Loss of positive allosteric interactions between neuronal nitric oxide synthase and phosphofructokinase contributes to defects in glycolysis and increased fatigability in muscular dystrophy

Abstract

Duchenne muscular dystrophy (DMD) involves a complex pathophysiology that is not easily explained by the loss of the protein dystrophin, the primary defect in DMD. Instead, many features of the pathology are attributable to the secondary loss of neuronal nitric oxide synthase (nNOS) from dystrophin-deficient muscle. In this investigation, we tested whether the loss of nNOS contributes to the increased fatigability of mdx mice, a model of DMD. Our findings show that the expression of a muscle-specific, nNOS transgene increases the endurance of mdx mice and enhances glycogen metabolism during treadmill-running, but did not affect vascular perfusion of muscles. We also find that the specific activity of phosphofructokinase (PFK; the rate limiting enzyme in glycolysis) is positively affected by nNOS in muscle; PFK-specific activity is significantly reduced in mdx muscles and the muscles of nNOS null mutants, but significantly increased in nNOS transgenic muscles and muscles from mdx mice that express the nNOS transgene. PFK activity measured under allosteric conditions was significantly increased by nNOS, but unaffected by endothelial NOS or inducible NOS. The specific domain of nNOS that positively regulates PFK activity was assayed by cloning and expressing different domains of nNOS and assaying their effects on PFK activity. This approach yielded a polypeptide that included the flavin adenine dinucleotide (FAD)-binding domain of nNOS as the region of the molecule that promotes PFK activity. Smaller peptides in this domain were then synthesized and used in activity assays that showed a 36-amino acid peptide in the FAD-binding domain in which most of the positive allosteric activity of nNOS for PFK resides. Mapping this peptide onto the structure of nNOS shows that the peptide is exposed on the surface, readily available for binding. Collectively, these findings indicate that defects in glycolytic metabolism and increased fatigability in dystrophic muscle may be caused in part by the loss of positive allosteric interactions between nNOS and PFK.

INTRODUCTION

The complex pathophysiology of dystrophin-deficiency is now understood as a consequence of multiple molecular defects that arise from the mutation of a single gene. Although the inability to express normal dystrophin is the primary defect of Duchenne muscular dystrophy (DMD) in humans and mdx dystrophy in mice, defects that arise from the secondary loss of other members of the dystrophin glycoprotein complex (DGC) are now known to contribute significantly to dystrophinopathy. For example, the discovery that the concentration of neuronal, nitric oxide synthase (nNOS) is dramatically reduced in dystrophin-deficient muscle (1,2) led to a more mechanistic understanding of how loss of dystrophin from muscle could contribute to muscle ischemia because of defects in normal NO-modulation of vasoconstriction (3,4). In addition, the secondary loss of nNOS from dystrophin-deficient muscle contributes importantly to exacerbating inflammation of dystrophic muscle (5,6), and its loss contributes to increased cardiac fibrosis, defects in neuromuscular junction (NMJ) structure, disruptions in the regenerative capacity of dystrophic muscle and defects in neurogenesis in the brains of mdx mice (7–10). In addition to these well-delineated features of the dystrophic pathology that result from the loss of normal NO production by dystrophic muscle, NO can influence the expression of numerous genes that may be dysregulated in muscular dystrophy and contribute to the pathology in unknown ways. For example, expression profiling studies show that NO is a strong inducer of the expression of genes in non-muscle cells that encode proteins that regulate transcription, influence the cell cycle, function in cell signaling and affecting cell motility (11).

DMD males and mdx mice also show increases in muscle fatigability that cannot be attributed directly to the loss of dystrophin, and may reflect loss of function of other proteins in the DGC. The increase in fatigability in DMD can be substantial. For example, DMD boys who were tested between the ages of 8 and 10 years were able to maintain a contraction of the biceps brachii at 60–70% of maximum voluntary contraction (MVC) for only 10 s (12). In contrast, healthy boys could maintain contraction at 60–70% MVC for 45 s. Because these data were expressed as % MVC maintained over time, they would reflect fatigability of the muscles rather than defects in force production caused by the loss of muscle mass in DMD. Generally similar findings have been reported in mdx mice. Tetanic stimulation of fast, glycolytic, muscles for 280 ms every 5 s produced much more fatigue in mdx muscle than in age-matched C57, expressed as decline in force production relative to pre-tetanic stimulation levels (13). This latter experimental design would preferentially reflect changes in fatigability of fast, glycolytic fibers during anaerobic contraction, suggesting that the increased fatigability could result from defects in glycolysis.

Recent investigations have shown that muscles of nNOS null mice experience increased fatigability (14) and that nNOS-null mice and mdx mice have reduced cage activity after bouts of mild exercise (15), suggesting that NO can reduce fatigability. Loss of muscle nNOS in mdx or nNOS null mice leads to an inability of muscle-derived NO to oppose vasoconstriction caused by increased sympathetic outflow (3,4) and this may be sufficient to increase muscle fatigue under some conditions. Indeed, vascular narrowings in muscle tissue occur in nNOS null mice and mdx mice but not in wild-type mice, supporting the interpretation that misregulation of vascular perfusion in muscles deficient in nNOS underlies the reduction in cage activity after exercise. However, other defects in NO-mediated signaling could contribute to fatigue, although they have not been explored. For example, NO modulates neuromuscular transmission in healthy muscle, and loss of this regulation could contribute to muscle fatigue (16). NO can also promote mitochondrial biogenesis (17); thus, loss of nNOS from dystrophic muscle could potentially increase fatigability by reducing mitochondria numbers. NO also promotes GLUT-4 expression and transport in muscle (18) and defects in that process may also be an unexplored feature of increased fatigability of dystrophic muscle.

