I. NITRIC OXIDE SYNTHASES

Top Previous Next References

Nitric oxide synthases (NOSs) are the products of three genes (79, 145). They are named after the cells or systems from which they were originally purified in order of their discovery and are thus designated neuronal (n) NOS (NOS1), macrophage (immune)/calcium calmodulin-independent or "inducible" (i) NOS (NOS2), and endothelial (e) NOS (NOS3). The numerical designation has come into vogue with the more recent appreciation that each of the NOSs in fact has a wide tissue distribution. In contrast, the classification of NOSs as "constitutive" or inducible has turned out to be unreliable because each of the isoforms may be regulated dynamically; inducibility is a function of the stimulus rather than the gene product. Alternative splicing factors are important in the pattern of NOS expression. In particular, nNOS in mature skeletal muscle and heart contains a 34-amino acid insert that arises from alternative splicing of nNOS pre-RNA between exons 16 and 17 (36, 178). The muscle isoform is called nNOSµ. Three distinct cDNAs, exon 1 variants arising from alternative promoters of a single nNOS gene, are robustly expressed in skeletal muscle (211).

NOS activity is regulated at transcriptional, translational, and posttranslational levels. Typically, iNOS is transcriptionally upregulated by cytokines (79), but so too are nNOS and eNOS in skeletal muscle (61, 105). All isoforms may be transcriptionally regulated by hypoxia. nNOS expression is increased by crush injury (167), muscle activity (15), and the ageing process (40), whereas it decreases following denervation (193). Both vascular and skeletal muscle eNOS are upregulated by chronic exercise (stimulation) (15, 199); eNOS in blood vessels is also upregulated by sheer stress (179). nNOSµ and iNOS expression coincide with myotube fusion in culture (109, 125), indicating that the expression is developmentally regulated. Moreover, cell phenotype (myotube vs. fused myoblast) determines the translational efficiency of selected nNOS mRNAs, transcribed from alternative promoters and differing only in exon 1, that is, exhibiting unique 5'-ends (211) (see sect. IVA2 for details).

nNOS and eNOS are activated by the interaction with calcium and calmodulin, whereas calmodulin remains bound to iNOS at resting levels of calcium (145). eNOS is co- and posttranslationally modified by myristoylation and palmitoylation, respectively, which target it to caveolae (168). In contrast, nNOS has adopted the use of an NH₂-terminal PDZ domain through which it interacts with ₁-syntrophin in the membrane cytoskeleton-dystrophin complex of skeletal muscle or with postsynaptic density proteins (PSD95 and PSD93) in synaptic membranes and perhaps neuromuscular junctions (33, 36, 52). In skeletal muscle, nNOS and iNOS coimmunoprecipitate with caveolin-3 (81, 204), but the functional implications are unclear (85) (see sects. IVC and VB). Neither PDZ domains nor fatty acylation motifs are found in iNOS.

Both the activity and subcellular localization of NOS are dictated by a variety of isoform-specific protein-protein interactions (see sect. VB for additional details). nNOS or nNOSµ may be inhibited through an interaction with caveolin-3, a component of the dystrophin complex that might displace the activator calmodulin from nNOS by analogy to the effect of caveolin-1 on eNOS (204). Proteins called PIN (107) and CAPON (106) are endogenous inhibitors of nNOS, the former being highly expressed in skeletal muscle (91). eNOS is localized in plasmalemma caveoli of endothelial cells and cardiac myocytes through association with caveolin-1 and -3, respectively. eNOS is activated through association with 90-kDa heat shock protein (77). iNOS is inhibited by its association with kalirin or NAP110, which evidently function much like PIN to destabilize the active dimer (156, 157). NOSs are also susceptible to posttranslational modification by phosphorylation, nitrosylation (cysteine and iron), and oxygen concentration (2, 3, 58, 60, 75, 99, 134, 146, 149, 151, 170). In particular, nNOS (2) and iNOS (60) activities are directly coupled to oxygen concentration throughout the physiological range.

All three NOSs share a similar catalytic scheme in which the amino acid L-arginine is transformed into a variety of nitric oxide (NO)-related molecules. The term NO is thus used in a generic or collective sense (179). Molecules most closely identified with NO-related activity include NO itself, S-nitrosothiols (SNOs), metal NO complexes, and higher oxides of nitrogen (NO_x), including peroxynitrite (184). nNOS may also produce nitroxyl anion (172, 184). All NOSs generate NO, SNO, and probably small amounts of peroxynitrite (NO/O₂) (179). There is good reason to believe that the availability of cofactors and cosubstrates as well as subcellular localization can influence the identity of the enzymatic product (185).

	II. MECHANISM OF ACTION

Top Previous Next References

NO is a signaling molecule. Its responses are mediated through reactions with thiol or transition metal centers in proteins or both (180, 184, 185). Specifically, NO responses are largely mediated by S-nitrosylation of cysteine redox centers or by coordinate interactions with heme or nonheme iron and copper. A motif for S-nitrosylation has been identified in ion channels, receptors, enzymes, transcription factors, and small G proteins: (R,H,D,E) C (D,E) (185). The acid and base residues act as catalysts for S-nitrosylation/denitrosylation of the cysteine and may further help to maintain proteins in NO-sensitive conformations. NO or related molecules can also oxidize cysteines to disulfides and sulfinic acids (182), and metals [Cu(I) and Fe(II)] to higher oxidation states (136, 180, 184); however, the importance of such oxidation in regulation of protein function are less clear. More generally, a continuum of nitrosative and oxidative modifications constitute redox-related signaling events on the one hand and underlie damage to proteins on the other hand. Hazardous levels of nitros(yl)ation are termed nitrosative stress (94, 182), a situation analogous to oxidative stress, which is imparted by elevated levels of reactive oxygen species and which has been implicated in muscle fatigue and cell injury (16, 158).

The interaction of NO with heme-containing proteins is exemplified in the binding to cytochrome-c oxidase, which reversibly inhibits the enzyme and thus cell respiration (31), and to guanylate cyclase, which activates the enzyme and thereby raises cGMP levels (140). Many actions of cGMP are elicited by cGMP-dependent (G) kinase. However, the downstream targets of G kinase are, with rare exception, not known. cGMP can also directly gate certain ion channels and regulates the activity of specific classes of phosphodiesterases (173).

	III. NITRIC OXIDE TARGETS IN SKELETAL MUSCLE

Top Previous Next References

In any discussion of NO targets, it is important to distinguish those that have been identified in vivo or whose function is regulated by endogenous NOS activity from those that can be modified in vitro or whose function is altered by treatment with NO donors. Two proteins fall into the first category, namely, the ryanodine receptor-calcium release channel and guanylate cyclase. Xu et al. (218) showed that ryanodine receptor purified from canine hearts was endogenously S-nitrosylated. These nitrosothiols (SNOs) in the protein are in equilibrium with low-mass compounds such as S-nitrosoglutathione. Xu et al. (218) also showed that S-nitrosylation of purified channels incorporated into planar lipid bilayers increased open channel probability and that channel activation was reversed by denitrosylation. We have more recently confirmed that the ryanodine receptor (RyR) from rabbit skeletal muscle is S-nitrosylated at a single cysteine (65), and others have reported that channel function is reversibly altered by exposure to various NO donors, albeit at relatively high concentrations (5, 135, 188). Both activation and inhibition of the RyR have been reported, suggesting that NO and related molecules can interact with more than one regulatory site (5, 66, 135, 188). In particular, Meszaros et al. (135) demonstrated that a skeletal NOS can inhibit RyR activity, and our own studies (65) strongly suggest that skeletal RyR activation by S-nitrosylation is physiologically relevant and functionally significant (also see sect. VIII).

Almost a quarter of a century has passed since Murad and co-workers (112) reported on low-level guanylate cyclase activity in skeletal muscle. Only very recently, however, did evidence for a NO-related function come to light, including a remarkable colocalization of nNOS and cGMP in the vicinity of the sarcolemma (118). These studies called attention to a cGMP-mediated inhibition of force production by showing that all of the following inhibit excitation-contraction coupling in skeletal muscle: 1) cell-permeable cGMP analogs; 2) agents that prevent cGMP breakdown (phosphodiesterase inhibitors); and 3) NO donors that increase cGMP synthesis (1, 118). Proof of a functional linkage between cGMP and muscle activity was provided by Lau et al. (124), who found that extensor digitorum longus cGMP content increased ~250% with electrical stimulation and that these increases did not occur in mdx mice that are deficient in nNOS. In addition to effects on contractility, there is evidence that cGMP may promote glucose uptake and/or utilization (see sect. X).

It is important to note, however, that cGMP effects are generally modest in skeletal muscle, and guanylate cyclase activity is relatively low compared with other tissues (112). The mechanism(s) by which cGMP operates in skeletal muscle is also not well understood. In this regard, it is particularly interesting that nNOS and cGMP-dependent protein kinase, a primary effector of NO, concentrate at the neuromuscular end plate (45) and cGMP-dependent kinase may phosphorylate the nicotinic acetylcholine receptor subunits (or associated proteins) in cultured myocytes (101, 209). Moreover, endogenous cGMP-dependent kinase has been demonstrated to phosphorylate dystrophin (131).

NOS activity has been shown to inhibit mitochondrial respiration in skeletal muscle (114, 119). Putative NO targets that have been studied in vitro and/or which may contribute to this effect include cytochrome-c oxidase (31, 119), creatine kinase (216), and the sarcoplasmic reticulum (SR) Ca²⁺-ATPase in fast-twitch (SERCA1a) and slow-twitch (SERCA2a) skeletal muscle (117, 205, 206). Evidence acquired in other tissues indicates that cytochrome-c oxidase is particulary sensitive to physiological amounts of NO (49, 83, 84, 224). In contrast, the SR-bound creatine kinase from rabbit skeletal muscle is exquisitely sensitive to inhibition by S-nitrosoglutathione (216). S-nitrosylation of a critical cysteine residue leads to a decrease in calcium uptake that is dependent on ATP generated from phosphocreatine and ADP. There is also circumstantial evidence for NO inhibition of the Ca²⁺-ATPase. In rabbit fast-twitch muscle, sustained contractile activity leads to significant (40-50%) inactivation of the ATPase, which is preceded by a SERCA1a to SERCA2a transition (117). Immunoblotting with antidinitrophenyl and antinitrotyrosine antibodies and spectrophotometric assays showed significant increases in the protein carbonyl content of the ATPase and "generally higher" labeling for nitrotyrosine, suggesting that increased oxidation and peroxynitrite-mediated nitration contributed to the inactivation. In cardiac muscle, the SR Ca²⁺-ATPase (SERCA2) is evidently regulated by an SR nNOS (217). Nitrotyrosine is also formed coincident with aging. Viner and co-workers (205, 206) showed nitration of one tyrosine in SERC2a of rat slow-twitch muscle at 5 mo and as many as four tyrosines by 28 mo. However, the case made for nitration in general and peroxynitrite-induced nitration in particular in the loss of SERCA function with ageing of slow-twitch muscle is less convincing, because nitration is rarely, if ever, the sole redox-related modification of a protein. Rather, it is a footprint for NO-related reactions, including oxidations of thiols, methionine, tyrosine, and amines. Although one is thus hard pressed to correlate any specific redox-related modification of the SERCA with the loss of activity, the reactions with critical SH groups may provide the best explanation (25, 86, 117); that is, peroxynitrite treatment of the rabbit skeletal Ca²⁺-ATPase (SERCA1a) was correlated with oxidative and nitrosative modifications of cysteines at positions 344, 349, 471, 498, 532, and 614, of which those with cysteine 349 were deemed responsible for enzyme inhibition (207). In addition, NO preferentially reacted with thiols of SERCA2a (206) coincident with a loss of enzyme activity.

Glyceraldehyde-3-phosphate dehydrogenase, aconitase, and complex 1 of the respiratory chain are additional NO targets (48, 180) whose inhibition might influence cell respiration or calcium homeostasis, if not under normal physiological conditions then in cells exposed to nitrosative stress. Treatment of muscles with NO donors can inhibit actomyosin ATPase activity (actin-myosin cross-bridge cycling) (152) and thereby reduce skeletal muscle force (152). However, it is difficult to imagine how physiological concentrations of NO or SNO might achieve this effect.

