Regulation of Neuronal Nitric-oxide Synthase by Calmodulin Kinases*

Phosphorylation of neuronal nitric-oxide synthase (nNOS) by Ca2+/calmodulin (CaM)-dependent protein kinases (CaM kinases) including CaM kinase Iα (CaM-K Iα), CaM kinase IIα (CaM-K IIα), and CaM kinase IV (CaM-K IV), was studied. It was found that purified recombinant nNOS was phosphorylated by CaM-K Iα, CaM-K IIα, and CaM-K IV at Ser847 in vitro. Replacement of Ser847 with Ala (S847A) prevented phosphorylation by CaM kinases. Phosphorylated recombinant wild-type nNOS at Ser847 (≈0.5 mol of phosphate incorporation into nNOS) exhibited a 30% decrease ofV max with little change of both theK m for l-arginine andK act for CaM relative to unphosphorylated enzyme. The activity of mutant S847D was decreased to a level 50–60% as much as the wild-type enzyme. The decreased NOS enzyme activity of phosphorylated nNOS at Ser847 and mutant S847D was partially due to suppression of CaM binding, but not to impairment of dimer formation which is thought to be essential for enzyme activation. Inactive nNOS lacking CaM-binding ability was generated by mutation of Lys732-Lys-Leu to Asp732-Asp-Glu (Watanabe, Y., Hu, Y., and Hidaka, H. (1997) FEBS Lett. 403, 75–78). It was phosphorylated by CaM kinases, as was the wild-type enzyme, indicating that CaM-nNOS binding was not required for the phosphorylation reaction. We developed antibody NP847, which specifically recognize nNOS in its phosphorylated state at Ser847. Using the antibody NP847, we obtained evidence that nNOS is phosphorylated at Ser847 in rat brain. Thus, our results suggest that CaM kinase-induced phosphorylation of nNOS at Ser847 alters the activity control of this enzyme.

Calcium ion acts as a ubiquitous second messenger, especially in neural tissues where many of its physiological responses are mediated by the Ca 2ϩ -binding protein calmodulin (CaM). 1 The actions of CaM are mediated by its association with specific target proteins, some of which are known as CaM-binding proteins, which include kinases such as CaM kinase I (CaM-K I), CaM kinase II (CaM-K II), and CaM kinase IV (CaM-K IV) (1,2). CaM also associates with flavoproteins such as the nitric-oxide synthases (NOSs), which catalyze the formation of nitric oxide (NO) and L-citrulline from L-arginine (3,4). NO is a major cellular signaling molecule and has been implicated in synaptic plasticity in the brain (5,6). The different NOS isoforms are involved in biological processes as diverse as neurotransmission (neuronal NOS (nNOS)), blood pressure homeostasis (endothelial NOS (eNOS)), and cytotoxicity (inducible NOS (iNOS)). These three NOS isoforms are the products of distinct genes located on different chromosomes and show distinct tissue-specific patterns of expression (7). nNOS is found in neurons and contains an NH 2 -terminal PDZ domain that links the enzyme to PSD-95 (8,9), as well as CAPON (10), which controls the subcellular localization and sensitivity to Ca 2ϩ -dependent activation of the enzyme in some neurons. It has been confirmed that not only cyclic AMP-dependent protein kinase (PKA) but also cyclic GMP-dependent protein kinase, protein kinase C (PKC), and CaM kinase can phosphorylate nNOS, although the physiological significance of this phosphorylation remains uncertain, as there is either no detectable effect on enzyme activity or else the effect is controversial (11)(12)(13)(14). More recently, it was determined that Ca 2ϩ / CaM-dependent NOSs (cNOSs) contain a unique polypeptide insertion, identified as amino acids 598 -642 for bovine aortic endothelial cell eNOS and 831-872 for rat brain nNOS, which serves as a control element (15). The insert peptide of nNOS contains three serine residues, which are potential sites for phosphorylation by CaM-K II. Since a synthetic peptide derived from the insertion of eNOS interferes with CaM binding to cNOSs, while a peptide derived from nNOS does not (15), phosphorylation/dephosphorylation of the serine and threonine residues found at amino acids 831-872 of nNOS would seem to influence its activity. Thus, we suspected that one of the three serine residues in this element of nNOS may be the phosphorylation site for CaM-K II.