If increased fatigability in fast-twitch, dystrophin-deficient muscle were to result from defects in glycolytic metabolism, phosphofructokinase (PFK) could provide a likely site for the regulatory defect because PFK is the rate-limiting enzyme in glycolysis and its activity is subject to extensive allosteric regulation. Interestingly, previous investigators showed that NOS binds to PFK (19) and they speculated that the interaction could influence the intracellular localization of the enzymes. Although all known regulatory roles played by NOS involve signaling via NO, we speculated that NOS could affect PFK activity by direct interactions between the enzymes, independent of NO production. The previous finding that the fast glycolytic fibers of DMD patients have a 45% reduction in PFK activity (20) supported our speculation and suggested loss of nNOS could contribute to increased muscle fatigability by disregulating PFK activity, in addition to causing defects in vascular perfusion. We also found suggestive similarities between some features of dystrophinopathy and other diseases that result from defects in glycolytic pathways, which were consistent with our speculations that misregulation of PFK could contribute to the dystrophic pathology. For example, PFK deficiency leads to impaired muscle function that typically has an onset during childhood, and is characterized by muscle fatigue, muscle soreness, muscle necrosis and elevated serum creatine kinase (21–23). Importantly, DMD and PFK deficiency diseases are both characterized by increased fatigability of muscle during moderately strenuous, anaerobic exercise, with no apparent increase in fatigability in low-intensity, aerobic exercise.

In the present investigation, we test whether increased fatigability of mdx mice is associated with decreases in the specific activity of PFK, and whether genetic manipulation of nNOS levels in mdx muscle can affect fatigability and the specific activity of PFK. We further examine whether nNOS is a positive allosteric regulator of PFK, to test a novel regulatory role of NOS that is independent of NO production.

RESULTS

nNOS transgene expression greatly increases treadmill running endurance in dystrophin-deficient mice

Five-month-old mdx mice that expressed the nNOS transgene in skeletal muscle (Tg+/mdx) showed no significant differences in their total mass or the mass of individual muscles or heart, compared to mdx mice that did not express the transgene (Tg−/mdx) (Fig. 1). They also did not differ significantly in grip strength (Fig. 1). However, Tg+/mdx mice performed continuous treadmill running for nearly six-times longer than Tg−/mdx mice (Fig. 1). Indeed, the difference between running times of Tg+/mdx and Tg−/mdx mice would have been greater, but we stopped running for 5 of the 12 Tg+/mdx mice at the 7 h time-point, although those mice would have continued voluntary running. Although identifying the time-point at which mice are delineated as fatigued is affected by subjective judgments by the experimentalist, all assays were performed by the same investigator who used consistent criteria to identify the time at which mice reached the point of fatigue. The large difference in running times of Tg+/mdx and Tg−/mdx mice were safely beyond differences that could be attributed to differences in the subjective judgment of fatigue.

Figure 1.

nNOS transgene expression in mdx mouse muscles increases endurance in treadmill running without affecting muscle mass but producing a small shift in fiber-type profile in some muscles. Body mass (A), muscle mass (B and C) and heart mass (D) in 5-month-old mice did not differ in Tg+/mdx and Tg−/mdx mice. Transgene expression also had no effect on the proportion of total fibers that were Type 2 fibers in the gastrocnemius (E) or quadriceps (F) muscles, although there was a small shift from type 2B to type 2A in gastrocnemius (G) but not in quadriceps (H). Although transgene expression did not affect grip strength of the mice (I), expression caused a large increase in endurance in treadmill running time (J). Seven Tg−/mdx mice and 12 Tg+/mdx mice were analyzed for each comparison. * indicates a significant difference at P < 0.001. Error bars, standard error of the mean. Where no error bars are evident, the errors were too small to appear on the scale used for the graph.

nNOS transgene expression has little effect on the fiber-type profile of mdx muscles

The tremendous increase in treadmill running capacity of mdx mice that expressed the nNOS transgene suggested that the transgene expression could affect muscle metabolism. We first assayed whether transgene expression increased the proportion of muscle fibers that were type 1, oxidative fibers, but found no difference in the proportions of type 1 and type 2 fibers (glycolytic) in either the gastrocnemius or quadriceps muscles in Tg+/mdx and Tg−/mdx samples (Fig. 1). We then assayed the proportion of fibers that were type 2A and found that there was less than 5% increase in the proportion of type 2 fibers that were type 2A in the gastrocnemius muscles of Tg+/mdx mice compared with Tg−/mdx samples. However, the expression of the transgene in mdx muscles did not affect the proportion of type 2A fibers in quadriceps muscles (Fig. 1).

We also assayed the relative concentrations of cytochrome C by western blots as an index of mitochondria concentration in the muscles because previous investigators have shown that NO could increase mitochondria biogenesis (17), but we found no difference in cytochrome C concentration between Tg+/mdx and Tg−/mdx muscles (Fig. 2). Similarly, no difference in the concentration of SDH was observed between Tg+/mdx and Tg−/mdx muscles (Fig. 2). We also tested whether transgene expression caused changes in GLUT4 levels in mdx muscle, because previous investigators found that NO elevates the GLUT4 expression in muscle (18) and this could contribute to differences in fatigability. However, GLUT4 levels in Tg+/mdx and Tg−/mdx muscles did not differ (Fig. 2).

Figure 2.

nNOS transgene expression in skeletal muscle does not affect concentration of mitochondrial enzymes or GLUT4, and there is no difference in the hemoglobin content of Tg+/mdx and Tg−/mdx muscles after running. (A) Immunoblots of whole muscle extracts from three, Tg+/mdx mice and three, Tg−/mdx mice were probed with anti-cytochrome C, anti-SDH or anti-GLUT4. The same blots used for incubation with antibodies in (A) were first stained with Ponceau red to show relative loading of each sample (‘Loading’ shown below). (B) Measurements of hemoglobin concentration in muscles of Tg+/mdx and Tg−/mdx mice show that there are no differences in hemoglobin content in the two lines following treadmill running. * indicates significant difference from wild-type controls.

nNOS transgene expression in mdx muscle does not affect vascular perfusion after treadmill running

