Yan XB et al / Acta Pharmacol Sin 2004 May; 25 (5): 617-622

Brain ischemia induces serine phosphorylation of neuronal nitric oxide synthase by Ca²⁺/calmodulin-dependent protein kinase II in rat hippocampus¹

Xue-bo YAN, Fan-jie MENG, Bo SONG, Guang-yi ZHANG²

Research Center for Biochemistry and Molecular Biology, Xuzhou Medical College, Xuzhou 221002, China

¹ Project supported by the National Natural Science Foundation of China, No 30070182; 30170220.

² Correspondence to Prof Guang-yi ZHANG. Phn 86-516-574-8423. Fax 86-516-574-8486. E-mail gyzhang@xzmc.edu.cn

Received 2003-09-09 Accepted 2004-01-12

KEY WORDS serine; phosphorylation; nitric oxide synthase; calmodulin kinase; KN 62; brain ischemia; hippo-campus

ABSTRACT

AIM: To investigate whether brain ischemia induces serine phosphorylation of neuronal nitric oxide synthase (nNOS) by Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) and the interaction between CaMKIIa and nNOS in rat hippocampus. METHODS: Brain ischemia was induced by bilateral carotid artery occlusion procedure. Phosphorylation and the interaction of proteins were studied by immunoprecipitation and immunoblotting. We investigated during brain ischemia serine phosphorylation and amount of nNOS in crude membranes fraction (P) and cytosolic fraction (S), interaction between CaMKIIa and nNOS, and the effects of 1-[N,O-bis-(5-isoquinolinesulfonyl)- N-methyl-L-tyrosyl]-4-phenylpiperazine (KN-62, a selective inhibitor of CaMKII) on phosphorylation and the interaction of proteins in P. RESULTS: Serine phosphorylation of nNOS in P increased persistently during brain ischemia, and 15 min ischemia-induced serine phosphorylation of nNOS was attenuated significantly by KN-62. But there was no serine phosphorylation of nNOS in S. The distributions of nNOS were not affected by ischemia and KN-62. However, the binding levels of both CaMKIIa with nNOS and Thr²⁸⁶ autophosphorylated CaMKIIa with nNOS increased after ischemia, and were diminished by KN-62. CONCLUSION: CaMKII interacted with nNOS and regulated serine phosphorylation of nNOS during brain ischemia.

INTRODUCTION

Constitutively expressed nitric oxide synthase (NOS) has three isoforms, including neuronal NOS (nNOS), inducible NOS (iNOS), and endothelial NOS (eNOS), which has been isolated and cloned^[1]. NOS catalyzed L-arginine to generate L-citruline and nitric oxide (NO), a gaseous messenger molecule, the overproduction of which plays a critical role in glutamate-induced neurotoxicity after ischemia^[2]. In mutant mice deficient in nNOS and subsequent NO production, infarct size becomes smaller and neuronal injury is less than those in normal mice after brain ischemia^[3]. nNOS locates on the membrane of mitochondria, endoplasmic reticulum, in pre- and postsynaptic profiles and in the cytosol^[4]. Posttranslational myristoylation presu-mably is a determinant for the subcellular localization, but may not be the only factor. No significant difference was detected between the soluble and particulate enzymes including their specific activity^[5]. It has been reported that phosphorylated nNOS at Ser⁸⁴⁷ decreases enzyme activity to approximate 70 % of unphosphoryl-ated enzyme^[6]. The Ca²⁺/CaM-dependent nNOS is coupled to N-methyl-D-aspartate receptor (NMDAR) on the membranes by PSD95 protein^[7]. Ca²⁺ influx through NMDAR stimulates nNOS and other signaling molecules including the Ca²⁺/calmodulin(CaM)-dependent protein kinaseII (CaMKII)^[8] .