Here, we report that nNOS is phosphorylated at Ser 847 by CaM kinases, including CaM-K I␣, CaM-K II␣, and CaM-K IV in vitro. The effects of phosphorylation at Ser 847 on activation of the enzyme activity are also documented. Furthermore, we present evidence that CaM kinases phosphorylate nNOS at Ser 847 in vivo. These findings demonstrate a novel regulatory mechanism for nNOS in neural cells via Ser 847 phosphorylation by CaM kinases.

EXPERIMENTAL PROCEDURES
Materials-CaM-K I␣, CaM-K II␣, CaM-K IV, and CaM kinase kinase (CaM-KK) were obtained from a rat brain cDNA library. The cDNA for rat brain nNOS was a generous gift from Dr. Solomon H. Snyder (Johns Hopkins University School of Medicine, Baltimore, MD). pGroESL, containing groEL and groES cDNAs, was from DuPont. pC-W oriϩ was provided by Dr. Michael Waterman (Vanderbilt University, Nashville, TN). L-[ 3 H]Arginine, [␥-32 P]ATP (6,000 Ci/mmol), and ECL Western blotting detection reagents were purchased from Amersham Pharmacia Biotech. Restriction enzymes and DNA-modifying enzymes were obtained from Takara Shuzo. Electrophoresis reagents and Bradford protein dye reagents were products of Bio-Rad. All other materials and reagents were of the highest quality available from commercial suppliers.
cDNA Construction and Mutagenesis-The cDNA for an inactive mutant nNOS was generated by mutation of Lys 732 -Lys-Leu to Asp 732 -Asp-Glu, as described previously (16). pCWnNOS, the plasmid for the expression of nNOS in Escherichia coli, was constructed as described preciously (17). The EspI/XbaI fragment of the cDNA for nNOS was inserted into M13mp18 to produce single-stranded DNA for mutagenesis. Four different oligonucleotides (5Ј-AGGAGGAGACGGCGTTG-AATCGGA-3Ј, 5Ј-AGGAGGAGACGTCGTTGAATCGG-3Ј, 5Ј-TCGGGA-GTCGGCATAGGAGGA-3Ј, and 5Ј-TCCGTCGCCGGCTGACTTTCGG-GA-3Ј) were synthesized to provide mutations of Ser 847 to Ala, Ser 847 to Asp, Ser 852 to Ala, and Ser 858 to Ala (underlined codons), respectively. Mutation was carried out using the Sculputor TM in vitro mutagenesis system (Amersham Pharmacia Biotech). Mutant clones were isolated, and the presence of site-directed mutations was confirmed by DNA sequencing. The mutant cDNAs were ligated into pCW oriϩ to generate pCWnNOS S847A, S847D, S852A, and S858A. Wild-type or mutant pCWnNOSs and pGroESL were co-transformed into the protease-deficient E. coli strain BL21 (Stratagene).
The CaM kinase I␣1-293 fragment was constructed by polymerase chain reaction from rat CaM kinase I␣ with BamHI and EcoRI restriction sites. This fragment was subcloned into the pGEX-2T vector containing glutathione S-transferase (GST) (Amersham Pharmacia Biotech), and the fusion protein (GST/CaM kinase I ␣1-293) was expressed in E. coli (BL21 cells, Stratagene) and purified by affinity chromatography on glutathione-Sepharose (Amersham Pharmacia Biotech) (18).
