Interaction of Neuronal Nitric-oxide Synthase with Caveolin-3 in Skeletal Muscle

Neuronal nitric-oxide synthase (nNOS) has been shown previously to interact with α1-syntrophin in the dystrophin complex of skeletal muscle. In the present study, we have examined whether nNOS also interacts with caveolin-3 in skeletal muscle. nNOS and caveolin-3 are coimmunoprecipitated from rat skeletal muscle homogenates by antibodies directed against either of the two proteins. Synthetic peptides corresponding to the membrane-proximal caveolin-3 residues 65–84 and 109–130 and homologous caveolin-1 residues 82–101 and 135–156 potently inhibit the catalytic activity of purified, recombinant nNOS. Purified nNOS also binds to a glutathione S-transferase-caveolin-1 fusion protein inin vitro binding assays. In vitro binding is completely abolished by preincubation of nNOS with either of the two caveolin-3 inhibitory peptides. Interactions between nNOS and caveolin-3, therefore, appear to be direct and to involve two distinct caveolin scaffolding/inhibitory domains. Other caveolin-interacting enzymes, including endothelial nitric-oxide synthase and the c-Src tyrosine kinase, are also potently inhibited by each of the four caveolin peptides. Inhibitory interactions mediated by two different caveolin domains may thus be a general feature of enzyme docking to caveolin proteins in plasmalemmal caveolae.

first localized in neurons (1). The nNOS isoform, however, is also highly expressed in skeletal muscle (2,3), where it appears to be involved in modulating contractile force (3). Proteinprotein interactions of nNOS function both to regulate enzyme activity and to target the protein to specific subcellular locations. nNOS is activated by interaction with Ca 2ϩ /calmodulin (CaM) (4) and is inhibited by interaction with a 10-kDa protein designated PIN (5). The nNOS protein is targeted to neuronal postsynaptic densities by interaction with the postsynaptic density proteins PSD-95 and PSD-93 (6,7) and to the sarcolemma of skeletal muscle cells by interaction with ␣ 1 -syntrophin in the membrane cytoskeleton dystrophin complex (6,8).
Interactions with PIN, PSD-95, PSD-93, and ␣ 1 -syntrophin are mediated by a 230-amino acid N-terminal extension of the enzyme that contains a PDZ domain. This domain is not found in either eNOS or iNOS (9). eNOS activity and subcellular localization is also modulated by protein-protein interactions. eNOS is localized in plasmalemmal caveolae of endothelial cells and cardiac myocytes through association with the caveolae integral-membrane structural proteins, caveolins-1 and -3, respectively (10,11). We have recently shown that caveolin-1 interacts directly with eNOS and inhibits the catalytic activity of the enzyme (12). eNOS and nNOS are structurally similar, sharing Ͼ60% amino acid sequence identity outside of the nNOS N-terminal extension (13). Caveolins-1 and -3 are also Ͼ60% identical (14). Interestingly, it has been shown recently that caveolin-3 in skeletal muscle is a component of the dystrophin complex (15). Therefore, in the present study, we have examined whether skeletal muscle nNOS, in addition to interacting with ␣ 1 -syntrophin, also interacts with caveolin-3.

EXPERIMENTAL PROCEDURES
Materials-Monoclonal anti-nNOS antibody (clone 16), polyclonal anti-caveolin antibody, monoclonal anti-caveolin-1 antibody (clone 2297), and polyclonal anti-caveolin-3 antibody were obtained from Transduction Laboratories (Lexington, KY). L-[ 14 C]Arginine and [␥-32 P]ATP were purchased from NEN Life Science Products. ECL detection reagents came from Amersham Corp. Synthetic peptides were obtained from Research Genetics, Inc. (Huntsville, AL) or from the Medical College of Georgia Biochemistry Core Facility and were Ͼ95% pure as assessed by high performance liquid chromatography. Purified recombinant human c-Src kinase (1,000,000 units/mg), purified rat forebrain Ca 2ϩ /CaM-dependent protein kinase II (0.6 mol of phosphate incorporated into Auto Camtide II substrate peptide/min/mg) and Ca 2ϩ /CaM-dependent protein kinase II assay kit were purchased from Upstate Biotechnology (Lake Placid, NY). Protein A/Protein G-agarose was obtained from Life Technologies Inc. Wood 46 was purchased from Zymed Laboratories Inc. AG 50W-X8 cation exchange resin, protein assay kit, and peroxidase-conjugated secondary antibodies came from Bio-Rad. Bovine CaM, NADPH, FAD, and FMN were obtained from Sigma, and tetrahydrobiopterin was obtained from Research Biochemicals International. CaM-Sepharose 4B was purchased from Pharmacia Biotech Inc.