NO generated by endothelial NOS (eNOS) or nNOS can cause vasodilation and increase vascular perfusion during muscle contraction (24) which led us to test whether expression of a nNOS transgene in mdx muscle could similarly increase blood supply to muscle during exercise. We have shown previously that the number of blood vessels and the total volume fraction of skeletal muscle comprised of blood vessels is not affected by the expression of a nNOS transgene in mdx muscle in sedentary animals (5), but those findings did not prove that blood supply to muscle during exercise was not elevated in Tg+/mdx mice. We tested this possibility by measuring total hemoglobin content in skeletal muscle of Tg+/mdx and Tg−/mdx mice that were not exercised or that experienced 22 min of running on a 20° uphill grade. Our data show that there is no significant difference in hemoglobin content of muscles of Tg+/mdx and Tg−/mdx mice after treadmill running (Fig. 2), indicating that differences in endurance during treadmill running are not attributable to greater vasodilation in the Tg+/mdx muscles. This observation is consistent with previous findings which showed that the increases in blood flow velocity and vascular conductance that were caused by muscle contraction did not differ significantly between mdx and wild-type mice (3).

PFK colocalizes with nNOS in skeletal muscle

Because our observations indicated that there was little effect of the nNOS transgene expression on indices that would reflect an increase in the oxidative phenotype or vascularization of mdx muscle, we tested whether nNOS transgene expression could influence glycolysis in mdx muscles. We focused our attention on PFK because previous investigators demonstrated the ability of PFK to bind to nNOS (19), and PFK is the rate-limiting enzyme in glycolysis. Fractionation of cell membranes and cytosolic proteins from skeletal muscle fibers followed by assaying for PFK in western blots showed that PFK is present in both the membrane and cytosolic fractions of muscles, and that the nNOS transgene expression did not affect levels of the expression of PFK in either fraction (Fig. 3A). The quality of fractionation was assessed by assaying for talin as a marker of cell membrane-associated proteins and PIN as a marker for cytosolic proteins (Fig. 3A). We then assayed whether PFK was enriched at the NMJs and myotendinous junctions (MTJs) of skeletal muscle fibers, where nNOS is highly concentrated (2,25) The locations of NMJs were identified by labeling cross-sections of muscle fibers with alpha-bungarotoxin followed by immunolabeling with anti-PFK, which showed PFK enrichment at NMJs (Fig. 3B and C). In addition, anti-PFK labeling of longitudinal sections of muscle fibers showed elevated concentrations of PFK at MTJs (Fig. 3D and E).

Figure 3.

PFK distribution and activity in muscle. (A) nNOS transgene expression does not affect the concentration of PFK at the cell membrane or in the cytosol of mdx muscle. Immunoblots of soluble and particulate fractions of muscle extracts that were probed with anti-PFK, anti-talin or anti-PIN. The seven left lanes contained the cytosolic fraction, confirmed by the enrichment of PIN and absence of talin in the samples. The seven right lanes contained the particulate fraction enriched in cell membrane, confirmed by the absence of PIN and enrichment with talin in the samples. Lanes 1–4 were samples prepared from Tg−/mdx hindlimb muscles. Lanes 5–7 were prepared from Tg+/mdx muscle extracts. (B) PFK is present in the cytosol of muscle fibers and enriched at the muscle cell membrane. A cross-section of wild-type muscle was immunolabeled with anti-PFK and then second antibody conjugated with a red chromogen (B) and also incubated with FITC-conjugated α-bungarotoxin to label NMJs (arrows) (C). A longitudinal-section of wild-type muscle was immunolabeled with anti-PFK and then second antibody that was conjugated with a red chromogen (D) and an adjacent section similarly treated, except anti-PFK was replaced by PBS only, as a negative control (E). Arrows in Figure 1D indicate the MTJs of two muscle fibers. Bar=50 µm. (F) The specific activity of muscle PFK is positively regulated by nNOS. The loss of nNOS in nNOS null mutants (B) results in a significant reduction in the specific activity of muscle PFK, relative to wild-type muscles (A). Mdx mice (C), which have greatly reduced concentrations of nNOS, also display significant reductions in PFK specific activity in muscle. Expression of a nNOS transgene in wild-type mice (D) or mdx mice (F) causes a significant increase in PFK-specific activity compared with wild-type mice. *significantly differs from wild-type at P < 0.05. (G–I) nNOS, but not eNOS or iNOS, is a positive allosteric regulator of PFK. Activity of PFK was assayed under allosteric conditions in the presence or absence of nNOS (G), eNOS (H) or iNOS (I). Reactions were performed for 2 min at a NOS/PFK molar ratio of 1:1 (solid red line), 1:5 (solid black line), 1:10 (solid blue line), 1:100 (dotted red line) or without NOS (dotted black line). Data shown are representative of three experiments.

Genetic modulation of nNOS expression in skeletal muscle significantly affects the specific activity of muscle PFK

The ability of PFK to bind nNOS (19) and the co-distribution of the two enzymes in muscle suggested that nNOS could modulate the activity of PFK in muscle. We tested this possibility by measuring the specific activity of PFK in muscles collected from mice in which the nNOS expression was ablated (nNOS null mice), reduced (mdx mice) or enhanced (nNOS transgenic mice and Tg+/mdx mice). We observed a strong, significant relationship between the level of the nNOS expression and the specific activity of PFK when measured under allosteric conditions (Fig. 3F). Furthermore, we found that the reduction of PFK specific activity that occurred in mdx muscles could be reversed by the expression of the muscle-specific, nNOS transgene (Fig. 3F).

nNOS, but not eNOS or iNOS, has a strong, positive allosteric effect on PFK activity

We tested whether purified nNOS would modify the activity of PFK in assays using purified nNOS and PFK in the absence of other proteins, and observed a strong, dose-dependent increase in PFK activity in the presence of nNOS (Fig. 3G). Notably, these assays were performed in the absence of the nNOS substrate, arginine, so the positive regulatory effect was not attributable to NO production. Despite the positive allosteric effect of nNOS on PFK, neither iNOS or eNOS had a significant affect on PFK activity (Fig. 3H and I).