CaMKII, a ubiquitously expressed Ca²⁺/CaM-dependent protein kinase with broad substrates, is enriched in the brain and constitutes approximate 1 %-2 % of the total protein in the hippocampus^[9]. Autophosphoryl-ation at Thr²⁸⁶ of the a isoform of CaMKII modifies its function, and converts the kinase into a Ca²⁺/CaM independent state^[10]. After decapitation of rats, the translocation to postsynaptic density from cytosol of CaMKII increases to approximate 10 % of the protein content in the postsynaptic density fraction^[11]. CaMKII translocates immediately and completely to crude membranes fraction from cytosolic fraction during brain ischemia^[12]. Substrates phosphorylated by CaMKII are involved in many aspects of cellular functions including membrane current, neurotransmitter synthesis and release^[10]. Mutant mice knocked out CaMKIIa are deficient in long-term potentiation (LTP), which supports that CaMKIIa is necessary for LTP^[13]. It has been documented that in the brain nNOS is one of the substrates of CaMKII, and the phosphorylation of nNOS at Ser⁸⁴⁷ by CaMKII peaks at 15 min ischemia with 0.5 h reperfusion in the hippocampus^[14].

In this report, we further studied the subcellular distributions and serine phosphorylation of nNOS in the hippocampus during the early part of brain ischemia, and whether CaMKII was involved in nNOS serine phosphorylation after its autophosphorylation and translocation, and the functional interaction between nNOS and CaMKII.

MATERIALS AND METHODS

Antibody Anti-CaMKIIa antibody (Product Number: C6974) and anti-nNOS antibody for immunoblotting (Product Number: N7280) were purchased from Sigma, anti-nNOS antibody for immunoprecipitation (Catalog: sc-648) was from Santa Cruz; anti-p-CaMKIIa (Thr286 phosphorylated, Catalog: V1111) antibody was from Promega (Madison, WI, USA); anti-p-Ser antibody was prepared by our laboratory.

Animal surgical procedures Adult male Sprague-Dawley rats (Shanghai Experimental Animal Center, Chinese Academy of Science) weighing 250-300 g were subjected to 3, 5, 15, and 30 min brain ischemia by four-vessel occlusion method as described before^[15]. Briefly, under anesthesia with chloral hydrate (0.3 g/kg, ip), vertebral arteries were electroncauterized then common carotid arteries were exposed. Ischemia was induced by occluding common carotid arteries with aneurysm clips after rats recovered for 24 h, except that the arteries of sham control rats were not occluded other operations were the same. Rats lost their righting reflex within 30 s and those whose pupils were dilated and unresponsive to light were selected for the experi-ments. Rectal temperature was maintained about 37 ℃ in the whole process. When necessary, 20 min before 15 min ischemia, a volume of 5 μL KN-62 which was dissolved in Me₂SO at a concentration of 1 g/L or Me₂SO as vehicle control was injected into the left cerebral ventricle (anteroposterior, -0.8 mm; lateral, 1.5 mm; depth, 3.5 mm from bregma) using a stepper-moto-rized microsyringe (Stoelting, Wood Dale, IL) within 5 min.

Tissue preparation Rats were decapitated immediately after ischemia. Both sides of hippocampus were removed, and frozen quickly in liquid nitrogen. Samples were homogenized in 1:10 (W/V) ice-cold homogenization buffer (HB) containing 50 mmol/L 3-(N-morpholino) propanesulfonic acid (pH 7.4), 100 mmol/L KCl, 320 mmol/L sucrose, 0.5 mmol/L MgCl₂, 0.2 mmol/L dithiothreitol, inhibitors of phosphatase and protease (20 mmol/L b-glycerophosphate, 20 mmol/L sodium pyrophosphate, 50 mmol/L NaF, 1 mmol/L each of edetic acid, egtazic acid, sodium orthovanadate, p-nitrophenyl phosphate, phenylmethylsulfonyl fluoride, and benza-midine and 5×10^-3 g/L each of aprotinin, leupeptin, and pepstatin A). The homogenates were centrifuged at 800×g for 15 min at 4 ℃, then supernatants were centrifuged at 100 000×g for 30 min at 4 ℃ in a TL 100.2 rotor (Backman, Palo Alto, CA), the supernatant cytosolic fraction (S) was collected, the pellet crude membranes fraction (P) was re-suspended in HB containing 0.1 % Triton X-100 for 10 s by sonication. The samples were stored at -80 ℃ before use.