Purification of Expressed Proteins-Recombinant wild-type and mutants rat nNOSs were expressed in E. coli or Sf9 cells and purified by 2Ј-5Ј-ADP-agarose (Sigma), as described previously (17,19). Recombinant rat CaM-K I␣, CaM-K II␣, and CaM-K IV were expressed in Sf9 cells and purified by CaM-Sepharose chromatography. Recombinant CaM-KK was expressed in E. coli and purified by CaM-Sepharose chromatography. Protein concentrations were determined by the method of Bradford using BSA as the standard (20).
In Vitro Activation of CaM-K I␣ and CaM-K IV and Phosphorylation of nNOS-Recombinant CaM-K I␣ and CaM-K IV (1 M each) were preincubated with recombinant CaM-KK (350 nM) at 30°C for 10 min in 40 mM HEPES (pH 7.0), 10 mM MgCl 2 , 0.4 mM dithiothreitol, 1 mM CaCl 2 , 1 M CaM, 100 M ATP, and 100 g/ml BSA before phosphorylation of nNOS. The standard assay conditions were 40 mM HEPES (pH 7.0), 10 mM MgCl 2 , 0.4 mM dithiothreitol, 1 mM CaCl 2 , 1 M CaM, 100 M [␥-32 P]ATP, 100 g/ml BSA, 53 g/ml nNOS, and 100 nM CaM kinases at 30°C in a final volume of 30 l. The final concentrations of protein kinases were 6.8, 4, and 6.6 g/ml for activated CaM-K I␣, CaM-K II␣, and activated CaM-K IV, respectively. Reactions were stopped by the addition of sample buffer, followed by electrophoresis on 6 -12.5% sodium dodecyl sulfate-polyacrylamide gel (21). Gels were stained with Coomassie Brilliant Blue, dried, and subjected to autoradiography. Quantitation of 32 P incorporation into nNOS was achieved by cutting out the appropriate gel pieces and determining their radioactivity by liquid scintillation counting.
Assay of nNOS Activity-NOS activity was determined by measuring the conversion of L-[ 3 H]arginine to L-[ 3 H]citrulline as described previously (17,19). Briefly, incubation was performed for 2 min at 25°C (E. Reductase Activities of nNOS-nNOS-catalyzed reduction of cyto-chrome c and DCPIP was determined photometrically in 0.2 ml of a 50 mM triethanolamine/HCl buffer, pH 7.0, containing 0.1 mM NADPH, 4 M FAD, 4 M FMN, and 30 nM nNOS either in the presence (cytochrome c) or absence (DCPIP) of Ca 2ϩ /CaM. The concentrations of the electron acceptors in the incubations were 100 M, and the initial rates of their reduction were calculated using the following extinction coefficients, given as mM Ϫ1 ϫ cm Ϫ1 : cytochrome c, 21 at 550 nm; and DCPIP, 20.6 at 600 nm, respectively. CaM Overlays-Wild-type and mutant nNOS proteins were resolved by 7.5% SDS-PAGE and electrophoretically transferred onto PVDF membranes as described by Towbin et al. (22). The membranes were blocked with 100 mM NaCl, 50 mM Tris-HCl (pH 7.5), 1% nonfat milk, and 0.05% Tween 20 in the presence of 1 mM CaCl 2 for 30 min at room temperature. Biotinylated CaM was then added at a final concentration of 0.1 g/ml in the same buffer for 1 h. After washing with the buffer, membranes were incubated with avidin-biotin-peroxidase (Vector Laboratories, Inc.), washed extensively, and then developed with the chemiluminescence reagent (Amersham Pharmacia Biotech).
In Vitro Phosphorylation of nNOS by PKA and PKC-Recombinant nNOS (0.5 g) expressed in E. coli was incubated for 10 min at 30°C in 40 mM HEPES (pH 7.0), 10 mM MgCl 2 , and 1 mM ATP. The final concentrations of PKA and PKC were 100 nM, respectively. For phosphorylation by PKC, 100 g/ml phosphatidylserine and 1 mM CaCl 2 were added. Reactions were stopped by the addition of sample buffer followed by 6% SDS-PAGE and electrophoretically transferred onto PVDF membranes. For immunodetection of the transferred proteins, the procedure of Burnette (23) was used, except that the second antibody was linked to horseradish peroxidase.