Tissue Extraction, Immunoprecipitation, and Immunoblotting-Rat quadriceps skeletal muscle was minced into small pieces with a razor blade and homogenized in 10 volumes of ice-cold buffer containing 20 mM Tris-HCl, pH 7.4, 2.5 mM EDTA, 50 mM NaF, 10 mM Na 4 P 2 O 7 , 1% Triton X-100, 1% phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml pepstatin A, and 5 g/ml aprotinin. Homogenates were centrifuged at 10,000 ϫ g for 20 min to remove insoluble material. Immunoprecipitation with anti-nNOS antibody (0.5 g) and anti-caveolin antibody (0.5 g) was carried out as described previously (10,11). Proteins in anti-nNOS immunoprecipitates were run on 12.5% gels, and proteins in anti-caveolin-3 immunoprecipitates were run on 7.5% gels. Proteins were transferred to nitrocellulose membranes by electroblotting, and membranes were probed with anti-caveolin antibody or anti-nNOS antibody followed by a peroxidase-conjugated secondary antibody. Bound antibody was visualized by the ECL chemiluminescent detection system and autoradiography.
Determination of the Effects of Synthetic Peptides on nNOS, eNOS, Ca 2ϩ /CaM-dependent Protein Kinase II, and c-Src Catalytic Activities-Synthetic peptides corresponding to bovine caveolin-1 residues 61-81 (DDVVKIDFEDVIAEPEGTHSF), 82-101 (DGIWKASFT-TFTVTKYWFYR), 135-156 (KSFLIEIQCISRVYSIYVHTFC), and 157-178 (DPLFEAIGKIFSNIRINTQKEI) (12) and rat caveolin-3 residues 65-84 (DGVWRVSYTTFTVSKYWCYR) and 109 -130 (KSYLIEIQ-CISHIYSLCIRTFC) (14) were tested for their effects on nNOS, eNOS, Ca 2ϩ /CaM-dependent protein kinase II, and c-Src catalytic activities. Bovine eNOS and human nNOS (100 pmol each) expressed and purified to Ͼ90% homogeneity from a baculovirus system (16 -18) were mixed on ice with various concentrations of peptides. NOS activity was then determined as described previously (16). c-Src activity was measured in an autophosphorylation assay. Three units of purified, human recombinant c-Src was incubated with various concentrations of caveolin peptides in 50 l of kinase reaction buffer (50 mM Tris-HCl, pH 7.4, 10 mM MgCl 2 , and 2 mM MnCl 2 ). Reactions were initiated by the addition of 10 Ci of [␥-32 P]ATP and allowed to proceed for 15 min at 25°C. Reactions were terminated by the addition of 50 l of 2 ϫ SDS sample buffer and boiling of samples for 5 min. Samples were analyzed by SDS-polyacrylamide gel electrophoresis (10% gels) followed by autoradiography. Ca 2ϩ /CaM-dependent protein kinase II activity was assayed using a CaM-dependent protein kinase II assay kit to phosphorylate Auto CaM II with [␥-32 P]ATP. 100 ng of purified enzyme was used to phosphorylate the Auto CaM II substrate according to the manufacturer's instructions in the presence of the same final concentration of Ca 2ϩ /CaM as used in the nNOS activity assays. To determine the stoichiometry of Ca 2ϩ /CaM-dependent protein kinase II autophosphorylation under these assay conditions, the 32 P-labeled kinase protein was separated from [ 32 P]ATP on SDS-polyacrylamide gels. The number of moles of phosphate incorporated into the enzyme in vitro was then calculated based on the based on the amount of 32 P present in the excised protein band and the known specific radioactivity of the [␥-32 P]ATP used, as described previously (19).