Expression of a nNOS transgene in mdx muscle increases glycogen utilization during treadmill running

The positive allosteric effect of nNOS on PFK suggested that increases in nNOS concentration in muscle could elevate glycogen metabolism during exercise. We tested this possibility by comparing glycogen in muscles from non-exercised mice and muscles collected from mice immediately following completion of a bout of treadmill running, analyzing both Tg+/mdx and Tg−/mdx mice. Running caused a reduction in glycogen concentration in Tg+/mdx muscle by nearly 60%, but did not reduce glycogen concentration in Tg−/mdx mice run under the same conditions and for the same length of time (Fig. 4A). This finding indicates that the expression of the nNOS transgene increased glycogen metabolism. We also noted that the expression of the transgene in muscle increased glycogen concentrations in muscle to levels that occur in wild-type mice. This increase may reflect the previously reported, large reduction in leakiness of the sarcolemma of mdx mice that is produced by the nNOS transgene expression (5).

Figure 4.

Expression of a nNOS transgene in mdx muscle increases glycogen metabolism during treadmill running. (A) Glycogen content was reduced by nearly 60% in Tg+/mdx mice during treadmill running. In contrast, there was no significant change in glycogen content in Tg−/mdx mice during treadmill running under the same conditions. * indicates significantly different from sedentary mice of same genotype. (B) Lactate content was increased by nearly 50% in Tg+/mdx mice during treadmill running. In contrast, there was no significant change in lactate content in Tg−/mdx mice during treadmill running under the same conditions. * indicates significantly different from sedentary mice of same genotype.

Effects of the nNOS transgene expression on glycogen utilization were also assessed by measuring changes in lactate accumulation in the muscles of Tg+/mdx and Tg−/mdx mice during exercise. Corresponding to the reduction in glycogen concentration in Tg+/mdx muscles, but not in Tg−/mdx muscles, lactate levels increased by 50% in Tg+/mdx muscles, but did not change in Tg−/mdx muscles (Fig. 4B).

The positive allosteric activity of nNOS for PFK resides in the FAD-binding region of the reductase domain of nNOS

We expressed the N-terminal domain of nNOS, from the start codon through the calmodulin (CaM)-binding site, and the C-terminal domain of nNOS, from the tetrahydrobiopterin (BH4)-binding domain through to the end of the coding sequence and assayed these polypeptides for their effects on PFK activity under allosteric binding conditions. The C-terminal half of the nNOS protein showed a positive, dose-dependent effect on PFK activity, but the N-terminal domain showed no significant effect on PFK activity (Fig. 5). We next assayed for allosteric activity of three polypeptides that represented the BH4-binding domain and the flavin monoamine nucleotide (FMN)-binding domain (Peptide 1), the flavin adenine dinucleotide (FAD)-binding domain (Peptide 2) and the nicotinamide adenine dinucleotide phosphate (NADPH)-binding domain (Peptide 3) within the C-terminal domain (Fig. 5). A strong, dose-dependent, positive, allosteric effect was observed for Peptide 2 in PFK activity assays, with no effect of either Peptide 1 or Peptide 3 on PFK activity (Fig. 6A).

Figure 5.

(A) Graphic representation of the domains of nNOS that used in PFK activity assays. The N-terminal and C-terminal domains and peptides 1, 2 and 3 were cloned and expressed in BL21 (DE3) cells. Peptides 4 through 8 were synthesized commercially. (B) The positive allosteric activity of nNOS lies in the C-terminal domain. Addition of the C-terminal domain of nNOS, but not the N-terminal domain, shows a positive, dose-responsive effect on PFK activity under allosteric conditions. * indicates significantly differs from PFK activity in the absence of nNOS C-terminal or N-terminal domain at P < 0.05.

Figure 6.

The positive allosteric activity of nNOS lies primarily in a 36-amino acid peptide in the FAD-binding domain. Addition of a peptide containing the FAD-binding domain, but not the BH₄/FMN- or NADPH-binding domain, shows a positive, dose-responsive effect on PFK activity under allosteric conditions (A). Peptide 7, a 36-amino acid peptide in the FAD-binding domain of nNOS shows a strong, dose-responsive, positive allosteric effect on PFK activity under allosteric conditions. * indicates significantly differs from PFK activity in the absence of nNOS peptides at P < 0.05.

We further assessed the region of nNOS that exhibited a positive allosteric effect on PFK by assaying for influences of smaller, synthetic peptides within the FAD-binding domain of nNOS. Although a slight, but significant, positive allosteric effect was exhibited by Peptide 4, a strong, dose-dependent positive influence on PFK activity was exhibited by Peptide 7 (Fig. 6B). Peptide 7, which corresponds to amino acids 1106–1141 in nNOS, consists of PLQLQQFASL ATNEKEKQRL LVLSKGLQEY EEWKWG and is exposed on the surface of the FAD-binding domain of nNOS, according to the structure determined by Zhang et al. (26) (Fig. 7). The peptide is 43.9% identical to the homologous region in mouse eNOS (27) and 44.1% identical to the homologous region of mouse iNOS (28). Note that although Peptide 7 is in the FAD-binding domain of nNOS, the presence of FAD did not prevent the interaction between Peptide 7 and PFK; activity assays in which the positive allosteric activity of Peptide 7 was observed included FAD in the reaction buffer.

Figure 7.

Peptide 7 in nNOS lies on the exposed surface of nNOS. The location of the 36-amino acid peptide on nNOS that promotes PFK activity is highlighted in yellow on the image obtained from World Wide Protein Data Bank (http://www.wwpdb.org/) based on the structure determined by Zhang et al. (26).