Preparation of Antiphosphopeptide Antibodies-We developed antibody recognizing the phosphorylation at Ser 847 on nNOS as described previously (24). Briefly, 1 g of the phosphorylated synthetic peptide, CKVRFNS(phosphoS)VSSYS, coupled with keyhole limpet hemocyanin, emulsified in Freund ' s complete adjuvant, was injected into a rabbit subcutaneously. Every 2 weeks, the same portion of the antigen solution emulsified in Freund ' s incomplete adjuvant was boosted. The rabbit was bled 5 weeks after the initial injection. Immunoglobulin G was affinity-purified by tandem column chromatography using columns coupled to phosphorylated synthetic peptide and non-phosphorylated synthetic peptide, CKVRFNSSVSSYS.

RESULTS AND DISCUSSION
Phosphorylation of nNOS by CaM Kinases-We first determined whether CaM kinases could phosphorylate recombinant nNOS expressed in the E. coli system and found that nNOS was phosphorylated by these kinases in the presence of Ca 2ϩ / CaM in vitro (Fig. 1). CaM-K II␣ caused the most rapid phosphorylation with half-maximal phosphorylation apparent at 3 min and a plateau level at 10 min. CaM-K I␣ and CaM-K IV activated by CaM-KK also phosphorylated nNOS, but the plateau level was not reached until 100 min (ϳ0.4 and ϳ0.7 mol of 32 P/mol of nNOS at 100 min, respectively). nNOS phosphorylation was not observed if Ca 2ϩ /CaM or the CaM kinase was omitted from the reaction mixture and CaM-KK (140 nM) could not phosphorylate nNOS (data not shown). The maximal phosphorylation of nNOS by CaM-K II␣ was observed at ϳ0.4 mol of 32 P/mol of nNOS under the conditions employed. To clarify whether the phosphorylation sites of the CaM kinases were the same, nNOS was sequentially phosphorylated by each of the two kinases. nNOS was initially phosphorylated by CaM-K II␣ for 30 min and was then exposed to activated CaM-K IV or CaM-K I␣ for an additional 30 min. No significant additional phosphate was incorporated into nNOS by addition of activated CaM-K IV or CaM-K I␣ (data not shown), indicating that CaM-K II␣, CaM-K IV, and CaM-K I␣ phosphorylated the same amino acid residue of nNOS.
In order to determine whether Ca 2ϩ /CaM binding to CaM kinases, nNOS, or both, was essential for the phosphorylation reaction, we needed a substrate of CaM kinases that did not bind Ca 2ϩ /CaM. We have previously shown that mutation to Glu or Asp of several residues (Lys 732 -Lys-Leu) of nNOS blocks the binding of CaM and activation of nNOS. Therefore, it was of interest to determine whether inactive nNOS lacking CaM binding ability, generated with mutation of Lys 732 -Lys-Leu to Asp 732 -Asp-Glu could be phosphorylated by CaM kinases. As shown in Fig. 2, mutant nNOS could be phosphorylated by CaM kinases like the wild-type enzyme in the presence of Ca 2ϩ /CaM. We also constructed a fusion protein of GST and a fragment of CaM kinase I␣ (residues 1-293) (GST/CaM kinase I ␣1-293), which did not contain the autoinhibitory domain or the CaM-binding domain (18). GST/CaM kinase I ␣1-293 phosphorylated wild-type nNOS either in the presence or absence of Ca 2ϩ /CaM (data not shown). These results demonstrated that the binding of Ca 2ϩ /CaM to nNOS was not required for the phosphorylation of nNOS by CaM kinases.
