM and LDH of red gastrocnemius skeletal muscle fibers. Antibodies against proteins of the mitochondrial respiratory chain were used as a label for the IMM, along with antibodies against LDH. A, labeling of the intact fibers. B, labeled permeabilized fibers exhibiting attenuated LDH signal in the presence of the IMM signal. Note that the LDH signal was preserved in the SS mitochondria, which appear as larger elongated organelles at the cell edge. C, augmentation of the LDH signal by addition of the exogenous LDH (approximately 25 units/ml of LDH) + gentle wash. D, loss of the LDH staining in mitochondria with trypsin treatment.

FIGURE 6.

Confocal laser scanning microscopic imaging of immunolabeled Bcl-2 and LDH of red gastrocnemius skeletal muscle fibers. Antibodies against Bcl-2 were used as a label of the outer mitochondrial membrane, along with antibodies against LDH. A, labeling of the intact muscle fibers. The mitochondrial network is easily recognizable, as is the colocalization of Bcl-2 (Aa) and LDH staining (Ab). An overlay of the images are shown in Ac. B, permeabilized fibers, showing preservation of the colocalization between the Bcl-2 (Ba) and LDH (Bb). C, area of the same preparation shown in panel B, but showing the region with the lost Bcl-2 signal (Ca), whereas the LDH signal is present (Cb).

Immunoblotting

Following permeabilization and treatment with either trypsin or vehicle (as described previously under “Muscle Tissue” “Permeabilized Fibers”), rat red gastrocnemius muscles were collected in cold buffer X from euthanized male adult Sprague-Dawley rats (n = 3). Samples were homogenized in modified RIPA buffer (50 mm Tris-HCl, 150 mm NaCl, 1% Triton X-100, 1% sodium deoxycholate, and 0.1% SDS, pH 7.5). Following protein quantification (DC Protein Assay, Bio-Rad, catalog number 500-0116), 100 μg of protein was separated in 12% SDS-PAGE gels and transferred to polyvinylidene fluoride membranes (catalog number IPFL00010 Immobilon® FL, Millipore, Billerica, MA). Membranes were stained with Ponceau red to assess loading; after they were rinsed with Tris-buffered saline/Tween 20 (TBST; 20 mm Tris base, 137 mm NaCl, and 0.1% Tween 20, pH 7.6), the blots were blocked in 5% skim milk/TBST for 1 h. Blots were incubated with antibodies for LDH (catalog number ab47010, Abcam Inc., Cambridge, MA), and MitoProfile Total OXPHOS Human WB Antibody Mixture (catalog number ab110411, Abcam Inc.) overnight at 4 °C with constant agitation. Following 3 washes with TBST, immunoblots were incubated with either Alexa Fluor 680 goat anti-mouse IgG (catalog number A21057, Invitrogen) or Alexa Fluor 750 goat anti-rabbit IgG (catalog number A21039, Invitrogen) depending on the primary antibody used. The bands were detected with Odyssey version 3.0, and multiple exposures were generated to ensure the linearity of the fluorescent signal.

Statistical Analysis

Results are expressed as mean ± S.D. Two-way analysis of variance with repeated measures were carried out to determine whether an interaction between the levels of variables exists. Additional within groups one-way analysis of variances with repeated measures (Bonferroni post hoc analyses) were performed to examine differences among fiber types or isolated mitochondrial subpopulations. The software package SPSS (Statistical Package for Social Sciences 15.0, SPSS Inc., Chicago, IL) was used for all statistical calculations. The α-level for statistical significance was set a priori at 0.05. Enzyme kinetic curves were fit to the Michaelis-Menten model (i.e. Y = V_max × X/(K_m + X), where X = [substrate], and Y = JO₂) using Prism 5.0 (GraphPad Software, Inc., La Jolla, CA).

RESULTS

Before examining possible lactate oxidation in either the red gastrocnemius (RG) or white gastrocnemius (WG), respiration in the two muscle types was compared (Fig. 1A). Respiration in permeabilized red fibers was greater than white muscle fibers. This is in agreement with previous studies and is due to greater mitochondrial content in red muscle. Subsequent experiments were therefore conducted in RG fibers. Using RG presented an additional advantage of responding to exogenous cytochrome c, suggesting damage to the outer mitochondrial membrane (OMM) of at least some fibers in the sample. The heterogeneous permeabilization pattern that we subsequently confirmed using confocal imaging would allow us to examine mitochondrial LDH localization in instances in which mitochondria appear to have lost the OMM.