Immunoprecipitation and immunoblotting Samples (400 μg) which were diluted four fold in immunoprecipitation (IP) buffer containing 50 mmol/L N-[2-hydroxyethyl]piperazine-N'-[2-ethanesulfonic acid] (HEPES, pH 7.1), 150 mmol/L NaCl, 1 mmol/L ZnCl₂, 1.5 mmol/L MgCl₂, 10 % glycerol, 1 % Triton X-100, and inhibitors of phosphatase and protease as indicated above. Samples were preincubated with 20 μL protein A sepharose CL-4B (Amersham, Uppsala, Sweden) for 1 h. Nonspecific protein A was removed by centrifug-ing. The supernatants were incubated with anti-p-Ser or anti-CaMKIIa or anti-nNOS antibody for 4 h or overnight at 4 ℃, and the reaction system continued incubating with protein A for 2 h. Samples were centrifuged at 10 000×g, and the pellets were washed three times with IP buffer. Pellets containing bound proteins were eluted by adding 4-fold SDS-PAGE sample buffer (12 μL) and boiled in 100 ℃ water for 5 min. Samples were centrifuged, and supernatants were used for immmunoblotting. For immunoblotting, proteins (50 μg) were separated by 10 % SDS-PAGE and were electrotransferred onto nitrocellulose membranes (NC, pore size, 0.45 μm). The membranes were blocked with blocking buffer for 3 h, then probed with anti-nNOS or anti-CaMKIIa or anti-p-CaMKII antibody overnight at 4 ℃. Detection was performed using alkaline phosphatase conjugated goat anti-rabbit IgG
(1:10000) and developed using BCIP/NBT color sub-strate. The bands on the membranes were scanned and analyzed by an image analyzer (Labworks Software, UVP Inc, Upland, CA).

Statistical analysis Values were expressed as the mean±SD. Statistical analysis of the results was carried out by one way analysis of variance (ANOVA) followed by the Duncan's new multiple range method. P<0.05 was considered significant.

RESULTS

Effect of brain ischemia on nNOS serine phosphorylation and the distributions of nNOS nNOS serine phosphorylation in P increased persistently and significantly after 3 min ischemia (P<0.05, Fig 1A). Interestingly, there was no serine phosphorylation of nNOS in S (Fig 1B). nNOS was found in both P (Fig 1C) and S (Fig 1D), and the distributions of nNOS did not change during brain ischemia (P>0.05).

Fig 1. Immunoblotting analysis of serine phosphorylation and the distributions of nNOS in P and S during brain ischemia. (A, B) IP with anti-p-Ser antibody and (A, B, C, D) IB with anti-nNOS antibody. (E) Quantitative representation of nNOS in P and S, and nNOS serine phosphorylation in P. n=3. Mean±SD. ^aP>0.05, ^bP<0.05 vs sham. P, crude membranes fraction; S, cytosolic fraction; IP, Immunoprecipitation; IB, immunoblotting.

Effect of KN-62 on the increased nNOS serine phosphorylation KN-62, a selective inhibitor of CaMKII, was administered 20 min before 15 min ischemia through left cerebral ventricle. The elevation of nNOS serine phosphorylation induced by 15 min ischemia was inhibited significantly by KN-62 in P (P<0.05), but not by Me₂SO (P>0.05, Fig 2A). The distributions and amounts of nNOS in P were not affected by KN-62 (P>0.05, Fig 2B).

Fig 2. Effect of KN-62 on the elevation of nNOS serine phosphorylation induced by 15 min-ischemia in P. (A) Sample proteins were immunoprecipitated with anti-p-Ser antibody and (A, B) immunoblotted with anti-nNOS antibody. (C) Quantitative representation of nNOS and nNOS serine phosphorylation in P. n=3. Mean±SD. ^aP>0.05, ^bP<0.05 vs sham. ^dP>0.05, ^eP<0.05 vs 15 min-ischemia. P, crude membranes fraction; IP, immunoprecipitation; IB, immunoblotting; K, KN-62; M, Me₂SO.