	IV. LOCALIZATION AND EXPRESSION OF NITRIC OXIDE SYNTHASE IN SKELETAL MUSCLE

Top Previous Next References

A.  nNOS

1.  Mature skeletal muscle

nNOS has been identified in skeletal muscles by immunological, histochemical, and cellular fractionation methods. All three methods have provided valuable information on the expression, location, and activity of nNOS. Immunocytochemistry and fluorescence have provided the most accurate information on the cellular distribution and location of NOS isoforms in skeletal muscle. An advantage of this method is that highly specific, commercially available antibodies can distinguish between the three major NOS isoforms (see sect. I). A limitation is its relative insensitivity in detecting the enzymes. An often-used method to overcome this problem is to stain cells concurrently for NADPH diaphorase activity of nNOS. The method is more sensitive but less specific, i.e., staining does not always correlate with the presence of one or more of the three NOS isoforms, as determined by immunocytochemistry (220). Also, staining for NADPH diaphorase activity cannot distinguish among the three isoforms. Cellular fractionation studies have been useful in determining whether NOS activity is present in cytosolic or membrane-bound forms.

nNOS has been detected in various skeletal muscles including human gastrocnemius, omohyoideus, quadriceps, sternocleidomastoideus, urethral sphincter and vastus lateralis muscle (42, 72, 89, 98, 144), rat and/or mouse diaphragm, deltoideus, extensor digitorum longus (EDL), gastrocnemius, levator labii, soleus, quadriceps and tibialis anterior muscle (28, 35, 42, 44, 47, 53, 62, 91, 111, 118, 122, 193, 199), and in the skeletal muscles in a variety of other species including gerbils, guinea pigs, hamsters, turtles, chicken, pigeons, and goldfish (80, 85, 88, 90).

In adult skeletal muscle, nNOS is largely targeted to specialized structures at the surface membrane. Nakane et al. (144) found that homogenates of human skeletal muscle tissue displayed a higher NOS activity than human brain. Much of this activity was recovered in a pellet fraction but could be released to a large extent by two 1 M KCl washes, indicating that it was loosely bound to membranes or cytoskeletal components. Western blot analysis indicated that the activity could be ascribed to the nNOS and not the eNOS isoform. Interestingly, Nakane et al. (144) failed to detect significant NOS activity in rat skeletal muscle homogenates, suggesting a higher level of nNOS activity in human than rat skeletal muscle. This last finding remains to be confirmed, however, in light of studies (118) establishing the presence of significant nNOS activity in rat skeletal muscle, the vast majority similarly residing in the particulate fraction. In fact, Kobzik et al. (118) found that nNOS activity was significantly higher in rat brain than in either rat limb or diaphram muscles. Since these pivotal studies of Nakane and Kobzik and their colleagues (118, 144), additional reports have appeared describing the subcellular distribution of nNOS. In these studies, few attempts were made to distinguish between loosely and tightly bound nNOS (as done by Nakane et al., Ref. 144). Chang et al. (42) used a low ionic strength buffer and found that ~80% of NOS activity was recovered in the pellet fraction of human skeletal muscles compared with ~50% in mouse skeletal muscles. Western blot analysis corroborated the activity studies by showing an nNOS band in the pellet and supernatant fractions of mouse skeletal muscle. Significant nNOS also remained with a pellet fraction after sequential extraction of mouse skeletal muscle homogenates with 0.1 M NaCl and 0.5 M NaCl in 0.5% Triton X-100, as determined by immunoblot analysis (35). In guinea pig skeletal muscle, almost all nNOS activity was recovered in the particulate fraction (80).

Rodents and other species including primates showed distinct nNOS immunostaining in a subset of skeletal muscle fibers (35, 88, 90, 111, 118). nNOS was concentrated at the surface membrane of type II (fast-twitch) fibers, whereas type I (or slow-twitch) fibers showed no or only a weak reaction (Fig. 1). These observations are supported by biochemical studies that have shown a higher NOS activity in muscle enriched in type II (EDL, gastrocnemius, plantaris) than type I (soleus) fibers (62, 111, 118) (see sect. VIIB). However, nNOS immunoreactivity was not restricted to type II fibers in all studies. In rat facial skeletal muscles (193) and extrafusal and intrafusal fibers of sheep (85), sarcolemmal nNOS immunostaining was observed in both type I and type II fibers. nNOS location at neuromuscular junctions was not related to skeletal muscle fiber type as nNOS immunostaining and NADPH diaphorase activity were present at all rat EDL and soleus end plates (122).

View larger version (111K): [in this window] [in a new window]   Fig. 1. Localization of neuronal nitric oxide synthase (nNOS) in rat skeletal muscle. A: immunohistochemical analysis of rat diaphram muscle cryosections, using a polyclonal anti-nNOS antibody, shows distinct labeling of the surface membranes of type II (fast) but not type I (slow, astericks) muscle fibers. B: staining of adjacent serial sections for myosin ATPase identified fibers as type II (nNOS positive) and type I (nNOS negative). [From Kobzik et al. (118).]

In humans, nNOS is expressed in both type I and type II fibers. Immuno- and histochemical studies have indicated that nNOS is expressed about equally in type I (slow-twitch) and type II (fast-twitch) fibers of some human skeletal muscles (85, 89). On the other hand, Frandsen et al. (72) observed a greater expression of nNOS in the cytoplasm and at the sarcolemma of type I than type II fibers of human vastus lateralis muscle. However, whereas nNOS expression and activity generally correlate well in rodents, the relationship has not been determined in humans (see sect. V).

nNOS is targeted to specialized structures at the surface membrane of skeletal muscle cells, by binding to ₁-syntrophin, a dystrophin-associated protein. A consequence of lack of dystrophin in patients with Duchenne muscular dystrophy and an animal model of this disease (mdx mice) is therefore the loss of nNOS from the sarcolemma (discussed in sects. IVA3 and XV). In normal adult skeletal muscles, the enzyme is highly enriched at neuromuscular junctions (34, 45, 47, 80, 88, 122, 148) and myotendinous junctions and costamers (42). Immunohistochemical and fluorescence studies further show a rather uniform distribution of nNOS along the sarcolemma of type I and II fibers in humans and type II fibers in rodents and other species (34, 35, 42, 47, 72, 88-90, 118). Wakayama et al. (208) analyzed the localization of nNOS, ₁-syntrophin and dystrophin in normal human skeletal muscle fibers by double immunogold-labeling electron microscopy. An association of the three proteins at the inner surface of the sarcolemma and subsarcolemmal areas near mitochondria was observed. In addition to ₁-syntrophin, nNOS binds caveolin-3 (204), a component of specialized invaginations of the sarcolemma, called caveolae. It is therefore possible that the subsarcolemma areas observed by Wakayama et al. (208) represented caveolae. However, caveolin-3 is unlikely to play a primary role in targeting nNOS to the sarcolemma because nNOS was absent from the sarcolemma of muscle fibers lacking dystrophin (see below).

It is unclear whether nNOS is targeted to intracellular organelle in skeletal muscle. Immunoelectron microscopy showed labeling of isolated cardiac muscle SR vesicles using an anti-nNOS gold-labeled antibody (217). However, no labeling was detected in skeletal muscle SR vesicles, which suggests that localization of nNOS to SR may be specific to cardiac muscle. In the soluble fraction of skeletal muscle homogenates, significant amounts of nNOS are, however, associated with the muscle phosphofructokinase isoform (70). Although the physiological function of this interaction in skeletal muscle is unclear, it is of interest that an association between the two proteins has been suggested to contribute to neuroprotection of nNOS positive cells in brain, as the product of the kinase, fructose-1-phosphate, is neuroprotective. In hamster skeletal muscle, nNOS immunolabeling exhibits a heterogeneous pattern reminiscent of mitochondria (174).

The nNOS gene has 29 exons that span a region of >240 kb at 12q24.2 and produce multiple mRNA transcripts by various mechanisms including the use of alternative promotors and the alternative splicing of nNOS pre-mRNAs (for review, see Wang et al., Ref. 210). Alternative splicing provides a means for the tissue-specific expression of nNOS. Skeletal muscle expresses an alternatively spliced nNOS isoform (nNOSµ) that is slightly larger than the brain isoform due to an alternatively spliced 102-bp (34-amino acid residues) segment between exons 16 and 17 (178). Wang and co-workers (210, 211) demonstrated the existence of additional mRNA transcripts that differ in the first exon upstream of the coding sequence and therefore do not result in different encoded protein variants.

nNOSµ is developmentally regulated as its expression coincides with myotube fusion in culture (also see sect. XIV). An isoform-specific antibody detected nNOSµ in skeletal and cardiac muscle but not other tissues (178). Immunocytochemical and histochemical evidence has been provided for the presence of all three NOS isoforms in differentiating C₂C₁₂ cells and developing muscle (28). In both C₂C₁₂ myoblasts and myotubes, diffuse NOS immunoreactivities were localized in the cytoplasm and/or at perinuclear regions. Similarly, in prenatal embryonic day 19 upper limb skeletal muscle tissue, a diffuse cytosolic staining could be detected in both early myotubes and myofibers (28). nNOS immunoreactivity was absent from the sarcolemma until postnatal day 5 when it could be detected in some myofibers; in adult skeletal muscle, it was clearly present at the sarcolemma.

In the skeletal muscle of mice, nNOS protein levels increased greatly from 2 to 52 wk (42). Capanni et al. (40) determined that the amounts of nNOS detected by immunoblotting were higher in the upper limb skeletal muscles of 24- than 6-mo-old rats. Immunocytochemistry in 6-mo-old rats showed a fluorescence that was localized exclusively at the sarcolemma, whereas some muscle fibers of 24-mo-old rats showed an additional granular and diffuse cytoplasmic fluorescence. Most of this activity likely remained membrane-bound, since a similar high fraction (~90%) of nNOS was recovered in a heavy microsomal fraction.

Developmental regulation in ventilatory muscles differs from that in limb skeletal muscle. NOS activity as well as nNOSµ and eNOS expression were high in fetal and neonatal diaphragm, with a gradual reduction in the expression levels seen during late postnatal development (62). In this study, nNOS was localized at the sarcolemma of both neonatal and adult diaphragm.

Denervation of skeletal muscle resulted in no or only a slow decrease in nNOS immunoreactivity. Chao et al. (45) compared the distribution of nNOS and cGMP-dependent protein kinase type I, a primary effector for NO, in mouse and rat fast-twitch (EDL) and slow-twitch (soleus) muscles. In rat EDL, nNOS was present at the extrajunctional sarcolemma and concentrated at the neuromuscular junction. In soleus muscle, nNOS staining was considerably weaker and was only readily detectable at the neuromuscular junction. In both types of muscle, cGMP-dependent protein kinase type I staining was detected only at the neuromuscular junction. The protein levels of nNOS and the distribution and density of nNOS and the kinase did not change after denervation for 14 days. Kusner et al. (122) showed that nNOS immunoreactivity at the neuromuscular junctions of both rat soleus and EDL muscles persisted 14 days after denervation; however, the intensity of staining was reduced compared with the control. Electron microscopy studies showed a postsynaptic colocalization of nNOS with dystrophin and desmin. iNOS was also found postsynaptically; however, it did not colocalize with desmin (219). Denervation of rat sciatic nerve did not decrease nNOS and iNOS immunoreactivity after 7 days but did so after 15 days. A complete loss of immunoreactivity was observed 30 days after denervation. In rat facial muscles, a significant downregulation of nNOS occurred after 14 days of denervation, as indicated by a lack of nNOS immunoreactivity in 30-60% of the muscle fibers (193). Downregulation was observed in >60% of the fibers 7-24 wk after denervation. There were no changes in the expression of eNOS, and iNOS was not detected. Denervated and immediately renervated muscles showed a transient downregulation of nNOS at weeks 2-7 followed by a return to control levels at weeks 10-20.

Changes in muscle activity and mechanical load affect the fiber composition and expression of nNOS in skeletal muscle (see sect. V for details). In particular, prolonged treadmill training increased protein expression of both nNOS and eNOS in soleus muscle homogenates (15). Reiser et al. (162) used chronic in vivo electrical stimulation of tibialis anterior and EDL muscles of rabbits to induce a fast to slow fiber type transformation. nNOS activity and protein levels doubled after 3 wk of stimulation, despite the fact that normal (unstimulated) fast-twitch skeletal muscle expresses higher levels of nNOS than slow-twitch muscle. Fujii et al. (74) determined the acute effects of 3 h of moderate and severe inspirative resistive loads on nNOS activity and protein expression in ventilatory (diaphragm, intercostal and transverse abdominis) and limb (gastrocnemius) muscles. nNOS activity decreased significantly in diaphragm and intercostal muscle, whereas activity remained unchanged in transverse abdominis and gastrocnemius muscles. The use of several in vivo and in vitro models confirmed that muscle activity regulates NO production acutely and nNOS expression in the long term (199).

In skeletal muscle of patients with Duchenne muscular dystrophy, nNOS is absent from the sarcolemma and diffusely distributed at reduced levels in the cytoplasm (30, 35, 42, 89, 225) (Fig. 2). nNOS is also absent from the sarcolemma in mdx mice, a dystrophic mouse model (35, 42, 89). Brenman et al. (35) found that total nNOS activity was lower in mdx mice, although NOS activity was higher in a soluble fraction of mdx than wild-type mice, which led them to suggest that nNOS dissociates from the sarcolemma and accumulates in the cytosol of mdx mice muscle. On the other hand, Chang et al. (42) noted that both the particulate and soluble nNOS levels were greatly reduced in mdx mice, as they are in muscles of patients with Duchenne muscular dystrophy (35). Chang et al. further noted that nNOS mRNA levels were also greatly reduced in mdx muscle, suggesting a defect in nNOS expression (42). Interestingly, nNOSµ is a calpain-sensitive protein, and elevated (20-fold) levels of the protease are expressed in Duchenne muscular dystrophy (123). An increased rate of proteolytic degradation may therefore contribute to the diminished nNOS activity of dystrophic muscle.