FIGURE 1.

Mitochondrial respiratory O₂ flux (JO₂) in red and white saponin-permeabilized rat gastrocnemius muscle fibers supported by components of the LDH reaction. A, addition of 5 mm lactate to either red or white fibers (5 mm ADP + 4 mm malate) had no affect on JO₂. Addition of NAD⁺ significantly increased JO₂. 10 μm cytochrome c further increased JO₂ in RG. The rate was subsequently abolished to pre-NAD⁺ levels following addition of the pyruvate transporter inhibitor, CHC. n = 6. ***, significantly different, p < 0.001. B, respiration normalized as a percentage of glutamate-stimulated JO₂ indicate no differences in proportional respiration between RG and WG. There was a non-significant trend for increased JO₂ after the addition of LDH to the fibers. n = 3. WG, significantly greater than ADP + malate, lactate, and CHC JO₂; aa, p < 0.01; aaa, p < 0.001; greater than CHC JO₂; bb, p < 0.01; bbb, p < 0.001. RG, significantly greater than ADP + malate, lactate, and CHC JO₂; c, p < 0.05; ccc, p < 0.001; greater than CHC JO₂; d, p < 0.05; ddd, p < 0.001. Data are mean ± S.D.

Contrary to many (e.g. (10)), but not all (11) previous studies, we observed greater respiration in SS mitochondria compared with IMF mitochondria. This was likely due to the exclusion of protease in the isolation of IMF mitochondria, unlike other studies examining lactate oxidation in isolated skeletal muscle mitochondria (3, 12). As mentioned, protease was excluded to address concerns about potential degradation of mitochondrial-bound LDH with protease treatment (6). To better compare the two subpopulations in the present study, respiration is expressed as a percentage of glutamate-stimulated JO₂ (glutamate added at the end of the protocol; Fig. 3).

FIGURE 3.

Lactate oxidation in isolated SS and IMF mitochondria from rat red gastrocnemius. Isolated mitochondrial respiratory O₂ flux (JO₂ as a % of glutamate-stimulated JO₂) was supported by 5 mm ADP + 4 mm malate. Addition of 5 mm lactate did not increase JO₂ in either SS or IMF mitochondria. However, 500 μm NAD⁺ increased JO₂ in SS mitochondria only. 10 μm cytochrome c further increased JO₂. The NAD⁺-dependent increase in JO₂ was abolished with 500 μm CHC. n = 3. SS, significantly greater than ADP + malate, lactate, and CHC JO₂; aaa, p < 0.001, significantly greater than all other conditions; bbb, p < 0.001, significantly greater than lactate JO₂; cc, p < 0.01. IMF, significantly greater than all other conditions; dd, p < 0.01. Data are mean ± S.D.

NAD⁺ Dependence of Lactate Oxidation in Permeabilized Fibers

To test whether mitochondria in situ are capable of oxidizing lactate, 5 mm lactate was added to both permeabilized red and white fibers atop 5 mm ADP + 4 mm malate. Addition of lactate resulted in no change to JO₂ in either RG or WG (Fig. 1). As hypothesized, addition of 500 μm NAD⁺ resulted in a significant increase in JO₂ in both RG and WG (p < 0.001), suggesting that extramitochondrial NAD⁺ is a requisite cofactor for the mitochondrial LDH reaction in both muscle fiber types. This increase in NAD⁺-dependent JO₂ represented a 74.6 and 94.7% increase in RG and WG, respectively, above JO₂ prior to NAD⁺ addition. To further determine whether this lactate oxidation occurred outside of the mitochondrial matrix, 500 μm CHC, an inhibitor of mitochondrial pyruvate transport across the inner mitochondrial membrane (IMM) (13), was added. As expected, the NAD⁺-induced increase in JO₂ was abolished to pre-NAD⁺ rates by CHC in RG and WG (p < 0.001; Fig. 1A). However, 500 μm CHC is high enough to raise concerns about whether MCT1, the lactate transporter, may have been inhibited as well. Therefore, a 100-fold lower concentration (5 μm) was used in another set of experiments (Fig. 1B). Similar, robust inhibition of lactate + NAD⁺-supported JO₂ was observed in RG and WG (p < 0.001). This implies that lactate was first oxidized to pyruvate, and then transported across the IMM for subsequent oxidation. As a positive control, the lactate protocol was repeated in permeabilized fibers, with the exception that pyruvate was added in place of lactate (data not shown). As with the lactate experiment (Fig. 1), addition of CHC inhibited respiration to pre-pyruvate rates. This further supports that pyruvate, but not lactate per se, enters the mitochondrial matrix.