CaM-Sepharose Chromatography-Purified human nNOS (1 g) was incubated with or without synthetic peptides (100 M) in 50 mM Tris-HCl, pH 7.5 buffer for 5 min at 37°C and then subjected to CaM-Sepharose chromatography as described previously (12,17). Bound nNOS was eluted from the column with 2 mM EGTA and quantitated by immunoblotting with anti-nNOS antibody as described previously (12,20).
Interaction of nNOS with a GST-Caveolin-1 Fusion Protein-In vitro binding assays of baculovirus-expressed, purified human nNOS and a full-length bovine GST-caveolin-1 fusion protein were carried out under the same conditions described previously for eNOS binding to GSTcaveolin-1 fusion proteins (12).

RESULTS AND DISCUSSION
Rat quadriceps muscle homogenates were immunoprecipitated with either an anti-nNOS or an anti-caveolin-3 antibody. Anti-nNOS immunoprecipitates were subjected to immunoblotting with anti-caveolin-3. Anti-caveolin-3 immunoprecipitates were subjected to immunoblotting with anti-nNOS. As shown in Fig. 1, caveolin-3 (20 kDa) was specifically coimmunoprecipitated by the anti-nNOS antibody and nNOS (160 kDa) was specifically coimmunoprecipitated by the anti-caveolin-3 antibody. A different anti-caveolin antibody that recognizes both caveolins-1 and -3 also coprecipitated nNOS. However, a caveolin-1-specific antibody (clone 2297) did not react with the caveolin protein precipitated by the anti-caveolin-3 antibody.
Caveolin-3 is structurally and functionally similar to caveolin-1, a 178-residue protein containing three distinct domains: a 101-residue N-terminal cytoplasmic domain, a 44-residue C-terminal cytoplasmic domain, and a 33-residue membranespanning domain (21)(22)(23). We have shown previously that interaction of eNOS with either cytoplasmic domain of caveolin-1 significantly inhibits eNOS catalytic activity. The eNOS-inhibitory region of the caveolin-1 N-terminal cytoplasmic domain is identified as a membrane-proximal region containing residues 82-101 (12). This 82-101 region has also been shown to inhibit the catalytic activities of G␣ subunits, Ha-Ras, and Src family tyrosine kinases (24 -26). It has been proposed, therefore, that this 20-amino acid sequence of caveolin-1 represents the caveolin-1 scaffolding domain. Other regions of caveolin-1 have not been suspected of participating in caveolin-1 protein-protein interactions (24 -26). In the present study, however, we have identified a second membrane-proximal caveolin scaffolding/ inhibitory domain that exists in the C-terminal cytoplasmic tails of both caveolins-1 and -3. The C-terminal cytoplasmic tail of caveolin-1 is comprised of residues 135-178. We therefore prepared synthetic peptides corresponding to bovine caveolin-1 residues 135-156 and 157-178 (12). Peptides were tested for their abilities to inhibit recombinant bovine eNOS expressed and purified from a baculovirus system (16 -18). Enzyme activity was determined by arginine-to-citrulline conversion assay in the presence of excess cofactors and in the absence or the presence of varying concentrations of the peptides. The 135-156 caveolin-1 peptide inhibited eNOS with a potency similar to that previously reported for inhibition of eNOS by the 82-101 peptide (IC 50 ϭ 1.0 M). In contrast, the 157-178 peptide had no effect on activity. The effects of caveolin-1 peptides on the activity of human recombinant nNOS expressed and purified from a baculovirus system (16,18) were also determined. The 82-101 peptide inhibited activity of purified nNOS ( Fig.  2A) with a potency (IC 50 ϭ 1.2 M) similar to that reported previously for inhibition of eNOS (12), suggesting that nNOS, like eNOS, can interact directly with caveolin proteins. The caveolin-1 peptide corresponding to residues 135-156 was also effective in inhibiting nNOS (IC 50 ϭ 0.9 M) (Fig. 2B). In contrast, peptides corresponding to caveolin-1 residues 61-81 and 157-178 had no effect on nNOS activity.