Effects of brain ischemia on the interaction of nNOS with CaMKIIa or p-CaMKIIa The associations of nNOS with CaMKIIa and p-CaMKIIa were examined by co-immunoprecipitation. nNOS and CaMKIIa can coimmunoprecipitate with each other (Fig 3A). Moreover, the nNOS and CaMKIIa complex level was up regulated remarkably after a 3-min ischemia (P<0.05). CaMKIIa level increased significantly during ischemia (P<0.05, Fig 3B). The phosphorylated CaMKIIa and that collocating with nNOS increased markedly (P<0.05, Fig 3D, 3E).

Fig 3. Interactions of nNOS with CaMKIIa or p-CaMKIIa in P during brain ischemia. (A, D) IP with anti-nNOS or anti-CaMKIIa antibody, and IB with anti-CaMKIIa or anti-nNOS or anti-p-CaMKIIa antibody. (B, E) IB with anti-CaMKIIa antibody or anti-p-CaMKIIa antibody. (C, F) Quantitative representation of interaction of nNOS with CaMKIIa or p-CaMKIIa, amounts of CaMKIIa and p-CaMKIIa in P. n=3. Mean±SD. ^bP < 0.05 vs sham. P, crude membranes fraction; IP, immunoprecipitation; IB, immuno-blotting.

Effects of KN-62 on the interaction of nNOS with CaMKIIa or p-CaMKIIa Following administration with KN-62, the up-regulation of nNOS and CaMKIIa complex level induced by 15 min ischemia was blocked (P<0.05, Fig 4A). Importantly, the increase of p-CaMKIIa collocating with nNOS was diminished accordingly (P<0.05, Fig 4B). Both were not affected by Me₂SO (P>0.05, Fig 4A, 4B).

Fig 4. Effect of KN-62 on the interaction of nNOS with CaMKIIa or p-CaMKIIa in P. (A, B) IP with anti-nNOS antibody and IB with anti-CaMKIIa or anti-p-CaMKIIa antibody. (C) Quantitative representation of the interaction of nNOS with CaMKIIa or p-CaMKIIa. n=3. Mean±SD. ^bP < 0.05 vs sham. ^dP > 0.05, ^eP < 0.05 vs 15 min-ischemia. P, crude membranes fraction; IP, immunopreci-pitation; IB, immunoblotting; K, KN-62; M, Me₂SO.

DISCUSSION

nNOS, a soluble and membrane-bound enzyme, appears to localize in specific populations of neurons in the brain and in the nonadrenergic, noncholinergic, autonomic nervous system^[16]. In the present study, nNOS was found in both P and S. nNOS serine phosphorylation in P increased markedly and there was no serine phosphorylation of nNOS in cytosolic fraction during brain ischemia, suggesting that nNOS in P was phosphorylated by the protein kinase of the membranes. Although little is known about the factors determining its subcellular distribution, the unchanging levels of nNOS in P and S indicated that brain ischemia was not a factor affecting the distributions of nNOS.

CaMKII constitutes approximately 1 %-2 % of the total protein in the hippocampus^[9]. It translocates immediately and completely to crude membranes fraction from cytosolic fraction during brain ischemia^[12]. The CaMKIIa level in P increased rapidly and remained high after 3 min and the CaMKIIa level in S decreased rapidly and thoroughly^[12]. In this study, nNOS serine phosphorylation in P increased significantly and that in S was not detected. KN-62 (a selective CaMKII inhibitor) diminished the increase of nNOS serine phosphorylation, and did not affect the amount of nNOS in P. These results supported that CaMKIIa was predominantly implicated in nNOS serine phosphorylation in hippocampus during brain ischemia. The inhibition of the Ser/Thr protein kinase CaMKIIa did not influence the distributions of nNOS.