DISCUSSION

The results of the present investigation show that the expression of a nNOS transgene in dystrophin-deficient muscle tremendously reduces muscle fatigability and increases in vivo the specific activity of PFK and elevates glycogen utilization by muscle. Our findings also show that a 36-amino acid peptide in the FAD-binding domain of nNOS interacts with PFK to increase PFK activity significantly. Because PFK is the rate-limiting enzyme in glycolysis, our findings indicate that the loss of positive allosteric interactions between nNOS and PFK can contribute to increased fatigability of dystrophin-deficient muscle. The pathophysiology of other, human, neuromuscular diseases supports this interpretation. For example, humans who are genetically deficient in muscle PFK exhibit a lifetime of muscle weakness (29) that is exacerbated by exertion (30) and associated with the elevation in the concentration of muscle creatine kinase in the serum and diminished glycogen metabolism in muscle during exercise (31). Each of these pathological features resembles DMD and mdx dystrophies.

Although these findings show that elevating nNOS expression in muscle increases PFK activity and reduces fatigability, nNOS-derived NO is also likely to affect muscle fatigability through mechanisms that do not involve modulation of PFK activity. Loss of normal regulation of vascular perfusion during exercise by NO provides a mechanism that may contribute to increased muscle fatigue (3,4,15). For example, treating mdx mice with phosphodiesterase 5A (PDE5A) inhibitor before exercise reduced the number of sites of constricted blood vessels in muscle-observed post-exercise and also increased the treadmill running time of mdx mice compared with untreated controls (15). That correlation suggested that muscle-derived NO may increase endurance by improving vascular perfusion because treatments with PDE5A inhibitor can prevent degradation of cGMP, a downstream signaling molecule that is generated at higher levels following NOS activation. However, systemic treatments with PDE5A inhibitor would also prevent degradation of cGMP that was generated by eNOS activity in endothelial cells; NO generated by eNOS increases vasodilation during muscle contraction (24). Nevertheless, under some special circumstances, muscle-derived NO can contribute to increases in vasodilation when there is increased α-adrenergic stimulation during exercise (3), but in the absence of increased sympathetic outflow, baseline mean arterial blood pressures, femoral blood flow velocities and vascular conductances do not differ significantly in mdx and C57 mice (3). Furthermore, increases in blood flow velocity and vascular conductance caused by muscle contraction do not differ significantly between mdx and wild-type mice (3). Similarly, our data show that the expression of the nNOS transgene in mdx muscles does not affect vascular perfusion of muscle that occurs during treadmill running, as shown by the mass fraction of muscle that is comprised by hemoglobin.

Previous investigators have shown that DMD and mdx dystrophy cause defects in muscle metabolism that are reflected in low levels of ATP and high levels of ADP in muscles at rest (32–38). Although expression profiling studies have shown that dystrophinopathies are associated with perturbations in the levels of the expression of genes that encode mitochondrial proteins (39,40) and substantial disruptions in calcium-levels in mitochondria occur (41,42), functional studies of the energetics of dystrophic muscle have shown a surprising lack of mitochondrial dysfunction in vivo. Metabolism in dystrophin-deficient muscle was carefully explored by Cole et al. (43) who used magnetic resonance spectroscopy (MRS) to analyze muscle at rest and following contraction in wild-type, mdx and utrophin-deficient/mdx mice. The investigators demonstrated that the initial recovery rates of phosphocreatine concentration and ADP concentration after contraction did not differ in the three lines of mice, reflecting normal mitochondrial production of ATP after contraction. Previous investigators had similarly noted that dystrophin-deficient humans show nearly normal oxidative metabolism during recovery from exercise, when ATP turn-over is rapid (34,44) and concluded that mitochondrial function in DMD patients is normal. Cole et al. (43) also observed that mdx muscles experienced less acidification during contraction than wild-type muscles, reflecting the reduced production of lactic acid, and interpreted the findings as evidence for glycolytic defects in mdx muscle. Similar observations were made previously on DMD patients where MRS showed less acidification of DMD muscles after exercise than occurred in healthy subjects (34). Similarly, MRS studies of muscle energetics in Becker's muscular dystrophy patients showed reduced cytosolic acidification during exercise than observed in healthy subjects, indicating defects in glycolysis, while mitochondrial ATP production was unaffected (44). More recently, NMR spectroscopy of a large spectrum of metabolites in DMD muscle again noted reduced lactate production, indicating a defect in glycolysis or substrate limitation of glycolytic metabolism (45). Instead of finding defective oxidative metabolism, this latter study noted high levels of the Kreb's cycle intermediate α-ketoglutarate, which has been shown previously to indicate elevated oxidative metabolism (46). Collectively, these findings show that metabolic defects in dystrophin-deficient muscles are primarily attributable to disruption of glycolysis, and introduce the possibility the metabolic defects may be caused in part by the loss of normal, positive allosteric regulation of PFK by nNOS. Notably, our findings show that the expression of the nNOS transgene in mdx muscle increased glycogen utilization and increased lactate production, correcting two of the most commonly noted indicators of defective metabolism in DMD or mdx muscles.

The present finding that the increased expression of nNOS in mdx muscle increases endurance provides further evidence that at least some of the normal functions of nNOS in muscle do not require binding to the DGC. This finding has therapeutic importance because truncated dystrophin molecules designed for the expression in DMD or mdx muscles frequently do not restore nNOS to the membrane, and whether these designs will be sufficient for optimal recovery is uncertain. Findings by others also indicate that nNOS binding to the DGC or localization to the cell membrane is not required to reduce muscle fatigue. For example, over-expression of ε-sarcoglycan (εSG) in α-sarcoglycan (αSG) null-mutant muscle restores the DGC to the muscle cell membrane and greatly increases nNOS in the cytosol, without detectable restoration of nNOS at the cell membrane (15). Nevertheless, PDE5A inhibitor treatment of εSG over-expressing αSG mice restores cage activity of mice following exercise to levels that do not differ significantly from wild-type mice (15). Similarly, PDE5A inhibitor treatment of mdx mice restores cage activity after exercise to wild-type levels although the treatment does not restore dystrophin expression. In contrast, PDE5A inhibitor treatment of nNOS null mutant mice has no ameliorative effect on cage activity levels. Collectively, these data indicate that nNOS localization at the DGC is not required to reduce muscle fatigue, and suggest that the greatly reduced concentration of nNOS in the cytosol as well as at the cell membrane (2) underlies defects in NO-mediated regulation of muscle fatigue.