View larger version (82K): [in this window] [in a new window]   Fig. 2. Absence of nNOS at the sarcolemma of muscle fibers in Duchenne muscular dystrophy. Skeletal muscle cryosections of normal (N) and Duchenne muscular dystrophy (D) skeletal muscle were immunostained with antibodies to dystrophin and nNOS. [From Brenman et al. (35). Copyright 1995 Cell Press.]

Loss of sarcolemmal nNOS is caused by the absence of dystrophin, a 427-kDa cytoskeletal protein localized at the sarcolemma, due to a mutation in the gene for the protein (120). Dystrophin gene replacement leads to the correct localization of nNOS in mdx mice, clearly pointing out its importance in targeting nNOS to the sarcolemma (43, 56). However, nNOS does not bind to dystrophin directly, but rather is targeted to the sarcolemma via ₁-syntrophin, a dystrophin-associated protein that is predominantly expressed in skeletal and cardiac muscle. Immunocytochemical studies have shown that this protein is particularly enriched at the neuromuscular junction of skeletal muscles but is also present at the extrasynaptic sarcolemma (110, 153). Direct interaction between nNOS and ₁-syntrophin has been shown using the yeast two-hybrid system, in pull-down and protein overlay assays (34), and in an ₁-syntrophin knock-out mouse (110). As with the absence of dystrophin (in Duchenne muscular dystrophy and the mdx mouse), loss of ₁-syntrophin caused a redistribution of nNOS from sarcolemma to the cytosol in muscle cells. Analogous analyses and gene replacement studies in Becker muscular dystrophy, which like Duchenne muscular dystrophy is an X-linked disease due to mutations in the dystrophin gene, were less straightforward. These data are covered in section XV.

In brain, nNOS interacts with postsynaptic density 95 protein (PSD-95) and the related protein PSD-93 (34). These interactions are mediated by PDZ protein motifs present in nNOS and the two PSD proteins. Similarly, binding of nNOS to ₁-syntrophin is mediated by PDZ protein motifs near the NH₂ termini of the two proteins. A large exon 2 encodes the first 236 amino acid residues of mouse nNOS. This region is not required for enzymatic activity and is not conserved in the two other NOS isoforms (eNOS and iNOS); rather, it includes a PDZ motif that heterodimerizes with the PDZ domains of postsynaptic density protein PSD-95 in brain and ₁-syntrophin in muscle (34). Crystallographic analysis has revealed that the PDZ domains in the nNOS-syntrophin complex associate in an unusual linear head-to-tail arrangement involving an internal -hairpin "finger" in nNOS that links up with the peptide binding groove of the syntrophin PDZ domain (96). In the complex, the peptide-binding groove of the nNOS PDZ domain remains accessible to potentially recruit additional proteins to the surface membrane of skeletal muscle.

Some observations suggest that the presence of dystrophin and ₁-syntrophin are not sufficient for surface membrane association of nNOS. In several neurogenic diseases, nNOS was absent from or greatly reduced at the sarcolemma despite the presence of dystrophin and the dystrophin complex (87). Furthermore, nNOS did not bind to ₁-syntrophin at neuromuscular junctions of transgenic mdx mice expressing truncated dystrophin proteins (43). Other studies indicate that nNOS may interact with additional proteins at the surface membrane of skeletal muscle. In neurons, PSD-95 links nNOS to N-methyl-D-aspartate (NMDA) receptors (121). A comparable interaction may take place in skeletal muscle as NMDA receptors are present at the neuromuscular junction in rat diaphragm (24), and colocalization of NMDAR-1 and nNOS in diverse skeletal muscles of rats and mice has been reported (88). In mdx mice, nNOS was essentially absent (43) but could be detected when a sensitive two-step histochemical method was used (88). A strong nNOS immunoreactivity was detected at the neuromuscular junction of Duchenne muscular dystrophy patients (219). These results raise the possibility that different mechanisms exist for anchoring nNOS to junctional and extrajunctional muscle surface membranes.

eNOS is a dually acylated peripheral membrane protein that is targeted to endothelial plasmalemmal caveolae through an interaction with the caveolae structural protein, caveolin-1 (78) and sarcolemmal caveolae (in the heart) through a similar interaction with caveolin-3 (69). Although immunoblot analysis shows that rat EDL and soleus muscles express eNOS (111), its low level of expression has made it more difficult to localize and study. Immunolocalization showed that eNOS colocalized with mitochondrial markers in a subset of fibers within rat diaphragm, EDL, gastrocnemius, and soleus (119). Mitochondria isolated from rat skeletal muscle showed a silver enhanced gold immunolabeling for eNOS, strongly suggesting a specific localization of eNOS to skeletal muscle mitochondria (21) (see sect. XII) (Fig. 3). In rat facial skeletal muscles, a patchy sarcolemmal and weak sarcoplasmic immunolabeling pattern was observed (193). eNOS is also expressed, along with the other two isoforms, in differentiated mouse C₂C₁₂ myoblast cultures and adult rat skeletal muscle tissues (28). In myoblasts and fused myotubes, the three isoforms were all present in the sarcoplasm. In guinea pig gastrocnemius muscle, significant eNOS immunoreactivity was detected only in the vascular endothelium (80).

View larger version (161K): [in this window] [in a new window]   Fig. 3. Electron micrograph of mitochondria from skeletal muscle labeled for endothelial nitric oxide synthase (eNOS) by enhanced gold immunolabeling. Immunodeposits (arrows) are seen over the membranes and the matrix space of mitochondria treated with anti-eNOS monoclonal antibody and gold-labeled anti-mouse IgG (A), but not in control preparations (eNOS antibody omitted) (B). ms, Matrix space; cr, cristae. Bars, 0.5 µm. [From Bates et al. (21).]

iNOS activity varies in skeletal muscles, depending on disease state and species investigated. iNOS expression is greatly increased in skeletal muscle cells of patients with chronic heart failure (4, 164) or autoimmune inflammatory myopathies (192) and in animals and skeletal muscle cell cultures after exposure to bacterial lipopolysaccharides (LPS) or inflammatory cytokines (29, 62).

iNOS mRNA is absent or present at very low levels in normal skeletal muscles of rats (105, 143) and mice (198). Similarly, immunoblot and immunocytochemical analyses showed no or only minimally detectable iNOS immunoreactivity in various skeletal muscles of rat (29, 61, 105, 111, 193) and mouse (198). On the other hand, iNOS is constitutively expressed in skeletal muscle fibers of pathogen-free guinea pigs (80). RNA protection analysis of gastrocnemius muscle and immunoblots of extracts of mouse gastrocnemius and diphragm muscles showed a weak basal expression of iNOS mRNA and protein (81), reminiscent of findings by Blottner et al. (28) in rat tissues. iNOS copurified with a pooled particulate protein fraction of guinea pig diaphragm and gastrocnemius muscles and showed a spotty intracellular appearance in type I fibers of pathogen-free animals (80) (Fig. 4). Both the mRNA and protein levels increased in mice treated with bacterial LPS (81). Specificity was ascertained by the demonstration that iNOS mRNA and protein were absent in the iNOS knock-out mice after treatment with LPS.

View larger version (83K): [in this window] [in a new window]   Fig. 4. Immunohistochemical localization of inducible nitric oxide synthase (iNOS) in gastrocnemius muscle of normal and pathogen-free guinea pigs. A: staining for myosin ATPase after preincubation in alkaline solution. B: serial section of A showing a spotty distribution of iNOS immunolabeling throughout the cytoplasm of muscle fibers containing alkali-labile myosin ATPase (type I fibers) (arrows). [From Gath et al. (80).]

Treatment of animals with LPS typically results in a transient appearance of iNOS in skeletal muscles, with a peak of activity ~8-12 h after LPS injection (29, 62). (The functional implications are discussed in sects. VIIE and XIII.) The initial rise in iNOS activity in rat diaphragm was due to the infiltration of the perivascular spaces of the diaphragm by inflammatory cells that expressed iNOS (29). Twelve and 24 h after LPS injection, a diffuse pattern of iNOS staining was observed in variable portions of the muscle fibers of diaphragm. iNOS expression was absent in abdominal, intercostal, and peripheral skeletal muscles. El-Dwairi and colleagues (61) also observed an intense diffuse staining after 12 h of LPS injection; however, in contrast to the findings of Boczkowski et al. (29), iNOS expression was found both in rat diaphragm and intercostal muscles. nNOS and eNOS expression also increased in diaphragm, although to a lesser extent than iNOS and in contrast to iNOS remained elevated 24 h after LPS injection. Thompson et al. (198) found that LPS treatment greatly increased iNOS mRNA levels after 4 h and protein levels after 8 h in mice diaphragm and skeletal muscles (198). Treatment of diaphragmatic explants with LPS resulted in a significant increase of cytoplasmic iNOS immunoreactivity. Thompson et al. (198) failed, however, to detect iNOS immunostaining in diaphragm myocytes 24 h after injection of LPS into mice. This absence in vivo was ascribed to the insensitivity of the immunocytochemical technique.

Normal human skeletal muscles express low levels of iNOS (150) that are markedly increased in patients with chronic heart failure, as determined by Northern blot or RT-PCR analysis (92, 164). Immunogold labeling (164) and immunohistochemistry (4) localized iNOS predominantly over the myofibrils of skeletal muscle in patients with heart failure. In control subjects, no staining could be observed. Tews et al. (192) reported an upregulation of iNOS and nNOS in muscle fibers of patients with autoimmune inflammatory myopathies. nNOS showed a typical sarcolemmal pattern with the exception of some atropic and necrotic fibers that displayed an enhanced cytoplasmic nNOS distribution. iNOS was also localized at the sarcolemma rather than over the myofibrils. In muscle fibers of patients with autoimmune inflammatory myopathies, an intense immunocytochemical labeling was observed along large stretches of the sarcolemma, in contrast to small spots seen in normal muscle.

It was recently reported that iNOS coimmunoprecipates with the sarcolemmal caveolae membrane protein, caveolin-3, thus providing a molecular basis for a sarcolemmal localization of iNOS (81). An iNOS antibody coprecipitated caveolin-3 from a detergent extract of mice treated and not treated with LPS and from C₂C₁₂ skeletal muscle cells treated and not treated with LPS and interferon (IFN)- (81). Conversely, a caveolin-3 specific antibody coprecipitated iNOS from the muscle samples. Accordingly, caveolin-3 can target iNOS to sarcolemmal caveolae in skeletal muscle. Inhibition of nNOS activity by caveolin-3 (78, 204), presumeably by displacement of calmodulin, raises the interesting possibility of a regulation of iNOS by the same mechanism.

Upregulation of iNOS by cytokines and LPS has been investigated in cultured rat L6 and mouse C₂C₁₂ skeletal muscle cells (22, 215). Neither LPS, interleukin (IL)-1, tumor necrosis factor (TNF)-, nor IFN- was able to stimulate nitrite production by C₂C₁₂ (215) myocytes when added alone. However, combinations of IFN with either or both of the two other cytokines increased nitrite production by C₂C₁₂ skeletal muscle cells and yielded a mRNA band homologous to mouse macrophage iNOS (215). In LPS/cytokine-treated L6 skeletal muscle cells, insulin reduced NO production by a mechanism that did not involve a suppression of iNOS gene transcription at the iNOS mRNA and protein level (22) (see also sect. XB). Okuda et al. (147) reported a complex regulation of iNOS in C₂C₁₂ cells by cytokines involving a tyrosine-dependent pathway and positive modulation by protein kinase A- and protein kinase C-dependent mechanisms.