Construction and Expression of nNOS Mutants-Analysis of a series of truncated mutants of nNOS suggested that the residues phosphorylated by CaM kinases are included within residues 825-884 (data not shown). It is known that CaM kinase II recognizes site that have an Arg located three residues toward the NH 2 terminus of the phosphorylated Ser or Thr. Since phosphoamino acid analysis revealed that all three CaM kinases phosphorylated nNOS at amino acids that comigrated with phosphoserine (data not shown), we made single mutants by introducing Ala residues in place of Ser 847 (S847A) and Ser 858 (S858A), which were at the P-0 positions on the COOH-terminal sides of Arg 844 and Arg 855 , respectively. We also wanted to place a Ser at three residues toward the NH 2 terminus of Arg 855 (S852A), because the orientation of the consensus phosphorylation sequence for CaM-K II (25) may be opposite to that of the normal substrate. Wild-type and single mutants were expressed using the E. coli or the baculovirus/Sf9 cell system and purified on 2Ј-5Ј-ADP-agarose as described under "Experimental Procedures." All of the recombinant nNOSs were at least 90% pure and gave a major band at 160 kDa on SDS-PAGE with Coomassie Brilliant Blue staining (Fig. 3A). Mutants S852A and S858A were phosphorylated by CaM kinases as was the wild-type enzyme, but the S847A mutant was not phosphorylated by activated CaM-K I␣ or CaM-K IV. In contrast, CaM-K II␣ did phosphorylate this mutant, but only very slightly (Fig. 3B).
Effect of Phosphorylation on nNOS Activity-The above results demonstrated that Ser 847 was a major phosphorylation site for CaM kinases. Mutant S847D was also constructed and expressed in the E. coli system (Fig. 3A). We tested the ability of Ser 847 residue phosphorylation to affect NOS enzyme activity using the mutant S847D. Enzyme activity was determined from the rate of conversion of L-arginine to L-citrulline. The mutant S847D exhibited NOS enzyme activity that was approximately 40% of the wild-type level in the presence of Ca 2ϩ / CaM (Fig. 4A). Since pretreatment of the nNOS preparation from E. coli caused instability of NOS enzyme activity (data not shown), wild-type nNOS was also expressed in the baculovirus/ Sf9 cell system (Fig. 3A). Most properties of the purified overexpressed enzyme were identical to those previously reported, including the specific activity (195 nmol/min/mg) (19). However, preincubation of Ca 2ϩ /CaM with nNOS from the Sf9 cells for 10 min markedly attenuated NOS enzyme activity to less than 10% of that without preincubation, and we did not detect any changes in NOS activity following in vitro phosphorylation by CaM kinases in the presence of Ca 2ϩ /CaM (data not shown). We employed a constitutive active form of CaM-K I␣ (GST/CaM kinase I ␣1-293) for phosphorylating nNOS. Wild-type nNOS was incubated either with (Ϸ0.5 mol of phosphate incorporation into nNOS) or without activated GST/CaM kinase I ␣1-293 at 30°C with 100 M ATP in the presence of EGTA, and NOS activity was determined from the rate of conversion of L-arginine to L-citrulline. The preincubation under the conditions employed also attenuated NOS enzyme activity to approximately 50% (85.8 Ϯ 1.7 nmol/min/mg) of that without incubation. However, phosphorylated nNOS exhibited NOS enzyme activity that was approximately 70% compared with unphosphorylated enzyme (Fig. 4B). Neither mutant S847D nor nNOS exhibited Ca 2ϩ /CaM-independent NOS activity. All NOS isoforms are known to catalyze electron transfer to a variety of artificial electron acceptors (26 -29). Phosphorylated nNOS exhibited CaM-dependent reduction of cytochrome c that was approximately 65% compared with unphosphorylated enzyme. Meanwhile, CaM-independent DCPIP reduction of phosphorylated nNOS was essentially indistinguishable from that of the unphosphorylated enzyme (Table I). Interestingly, the reductase activities of preincubated nNOS was similar to that of the enzyme without incubation (data not shown). Thus, Ser 847 in nNOS appears to represent an essential determinant for transducing the nNOS-CaM interaction into stimulation of enzyme activity via its phosphorylation.