Other mechanisms exist through which modulation of NOS expression could affect muscle fatigue caused by defects in glycolysis. Treatment of astrocytes with exogenous NO reduced the concentration of fructose-6-phosphate (the substrate of PFK) and increased fructose-1,6-biphosphate, the product of PFK activity (47), thereby indicating an increase in PFK activity and glycolysis. Although this NO effect has not been tested in muscle, it has the potential to promote glycolysis in muscle and thereby accelerate ATP replenishment and reduce fatigability. The activation of glycolysis by NO in astrocytes was unaffected by treatment with a guanylate cyclase inhibitor, which indicated that the NO-mediated increase in glycolysis is cGMP independent (14), and that it would be unaffected by PDE inhibitors. Instead, the NO-mediated activation of glycolysis in astrocytes was dependent on PFK activation by AMP protein kinase. While this mechanism may feasibly affect the reduction in fatigability of nNOS Tg+ muscles reported in the present investigation, this mechanism would not contribute to the increases in specific activity of PFK that were produced by reacting the enzyme with nNOS under allosteric conditions. Those experiments were performed in the absence of arginine, the nNOS substrate, so no NO production occurred in the reaction.

Other possible mechanisms through which changes in nNOS concentration in muscle could affect muscle fatigability remain unexplored, but current knowledge suggests that they are not important features in the pathophysiology of dystrophin-deficient muscle. For example, NO can be a positive regulator of mitochondrial biogenesis (17), so that the loss of nNOS could cause a reduction of ATP production by oxidative metabolism and thereby increase fatigability. However, numerous functional studies have been unable to demonstrate a significant defect in mitochondrial function in dystrophic muscle in vivo (32–38,43–45) and expression of a nNOS transgene in mdx muscle reduces fatigability without increasing cytochrome C or SDH concentrations, indices of mitochondrial concentration (present study). Alternatively, the loss of nNOS could feasibly contribute to increases in fatigability because NO can modulate neuromuscular transmission (16) and nNOS modulates NMJ structure in dystrophic muscle (8). However, mdx mice show only slight disruptions of neuromuscular transmission at the NMJ that appear as small changes in membrane capacitance (48) that would not affect fatigability. Thus, loss of the positive allosteric interactions between nNOS and PFK provide the only current explanation for the defects in glycolytic metabolism that occur in dystrophic muscle. This mechanism is the first regulatory role identified for nNOS that is not mediated by NO.

MATERIALS AND METHODS

Mice

All experiments using animals were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and the University of California, Los Angeles Institutional Animal Care and Use Committee. C57BL/6J mice (wild-type), mdx mice [C57BL/10ScSn-Dmd(mdx)] and nNOS–/– mice (B6;129S4-NOS1^tm1P1h/J) were purchased from The Jackson Laboratories (Bar Harbor, ME, USA). Mice with a muscle-specific, over-expression of a nNOS transgene (Tg+ mice) were generated as described previously (5). Mdx mice were crossed with nNOS Tg+ mice to produce dystrophin-deficient, nNOS Tg+ mice (Tg+/mdx mice) (5). Null mutation for dystrophin was confirmed by mdx-amplification-resistant mutation system PCR (49). Null mutation for nNOS was confirmed by western analysis of muscle extracts. Over-expression of nNOS was also confirmed by western blots of muscle extracts. Tg+/mdx mice exhibit approximately a 50-fold over-expression of nNOS protein, but the levels of NO in the serum of Tg+/mdx mice were elevated over wild-type serum levels by only, approximately, 2-fold (10).

Materials

The sources of the enzymes, substrates and antibodies used in the investigation were: disodium fructose 6-phosphate (Na₂F6P; Sigma Chemical Co., St Louis, MO; #F3627), β-nicotinamide adenine dinucleotide (DPNH; Sigma; #340-101), PFK (Sigma; #F0137), aldolase (Roche Pharmaceuticals, Pleasanton, CA; #102-652), triose phosphate isomerase (Roche; #109-762), α-glycerophosphate dehydrogenase (Roche; #127-752), neuronal NOS (nNOS; Calbiochem, San Diego, CA; #482721), endothelial NOS (eNOS; Calbiochem; #482732), inducible NOS (iNOS; Calbiochem; #482777), hemoglobin (Sigma; #H-0267), lactate dehydrogenase (Sigma; #L-3916), mouse anti-myosin fast type heavy chain (MHCf) (clone WB-MHC; Novacastra Labs, Newcastle-upon-Tyne, UK), rabbit anti-rat nNOS (Serotec, Raleigh, NC, USA), goat anti-rabbit PFK (Biodesign, Saco, ME, USA), rabbit anti-rat nNOS (Transduction Labs, Lexington, KY, USA), goat anti-rabbit PFK (Biodesign) mouse anti-chicken talin (clone 8D4; Sigma), mouse anti-protein inhibitor of NOS (Transduction Laboratories), mouse anti-cat myosin heavy chain 2A (MyHC 2A; clone 2F7; Developmental Studies Hybridoma Bank, Iowa City, IA, USA), rabbit anti-human succinate dehydrogenase (SDH; Santa Cruz Biotechnology, Santa Cruz, CA, USA), mouse anti-human cytochrome C (BD Pharmingen, San Jose, CA, USA) and goat anti-human GLUT4 (Santa Cruz). Inorganic reagents were obtained from Sigma Chemical Co (St Louis, MO, USA).

Function tests

We assayed whether the expression of the nNOS transgene in muscles of mdx mice affected the strength or treadmill running endurance of 5-month-old mice. Forearm, grip strength of the mice was tested using a digital grip strength meter (Columbus Instruments, Columbus, OH, USA). The mice grasped a horizontal bar attached to a force transducer and were gently pulled by the tails until they released their grip. The peak pull force was digitally displayed on the grip strength meter. The mean peak force was calculated from three trials with 1 min of rest between each trial.