In the present work, the levels of nNOS and CaMKIIa complex and the phosphorylated CaMKIIa collocating with nNOS were up-regulated during brain ischemia. KN-62 decreased the up-regulations. Our results suggested that nNOS could interact with CaMKIIa, and CaMKIIa collocating with nNOS was autophosphorylated and auto-activated. KN-62 diminished the association of CaMKIIa with nNOS, and decreased the activation of CaMKIIa. Brain ischemia results in rapid increase in Ca²⁺ concentration of cytoplasma in cell that comes from extracellular Ca²⁺ and intracellular store. CaM, an EF-hand family of Ca²⁺-binding proteins, is a predominant intracellular receptor for Ca²⁺. Ca²⁺/CaM binds to CaM binding domain of CaMKIIa that lays C-terminal to the autoregulatory regions, removing autoinhibitory domain, then exposes ATP-binding sites, protein substrate binding sites and Thr²⁸⁶ for autophosphorylation^[10]. After translocating to postsynaptic density, autophosphorylation of CaMKIIa at Thr²⁸⁶ enables substrates such as nNOS to interact with its protein substrate binding sites in the catalytic core of the kinase, thus serine residues of nNOS may be phosphorylated by active p-CaMKIIa^[10]. KN-62, a drug belonging to quinolonesulfonamides, functions as a competitive ATP inhibitor^[17] . It disabled substrate phosphorylation and autophosphorylation at Thr²⁸⁶ by competing ATP-binding sites of CaMKIIa. Moreover, there is a reciprocal facilitation between binding of Ca²⁺/CaM and ATP^[10] , therefore Ca²⁺/CaM binding was attenuated by KN-62 and the regulatory domain as a pseudosubstrate inhibited nNOS binding. Our results that the levels of CaMKIIa and phosphorylated CaMKIIa increased in P during brain ischemia were in agreement with previous study, supporting that CaMKIIa autophosphorylates and translocates to crude membranes fraction from cytosolic fraction during brain ischemia^[12].

nNOS catalytic activity is likely to be decreased by phosphorylation^[18], and Ser⁸⁴⁷ phosphorylated nNOS decreases enzyme activity to 70 % of unphosphorylated enzyme^[6]. nNOS catalytic activity regulates the production of NO that medicates not only biological actions but also neurotoxicity triggered by NMDAR activation through PSD95^[7]. In this current study, during brain ischemia nNOS serine phosphorylation was increased by CaMKIIa. Previous study showed that in the reperfusion, after the peak of the phosphoryation of nNOS at Ser⁸⁴⁷ at 15 min ischemia with 0.5 h reperfusion, the Ser⁸⁴⁷ phosphorylation decreased below sham level at 2 h reperfusion^[14]. Therefore, CaMKIIa might play different roles in the early part of brain ischemia and in the transient brain ischemia followed reperfusion. In the early part of brain ischemia, CaMKIIa might play a neuroprotective role by up-regulation of nNOS serine phosphorylation to reduce NO production.

nNOS can be phosphorylated by cylic AMP-dependent protein kinase (PKA), protein kinase C (PKC), and calcium-calmodulin protein kinases (CAM-K)^[1]. In the present work, KN-62 attenuated the elevation of nNOS serine phosphorylation induced by 15 min ischemia nearly to sham level in P. There were unchanging levels of nNOS in S during brain ischemia, but there was no serine phosphorylated nNOS although there were other protein kinases such as PKA, PKC. These results supported the idea that during brain ischemia, it was CaMKII, not other protein kinases, that phosphorylated nNOS at serine residues. Little is known concerning why other protein kinases such as PKA, PKC were not involved in serine phosphorylation of nNOS during brain ischemia, and the precise mechanism of the physical interaction between nNOS and CaMKIIa. It is obscure whether nNOS recruits to nNOS-CaMKIIa complex directly through PDZ domain or there is any other adapter proteins such as PSD95 between them like NMDAR-PSD95-nNOS complex^[19] .

Taken together, CaMKIIa may play a neuropro-tective role in the early part of brain ischemia by phosphorylating serine residues of nNOS to down-regulate the activity of nNOS and NO production.