Effect of Phosphorylation on nNOS Kinetic Parameters of nNOS-It was important to determine how the phosphorylation at Ser 847 altered the kinetics of nNOS. Table II summarizes some of the properties of the phosphorylated nNOS compared with the unphosphorylated enzyme. From the results shown, it is clear that the main effect of the phosphorylation at Ser 847 was to lower the V max of NOS activity, with little change of the K m for L-arginine and K act for CaM. We also analyzed the kinetics of the mutant S847D and wild-type enzyme. Mutation of Ser 847 to Asp resulted in attenuation of the specific activity (V max ) (304 Ϯ 24.5 versus 158.5 Ϯ 18.5 nmol/min/mg), with little change of the K m for L-arginine and K act for CaM.
In order to understand how CaM kinases regulate nNOS, the CaM-binding ability of the mutants was analyzed. The CaM binding of mutant S847D or phosphorylated nNOS was significantly decreased compared with that of wild-type or unphosphorylated enzyme, respectively, as assessed by the gel overlay technique (Fig. 5). Previous characterization of NOS has indicated that the native protein is a homodimer, and dimerization has been shown to be necessary for catalytic activity of the enzyme. We therefore explored the possibility that the mutated or phosphorylated residue (Ser 847 ) affected the dimerization of nNOS. It is known that the nNOS homodimer is stabilized by tetrahydrobiopterin and L-arginine during low temperature SDS-PAGE. In this assay, the dimerization of mutant S847D or phosphorylated nNOS was essentially indistinguishable from that of the wild-type or unphosphorylated enzyme, respectively (data not shown). These findings suggest that Ser 847 residue influenced the interaction between CaM and nNOS via its phosphorylation.
Identification of the in vivo phosphorylation at Ser 847 on nNOS-nNOS is phosphorylated on different serine site of the enzyme by PKA, PKC, and CaM-K II in vitro (13). To confirm in vivo phosphorylation at Ser 847 on nNOS, we used a phosphospecific antibody, NP847. This antibody was specific for Ser(P) 847 in nNOS, reacting with phosphorylated nNOS by CaM-K II␣ but not with that by PKA or PKC (Fig. 6A). Protein immunoblot analyses with NP847 of a partially purified rat brain nNOS detected an immunoreactive band corresponding to the 160-kDa nNOS (Fig. 6B). These data demonstrate that phosphorylation at Ser 847 on nNOS is catalyzed not only in vitro but also in vivo.
Conclusions-Although it has already been confirmed that  nNOS is phosphorylated by several protein kinases, such as PKA, PKC, and CaM kinase, the easy thermal denaturation and inactivation of NOS has limited our understanding of the activation mechanisms mediated via phosphorylation. By using truncated and site-directed nNOS mutants, we identified the amino acid residue phosphorylated by CaM kinases and found that this phosphorylation caused a decrease in NOS enzyme activity. We also employed a constitutive active form of CaM-K I␣ for phosphorylating nNOS, followed by analysis of the phosphorylated enzyme. This finding is consistent with the results of previous in vitro studies demonstrating a dramatic attenuation of nNOS enzyme activity following CaM-K II phosphorylation (12). Such down-regulation of nNOS activity by CaM kinases may represent an important component of the "cross-talk" between kinases and NO. All three tested CaM kinases incorporated 0.4 -0.7 mol of 32 P/mol of recombinant nNOS expressed in E. coli, in good agreement with the results using purified rat brain nNOS (13). Nakane et al. reported that 9 mol of 32 P/mol of purified rat brain nNOS were incorporated by CaM-K II (12). We also analyzed the 32 P stoichiometry using partially purified rat brain nNOS as a control experiment, which was similar to the results using recombinant nNOS expressed in E. coli under the conditions employed (data not shown).