Bovine caveolin-1 residues 82-101 and 135-156 (12) are equivalent to rat caveolin-3 residues 65-84 and 109 -130, respectively (14). The primary structures of caveolin-1 residues 82-101 and corresponding caveolin-3 residues 65-84 contain conservative substitutions at six positions. Caveolin-1 residues 135-156 and corresponding caveolin-3 residues 109 -130 contain conservative substitutions at seven positions. To determine whether these amino acid substitutions significantly alter the inhibitory potency of the two caveolin scaffolding domains, we prepared synthetic peptides corresponding to rat caveolin-3 residues 65-84 and 109 -130 and determined the effects of the peptides on nNOS activity. As shown in Fig. 2 (C and D), low micromolar concentrations of both caveolin-3 peptides inhibited nNOS, indicating that nNOS can interact in vitro with both inhibitory domains of both caveolin-3 and caveolin-1. The 109 -130 caveolin-3 peptide was 8-fold more potent than the 65-84 peptide in inhibiting nNOS (IC 50 ϭ 0.5 M versus IC 50 ϭ 4.0 M), suggesting that the second caveolin-3 scaffolding do- FIG. 1. Coimmunoprecipitation of nNOS and caveolin-3. Rat quadriceps muscle homogenates were immunoprecipitated (IP) with either anti-nNOS or anti-caveolin-3 antibody. Precipitated proteins were separated by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes by electroblotting. Anti-nNOS immunopreciptitates were probed with anti-caveolin-3 antibody, and anticaveolin-3 immunoprecipitates were probed with anti-nNOS antibody. The results shown are representative of three separate experiments. main identified in this study may contribute significantly to nNOS inhibition when nNOS and caveolin-3 interact in skeletal muscle cells. To determine whether inhibition of nNOS activity was due to binding of the caveolin-3 peptides to nNOS or to Ca 2ϩ /CaM, we tested the effects of the peptides on the activity of another Ca 2ϩ /CaM target enzyme, the Ca 2ϩ /CaMdependent protein kinase II. Concentrations of caveolin-3 peptides (10 M) that inhibited nNOS by 100% only partially inhibited this Ca 2ϩ /CaM-dependent enzyme. Inhibition of 56 Ϯ 2 and 50 Ϯ 3% were determined for the 65-84 and 109 -130 peptides, respectively (mean Ϯ S.E., n ϭ 4). Interactions of the peptides with Ca 2ϩ /CaM may thus contribute to the peptide inhibition observed for nNOS. However, because complete inhibition was not observed, this mechanism can only account for about half of the inhibitory effects of the peptides on nNOS. The ϳ50% residual CaM kinase II activity determined in these experiments cannot be explained by partial activation of the enzyme due to autophosphorylation, which yields a CaM-independent enzyme. This can be concluded from experiments in which the stoichiometry of autophosphorylation under identical assay conditions was quantitated and found to be Ͻ0.001 mol phosphate incorporated/mol of enzyme in the presence or the absence of the caveolin peptides (n ϭ 3). Moreover, the peptides could also inhibit Ca 2ϩ /CaM-dependent protein kinase II by direct interaction with the enzyme rather than with Ca 2ϩ /CaM. This interpretation is supported by the results of our previous study in which we reported that Ca 2ϩ /CaM does not bind to a GST-caveolin-1 fusion protein in in vitro binding assays (12).
Dose-response relationships for caveolin-3 peptide inhibition of eNOS were very similar to those shown for nNOS in Fig. 2 27), it is not likely that caveolin proteins will preferentially bind one of the two NOS isoforms and not the other.
To demonstrate unequivocally the capacity of nNOS to interact directly with caveolin proteins, we performed in vitro binding assays with purified human nNOS and a full-length bovine caveolin-1-GST fusion protein (12). The GST-caveolin-1 fusion protein (GST-cav 1-178) and a GST-nonfusion protein were expressed in Escherichia coli and purified by affinity chromatography on glutathione-agarose. The GST-caveolin-1 fusion or GST alone prebound to beads were then used in in vitro binding assays with recombinant nNOS expressed and purified from a baculovirus system (16,18). Purified nNOS was preincubated for 5 min at 37°C with or without the caveolin-3 65-84 and 109 -130 peptides (100 M). As a negative control nNOS was also preincubated with the caveolin-1 61-81 peptide. Following preincubation, nNOS was incubated overnight a 4°C with GST-caveolin-1 or GST prebound to beads. Beads were washed extensively (six consecutive washes), and bound proteins were eluted with reduced glutathione. Eluted proteins were separated on SDS-polyacrylamide gels, transferred to nitrocellulose, and immunoblotted with anti-nNOS antibody. As shown in Fig. 3, nNOS bound specifically to the GST-caveolin-1 fusion protein but not to GST alone, demonstrating that nNOS can interact directly with caveolin proteins. Furthermore, binding was completely blocked by preincubation of nNOS with the caveolin-3 65-84 and 109 -130 peptides. The 61-81 caveolin-1 peptide, however, had no effect on binding.