Endurance during uphill treadmill running was measured in mice by placing them on a treadmill (Columbus Instruments) with a 10° uphill incline for which the running speed was set at 15 m/min. An electrified grid through which a 1.2 mA current passed for 200 ms once each second provided incentive for the mice to continue running. Running continued until the mice would rest on the grid for 10 s or more, without returning to treadmill running. In some cases, mice continued running for 7 h without stopping, in which case they were removed from the treadmill because of our concern for their dehydration. In those cases, their running time was recorded as 7 h.

Histology

Muscles from 3- to 5-month-old mice were dissected, embedded in 10.24% polyvinyl alcohol 4.26% polyethylene glycol, frozen in liquid nitrogen-cooled isopentane and then stored at −80°C. Frozen sections were cut at 10 µm thickness, air-dried and then fixed in acetone for 10 min. Sections were processed as previously described (50), and then incubated in mouse anti-MHCf, mouse anti-cat MyHC 2A rabbit anti-rat nNOS or goat anti-rabbit PFK. The sections were then washed in PBS and incubated with biotinylated anti-rabbit, anti-mouse or anti-goat IgG (Vector Laboratories, Burlingame, CA, USA) for 1 h. Sections that were incubated with anti-mouse second antibodies were pre-treated with the mouse-on-mouse kit (Vector Laboratories) to block endogenous mouse immunoglobins. Sections were washed with PBS and incubated with streptavidin-conjugated horseradish peroxidase (Vector Laboratories) and the antigen–antibody–chromogen complex was localized using a peroxidase substrate kit (AEC; Vector Laboratories).

We also tested whether PFK was enriched at sites in muscle where nNOS is enriched by performing double-labeling for PFK and alpha-bungarotoxin (BTX) on muscle sections. BTX binds acetylcholine receptors with high specificity and affinity at NMJs, where nNOS is highly concentrated. In double-labeling assays, sections were prepared as described above for anti-PFK labeling, but sections were incubated with 1 µg/ml FITC-BTX for 1 h prior to their incubation in anti-PFK. We used this combination of fluorescent and bright-field imaging rather than two separate fluorophores so that there was no possibility that signal leakage from one wavelength to another could give a false positive result.

Relative proportions of muscle fiber types were assayed in cross-sections of gastrocnemius and quadriceps from mice in each strain after immunolabeling sections with anti-MHCf and counting the percentage of the total number of fibers in each cross-section that expressed MHCf.

Membrane isolation

Hindlimb muscles were rapidly dissected from mice following euthanasia. Adipose tissue, large neurovascular bundles and large pieces of connective tissue were removed, after which the muscles were minced and then homogenized (Sorvall Omnimixer) in 40 ml of homogenization buffer per gram of muscle [homogenization buffer: 50 mm Tris pH 7.5 containing 1 mm dithiothreitol (DTT), 1 mm ethylene diamine tetra-acetic acid (EDTA), 1 mm benzamidine, 0.5 mm phenylmethylsulfonyl fluoride, and 1 µg/ml of aprotinin, 2 µg/ml leupeptin and 10 µg/ml soybean trypsin inhibitor). The homogenate was filtered through gauze, and the filtrate was then centrifuged at 12 000g for 15 min at 4°C. The supernatant was then centrifuged at 100 000g for 30 min at 4°C. The final pellet contained the particulate fraction that was enriched in vesiculated membranes, and the supernatant was enriched in cytosolic proteins. Protein in the supernatant fraction was concentrated by centrifugation through a 10 000 molecular mass cut-off cellulose membrane (Centriprep). The pellet fraction was solubilized by boiling in reducing sample buffer (51). Protein concentration in the each fraction was measured using the technique of Minamide and Bamburg (52).

Western analysis

The relative concentrations of selected proteins in cytosolic and membrane fractions were assessed in western blots. Each lane of the gels was loaded with 30 µg of total protein. Uniform loading of the samples was confirmed by staining each membrane with 0.1% Ponceau S in 5% acetic acid. Antibodies used in western analyses were: rabbit anti-rat nNOS, goat anti-rabbit PFK, mouse anti-chicken talin, mouse anti-protein inhibitor of NOS, rabbit anti-human SDH, mouse anti-human cytochrome C and goat anti-human GLUT4. After incubation with primary antibodies, the membranes were washed extensively with 50 mm sodium phosphate, pH 7.2, containing 150 mm sodium chloride (PBS) and 0.1% Tween-20. The blots were then incubated with horseradish peroxidase conjugated-donkey anti-rabbit IgG (Amersham), washed with PBS/Tween and the reaction product was visualized using chemiluminescence.

Assay for muscle PFK activity

PFK activity in skeletal muscle extracts was measured using modifications of previously described techniques (53,54). Muscles were rapidly dissected from mice and then frozen in liquid nitrogen until used for analysis. Muscle samples for the assays were prepared by homogenizing approximately 200 mg of gastrocnemius muscle in a Dounce homogenizer with 10 volumes of 50 mm potassium phosphate buffer pH 7.5, containing 100 mm DTT and 5 mm EDTA. The homogenates were centrifuged at 10 000g for 15 min and then the supernatant collected, protein concentration measured by absorbance at 280 nm and then used for PFK activity measurements.

PFK activity was measured under allosteric (pH 6.9) and non-allosteric (pH 8.1) conditions. The reaction solution for the assays contained 33 mm Tris–HCl at either pH 6.9 or pH 8.1, 2 mm ATP, 5 mm MgSO4, 2 mm Na₂F6P, 0.16 mm DPNH, 50 mm KCl and 1 mm DTT. A standard curve was generated at pH 8.1 by adding PFK to the reaction mix and measuring rate of change in absorbance of the solution at 340 nm per unit of PFK activity. PFK activity in the muscle extracts was then determined. Specific activity of PFK in the muscle was expressed as the ratio of allosteric to non-allosteric activity per milligram of muscle protein.

Allosteric regulation of PFK activity by NOS

PFK activity was measured under allosteric conditions in the presence and absence of nNOS, eNOS or iNOS, at NOS/PFK molar ratios of 1:1, 1:5, 1:10 and 1:100. Activity was expressed as mU/ml of reaction solution. One unit of total PFK was used per reaction.