Sensitivity and Capacity of Red and White Permeabilized Fibers for Lactate-supported Mitochondrial JO₂

To determine the sensitivity to lactate (K_m,app), and capacity (V_max) of lactate-supported JO₂ in RG versus WG permeabilized fibers, lactate was titrated (0.5–15 mm) atop 5 mm ADP + 4 mm malate + 500 μm NAD⁺. As illustrated in Fig. 2A, respiration tended to be greater in RG mitochondria for a given concentration of lactate. Indeed, the K_m,app for lactate (Fig. 2B) was significantly lower, and the V_max for JO₂ supported by lactate (Fig. 2C) significantly greater (p < 0.05) in RG versus WG. Thus, RG exhibited both greater sensitivity to lactate, as well as a greater capacity for lactate-stimulated respiration.

NAD⁺ Dependence of Lactate Oxidation in Subsarcolemmal and Intermyofibrillar Isolated Skeletal Muscle Mitochondria

To further explore which subpopulations of skeletal muscle mitochondria may be responsible for lactate-stimulated JO₂, the lactate protocol was replicated in SS and IMF mitochondria isolated from rat RG. During ADP + malate-supported respiration in IMF mitochondria, addition of either lactate or lactate + NAD⁺ did not increase JO₂ (Fig. 3). However, SS mitochondria recapitulated the results of permeabilized fiber experiments. More specifically, whereas addition of lactate did not increase JO₂ in SS mitochondria, subsequent addition of NAD⁺ significantly increased JO₂ (p < 0.001; Fig. 3). Moreover, as in the permeabilized fiber experiments, CHC effectively abolished JO₂ to pre-NAD⁺ rates.

Immunofluorescence in Permeabilized Fibers

The results of the functional assays are consistent with the notion of close spatial positioning of LDH with mitochondria. To further investigate the potential colocalization between LDH and mitochondria, immunofluorescent labeling of LDH and mitochondrial membrane proteins were performed in RG fibers. The RG fibers used for immunostaining were prepared in parallel with the functional assay to ensure identical conditions for both experiments.

In intact fibers, scanning laser confocal microscopic imaging of antibodies specific for IMM proteins (mixture of antibodies against proteins of the respiratory chain, Mitosciences®) and LDH confirmed mitochondrial LDH colocalization (Figs. 4A and 5A, a–c). In an attempt to discern whether LDH was associated with the OMM or IMM, we studied with immunofluorescence the relative localization of LDH with respect to Bcl-2, which is bound to the OMM. In intact fibers, LDH colocalized strongly with Bcl-2 (Figs. 4B and 6A, a–c). Colocalization of LDH with Bcl-2 was largely preserved upon membrane permeabilization (Fig. 6B, a–c). There were rarely cells in which the Bcl-2 signal was lost (presumably due to compromised OMM integrity), but the LDH signal remained (Fig. 6C, a–c). This phenomenon is illustrated by the orange pseudo-color of the overlay image in Fig. 6C, c, a result of the markedly reduced Bcl-2 (green) signal. These observations suggest an association between LDH and mitochondria that is independent of the OMM.

We observed essentially three phases of intactness among the muscle fibers permeabilized with 50 μg/ml of saponin: 1) completely intact mitochondria (Fig. 6B, a–c); 2) a transitional phase, in which Bcl-2 is gone, but LDH remains colocalized to the IMM (orange signal; Fig. 6C, a–c); and 3) loss of LDH, with IMM still present (Fig. 5B). Indeed, in Fig. 5B, a–c, there are regions where the LDH signal was diminished, but the IMM signal remained. This pattern is consistent with the idea that LDH binding sites are on the outer side of the IMM. This was further confirmed by restoration of the LDH signal when the exogenous enzyme was added to the permeabilized fibers (Fig. 5C, a–c). Furthermore, in fibers with damaged OMM, the LDH signal was markedly lower in some of the fibers in samples treated with trypsin (Fig. 5D, a–c), suggesting that LDH in skeletal muscle mitochondria is localized outside of the matrix, but within the OMM. Although we observed a markedly lower LDH signal in IMF mitochondria of permeabilized fibers, the signal was largely preserved in the SS mitochondria (note insets in Fig. 5, B and C). This suggests a stronger association of LDH with SS compared with IMF mitochondria.