	V. MUSCLE NITRIC OXIDE SYNTHASE ACTIVITY

Top Previous Next References

NOS activity resides mainly in the particulate fraction of skeletal muscle homogenates (118, 144). By monitoring the conversion of [³H]arginine to [³H]citrulline, NOS activity has been estimated to range from 2 to 25 (average ~10) pmol·min¹·mg¹ (14, 15, 61, 80, 118, 125, 144, 161, 162, 199). The NOS activity in rat muscles (soleus, diaphragm, gastrocnemius, and EDL) correlates with percent type II (fast-twitch) fiber composition (118). The same fiber-type correlation may not hold as true in human skeletal muscle where nNOS is somewhat more evenly distributed among fiber types (33, 90). nNOS activity is increased up to severalfold by muscle stimulation, evidently under a variety of stimulation protocols [2 Hz/ min (199), 25 Hz/60 min (118), 100 Hz/5 min (14), and 10 Hz/3 wk (14)]. Isolated skeletal muscle mitochondrial preparations show calcium-dependent NOS activity of ~1 fmol·mg protein¹·min¹ (9 fmol/16 min) (119), which has been attributed to eNOS enrichment. Human cytosolic fractions contain a similarly low level of activity (144). Guinea pig is an exception to most of these rules (80). Only one-half of the particulate activity in guinea pig skeletal muscle derives from nNOS; the other half comes from iNOS. Specifically, using a cGMP reporter assay in RFL 6 fibroblasts to assay NOS activity, Gath et al. (80) measured 0.93 pmol cGMP/g calcium-dependent NOS activity and 1.17 pmol cGMP/g calcium-independent activity. In addition, significant calcium-independent NOS activity (iNOS) (0.14 pmol cGMP/g) was detected in the soluble fraction. In contrast to rats, NOS2 was localized selectively to type I (slow-twitch) fibers.

NOS activity has been measured directly in rat muscle bundles (14, 118, 161, 199). Resting rat diaphragm produced ~3-5 pmol NO equivalents·min¹·mg muscle¹ over 60-90 min as measured by two different assays (118, 161). NO production was increased approximately sixfold in actively contracting muscle (6 contractions/min at 25 Hz for 1 h) in a tissue bath (118, 161). Balon and Nadler (14) reported slightly lower levels of NO metabolites produced by cordofemoralis and quadratus femoralis muscles, stimulated in situ via the peroneal nerve and then incubated in vitro (14). Resting muscle produced ~1 pmol·mg¹·min¹ over the first hour, which was increased two- to threefold by prior electrical stimulation. In contrast, Tidball et al. (199) measured a resting activity of 17 pmol NO·mg¹·min¹ in excised rat soleus muscles, but only 20% increases after a single stretch or after 2 min of submaximal stimulation (at 2 Hz). They attributed these higher levels to an early burst in NOS activity that normalizes by 2 min. Although these authors used a more suspect electrophotometric technique that may have overestimated absolute NOS activity, others have likewise found disproportionately higher activity in the first minute (159) and a fall in activity over time (14).

Mouse diaphragm and soleus muscles produced ~13 pmol·mg¹·min¹ NO equivalents during passive incubation (97). No differences were seen between muscle types or between wild-type and eNOS-deficient animals, thus identifying basal NOS activity with nNOS. During active contraction, however, diaphram NOS activity increased eightfold, whereas soleus (slow-twitch muscle) only increased three- to fourfold. This difference in stimulated NOS activity between muscle types was preserved in eNOS-deficient animals.

Reiser et al. (162) used a model of chronic in vivo electrical stimulation to the tibialis anterior and EDL muscles of rabbits as a method of inducing fast or slow fiber-type transformation. They found that this protocol was associated with a doubling of membrane NOS activity, from 21 pmol·mg¹·min¹ in control muscles, which they attributed to increased expression of nNOS protein. [It should be noted, however, that very high concentrations of L-arginine substrate (20 mM) were used in these studies and that NOS activity was ~10-fold lower in assays using more typical (3 mM) arginine concentrations.] Tidball et al. (199) have also reported that mechanical activity positively regulates NOS activity and expression. Specifically, removal of weight bearing from rat hindlimb resulted in a 43% decrease in nNOS content of soleus muscle after 10 days (199). nNOS protein levels were restored by 2 days of weight bearing. It is unclear from the data if NOS activity correlates with protein content. Tidball et al. (199) further showed that myotubes subjected to cyclic stretching for 2 h similarly showed a 42% increase in NOS activity (basal 0.6 pmol NO·mg¹·min¹). Taken together with earlier studies, the findings are consistent with a picture in which muscle activation is associated with increased NOS activity.

NOS activity is developmentally regulated. Total activity in diaphragm homogenates averaged 40-45 pmol·min¹·mg protein¹ from embryonic day 18 to postnatal day 1 (62). NOS activity then declined to ~25 pmol·min¹mg protein¹ by day 7 and to the typical adult rates by day 30. By analogy, nNOS and iNOS activities have been detected in cultured embryonic myoblasts (109, 125). Activities peak acutely at the time of myoblast fusion (approximately day 13 and disappear by day 20). Levels of nNOS also increase with age (40). In Wistar rats, nNOS enrichment of both the microsomal and soluble fractions takes place from months 6-24.

Endotoxin has profound effects on the patterns of NOS isoform expression. In one series of studies, iNOS protein peaked 12 h after LPS injection and disappeared by 24 h (61), whereas eNOS and nNOS increased progressively over 24 h. More specifically, El-Dwairi et al. (61) measured a calcium-independent activity (iNOS) 12 h after LPS injection (~15 pmol·min¹·mg protein¹) in diaphragm, intercostals, and gastrocnemius and a calcium-dependent activity (eNOS and nNOS) (15-30 pmol· min¹·mg protein¹) at 24 h after LPS injection. Interestingly, soleus was unresponsive to LPS. The same group then concluded from studies in iNOS mutant mice that this isoform serves to downregulate the expression of nNOS (50). Inasmuch as nNOS may contribute more to overall NOS activity than iNOS, a yin-yang relationship appears to exist between the level of iNOS expression and total NOS activity in skeletal muscle; such a relationship has not been previously noted in other murine tissues.

Increases in whole muscle NOS activity that occur with electrical stimulation of skeletal muscle and/or its innervating nerves (above) are consistent with the idea of calcium activation of nNOS during contraction. In cardiac muscle, dynamic changes in the concentrations of NO take place on millisecond time scales that typify excitation-contraction coupling (154). These sorts of measurements, however, have not been made in skeletal muscle.

The majority of NOS activity in skeletal muscle is identified with nNOS. Several small proteins have been implicated in the regulation of its activity. Muscle nNOS is activated by extracellular signals that increase the cytosolic Ca²⁺ concentration and thereby facilitate the interaction of the enzyme with calmodulin. Two other proteins, caveolin-3 and a 10-kDa protein designated PIN, have been reported to inhibit nNOS activity in skeletal muscle. nNOS interacts with two scaffolding/inhibitory domains of caveolin-3 (78, 204), a component of specialized invaginations of the sarcolemma, called caveolae. A primary regulatory rather than targeting role of caveolin-3 is suggested by the absence of sarcolemmal nNOS staining in fibers lacking dystrophin (30, 35, 42, 89, 225) or ₁-syntrophin (110). Caveolin-3 interacts with other signaling proteins such as c-src, Ha-ras, and G_s, which suggests that nNOS may be part of a signaling complex in skeletal muscle (51). Caveolin-3 may also regulate iNOS and eNOS in skeletal muscles. Specifically, caveolin-3 coprecipitates iNOS from skeletal muscle cell homogenates exposed to LPS and IFN- (81), and eNOS associates with caveolin-3 in cardiac myocytes, perhaps rationalizing the patchy sarcolemmal immunolabaling for eNOS in facial skeletal muscles (193).

Jaffrey and Snyder (107) used the yeast two-hybrid system to show that a small (89 amino acid residues) protein designated inhibitor protein of nNOS (PIN) interacts with nNOS. PIN inhibited nNOS activity at low micromolar concentrations by destabilizing the nNOS dimer. PIN specifically associated with nNOS by binding to amino acid residues 161-245 that are absent from eNOS and iNOS. More recent NMR titration experiments have the PIN binding region mapped to amino acid residues 228-244 of nNOS (67). These studies also showed that PIN binds to nNOS in a 1:2 stoichiometry. It is unlikely that PIN interferes with the binding of nNOS to ₁-syntrophin, since the PIN and ₁-syntrophin binding domains of nNOS are separated by a flexible linker at least 68 amino acids long (67).

PIN is expressed at high levels in testis, intermediate levels in brain, and low levels in most peripheral tissues (107). Nevertheless, embryonic rat diaphram levels are evidently much higher than adult (91). Guo et al. (91) have determined that levels of PIN protein expression are similar in rat, mouse, and human diaphragms and unrelated to muscle fiber type. Immunohistochemistry showed that, in accordance with its proposed regulatory role, PIN was localized in close proximity to the sarcolemma and was particularly abundant in muscle spindles.

PIN is a highly conserved protein, with a high homology to a light-chain component in dynein of Chlamydomonas (115) and Drosophila (57), which suggests that its primary role may be to serve as a dynein component rather than as an inhibitor of nNOS. Indeed, a recent study (165) failed to demonstrate inhibition of nNOS by PIN. Although PIN bound to nNOS, it did not dissociate nNOS dimers into monomers, in contrast to the findings of Jaffrey and Snyder (107). The reasons for these discrepancies are not clear. Some additional datails on NOS-interacting proteins can be found in section IB.

	VI. RESTING POTENTIAL

Top Previous Next References

nNOS is enriched at neuromuscular junctions where it apparently colocalizes with the NMDA receptor (88). Functional coupling between these proteins has been demonstrated in the central nervous system (113, 169). Such coupling was thus explored in muscles. It was found that denervated muscle fibers display downregulation of nNOS and that enzyme level was restored upon reinnervation (163, 193) (see also sect. IVA2). The earliest change in the muscle postdenervation is depolarization of the membrane potential, which depresses muscle excitability and thus contractility. Urazaev et al. (202) reported that the endogenous NMDA agonist glutamate prevented postdenervation depolarization of rat diaphragm muscles. Both NMDA receptor antagonists and NOS inhibitors blocked the effect of glutamate. Conversely, treatment of the diaphragm with sodium nitroprusside (SNP) attenuated the early postdenervation depolarization (201). Taken together, these data indicate that NOS activity stimulated through NMDA-mediated influx of calcium regulates the resting potential in rat diaphragm. Activity-dependent synaptic suppression at developing neuromuscular synapses, both in culture and after denervation, is also mediated by postsynaptic-derived NO (163, 209). One mechanism may involve the inhibition of neurotransmitter release as shown in isolated synaptogenes (163) and the intact frog neuromuscular junction (127).

Inhibitors of guanylate cyclase [oxadiazoloquinoxalinone (ODQ) and methylene blue] were found to attenuate the effects of SNP on postdenervation depolarization of diaphragm muscles and blocked NO-mediated synaptic suppression induced by repetitive postsynaptic depolarization (201). It is difficult to interpret these data because both agents can inhibit NOS and/or scavenge NO. That being said, this circumstantial evidence, taken with the colocalization of cGMP-dependent protein kinase and nNOS at the neuromuscular endplate (45), tends to favor a role for cGMP.

	VII. CONTRACTILE FUNCTION

Top Previous Next References

Excitation-contraction coupling is the process by which chemical and/or electrical signals at the cell surface are coupled to the release of calcium from the SR, which in turn promotes actin-myosin interaction and thereby contracts the muscle fiber (12, 41). ATP-dependent pumps (Ca²⁺-ATPases) in the SR then resequester the calcium. When such events follow a single action potential, a twitch (contraction) results. Bursts of repetitive action potentials, on the other hand, sustain calcium release and produce more forceful (tetanic) contractions. Thus effects on the sarcolemma, SR, or myofilaments can modulate the force-frequency relationship that defines excitation-contraction coupling. NO has been shown to modulate excitation-contraction coupling at each of these levels (159). Exactly what the consequences are, however, seem to depend on the muscle preparation, protocol, NO concentration, and source.

Murrant et al. (141) used a whole blood, perfused skeletal muscle preparation to study the effect of NO. Specifically, mongrel dogs were intubated and anesthetized by intravenous administration of phenobarbital sodium. The arterial and venous drainage of the left gastronemius-plantaris muscle was isolated, and the muscle group was freed from surrounding tissue. Isometric twitch (8 V, 4.2-ms duration, at 0.5, 1.5, or 4 Hz) and tetanic (8 V, 0.2 ms at 12 or 40 contractions/min) contractions were induced by stimulation of the sciatic nerve. S-nitroso-N-acetylcysteine (100 µM) (SNAC) was infused intra-arterially. Force production was markedly attenuated by SNAC, but only at the higher frequency twitch and tetanic contractions. In contrast, SNAC had lesser effect on lower frequency stimulation trials. In a similar experimental design, King-VanVlack et al. (116) more directly assessed the endogenous NO effect by infusing the NOS inhibitor N^G-nitro-L-arginine methyl ester (L-NAME) (20 mg/kg) into the gastrocnemius. They found that L-NAME increased force production during slow-twitch contractions. It is clear from these studies that NO (or related molecules) of either endogenous or exogenous origin can attenuate force production of limb muscles in situ. It is not clear, however, if NO derives from the vasculature, innervating nerves, or the muscle itself.