We have shown previously that the mechanism by which the caveolin-1 82-101 peptide reversibly inhibits eNOS activity is through interference with the interaction of the enzyme with Ca 2ϩ /CaM (12). We now report that the caveolin-1 135-156 peptide inhibits eNOS by the same mechanism. This particular mechanism of inhibition was also observed for nNOS. Purified nNOS was preincubated in the absence or the presence of the caveolin-3 65-84 and 109 -130 peptides (100 M). The enzyme was then subjected to affinity chromatography on CaM-Sepharose. nNOS was allowed to bind to the column in the presence of 2 mM CaCl 2 and was specifically eluted with 2 mM EGTA. The amount of enzyme eluted in each condition was quantitated by immunoblotting with anti-nNOS antibody. As shown in Fig. 4, both of the caveolin-3 peptides completely blocked the binding of nNOS to CaM-Sepharose, demonstrating that both caveolin scaffolding/inhibitory domains inhibit nNOS by a common mechanism. The caveolin-1 82-101 and 135-156 peptides also blocked the binding of nNOS to CaM-Sepharose. Control peptides (caveolin-1 residues 61-81 and 157-178), however, had no effect on binding.
The results of our previous study and of the present study demonstrate that the NOS enzymes interact with both N-and C-terminal cytoplasmic domains of caveolin proteins. We have therefore suggested that the caveolin-NOS interaction may be fundamentally different from that of caveolin with G␣ subunits, Ha-Ras, and Src tyrosine kinases (12). These particular proteins are reported to interact with (and be inhibited by) caveolin-1 exclusively through residues 82-101 in the N-terminal cytoplasmic domain (24 -26). However, in the present study, we have found that c-Src activity is actually more potently inhibited by the caveolin-1 135-156 peptide than by the caveolin-1 82-101 peptide. Auto-activation of c-Src occurs through auto-phosphorylation of tyrosine 416 and has been used previously to measure the effects of the 82-101 peptide in inhibition of c-Src (26). We have used this same approach to determine the effects of the 135-156 peptide on c-Src activity. Purified human recombinant c-Src was preincubated in the absence or the presence of varying concentrations of the 82-101 and 135-156 peptides. Autophosphorylation was then assessed by 32 P labeling and autoradiography. As shown in Fig. 5, both peptides inhibited autophosphorylation, with the 135-156 peptide having a greater inhibitory potency than that of the 82-101 peptide. Complete inhibition required a 3 M concentration of the 82-101 peptide versus a 1 M concentration of the 135-156 peptide. Caveolin-3 65-84 and 109 -130 peptides in the low micromolar range also completely inhibited c-Src.
In summary, the results of this study demonstrate several important but previously unrecognized features of the proteinprotein interactions of both nNOS and the caveolins. Notable among these is the fact that nNOS in the skeletal muscle dystrophin complex interacts not only with ␣ 1 -syntrophin but also with caveolin-3. Interaction of the two proteins appears to be direct and to involve two distinct and physically separated caveolin scaffolding domains. Furthermore, interaction serves to suppress or inhibit nNOS catalytic activity. The capacity to interact with caveolin proteins is thus a general property of the Ca 2ϩ /CaM-dependent NOS enzymes and is not unique to eNOS. Two distinct caveolin scaffolding domains are also in-volved in caveolin inhibition of the c-Src tyrosine kinase. We propose, therefore, that caveolin proteins contain both N-and C-terminal scaffolding/inhibitory domains and suggest that these domains be referred to as scaffolding domains 1 and 2, respectively. Purified, baculovirus-expressed nNOS was preincubated with or without the indicated synthetic peptides and subjected to chromatography on CaM-Sepharose. Bound enzyme was eluted from CaM-Sepharose with EGTA and quantitated by immunoblotting with anti-nNOS antibody. Similar results were obtained in three separate experiments.