Effect of Protease on Mitochondrial Lactate Oxidation and LDH Protein

To further examine the localization of mitochondrial LDH, and allay concerns about LDH contamination and mitochondrial integrity, RG fibers were first permeabilized with a saponin concentration 40% lower than prescribed by standard protocols. This was to ensure intactness of the OMM, as confirmed with the cytochrome c test. RG fibers were treated with trypsin, and then assessed for NAD⁺-dependent lactate oxidation (Fig. 7). Despite a negligible reduction in NAD⁺-dependent lactate oxidation, and an increase in glutamate-stimulated JO₂, the mitochondrial LDH function persisted.

In an effort to further assess whether LDH was lost with trypsin treatment, RG fibers were first subjected to conditions paralleling the functional assays. Fiber proteins were then immunoblotted using antibodies for LDH and complex II (IMM protein). In contrast to the degradation of LDH standard with trypsin (Western blot lanes 3 and 4 in Fig. 8), LDH appears to have remained in fibers, despite trypsin treatment (Western blot lanes 1 and 2 in Fig. 8). This finding, along with the results of the functional assays, suggests that a substantial level of protease-inaccessible LDH remains in permeabilized skeletal muscle.

DISCUSSION

The major finding of the present study is that permeabilized muscle fibers and isolated SS mitochondria oxidize lactate in an NAD⁺-dependent manner. This was demonstrated by an increase in JO₂ following addition of NAD⁺ to permeabilized RG and WG fibers or isolated SS mitochondria, but not isolated IMF mitochondria (Fig. 3). These results demonstrate that lactate oxidation occurs in skeletal muscle mitochondria. However, transport of lactate into the mitochondrial matrix is not a requirement for its oxidation. This was evidenced by a return of JO₂ to pre-NAD⁺ rates following addition of the mitochondrial pyruvate transporter inhibitor, CHC, even in the presence of lactate (Figs. 1, 3, and 7).

The significant cytochrome c response (i.e. increased JO₂ upon addition of exogenous cytochrome c; Fig. 1) suggests that in RG, permeabilization compromised the OMM in some mitochondria. Published during the preparation of the current manuscript, the work of Jacobs et al. (14) similarly demonstrated mitochondrial lactate oxidation in permeabilized skeletal muscle fibers. In their study, although exogenous LDH did not increase respiration, no cytochrome c response was observed (14). The reasons for this discrepancy in cytochrome c response are unclear, but may be due to the differences in muscle samples (i.e. rat gastrocnemius versus human m. vastus lateralis), the different assay buffers used between studies (i.e. potassium salt versus sucrose buffer), or even the different sequence of additions (i.e. cytochrome c added after lactate + NAD⁺ versus after pyruvate). It should be noted that the same sucrose buffer (i.e. MiR05) was used in the current study to assess isolated mitochondria (Fig. 3); in WG, no cytochrome c response was observed (Fig. 1). Nevertheless, to allay concerns about potential LDH contamination in permeabilzed RG, we first reduced the saponin concentration to 30 μg/ml. This concentration of saponin was previously shown to preserve the OMM better in human m. vastus lateralis muscle fibers assessed in potassium salt buffer (15). Incubating permeabilized RG fibers in 10 μm trypsin for 15 min at 37 °C did not significantly affect the NAD⁺-dependent increase in lactate-supported mitochondrial JO₂ in permeabilized RG fibers (Fig. 7). These data are fully consistent with the “persistent” level of protease-inaccessible mitochondrial LDH in rat liver, kidney, and heart described by Brandt et al. (16) over 25 years ago. The data in the current study are also consistent with the notion of significant amounts of LDH bound to the OMM, as well as localized to the mitochondrial intermembrane space (16). The work of Brandt et al. (16) also demonstrated that with the exception of kidney, the LDH contained in the intermembrane space and bound to the OMM in rat tissues exhibit the same LDH isoform profile as the cytosolic fraction. A line of inquiry further exploring the relationship between LDH isoform quantities with the relative rates of mitochondrial lactate oxidation represents a logical extension for further studies as methodologies improve, as well as elucidating the physiological processes by which mitochondrial lactate oxidation contributes to cellular metabolic redox homeostasis in muscle.