Ambiel and Alves-Do-Prado (8) assessed NO involvement in the neuromuscular control of rat diaphragm. Phrenic nerve and diaphragm muscle were isolated from Wistar rats and placed in chambers containing Krebs buffer. A steady amplitude of muscle contraction was induced by stimulation of the phrenic nerve (0.2 Hz, 0.05 ms), after which agents were added to the organ bath. L-Arginine (4.7-9.5 mM) and sodium nitrite both increased the amplitude of muscular contraction, whereas the NOS inhibitor N^G-nitro-L-arginine (18 mM) had the reverse effect. The increases in twitch amplitude induced by L-arginine were lost, however, when the muscle was first treated with D-tubocurarine, suggesting that the potentiation of twitch contraction resulted from an action at the prejunctional level. Lindgren and Laird (127) used a frog neuromuscular junction to investigate this possibility. They found, in contrast to the above report, that SNP significantly attenuated the amplitude of end-plate potentials (synaptic transmission) when applied at concentrations as low as 10 nM. A further quantal analysis revealed that sodium nitroprusside blocked the release of the acetylcholine neurotransmitter without altering the sensitivity of the muscles to its effect. These results are in keeping with the presynaptic suppression mediated by NO (or related molecules) derived from postsynaptic myocytes that was seen in nerve muscle cocultures prepared from Xenopus embryos (209). It is not obvious whether the discrepancies among studies (in fact diametrically opposite conclusions) can be explained by the differences in muscle type, model, or design or whether NO will activate neuromuscular transmission in some cases and inhibit it in others. Such dual actions on synaptic transmission have been seen in the central nervous system.

Bisnett et al. (26) assessed in vitro diaphragmatic function in rats after inspiratory resistive loading in vitro. Infusion of L-NAME (100 mg/kg iv) did not affect ex vivo contractility or the decline in muscle function resulting from occlusion of the inspiratory limb. In contrast, both L-NAME (5 mg/kg iv) and SNP infusions caused a slight decrease in transdiaphragmatic pressure-frequency curves and muscle endurance in pigs (7). The authors attributed the effects of L-NAME to a compromise in diaphragmatic blood flow and those of nitroprusside to direct effects on muscle function. Taken together, these studies suggest that systemic administration of NO donors or inhibitors have only modest effects on diaphragm function in healthy animals.

Kobzik et al. (118) observed that force generated by soleus, diaphragm, and EDL was inversely correlated with their NOS activities (Fig. 5), and that the differences in the force-frequency relationships (i.e., the intrinsic differences in contractile function) among muscle groups were greatly diminished by application of NOS inhibitors. In other words, NOS activity evidently contributes to the intrinsic differences in force generation by type I and type II fibers. Multiple lines of pharmacological evidence further established that NO and/or related molecules attenuate both twitch and tetanic contractions of rat diaphragm without affecting maximal force production. Specifically, NOS inhibitors [7-nitroindazole, aminoguanidine, and nitro-L-arginine (100 µM to 1 mM)] and the NO scavenger hemoglobin increased force production on the one hand, whereas the NOS substrate L-arginine (1 mM) and the NO donors SNAC and SNP (100 µM each) depressed force production, on the other hand.

View larger version (16K): [in this window] [in a new window]   Fig. 5. NOS inhibitors enhance contractile function of rat diaphragm but not soleus muscle. A: contractile responses of soleus, diaphragm, and extensor digitorum longus (EDL) muscles to increasing stimulation frequency. B: addition of either of two NOS inhibitors, 7-nitroindazole (NIZ, 1 mM) and aminoguanidine (AMG, 0.5 mM), shifts the force-frequency relationship (lower dashed line) up and to the left in nNOS-positive (diaphragm) but not nNOS-deficient (soleus) muscles. [From Kobzik et al. (118).]

The relative selectivity of 7-nitroindazole for nNOS over other isoforms (118), the finding that the majority of NOS activity resides with nNOS in the sarcolemma (118), and the demonstration that contractile function is unchanged in mice deficient in eNOS (97) identifies the nNOS isoform with force-inhibitory activity. Moreover, the upregulation of nNOS in iNOS mutant mice exposed to endotoxin proved that nNOS can inhibit force generation (at least in rodent diaphragm in vitro) (50). Such inhibition of contractility by NO (or related molecules) has been confirmed by Abraham et al. in rat diaphragm (1), el-Dwairi et al. in both neonatal and adult rat diaphragm (62), Ambiel et al. in isolated rat diaphragm (8), Murrant et al. in an isolated mouse soleus preparation (142), and Gath et al. in guinea pig diaphram and gastrocnemius muscle (80). Noteworthy findings in these studies included significant augmentation of force following NOS inhibition in both adult and neonatal diaphragm, but larger augmentation of force in the adult using the isothiourea class of NOS inhibitor (1 mM S-methylisothiolurea) (1), and the ability of endogenous and exogenous NO to accelerate the time-dependent decline in force production (fatigue) over 4 h in both diaphragm and gastrocnemius muscle (80). Moreover, the inhibition of isometric force production (i.e., the shift in the force-frequency relationships) is evident for both twitch and tetanic contractions over the entire stimulation frequency range of 10-60 Hz (e.g., Fig. 6). More specifically, the absolute increases in contractility are greatest at 40 Hz; however, the relative increase compared with force in the absence of a NOS inhibitor is the same from ~10 to 40 Hz and ranges among studies from 20 to 50% change. One caveat with these data is that they were derived mainly from the rat diaphragm. Murrant et al. (142) have reported, for example, that SNP and endothelium-derived relaxing factor (EDRF) improved force maintenance in mouse soleus muscle stimulated at 50 Hz every 30 s (142). A second caveat is that these studies were done in standard organ chamber bioassays, typically 95% O₂, whereas tissue oxygen tension is much lower. Thus the effect of species, muscle type, oxygen tension, and protocol need further clarification.

View larger version (22K): [in this window] [in a new window]   Fig. 6. Diagram of modulation of isometric force development by endogenous reactive oxygen intermediates (ROI) and nitrogen oxides (NO). Control relationship illustrates that isometric forces developed by rat diaphragm increase as stimulation frequency increases. The diagram further shows that forces developed during submaximal stimulation are decreased by ROI depletion and increased by NO depletion. [From Reid (158).]

In contrast to the depression of isometric force production at submaximal stimulation frequencies, NO appears to be essential for optimal myofilament function during active shortening. In this context it should be appreciated that maximal velocity of shortening (MVS) is among the most important functional parameters of the skeletal muscle fiber, reflecting the maximum rate of cross-bridge cycling. Maximal velocity of shortening is independent of calcium concentration, maintained constant during tetanus, and is relatively resistant to physiological effectors (e.g., pH, temperature). The variation in MVS among muscles has been attributed to myosin ATPase activity (132): type 1 < type 2a = type 2x < type 2b, an activity profile strikingly similar to that of NOS (in rodents) (118); that is, MVS and NOS activity of muscle are tightly correlated.

Three studies by two groups establish that skeletal muscle NOS in mice and rats augments MVS. Morrison et al. (137, 138) studied NO effects in rat diaphragm. They examined both loaded (maximal shortening against a load) and unloaded shortening velocity. NOS inhibition decreased maximal shortening velocity and unloaded shortening velocity, whereas coadministration of SNP prevented the fall in these parameters. Marechal and Beckers-Bleukx (132) observed virtually identical effects in EDL muscles isolated from C57/BL1OScSc mice; that is, NOS inhibitors reduced MVS by ~20%. In addition, the NO donor S-nitrosyl-N-acetylpenicillamine (SNAP) (1 µM, i.e., a very low concentration) increased MVS by ~8%. They also observed that NOS inhibitors prolonged the time of the tetanic plateau. The only notable difference between the above studies is that Morrison et al. (137, 138) did not observe direct effects of the NO donor alone. Although differential effects of NO donors are not unusual, the much higher dose used by Morrison (~1 mM) may have obscured a highly selective and specific NO effect.

Studies in mutant mice deficient in the nNOS and eNOS genes have not been performed. Information gleaned from iNOS-deficient animals in which nNOS (and eNOS) are upregulated is covered in sections VA and VIIE.

Galler et al. (76) used skinned muscle fibers from both slow- and fast-twitch rat leg muscles to elucidate the mechanism of NO action. They found that 100 µM SNAP or 1 mM SNP caused a marked decline in both maximal calcium-activated force and maximal shortening velocity. These effects were attributed to inhibition of myosin ATPase, which in turn controls the rate of actin-myosin cross-bridge cycling. Perkins et al. (152) reached similar conclusions using rabbit soleus single fiber preparations. Specifically, they observed that SNP significantly inhibited isometric force, shortening velocity, and actomyosin ATPase activity. This study additionally demonstrated a marked decrease in the calcium sensitivity of SNP-treated fibers, results derived from a detailed analysis of force-calcium relationships in the fibers. Andrade et al. (10) investigated NO effects in single skeletal muscle fibers from a mouse foot muscle. They too observed decreases in myofibrillar calcium sensitivity, but force and maximum shortening velocity were largely unchanged. Increases in intracellular calcium transients during submaximal contraction were found to compensate for decreased myofibrillar calcium sensitivity. Andrade et al. (10) did not use skinned muscle fibers. Decreased access of the NO donor might explain why force was largely unchanged in this case.

These studies with NO donors added to single muscle fibers in high concentrations are unlikely to have a physiological correlate. In particular, changes in myofibrillar calcium sensitivity are attributeable to direct effects on myosin, which is present in muscle at millimolar concentrations. Notwithstanding, a potential for localized effects of NOS and high NO levels produced under pathological conditions, NO tends to operate at nanomolar to micromolar concentrations. It is thus unlikely that NO production on millisecond time scales characteristic of excitation-contraction coupling can directly affect myosin function.

D.  Mechanism

1.  cGMP

Effects of cGMP on the various mechanical parameters of skeletal muscle are generally modest. The one exception to this rule may be the case of shortening velocity. Analogs of cGMP increased MVS by 17% and intracellular increases in cGMP probably account for most of the NO effect (132); that is, phosphodiesterase inhibitors that raise intracellular levels of cGMP increased MVS by exactly the same amount as did an NO donor (8%), whereas guanylate cyclase inhibitors that lower intracellular levels of cGMP reduced MVS by the same amount as did inhibitors of NOS (~20%). cGMP also modestly attenuates submaximal force production in muscle. Specifically, Kobzik et al. (118) and Abraham et al. (1) found that cGMP accounted for at most 50% of the NO effect. This is a best-case scenerio because the agents used to manipulate levels of cGMP are notoriously nonspecific, and direct measurements of cGMP were in fact not performed. Instead, cGMP effects were inferred from use of guanylate cyclase inhibitors (e.g., methylene blue, LY83583) that may also inhibit NOS or scavenge NO and from phosphodiesterase inhibitors that can also elevate cAMP. Thus the involvement of cGMP rests in significant part with the finding that cell-permeable analogs of cGMP can slightly decrease submaximal tetanic force production.

Guanylate cyclase activity in skeletal muscle is lower than in most other tissues. Classical studies by Murad and colleagues (11, 112) measured only 6 pmol/mg protein in the resting state and 15-fold increases upon exposure to NO. Only one study has attempted to correlate cGMP formation with NOS activity. Lau et al. (124) found that cGMP increased in contracting EDL muscles from a resting level of 2 fmol/mg wet wt tissue (i.e., 1,000-fold lower than Murad et al.) to ~5 fmol with electrical stimulation. Although NOS inhibitors did not affect the resting level of cGMP, they blocked the increase due to electrical stimulation. (By the same token, mdx muscles, which evince a selective loss of nNOS activity from the dystrophin complex, were similarly resistant to electrical stimulation.) Moreover, addition of a NO donor increased cGMP content to 4 fmol. These changes in cGMP are extremely small and although difficult to put into physiological context, quite intriguing. In heart muscle, for example, cGMP levels have been reported to change on the millisecond time scale that is characteristic of excitation-contraction coupling (154). No such measurement or correlation with force production has been made in skeletal muscle.

Kobzik et al. (118) originally proposed that NO either directly modulates the activity of the RyR/calcium release channel or scavenges reactive oxygen species that do so, to rationalize a cGMP-independent inhibition of contractile function. This viewpoint was later corroborated by the work of Meszaros et al. (135), who found that calcium release from rabbit fast-twitch muscle homogenates was attenuated by NO or related species derived from L-arginine (the effect was blocked by inhibitors of NOS). In addition, they found that calcium release from isolated skeletal muscle SR (monitored by absorbance changes in arsenazo III) as well as open probability of single channels incorporated in planar lipid bilayers was reduced by NO donors. Meszaros's group subsequently demonstrated the same effects with cardiac SR (223), which contains nNOS (217). It is not clear, however, if the source of NO identified in SR preparations from rabbit fast-twitch skeletal muscle is from within the SR itself; that is, NOS has not been localized to skeletal muscle SR and sarcolemmal NOS may contaminate the SR preparations.