Glycogen depletion assay

Glycogen depletion was measured in mice subjected to treadmill running. Five Tg−/mdx mice were run at 15 m/min at a 20° incline until they reached exhaustion, which occurred at a mean time to exhaustion of 22 min. Five Tg+/mdx mice were then run at the same speed and incline for 22 min. Muscles from all mice were dissected immediately after the mice were removed from the treadmill at the end of the running period, and frozen in liquid nitrogen. Muscle glycogen content was determined following the acid hydrolysis method of Passonneau and Lauderdale (55). Approximately 10–15 mg of frozen quadriceps muscle were hydrolyzed in 2 N HCl for 2 h at 120°C and subsequently neutralized with an equal volume of 2 N NaOH. Samples were mixed with hexokinase reagent and incubated at room temperature for 10 min. The absorbance was measured at 340 nm and a standard curve was generated using glucose stock. Values were normalized to muscle mass.

Lactate assay

Lactate content was also measured in muscles that were used for glycogen depletion measurements that are described above. Skeletal muscles were prepared using a modification of Asllani et al. (56) and assayed for lactate content as described previously (57). Quadriceps muscles were manually homogenized in 10 volumes of 6% perchloric acid and extracts were centrifuged at 14 500g for 10 min at 4°C. The supernatant was neutralized with KOH and centrifuged at 3000g for 5 min to remove the insoluble precipitate. Glycine buffer, pH 9.2 (0.6 m glycine and 0.5 m hydrazine) containing 1.5 mg/ml nicotinamide adenine dinucleotide and 15 U/ml lactate dehydrogenase was added to samples and incubated at 37°C for 45 min. The increased absorbance at 340 nm was determined and lactate was calculated based on the Beer–Lambert law with 6.22 as the millimolar extinction coefficient for NADH. Values were normalized to the wet mass of the muscle sample.

Hemoglobin assay

The hemoglobin concentration in skeletal muscles was measured according to Ownby et al. (58) using the cyanmethemoglobin method. Approximately 50 mg of frozen quadriceps muscle were manually homogenized in ddH₂O. The muscle extract was cleared by centrifuging at 21 000g and 4°C for 30 min. Twice normal-strength Drabkin's reagent (Sigma) was added to the cleared muscle extract and incubated at room temperature for 15 min after which the absorbance was measured at 540 nm. The grams per 100 ml (g%) of hemoglobin were calculated from a standard curve generated using lyophilized human hemoglobin and corrected for weight of each muscle sample.

nNOS cloning and allosteric regulation of PFK by nNOS fragments

All nNOS fragments were generated using a rat brain nNOS template in pCMV5 (provided by Dr James T. Stull, University of Texas, Southwestern, Dallas, TX, USA) (Tables 1 and 2). The N-terminal half of nNOS (amino acids 1–739) was cloned into pCAL-c (Stratagene, La Jolla, CA, USA) at the Nco1 site and a Klenow-filled BamH1 site. The C-terminal half of nNOS (amino acids 740–1429), and the FMN (amino acids 754–938), FAD (amino acids 982–1222) and NADPH (amino acids 1242–1369) binding sites were cloned into pCAL-n (Stratagene) at the EcoRI and XhoI sites. BL21 (DE3) cells were transformed with the nNOS expression vectors. nNOS proteins of the expected masses were purified from cell lysates using calmodulin affinity resin (Stratagene) according to the manufacturer's instructions. Peptides within the nNOS FAD-binding site (amino acids 992–1036, 1040–1060, 1061–1084, 1106–1141, 1172–1214) were synthesized by Sigma-Genosys (The Woodlands, TX, USA) using FMOC (9-fluorenylmethoxycarbonyl) solid-phase peptide synthesis. The quality and purity of the peptides was measured using reverse phase chromatography and MALDI-TOF. Cloned, nNOS fragments or synthetic nNOS peptides were then used in nNOS activity assays under allosteric conditions. All reactions were performed for 2 min using one unit of total PFK per reaction.

Table 1.

Fragments of nNOS that were cloned or synthesized for use in PFK activity assays

Peptide	Nucleotide sequence (bp)	Amino acid sequence	Binding domain
N-terminal fragment	349–2565	1–739	PDZ, Arg, heme, CaM
C-terminal fragment	2566–4638	740–1429	BH4, FMN, FAD, NAD
Peptide 1	2608–3162	754–938	BH4, FMN
Peptide 2	3292–4014	982–1222	FAD
Peptide 4	3322–3456	992–1036
Peptide 5	3466–3528	1040–1060
Peptide 6	3529–3600	1061–1084
Peptide 7	3664–3771	1106–1141
Peptide 8	3862–3990	1172–1214
Peptide 3	4072–4455	1242–1369	NAD

Table 2.

Primers used for cloning nNOS fragments

nNOS domain	Forward primers	Reverse primers
N-term	CCTCTTCATGAGCATGGAAGAGAACACGTTTGGG	CCTCTTTTAAACCTTGACGGCCTCTGCCAA
C-term	CCTCTGAATTCTGTTCTCAGCCAAGCTAATGG	CCTCTCTCGAGTTAGGAGCTGAAAACCTCATC
FMN-binding	CCTCTGAATTCTGAAGGCGACCATTCTCTACGCCACA	CCTCTCTCGAGTTAGGCTGCCTTGAAGACTTTCTTGGC
FAD-binding	CCTCTGAATTCTGCAAGGTCTTTCCAATGTTCACAAA	CCTCTCTCGAGTTAATCGTCAGCCTGTATTCTGTTGAG
NADPH-binding	CCTCTGAATTCTGCCTTGCATCCTGGTTGGC	CCTCTCTCGAGTTACTGCTGGGTCATTATGCGCTG

FUNDING

This work was supported by grants from the National Institutes of Health (R01 AR40343 and R01 AR47721 to J.G.T.) and the Muscular Dystrophy Association (MDA 03018947 to J.G.T.).