Research aimed at determining whether mitochondria may be fueled by lactate in skeletal muscle has not resolved the issue (2–5, 12, 17–23). Histochemical (24, 25) and immunolabeling (19) techniques have been used to verify the presence of LDH in the mitochondria. Using stable carbon isostope-labeled lactate (i.e. [3-¹³C]lactate), Bertocci and Lujan (17) implied that lactate was oxidized in the mitochondria. Moreover, in isolated rat skeletal muscle mitochondria, Brooks et al. (19) demonstrated oxidation of lactate. However, several studies were unable to observe oxidation of lactate by mitochondria in vitro (3, 12, 22) or in situ (4). It was subsequently suggested that the results of Brooks et al. (19) may owe to cytosolic LDH contamination in the mitochondrial preparation and/or oxidation of malate (3, 26). However, it has been argued (6) and observed (20) that protease used in the procedure for isolating mitochondria from skeletal muscle may result in loss of mitochondrial LDH. In agreement with the results of the current study, the evident loss of mitochondrial LDH during mitochondrial isolation strongly suggests that the enzyme is located outside of the IMM. Indeed, we demonstrate that it is only with exogenous NAD⁺ that observable lactate-supported JO₂ can occur in permeabilized fibers or isolated skeletal muscle mitochondria (Figs. 1 and 3). Moreover, the inhibition of lactate-supported JO₂ by CHC (Figs. 1, 3, and 7) further supports the notion that mitochondrial LDH is located outside of the IMM. Supporting the functional data, immunofluorescent staining of highly permeabilized fibers (i.e. LDH absent, IMM proteins remaining) also puts the location of mitochondrial-associated LDH external to the matrix (Figs. 5 and 6).

Several previous studies lent support for the direct uptake and oxidation of lactate by the mitochondria (19, 23, 27, 28), as per the intracellular lactate shuttle hypothesis (ILSH). The ILSH postulates that lactate can be directly transported into the mitochondria where it is oxidized, to fuel respiration (1). The present study was unable to find evidence in support of the ILSH as originally proposed by Brooks et al. (19). If this shuttle exists as such, the addition of lactate to the experimental chambers in the current study should have induced an increase in JO₂. However, with the exception of isolated liver mitochondria (data not shown), evidence of direct lactate oxidation (i.e. lactate transported into the mitochondrial matrix and oxidized) was not observed in the current study. The results of the current study are, however, consistent with aspects of the more recently described lactate oxidation complex, in which mitochondrial LDH is anchored to the outside of the IMM (5). The results of the current study are also congruent with findings of Sahlin et al. (12), Rasmussen et al. (22), and Yoshida et al. (3), who have all carried out experiments aimed at investigating the ILSH. These studies were completed using isolated mitochondria and utilized similar methodologies to those used in the studies by Brooks et al. (19). The lack of support for the ILSH was attributed to methodological issues (29), including protease use when isolating mitochondria. This provided reasonable grounds for the present study to investigate this controversial concept with a different approach, using permeabilized fibers. In a study utilizing the saponin-permeabilized myofiber approach, Ponsot et al. (4) concluded that mitochondria in situ did not oxidize lactate. The present investigation built upon the study by testing the LDH cofactor NAD⁺ in addition to lactate, and inhibiting pyruvate transport with CHC. As depicted in Fig. 1, the addition of NAD⁺ to the chamber following lactate did elicit an increase in JO₂. As hypothesized, this rise in respiration would be due to the presence of newly formed pyruvate that resulted from the biochemical reaction requiring lactate and NAD⁺, and catalyzed by the enzyme LDH. Pyruvate, which can easily cross the IMM, was then transported into the mitochondria and oxidized by pyruvate dehydrogenase to acetyl-CoA, and further oxidized in the Krebs cycle for ATP synthesis.

The kinetics of lactate-stimulated JO₂ (NAD⁺ present) in the current study exhibited differences between RG and WG (Fig. 2), which reflect a metabolic strategy consistent with the intercellular- or “cell-cell,” lactate shuttle (29, 30). In the current study, the sensitivity toward lactate oxidation in the more oxidative RG (predominantly type IIA oxidative/glycolytic fibers) versus WG (predominantly type IIB glycolytic fibers) was significantly greater (i.e. lower K_m,_app for RG; Fig. 2B). This is in line with the work of Pagliassotti and Donovan (31, 32) on lactate removal and disposal among different skeletal muscle fiber types, which they determined to be favored among type IIA compared with type IIB fibers. The physiological consequences of kinetic differences among muscle fiber types for mitochondrial-associated lactate oxidation support the notion that the more oxidative fiber types make a larger contribution to both absolute (greater V_max) and relative (lower K_m,_app) lactate disposal/oxidation. The mechanisms by which these different sensitivities of LDH for its substrates may occur have been attributed to the interaction of LDH with other cellular proteins in muscle (e.g. actin, tubulin) (33); thus, conducting these experiments in permeabilized fibers provides a physiological representation of the in vivo sensitivity of LDH in different muscle types (i.e. glycolytic versus oxidative).