REFERENCES

• 1 Bredt DS, Hwang PM, Glatt CE, Lowenstein C, Reed RR, Snyder SH. Cloned and expressed nitric oxide synthase structurally resembles cytochrome P-450 reductase. Nature 1991; 351: 714-8.
• 2 Dawson VL, Dawson TM, London ED, Bredt DS, Snyder SH. Nitric oxide mediates glutamate neurotoxicity in primary cortical cultures. Proc Natl Acad Sci USA 1991; 88: 6368-71.
• 3 Huang Z, Huang PL, Panahian N, Dalkara T, Fishman MC, Moskowitz MA. Effects of cerebral ischemia in mice deficient in neuronal nitric oxide synthase. Science 1994; 265: 1883-5.
• 4 Batista CM, de Paula KC, Cavalcante LA, Mendez-Otero R. Subcellular localization of neuronal nitric oxide synthase in the superficial gray layer of the rat superior colliculus. Neurosci Res 2001; 41: 67-70.
• 5 Hecker M, Mulsch A, Busse R. Subcellular localization and characterization of neuronal nitric oxide synthase. J Neuro-chem 1994; 62: 1524-9.
• 6 Hayashi Y, Nishio M, Naito Y, Yokokura H, Nimura Y, Hidaka H, et al. Regulation of neuronal nitric-oxide synthase by calmodulin kinases. J Biol Chem 1999; 274: 20597-602.
• 7 Sattler R, Xiong Z, Lu WY, Hafner M, MacDonald JF, Tymianski M. Specific coupling of NMDA receptor activation to nitric oxide neurotoxicity by PSD-95 protein. Science 1999; 284: 1845-8.
• 8 Shen K, Meyer T. Dynamic control of CaMKII translocation and localization in hippocampal neurons by NMDA receptor stimulation. Science 1999; 284: 162-6.
• 9 Erondu NE, Kennedy MB. Regional distribution of type II Ca²⁺/calmodulin-dependent protein kinase in rat brain. J Neurosci 1985; 5: 3270-7.
• 10 Hudmon A, Schulman H. Neuronal CA²⁺/calmodulin-dependent protein kinase II: the role of structure and autoregulation in cellular function. Annu Rev Biochem 2002; 71: 473-510.
• 11 Suzuki T, Okumura-Noji K, Tanaka R, Tada T. Rapid translocation of cytosolic Ca²⁺/calmodulin-dependent protein kinase II into postsynaptic density after decapitation. J Neurochem 1997; 69: 435-7.
• 12 Meng F, Zhang G. Autophosphorylated calcium/calmodulin-dependent protein kinase II alpha induced by cerebral ischemia immediately targets and phosphorylates N-methyl-D-aspartate receptor subunit 2B (NR2B) in hippocampus of rats. Neurosci Lett 2002; 333: 59-63.
• 13 Silva AJ, Stevens CF, Tonegawa S, Wang Y. Deficient hippocampal long-term potentiation in alpha-calcium-calmodulin kinase II mutant mice. Science 1992; 257: 201-6.
• 14 Osuka K, Watanabe Y, Usuda N, Nakazawa A, Fukunaga K, Miyamoto E, et al. Phosphorylation of neuronal nitric oxide synthase at Ser847 by CaM-KII in the hippocampus of rat brain after transient forebrain ischemia. J Cereb Blood Flow Metab 2002; 22: 1098-106.
• 15 Pulsinelli WA, Brierley JB. A new model of bilateral hemispheric ischemia in the unanesthetized rat. Stroke 1979; 10: 267-72.
• 16 Huang PL. Mouse models of nitric oxide synthase deficiency. J Am Soc Nephrol 2000; (11 Suppl 16): S120-3.
• 17 Means AR. Regulatory cascades involving calmodulin-dependent protein kinases. Mol Endocrinol 2000; 14: 4-13.
• 18 Dinerman JL, Steiner JP, Dawson TM, Dawson V, Snyder SH. Cyclic nucleotide dependent phosphorylation of neuronal nitric oxide synthase inhibits catalytic activity. Neuropharmacology 1994; 33: 1245-51.
• 19 Christopherson KS, Hillier BJ, Lim WA, Bredt DS. PSD-95 assembles a ternary complex with the N-methyl-D-aspartic acid receptor and a bivalent neuronal NO synthase PDZ domain. J Biol Chem 1999; 274: 27467-73.