It is well known that the activities of several enzymes can be inhibited by phosphorylation of their CaM-binding domains, but the mutant residue (Ser 847 ) was not near the putative CaM-binding domain of nNOS and binding of Ca 2ϩ /CaM to nNOS did not inhibit the phosphorylation of nNOS by CaM kinases (Fig. 2). Furthermore, the replacement of Ser 847 by Asp and phosphorylated nNOS at Ser 847 had little effect on the Kact for CaM (Table II). However, the CaM binding of mutant S847D or phosphorylated nNOS at Ser 847 was significantly decreased compared with that of the wild-type and unphosphorylated enzyme, respectively (Fig. 5). The mutant S847D and phosphorylated nNOS at Ser 847 also showed loss of NOS enzyme activity in the presence of Ca 2ϩ /CaM (Fig. 4). Thus, the putative nNOS inhibitory element on nNOS may allosterically perturb the CaM binding domain through phosphorylation of the Ser 847 residue. Salerno et al. reported that cNOSs contain a unique polypeptide insert in the FMN binding domains that acts as an inhibitor of cNOS enzyme activity (15). The synthetic peptides derived from the insertion of eNOS interfered with CaM binding to cNOSs by interacting with cNOSs but not with CaM. However, nNOS-derived polypeptides had no effect on nNOS. Since the Ser 847 residue is located at the insertion of nNOS, phosphorylation of the Ser 847 residue may influence the affinity of the insertion for an internalized binding site on nNOS. It was surprising to find that inactive nNOS lacking CaM-binding ability was phosphorylated by CaM kinases to the same extent as wild-type nNOS and that constitutively active CaM kinase I␣ (residues 1-293) phosphorylated nNOS either in the absence or presence of Ca 2ϩ /CaM (Fig. 2), since CaM displaces the FMN binding domain insert peptide of nNOS that enhances exposure of the insert to trypsinolysis at Lys 856 (15). It is possible that exposure of Lys 856 is induced by CaM-nNOS binding, whereas that of Ser 847 is not.
We expected an Asp mutant but not an Ala one to mimic the negatively charged state of a phosphorylated Ser 847 on nNOS; however, both mutants exhibited decrease in NOS enzyme activity and similar characteristics (data not shown). It is possible that Ser 847 is crucial for catalytic activity of nNOS, perhaps through involvement in the binding of CaM. Since the Ser 847 residue is located in FMN binding domains of nNOS, it was of interest to determine whether the replacement of Ser 847 interferes with FMN incorporation and therefore results in lower specific activities. However, flavin content of the mutants S847A and S847D was essentially indistinguishable from that of the wild-type enzyme by the measuring fluorescence (data not shown). Since nNOS, CaM-K I␣, CaM-K II␣, CaM-K IV, and CaM-KK have very similar requirements for Ca 2ϩ /CaM (Table I) (18,30,31), it is puzzling why the adjacent steps in kinase and NO signaling should both require the same activator, Ca 2ϩ /CaM. Phosphorylation of nNOS at Ser 847 is thought to be physiological, since partially purified rat brain nNOS is phosphorylated at Ser 847 by a given amount, as assessed by immunoblot analysis with NP847 (Fig. 6). However, the physiological meaning of nNOS phosphorylation by CaM kinases, including which CaM kinase is essential for such phosphorylation in vivo, must be determined in the future. We are also FIG. 5. CaM overlays of mutant S847D or phosphorylated wildtype nNOS at Ser 847 . A, wild-type (WT) and mutant (S847D) nNOSs expressed in E. coli (0.5 g each) were resolved on SDS-PAGE using 6% gel. B, unphosphorylated (PϪ) and phosphorylated (Pϩ) nNOSs (0.5 g each) as in Fig. 4B were resolved on SDS-PAGE using 7.5% gel. These gels were electrophoretically transferred onto PVDF membranes, and analyzed by reaction with biotinylated CaM in the presence of 1 mM initiating a study to determine whether Ser 847 affects intramolecular protein protein interaction using peptide chemistry.