Although the results by Meszaros and co-workers (135, 223) provide a satisfying explanation for the inhibitory action of NOS in intact muscle fibers, the effect of NO on RyR function is not straightforward. Stoyanovski et al. (188) found that NO donors activate calcium release from SR and increase open-channel probability of skeletal RyR incorporated in the bilayers. Xu et al. (218) saw much the same effect studying the cardiac RyR. Perhaps these results are partly reconciled by the work of Aghdasi et al. (5) who found that NO can prevent activation of the RyR by oxidants, and by the work of Eu et al. (65), showing that the NO effect on RyR activity is critically dependent on redox state and calmodulin. A simplistic view of the data is that different molecular modifications (nitrosation vs. oxidation) and/or interactions with different classes of thiol explain the discordant effects. It is more likely, however, that the various investigators are working under very different "redox" conditions and/or with different complements of RyR-associated regulatory proteins. In particular, redox state (oxygen tension, receptor thiol content, buffer, and pH) has not been strictly controlled for in any study, nor has anyone come close to recapitulating the redox state of skeletal muscle in vivo.

In cardiac muscle, immunolabeling and functional analyses have correlated nNOS activity with inhibition of the SR Ca²⁺-ATPase (217). nNOS was not detected in rodent skeletal muscle SR. Notwithstanding this distinction, NO generated by chronic low-frequency stimulation in rabbit fast-twitch muscle may contribute to activity-dependent inactivation of the ATPase (117). Viner et al. (207) have shown that this loss of activity probably results in part from S-nitrosylation or oxidation of critical thiols. The ATPase is covered in more detail in section III.

Reid et al. (160) made the important discovery that superoxide generated by diaphragm muscle bundles in standard organ chambers bioassays (Krebs-Ringer solution equilibrated with 95% O₂-5% CO_2, pH 7.32, at 37°C) increases both twitch and tetanic forces developed during submaximal stimulation (0-150 Hz). The studies were simple but decisive: addition of superoxide dismutase was shown to decrease isometric force developed by rat diaphragm. Thus superoxide produced by skeletal muscle is obligatory for optimal force production in vitro and more generally promotes excitation-contraction coupling. NO scavenges superoxide three times more rapidly than superoxide dismutase, raising the possibility that it may have been adapted to serve a similar scavenging function. Indeed, NO opposes the effects of superoxide in other systems (55, 139, 160, 181). It is not clear, therefore, if the decrease in force elicited by NOS derives directly from an NO-mediated effect or indirectly from its scavenging of superoxide. It would be interesting to know what effect superoxide dismutase has on excitation-contraction coupling in mice deficient in nNOS or in wild-type muscles treated with NOS inhibitors.

Exposure of C₂C₁₂ skeletal muscle myocytes to endotoxin, IL-1, TNF-, and IFN- or the combination of IFN- and either TNF or IL-1 but not to the individual agents alone was able to induce the iNOS gene and generate NOS activity (215). The stage of myocyte differentiation did not appear to influence the degree to which nitrite accumulated in the media. The authors speculated that cytokine-induced NO production might serve to enhance the supply of oxygen during vigorous exercise through vasodilation (in keeping with elevated levels of cytokines found in the serum of normal volunteers after prolonged exercise). Alternatively, they speculated that NO might be involved in loss of contractile function associated with postexercise muscle inflammation or sepsis. Both ideas have been verified experimentally. Hussain (104) showed that diaphragmatic oxygen uptake was increased and phrenic vascular resistance was decreased during endotoxemia in mechanically ventilated dogs. Infusion of N^G-nitro-L-arginine methyl ester into the phrenic artery reversed the decline in vascular resistance, but not the increase in oxygen uptake. It is difficult, however, to know exactly what contribution NO makes to the endotoxin effect because L-NAME has complex systemic actions and does not inhibit NOS by >80%. Notwithstanding this caveat with the data, several groups have established that endotoxin injection into the peritoneal cavity of rats leads to diaphragmatic contractile dysfunction (29, 61, 177). Force-frequency curves are shifted to the left by L-NAME pretreatment.

Comtois et al. (50) argue that iNOS induction following endotoxin, at least in mice, might actually serve to protect from a loss of function by acting to decrease the expression of nNOS and eNOS, which in fact comprise the majority of NOS activity. They suggest that the inhibition of force production following endotoxin is the result of increased nNOS expression. The effects of endotoxin are therefore not straightforward.

Liu et al. (129) suggest that endotoxin (7.5 mg/kg) may even increase twitch and tetanic force (of mouse diaphrgam) in a NO-dependent manner (129). These effects were attributed to decreases in chloride conductance that depolarize and thereby activate muscle. Perhaps the differences between Liu et al. (who report increases in force, Ref. 129) and Comtois et al. (who show protection from a decrease, Ref. 50) can be explained by the endotoxin concentrations they used, which were four times higher (25 mg/kg) in the animals exhibiting the decline in force. Alternatively, the exact timing of the ex vivo measurement of force may have influenced the result, since the expression of eNOS, nNOS, and iNOS varied dramatically during the 24 h after LPS injection (50). Although NOS activity is increased by endotoxin in some peripheral muscles (61, 105, 198), the functional ramifications are not known. Thus although all data indicate that cytokine induction of NOS can have detrimental effects, it remains unclear whether there are not salutary effects as well.

	VIII. RYANODINE RECEPTOR

Top Previous Next References

RyRs are Ca²⁺ channels that control the levels of intracellular Ca²⁺ in skeletal muscle by releasing Ca²⁺ from an intracellular compartment, the SR. They are regulated in skeletal muscle by voltage-sensing surface membrane L-type Ca²⁺ channels (dihydropyridine receptors or DHPRs) via a direct physical interaction (41). Many endogenous modulators such as Mg²⁺, ATP, or calmodulin affect the channel activity of the skeletal muscle RyR (RyR1) isoform (133). RyRs and DHPRs are also known to be redox responsive. Skeletal muscle (and cardiac RyR) activity has been shown to be modulated by nitrogen-reactive species via a cGMP-independent pathway (for review, see Eu et al., Ref. 66). L-type Ca²⁺ channels in cardiac muscle (39, 68, 100) and carotid body (190) also contain sulfhydryls whose S-nitrosylation and/or oxidation modulates function; however, regulation of the skeletal muscle DHPR by NO or reactive nitrogen species has not been described.

RyRs are potential targets for NO and related compounds because they contain a large number of sulfhydryls; indeed, they contain the most of any protein studied to date. The tetrameric mammalian skeletal RyR has a total of 404 cysteines [100 cysteines per 560-kDa subunit (191) and 1 per FK506 binding protein (108)]. In the tetrameric skeletal RyR, up to ~200 cysteines or ~50 per subunit, respectively, are free, as determined by monobimane reactivity (65). A direct modulation of RyR activity by both skeletal muscle NOS and NO donors has been demonstrated in vesicle-Ca²⁺ flux and single-channel and [³H]ryanodine binding measurements. Both an activation (5, 188, 189) and inhibition (135, 189) have been observed (Fig. 7), which suggests that NO or NO-related species interact with the skeletal RyR in a complex way.

View larger version (30K): [in this window] [in a new window]   Fig. 7. NO donor molecules inactivate (A) or activate (B) skeletal muscle ryanodine receptor ion channels incorporated in planar lipid bilayers. A: inactivation of a single ryanodine receptor ion channel by S-nitro-N-acetylpenicillamine (SNAP) and return of channel activity close to that of the untreated channel by the reducing agent 2-mercaptoethanol (not shown). Single-channel currents (shown as upward deflections) were recorded in the presence of an activating ligand (20 µM Ca2+) before (top panel) and after the successive addition of 0.1 mM SNAP (bottom panel). [From Meszaros et al. (135). Copyright 1996 Elsevier Science, Amsterdam, The Netherlands.] B: activation of a Ca2+-activated single skeletal muscle ryanodine receptor ion channel by SNAP. Single-channel currents (shown as upward deflections) were recorded before (top trace) and after the addition of 0.5 mM SNAP (bottom trace). [From Stoyanovsky et al. (188).] The opposite results in A and B exemplify the poorly defined interaction of NO with the skeletal muscle ryanodine receptor ion channel (see text).

Recent evidence suggests that NO and related molecules may directly affect the cardiac Ca²⁺ release channel via covalent modifications of thiol groups (218). Both a nitrosative and an oxidative modification of RyRs were described, involving up to 12 thiols per subunit and leading to either a reversible or irreversible alteration of RyR ion channel activity, respectively. The activation of the RyR by multiple covalent attachments is perhaps revealing of RyR behavior in situations characterized by nitrosative or oxidative stress, where large amounts of reactive nitrogen and oxygen species are produced. In contrast, the skeletal muscle RyR is activated by S-nitrosylation of a single cysteine residue, but only at low PO₂ (~10 mmHg), i.e., at oxygen concentrations found in tissues, whereas neither activation nor S-nitrosylation of the channel occurs in room air (PO₂ ~150 mmHg, Ref. 65). The molecular basis of O₂ responsiveness is identified with 6-8 of 50 RyR thiols whose redox state is dynamically controlled. Among SR proteins, S-nitrosylation is specific to the RyR, and its effect on the channel is calmodulin dependent (65). Morever, the skeletal RyR, much like its cardiac counterpart (218), is endogenously S-nitrosylated (65). It thus appears that channel cysteine residues subserve coupled O₂ sensor and NO regulatory functions and that S-nitrosylation of the skeletal RyR meets criteria for being a regulatory modification in the strictest sense.

Intense skeletal muscle activity leads to a decline in muscle performance, which is known as fatigue. The extent of skeletal muscle fatigue and recovery from fatigue depend on the intensity and duration of exercise, muscle fiber type composition, fitness, and dietary state (38, 71). It is now generally recognized that the excitation-contraction coupling machinery plays a critical role in the fall of contractile force in fatigued skeletal muscle. However, there are many factors that may affect excitation-contraction coupling in exercised and fatigued muscle. These include a decline in intracellular K⁺, an increase in intracellular Na⁺, accumulation of H⁺ and P_i, and reduction in ATP phosphorylation potential (71). Reactive nitrogen (see sect. IX) and oxygen (159) species are increased in exercising and fatiguing muscle. Oxidants are clearly thought to have deleterious consequences, and at least in one study, NOS inhibition protected against a decline in function over time (80). Such nitrosative or oxidative stress is likely to directly impact RyR1 activity; however, the extent to which RyR1 is S-nitrosylated and/or oxidized in exercised or fatigued skeletal muscle is not known.

	IX. EXERCISE

Top Previous Next References

Balon and Nadler (15) reported that a chronic exercise regimen, running up to 90 min/day for 8 wk, with 1-min sprints every 10 min, increases the expression of both eNOS and nNOS in rats (15). By analogy, Reiser et al. (162) found that chronic stimulation of limb muscles, 3 wk of electrical stimulation of rat tibialis anterior and EDL muscles, induced nNOS expression and roughly doubled NOS activity. Thus, at least in the rat, nNOS expression can be modulated by muscle activity. The question arises as to where in muscle and in which fiber types is NOS expressed and whether the level of expression depends on species and exercise regimen. For example, Fujii et al. (74) measured total NOS activity in homogenates of ventilatory (diaphragm, intercostal, and transversis abdominas) and limb (gastrocnemius) muscles after 3 h of resistive loading, consisting of a moderate or severe inspiratory pressure load (30 and 70% of maximum tracheal pressure, measured by occluding the inspiratory port for 30 s, respectively). They reported that NOS activity declined by ~50% in ventilatory muscles (diaphragm and intercostals) but not in abdominal or limb muscles. Immunoblotting revealed that protein expression (and thus probably subcellular distribution) of nNOS and eNOS had not changed significantly, whereas NOS activity had declined. Thus the effect was evidently posttranslational. Hussain et al. (105) and Frandsen et al. (72) had more difficulty detecting eNOS in human, canine, and rabbit skeletal muscles. It is therefore premature to draw conclusions regarding eNOS expression, yet still tempting to speculate on potential roles for both isoforms in regulating blood flow, sustaining force over time, and as discussed below, matching metabolic needs to muscle activity (see also sect. IVB).

	X. GLUCOSE UPTAKE

Top Previous Next References

Two distinct pathways regulate glucose utilization in skeletal muscle: insulin and contractile activity. Insulin or insulin-like factors increase eNOS activity in endothelial cells and iNOS expression in rat and human myoblasts (109) through a phosphatidylinositol (PI) 3-kinase-dependent pathway. Contractile activity of skeletal muscle also increases NOS activity and the expression of eNOS and nNOS (15, 175, 199). Classic pharmacological studies by Baylon and Nadler (14, 15) in intact skeletal muscles (14, 15) and Baron (17, 18) in both rats and human subjects have established that NOS activity increases glucose transport. Specifically, infusion of N^G-monomethyl-L-arginine (L-NMMA) systemically into rats or humans or its addition to isolated muscle fiber bundles (rat EDL) after contractile activity, decreases total 2-deoxyglucose transport. Current evidence indicates, however, that insulin-mediated NO effects are mainly microvascular, not myocyte, in origin; that is, insulin-mediated GLUT-4 transport of glucose in isolated skeletal muscle is not dependent on NO.