Observable differences in morphology, physiology, and metabolism of mitochondrial subpopulations in muscle have long been recognized (e.g. Refs. 10 and 34–37); reports continue to reveal differences between the SS and IMF mitochondria of both tissues. The results of the current study demonstrate that isolated SS mitochondria are capable of oxidizing lactate, whereas isolated IMF mitochondria are not (Fig. 3). This is in slight contrast to the findings of Yoshida et al. (3) in which they observe greater, albeit negligible, lactate oxidation in IMF mitochondria from rat red and white skeletal muscle compared with isolated SS mitochondria. Because we did not observe lactate oxidation in the absence of NAD⁺, in either mitochondrial subtype, it is difficult to speculate as to how respiration might be stimulated with lactate alone in previous studies. Perhaps under certain conditions, a small amount of NAD⁺ can remain bound to LDH in skeletal muscle mitochondria, as it does with liver mitochondria (38). Indeed, in a separate set of experiments, we observed a marked increase in JO₂ following addition of lactate to isolated rat liver mitochondria without adding NAD⁺ (data not shown). However, based on the results of the current study, the mitochondria from rat skeletal muscle does not appear to be capable of lactate oxidation without supplying exogenous NAD⁺.

Our immunofluorescence investigation highlighted several important facts related to the mitochondrial-LDH relationship in the intact and permeabilized fiber preparations. First, although evidence of close localization of LDH and mitochondria in muscle fibers was suggested previously by multiple experiments, to our knowledge these data are the first direct demonstration of the colocalization between mitochondria and LDH in the permeabilized fiber preparation. Second, these data support that LDH is localized inside the OMM and likely physically associated with the IMM, as LDH was still present in fibers that lost Bcl-2 (Fig. 6C), but was lost in some fibers that still showed IMM proteins (Fig. 5C). Although Bcl-2 has been observed in the endoplasmic reticulum (39), the logical inference based on our observation of colocalization between LDH and both Bcl-2 and IMM proteins would seem to indicate that most of the Bcl-2 in our preparations is either localized to the OMM, or so close to the mitochondria that it cannot be resolved by confocal imaging. It is possible that endoplasmic reticulum-localized Bcl-2 could be limited to linkage sites of the endoplasmic reticulum-mitochondria interface (40). Nevertheless, the colocalization profile of LDH with IMM proteins indicates quite unambiguously that LDH is associated with the mitochondria, regardless of Bcl-2. Interestingly, we also found a stronger association of LDH with SS versus IMF mitochondria. Although the molecular basis of this association is not entirely clear, it is noteworthy that these data are consistent with our functional assays (Fig. 3).

The possibility of contaminating LDH infiltrating the mitochondrial intermembrane space, and thereby evading the proteolytic activity of trypsin is unlikely. The molecular masses of tetrameric LDH, trypsin, and cytochrome c are ∼140, 23, and 12 kDa, respectively. Without an increase in JO₂ following the addition of cytochrome c (indicating intact OMM; Fig. 7), and a plausible quick-acting mechanism for LDH transport across the OMM, LDH should be prevented from entering the mitochondrial intermembrane space. Moreover, assuming contaminating LDH was responsible for the NAD⁺-dependent lactate oxidation in the current study, trypsin should have access to any contaminating LDH. Indeed, trypsin treatment did not remove the NAD⁺-dependent mitochondrial lactate oxidation, nor did exogenous LDH significantly increase respiration in trypsin-treated fibers. Because exogenous cytochrome c failed to increase respiration, the results of experiments illustrated in Fig. 7 support that at least some of the LDH in skeletal muscle is located within the OMM, but outside the IMM.

Conventions of cellular metabolism are important to the knowledge base upon which applied research relies. The results presented herein help to reconcile seemingly incongruous findings regarding mitochondrial lactate oxidation. Indeed, it appears muscle fibers have the potential to use the same lactate they produce, but this lactate is converted first to pyruvate before entering the mitochondrial matrix. Addressing how challenges to metabolic homeostasis affect, and are affected by mitochondrial lactate oxidation in muscle may constitute topics for future research. Assessing mitochondrial lactate oxidation in permeabilized tissue may also serve as an outcome variable for testing applications aimed at modifying metabolism.