Insulin increases muscle blood flow in rats and humans (18-20, 166, 171, 187). This effect is probably mediated through the PI 3-kinase/AKT kinase pathway, a recently established mechanism of eNOS activation by receptor tyrosine kinases (58, 75). The increase in muscle perfusion (19) is thought to increase the delivery of insulin's major substrate, glucose, to the muscle cell. Indeed, Baron and co-workers (19, 20) suggest that ~30% of insulin's effect on glucose uptake can be accounted for by increases in muscle perfusion. NOS inhibition by L-NAME in the standard hyperinsulinemic-euglycemic clamp model reduces whole body insulin action and glucose disposal by ~15-30% in the conscious rat (166), in keeping with previous studies in isolated limbs of human subjects (187). Although Butler et al. (38a) reported that L-NMMA infusion increases rather than decreases whole body glucose transport in 16 healthy male subjects, the NOS inhibitor also increased rather than decreased calf blood flow and hardly affected blood pressure; that is, L-NMMA inexplicably produced diametrically opposite hemodynamic responses from those seen in numerous previous studies, including those in which it caused insulin resistance (20). Fryburg (73) and Scherrer et al. (171) have also challenged the strength of the relationship between insulin-mediated increases in blood flow and glucose transport, but good reasons have been forwarded to negate their results (166). Thus there is good evidence that insulin-mediated vasodilation is partly NO dependent and that such increases in perfusion contribute to increased skeletal muscle glucose transport (i.e., independent of insulin-mediated increases in the GLUT-4 transporter). EDRF thus amplifies insulin's action to stimulate glucose uptake.

B.  Isolated Muscles

1.  Constitutive NOS

Balon and Nadler (14) have reported that 1 µM L-NMMA decreases basal (resting) transport of glucose in rat EDL muscle by ~20-30% (14). Baron et al. (20), Kapur et al. (111), and Etgen et al. (64), however, found no effect of more realistic concentrations of L-NMMA (0.1-1 mM) in soleus, EDL, or epitrochlear muscle. Balon and Nadler's preparation was different and one cannot a priori exclude concentration-dependent effects of NOS inhibitors, but if there is any effect of NOS on basal glucose uptake in isolated muscle preparations, fiber type notwithstanding, it is probably modest at best. Balon and Nadler (15) showed more convincingly that a NOS inhibitor attenuates glucose uptake in muscle previously exposed in situ to an electrical stimulation protocol. In these studies, the common peroneal nerve was stimulated in vivo before the EDL muscles were isolated and prepared for incubation with a NOS inhibitor. In contrast, Etgen et al. (64) were unable to demonstrate that glucose transport is effected by L-NMMA or 7-nitroindazole (a more nNOS-selective agent) in electrically stimulated pinned epitropiaris muscle. Preparation and protocol differences might explain the conflicting results. Perhaps nerve stimulation in vivo facilitates subsequent glucose uptake by neuronal or eNOS. If so, contraction induced by direct electrical stimulation of the muscle ex vivo would not achieve the same effect. Accordingly, the jury is still out on the possibility that increases in contractility activate an eNOS or nNOS within the myocyte to facilitate glucose uptake. However, there is agreement among studies that NOS inhibitors do not alter the sensitivity of isolated muscle preparations to insulin or effect glucose uptake by insulin-stimulated muscles (15, 64, 111, 166).

The effects of NO donors on glucose uptake by muscle are also open to question. Several groups have reported that nitroprusside, at concentrations of 5-100 mM, increases glucose transport (15, 64) and glucose oxidation (i.e., glucose utilization) (222) in isolated muscle preparations. However, adequate controls for nitroprusside at these concentrations were not performed and more specific NO donors were not tested. In contrast, both iNOS induction (22) and two classes of NO donors (111) inhibited insulin-stimulated glucose transport in both isolated soleus and EDL muscles and cultured L6 muscle cells. iNOS also decreased GLUT-4 expression in L6 cells (22). These data suggest that iNOS and exogenous NO can block insulin-mediated glucose uptake by downregulation of its transporter. On the other hand, insulin may also downregulate iNOS through a posttranscriptional effect (23).

C.  Mechanism of Action

1.  cGMP

The mechanism by which endothelial-derived NO dilates blood vessels is partly cGMP dependent (139, 179). Inasmuch as insulin-mediated vasodilation involves NO (or related molecules), perfusion-stimulated glucose uptake is thus likely to involve guanylate cyclase. This is, however, only speculation on our part. Indeed, van Veen and Chang (203) have noted that prostaglandins contribute to NO-mediated insulin-induced vasodilation in the human forearm.

The effect of cGMP on glucose metabolism by skeletal muscle is controversial. Kapur et al. (111) reported that 50-500 µM 8-bromo-cGMP (a cell-permeable analog) had no effect on basal or insulin-stimulated glucose transport in isolated soleus muscles treated for ~1 h before uptake measurements were made (111). In contrast, Young and Leighton (222) found that 1 mM 8-bromo-cGMP significantly increased the rate of soleus muscle glucose oxidation by 50%, and Etgen et al. (64) found that the same cell-permeable cGMP analog accelerated epitrochlearis glucose transport activity sixfold over basal levels. Both groups found that high concentrations of SNP mimicked the 8-bromo-cGMP effect. Young and Leighton (222) have also reported that a selective cGMP phosphodiesterase inhibitor, zaprinast, increases basal glucose utilization in rat soleus muscle preparations and that obese Zacca rats that are insulin resistant show impaired response to both zaprinast and the NO donor SNP (221). However, these studies could not show that the SNP or endogenous NO effect on glucose transport was dependent on cGMP, and SNP did not increase the activity of protein kinase G (222), further weakening the NO link to cGMP. Additionally, studies on guanylate cyclase were done with a notoriously problematic reagent, LY83583 (93), and intracellular cGMP did not increase in muscles stimulated to transport glucose by insulin, by contraction, or by the calcium-releasing reagent N-(6-aminohexyl)-5-chloro-1-naphthalenesulfonamide (W-7) (64). Taken together, these data support the notion that the cGMP pathway can increase glucose utilization, but the physiological context in general, and the link to NOS/NO in particular, is not established.

SNP increased GLUT-4 expression on the cell surface of isolated rat epitrochlearis muscle (64). In contrast, the combination of cytokines and LPS, which are known to induce iNOS, caused a 50% increase in the levels of the GLUT-1 transporter but decreased by 50% the expression of GLUT-4 (without altering the expression of GLUT-3) in L6 myocytes (22). L-NAME treatment prevented the increase in GLUT-1 protein but failed to normalize GLUT-4 levels. It is possible that these effects of high doses of SNP and iNOS are due to secondary metabolic impairments, such as inhibition of mitochondrial respiration.

	XI. BLOOD FLOW

Top Previous Next References

Autoregulation of blood flow refers to the continuous adjustment in blood supply to various organs that is individualized in relation to metabolic demand. NO has been implicated in this microvascular response to muscle contraction (27, 37, 74, 95, 103, 200), hypoxia (212), vascular occlusion (reactive hyperemia) (27), and the reflex sympathetic discharge accompanying both resting and exercising skeletal muscles (54, 196, 197). These studies in cat, rat, dog, and human skeletal muscles have typically addressed the question of "NO involvement" by infusion of NOS inhibitors and therefore have been unable to tease out the neurogenic, vascular, and myogenic components of the response. That being said, it is believed that NO or nitrosothiol derived from microvascular endothelial cells factors importantly in control of skeletal muscle blood flow. This area of research is beyond the scope of our review.

Functional sympatholysis contributes to the hyperemic response of contracting skeletal muscle. Thomas, Victor, and co-workers (196, 197) have made an excellent case for skeletal muscle nNOS mediating the metabolic inhibition of sympathetic vasoconstriction that is inherent to quiescent skeletal muscle. Specifically, it has been shown (9, 195) that ₂- but not ₁-adrenergic vasoconstriction is inhibited during contraction of fast-twitch gastrocnemius muscle that expresses nNOS (90, 111, 118), but not in the slow-twitch muscle that is nNOS deficient (111, 118). Inhibition of NOS with L-NAME or the nNOS-selective inhibitor 7-nitroindazole reversed this attenuation of the vasoconstrictor response to lumbar sympathetic nerve stimulation (54, 197). Hindlimb contraction studies in nNOS mutant mice and mdx mice (196), which are deficient in functional nNOS (35), proved that this isoform is responsible for the inhibition of -adrenergic vasoconstriction. In particular, these studies revealed that contraction of rat hindlimb by intermittent stimulation of the sciatic nerve resulted in increased blood velocity and vascular conductance, whereas mice deficient in nNOS did not exhibit this relief of adrenergic vasoconstriction (Fig. 8). The one caveat with the data is that they do not delineate if the source of NO is muscle nNOS or the lumbar sympathetic cell bodies, postganglionic fibers, and varicosities in the femoral artery that may also produce vasodilation (32, 54). One would like to believe that the impairment in mdx mice identifies muscle as the source of NO; however, this interpretation is confounded by several factors (196): 1) additional metabolic defects in these animals, 2) dramatic differences in the force generated among the various animal groups used in the study, 3) genetic background differences that were not controlled for, and 4) a significant effect of L-NAME on the decline in force over time in both control and nNOS-deficient animals.

View larger version (34K): [in this window] [in a new window]   Fig. 8. Metabolic inhibition of -adrenergic vasoconstriction is defective in contracting mdx and nNOS knockout mice. A: segments of a recording from an experiment in one C57 mouse showing the arterial blood pressure (BP) and femoral blood flow velocity (FBF) responses to intra-arterial injection of norepinephrine (NE) in resting and contracting hindlimbs. B and C: segments of recordings from experiments in an mdx mice (B) and an nNOS knock-out mouse (C) showing the BP and FBF responses to intra-arterial injection of NE in resting and contracting hindlimbs. NE-mediated vasoconstriction was attentuated in A but not in B and C, indicating that metabolic modulation of -adrenergic vasoconstriction is impaired in mdx and nNOS knock-out mice due to a lack of nNOS-generated NO-related activity. A-C, right: summary data showing the NE-mediated decreases in calculated femoral vascular conductance (FVC) in resting and contracting (Contr) hindlimbs. [From Thomas et al. (196). Copyright 1998 National Academy of Sciences, USA.]

	XII. TISSUE RESPIRATION

Top Previous Next References

Oxygen utilization is a complex function of 1) the oxygen carrying capacity of hemoglobin, 2) blood flow, 3) oxygen extraction by the cell, and 4) its consumption, predominantly by mitochondria. NO integrates these functions: it is released from hemoglobin and eNOS in response to hypoxia (183), thereby improving blood flow, and it acts as a break at the level of the mitochondria (31, 49, 83, 114, 119, 130). The source of NO that regulates mitochondrial oxygen consumption in skeletal muscle is not entirely clear, but the data point to concerted actions of eNOS in the microvascular endothelial cells and nNOS and eNOS within the muscle itself.

In 1994, King et al. (114) tested the hypothesis that a reduction in EDRF would compromise tissue oxygen uptake in canine skeletal muscle. They infused L-NAME (20 mg/kg iv) into five dogs and the stereo-specific isomer D-NAME in another group of similar size. Measurements of both whole body and hindlimb oxygen consumption were made and well interpreted in the context of systemic hemodynamics and local perfusion. In contrast to expectation, they reported that hindlimb skeletal muscle oxygen uptake increased from 5.2 ± 4.4 to 7.4 ± 0.6 ml·kg¹·min¹ despite decreases in blood flow. Although systemic oxygen consumption did not change, cardiac output fell dramatically. D-NAME had no effect on either oxygen consumption or blood flow. Thus basal NO release was somehow inhibiting oxygen consumption independently of oxygen delivery. These experiments were not designed to investigate the possible mechanisms underlying the metabolic response. Shen et al. (176) made similar observations. In a follow-up study, they showed that endothelial agonists, bradykinin and carbachol, decreased tissue oxygen consumption in slices of dog triceps brachii. A NO donor reproduced the effect, and nitro-L-arginine blocked it. Interestingly, 8-bromo-cGMP decreased oxygen consumption, and the major site of NO action was suggested to be the mitochondria. NOS inhibitors have since been shown to increase oxygen consumption in healthy human volunteers. In one study, L-NMMA (4-16 µmol/min) was infused into the brachial artery of 18 subjects exercised by intermittent handgrip (82). L-NMMA increased exercise oxygen extraction by ~16%, even though it reduced exercise blood flow. In another, using a similar protocol (63), L-NMMA (4 µmol/min) infused interarterially increased oxygen consumption during static exercise from 47.7 ± 8.1 to 53 ± 16.3 ml·min¹·100 ml¹·10². The arteriovenous oxygen difference also increased from 27 to 31%, suggesting increased oxygen uptake. Thus not only does NO regulate basal oxygen consumption but evidently counters the exercise-induced increase as well.

It is difficult to understand how NO derived from endothelial cells accesses skeletal muscle rich in myoglobin to inhibit respiration. NO would need to escape the hemes of myoglobin and still be capable of targeting the hemes of the respiratory chain cytochromes. It is easier to rationalize a skeletal muscle source of NO. Indeed, the picture emerging in NO biology is one of subcellular localization of NOS isoforms. In particular, Kobzik et al. (119) and Kapur et al. (111) have shown by immunohistochemistry that skeletal muscle fibers express eNOS in a heterogeneous pattern. The immunolabeling of eNOS correlated tightly with histochemical stains for succinate dehydrogenase, indicating a mitochondrial relationship (119). Mitochondria-enriched fractions from these muscles, moreover, exhibited calcium-dependent NOS activity, which functioned to inhibit oxygen consumption. These immunohistochemical studies gained more credence when Bates et al. (21) showed by gold immunolabeling that eNOS was indeed present within mitochondrial preparations of skeletal muscle. One hundred thirty-two out of a total of 184 mitochondria displayed the gold immunodeposit against eNOS over the mitochondrial membranes and matrix space. Much evidence has since proven that NO, whatever its origin, regulates mitochondrial respiration by virtue of reversible interactions with cytochrome-c oxidase (49, 83, 84, 114, 130). Additional interactions at the level of complex 1 have also been identified (48).

The effects of NO on oxygen consumption will obviously depend on blood flow. Inhibition of endothelial function in the rabbit hindlimb has been associated with a reduction in limb oxygen consumption simply by virtue of vasoconstriction (155), thus emphasizing the importance of the model. Case in point, Ward and Hussain (213) showed that under conditions where oxygen delivery is limiting, infusion of a NOS inhibitor into the phrenic artery had no effect on oxygen extraction by contracting diaphragm.

	XIII. INJURY

Top Previous Next References

NO and its related oxidation products, typically when produced in "high amounts" by iNOS, are implicated in the pathophysiology of many inflammatory conditions and degenerative diseases. Their involvement in various forms of muscle injury and fatigue, however, has not been well studied. That said, a few reports deserve mention. Stangel et al. (186) have shown that high doses of NO donors added to rat skeletal muscle myoblasts induce apoptosis. Perhaps these observations are relevant to the model of experimental muscle crush injury, in which a remarkable induction of iNOS (at 72 h) is seen in the crushed limb (167) or to certain models of ischemia-reperfusion injury in which nNOS is the villain (194). Rubinstein et al. (167) have proposed that delayed induction of iNOS (and for that matter increased nNOS activity) may cause devastating damage to muscle and contribute to rhabdomyolysis. These authors also reported an induction of eNOS, which they associated with increased capillary perfusion and thus identified with the extreme vasodilatation that is known to occur after injury. It is not clear if these changes in eNOS expression are good or bad: increased perfusion may be required to maintain tissue viability on the one hand, whereas it may cause edema and aggravate hypovolemic shock on the other hand. All three isoforms may therefore contribute to various forms of skeletal muscle injury. The functional consequences of excessive reactive nitrogen production by cytokines in general, and as they may relate to muscle exercise and fatigue in particular, are covered in sections VIIE and VIIIB.

Chen et al. (46) have reported that infusions of the NO donor SNAC protect the contractile function of ischemic rat skeletal muscle from reperfusion injury. Notably, the protocol involved 3 h of ischemia followed by 3 h of reperfusion, and SNAC was protective at 100 nmol/min but not at 1 µmol/min. Transmission electron micrographs of rat EDL revealed that the severe mitochondrial swelling and degeneration caused by ischemia-reperfusion was significantly attenuated in the SNAC-treated animals. In a slightly modified protocol (5 h of ischemia and 90 min of reperfusion), 100 nmol/min SNAC not only preserved muscle contractility, but also maintained microvascular perfusion, which is otherwise compromised by the damage to endothelium (128). It has, in fact, been shown in rabbits that the insult caused by ischemia-reperfusion insult inhibits NO production and increases reactive oxygen generation, thereby leading to arterial vasoconstriction and neutrophil adherence (6). This happens despite evidence that NO is generated by a NOS-independent mechanism in ischemia-reperfused skeletal muscle (126). Moreover, infusion of L-arginine (4 mg·kg¹·min¹) for 1 h had the effect of further increasing NO generation and blocking formation of the injurious superoxide anion (102). These data would suggest that supplementing NO levels can protect muscle from ischemia-perfusion insult. A word of caution may be prudent, however, as nNOS inhibitors actually decreased infarct size in a rabbit model in which the distal aorta was occluded for 4 h and then reperfused for 3 h (194). Thus the timing, source, and form of NO delivery appear to factor importantly in the outcome. Ultimately, preservation of microcirculatory function has to be a critical determinant of whether the injured muscle will heal. One therapeutic approach, therefore, may be to supplement eNOS activity by L-arginine infusion and to concomitantly inhibit nNOS.

	XIV. MYOBLAST DIFFERENTIATION

Top Previous Next References

Differentiation of embryonic muscle cells involves fusion of mononucleated myoblasts into multinuclear myotubes. These myoblasts have NOS activity as measured by their ability to convert arginine to citrulline (125) (see also sect. IVA2). NOS activity increases dramatically in cells competent for fusion (cultured for 50-55 h) but not in rapidly proliferating myoblasts (e.g., cultured for 24-36 h) or in myotubes (cultured for more than 72 h). In other words, NOS activity in cultured myoblasts is critically dependent on time of differentiation. A similar temporal profile was seen in extracts of breast muscle tissues prepared from variously aged chick embryos: NOS activity increased acutely between days 10 and 12 in cells becoming competent to fuse. Levels of intracellular cGMP correlated well with NOS activity.

Lee et al. (125) also report that NOS inhibitors delayed the fusion of myoblasts, whereas the addition of SNP accelerated fusion. These effects, however, were very modest. Specifically, the mean time to fusion in the experiments described was ~54 h in control cells, 52 h in NO-treated cells, and 57 h in cells treated with NOS inhibitors. Moreover, under these conditions, only a minor fraction of cells was undergoing fusion events. The evidence for cGMP involvement is also rather weak, resting with effects of inhibitors (methylene blue and LY83583), which can block NO effects by other mechanisms (93).

The commitment of skeletal muscle cells to form multinucleated myotubes is triggered potently by insulin-like growth factors. This process is known to involve NFB. Kaliman et al. (109) showed that the insulin-like growth factor II (IGF-II) myogenic signaling cascade involving NFB activation (PI 3-kinase-mediated IB downregulation) leads to iNOS upregulation in rat L6E9 myoblasts (2- to 3-fold at differentiation day 1 compared with myoblasts maintained in culture media alone). Subconfluent myoblasts incubated for 4 days with IGF-II formed large multinucleated myotubes (66% of nuclei were in myotubes compared with only 9% in media alone) (109). Remarkably, both the NFB inhibitor pyrrolidinedithiocarbamic acid (PDTC) and the NOS inhibitor L-NAME blocked IGF-II myotube formation capacity entirely; that is, no myotube formation was seen in any of 10 independent experiments.

Remarkable too was the effect of 5 or 10 mM L-arginine added to L6E9 myoblasts, which had been grown to confluence in 10% fetal bovine serum, and then allowed to differentiate (109). After 4 days, 46% of 2,099 nuclei were in myotubes, a degree of differentiation that was comparable to that induced by IGF-II (but independent of it). L-Arginine also led to expression of muscle-specific protein markers including GLUT-4, caveolin-3, and myosin heavy chain. Evidently, these effects on biochemical markers of muscle differentiation were NOS dependent inasmuch as a NOS inhibitor blocked them (differentiation itself was not studied). The authors attributed the effects of L-arginine to increased constitutive NOS activity, but in fact did not measure directly the expression of NOS over the 4 days in culture. Kaliman et al. (109) also mentioned that they observed the same L-arginine-dependent effects in myoblast primary cultures derived from human skeletal muscle biopsies, in particular, L-arginine-, but not L-lysine (control)-induced multinucleated myotube formation. Interestingly, L-arginine has long been known to exhibit insulinomimetic effects. These data at least raise the possibility that NOS activation is the reason why.

	XV. MUSCULAR DYSTROPHIES

Top Previous Next References

ROS have been implicated in the pathogenesis of Duchenne muscular dystrophy (13). In principle, NO-related species could augment such pathological processes by reacting with (or together with) superoxide (O₂) or hydrogen peroxide to form highly reactive peroxynitrite (OONO)-related or other species, which, under the right circumstances, cause oxidation and cellular injury. nNOS may even generate peroxynitrite or related reactive species in high amounts, particularly if uncoupled from certain substrates (see sect. IC), whose access to NOS is tightly regulated. Accordingly, a redistribution of nNOS from the sarcolemma to the cytosol (despite a decreased total content, see sect. IVA3), has been hypothesized to contribute to fiber degeneration in dystrophic muscle.

Disatnik et al. (59) assessed the role of a nitrosative or oxidative stress in mdx mice before the onset of muscle cell death. An increased expression of genes encoding antioxidant enzymes suggested an early increase in the production of ROS; however, no evidence for an increased NO-mediated injury preceding the onset of muscle cell death was detected. Specifically, immunohistochemical and quantitative analysis indicated no differences in nitrotyrosine content (a marker of reactive nitrogen/oxygen interaction) or location in control and mdx mice. The effects of a redistribution and decreased level of nNOS in mdx mice were also assessed in comparative studies between muscles that are spared (extraocular) and not spared (tibialis anterior) from dystrophic pathogenesis (214). No differences in nNOS content and location relative to control mice could be detected, which again suggested that other mechanisms are responsible for the dystrophic pathology in mdx mice. A more direct test of the role of NO in dystrophic muscle involved the generation of mice devoid of both dystrophin and nNOS (44, 53). Deletion of nNOS from dystrophic-deficient mice did not cause detectable changes in muscle morphology and membrane integrity, suggesting that aberrations in nNOS location alone do not cause muscular dystrophy in mdx mice (44, 53). The effects of a genetic deletion of nNOS have been also studied in skeletal muscle of normal mice. Chao et al. (44) did not detect any noticeable changes in muscle morphology of nNOS-deficient mice.

Using dystrophin-deficient and nNOS knockout mice, Thomas et al. (197) found that the metabolic feedback modulation of -adrenergic vasoconstriction was impaired in the contracting skeletal muscle of nNOS-deficient muscle. In EDL muscle of mdx mice, muscle contractions failed to increase cGMP content, and vascular dilation was attentuated (124). Accordingly, working muscle appears to generate NO that causes vasodilation and thereby improves its own blood supply. In other words, contracting muscle uses NO produced by plasmalemmal nNOS to autoregulate oxygen delivery (see sect. IX for details). The above results also emphasize the well-known fact that function-regulating action by NO (or related molecules) does not require a close proximity of nNOS and target proteins.

Dystrophin gene replacement leads to the correct localization of nNOS in mdx mice, clearly pointing out the importance of dystrophin in targeting nNOS to the sarcolemma (43, 56). Analogous studies with a Becker-like human mini-dystrophin gene were less straightforward. Becker muscular dystrophy, like Duchenne muscular dystrophy, is an X-linked disease caused by mutations in the dystrophin gene. However, it is a less aggressive phenotype because the mutations lead to the formation of truncated proteins that accumulate at the sarcolemma of muscle fibers. nNOS was absent or present at reduced levels at the sarcolemma of human Becker patients (43, 87), and it did not bind to ₁-syntrophin (that remained) at neuromuscular junctions of mdx mice (43). Evaluation of four mdx trangenic mice expressing either full-length dystrophin or truncated dystrophins, lacking either the 330 nucleotides of exon 71-74 near the COOH terminus, or the whole complex except the COOH-terminal 71 kDa or exons 17-48, restored extrajunctional and junctional ₁-syntrophin staining. In contrast, sarcolemmal nNOS staining was only restored by the full-length dystrophin and the truncated dystrophin lacking exons 71-74 (43). These observations suggest that the proper assembly of the ₁-syntrophin-dystrophin complex is required for nNOS targeting. In agreement, injection of plasmid cDNAs encoding full-length dystrophin into tibialis anterior and diaphragm muscles of mdx mice resulted in the reappearance of nNOS at the sarcolemma in about half of the dystrophin-positive fibers (56). However, data by Decrouy et al. (56) have somewhat confused matters. They found that tibialis muscle fibers injected with the mini-dystrophin gene lacking exon 17-48 also displayed sarcolemmal nNOS immunoreactivity. The reasons for the conflicting results are unclear but might have been due to the use of transgenic mice in one study (43) and use of dystrophin cDNA injections in the other (56).