Journal of Neurochemislry Raven Press, Ltd., New York 0 1992 International Society for Neurochemistry
Ischemia-Induced Loss of Brain Calcium/CalmodulinDependent Protein Kinase I1 Hideyuki Yamamoto, Koji Fukunaga, *Kevin Lee, and Thomas R. Soderling Department of Molecular Physiology and Biophysics, Vanderbilt University, Nashville, Tennessee, and *Department of Neurosurgery, University of Virginia, Charlottesville, Virginia, U.S.A.
Abstract: Forebrain ischemia in gerbils, produced by brief bilateral carotid occlusion, induced the dramatic loss of Ca2+/ calmodulin-dependent protein kinase I1 (CaM-kinase 11) as determined by both kinase activity assays and western blot analysis. In cortex and hippocampus, cytosolic CaM-kinase I1 was completely lost within 2-5 min of ischemia. Particulate CaM-kinase I1 was more stable and decreased in level -40% after 10 min of ischemia followed by 2 h of reperfusion. CaMkinase I1 in cerebellum, which does not become ischemic, was not affected. The rapid loss of CaM-kinase I1 within 25 min was quite specific because cytosolic cyclic AMP kinase
and protein kinase C in hippocampus were not affected. These data indicate that cytosolic CaM-kinase I1 is one of the most rapidly degraded proteins after brief ischemia. Because the multifunctional CaM-kinase I1 has been implicated in the regulation of numerous neuronal functions, its loss may destine the neuronal cell for death. Key Words: Protein kinaseIschemia-Neuronal cell death-Hippocampus. Yamamoto H. et al. Ischemia-induced loss of brain calcium/calmodulindependent protein kinase 11. J. Neurochem. 58, 11 10-1 117 ( 1992).
Transient ischemic events such as stroke result in delayed neuronal cell death occumng over several days, and certain regions of the brain such as hippocampus appear particularly vulnerable (Kirino, I 982; Pulsinelli et al., 1982; Johansen et al., 1983; Suzuki et al., 1983). Several studies have suggested that prolonged elevation of intracellular calcium may initiate neuronal cell death, perhaps by activation of the Ca*+-dependent protease calpain (Choi and Rothman, 1990). Elevation of intracellular calcium levels could occur by the following multiple mechanisms: (a) inhibition of the Ca2+ATPase extrusion pump due to ischemic decreases in ATP levels (Kass and Lipton, 1986), (b) influx of extracellular calcium through voltage-gated or receptoroperated channels (Choi and Rothman, 1990),and/or (c) release of sequestered intracellular calcium. Benveniste et al. (1984) have reported that the extracellular content of glutamate is increased after ischemia, and the glutamate-gated N-methyl-paspartate ion channel has high permeability for calcium (Dingledine, 1983). This hypothesis is strengthened by the observations that
protease inhibitors ameliorate postischemic cell death after transient forebrain ischemia (Lee et al., 1991) and improve the recovery of synaptic transmission after hypoxia in hippocampal slices (Arai et al., 1990). Furthermore, levels of substrate proteins of calpain, such as spectrin (Seubert et al., 1989) and microtubule-associated protein 2 (Kitagawa et al., 1989),are decreased by ischemia in vivo, and spectrin levels are reduced by hypoxia in vitro (Arai et al., 1991). We wanted to investigate the effects of ischemia on the proteolysis of Ca2+/calmodulin (CaM)-dependent protein kinase I1 (CaM-kinase 11) for several reasons. First, CaM-lunase I1 is a substrate for calpain in vitro (Kwiatowski and King, 1989; Rich et al., 1990). Limited proteolysis by calpain generates a fully activated catalytic 30-kDa fragment of CaM-kinase I1 that, on prolonged treatment, is further degraded to inactive fragments (Rich et al., 1990). Second, CaM-kinase I1 is the most abundant protein kinase in brain, and immunohistochemical studies have demonstrated its concentration in hippocampal pyramidal neurons
Received May 14, 199I ; revised manuscript received August 13, 1991; accepted August 17, 1991. Address correspondence and reprint requests to Dr. T. R. Soderling at his present address: Vollum Institute for Advanced Biomedical Research, Oregon Health Sciences University, Portland, OR 97201, U.S.A. The present address of Dr. H. Yamamoto is Vollum Institute for
Advanced Biomedical Research, Oregon Health Sciences University, Portland, OR 9720 1, U.S.A. The present address of Dr. K. Fukunaga is Department of Pharmacology, Kumamoto University Medical School, Kumamoto 860, Japan. Abbreviations used: CaM, calmodulin; CaM-kinase 11, Ca2+/calmodulin-dependent protein kinase 11; IgG, immunoglobulin G.
1110
LOSS OF CaM-KINASE II IN ISCHEMIA (Fukunaga et al., 1988), where it constitutes 52% of protein (Erondu and Kennedy, 1985). Last, because CaM-kinase I1 phosphorylates multiple proteins and constitutes the major postsynaptic density protein at excitatory synapses in forebrain (reviewed by Nairn et al., 1985; Colbran et al., 1989), its destruction could be lethal for the cell. Taft et al. (1988) reported the reduction of Ca2+-dependentphosphorylation of some proteins after ischemia. These proteins could be substrates for or subunits of CaM-kinase 11. The studies reported in this article establish that CaM-kinase I1 in the cytosolic fraction from cortex and hippocampus is completely lost by periods of ischemia of 1 2 min. CaM-kinase I1 in the particulate fraction disappears more slowly, with a half-life of > l o min or -2 h after 10 rnin of occlusion. These data strongly suggest that very brief ischemia causes immediate proteolysis of soluble CaM-kinase 11, which would be deleterious to neuronal cell function. MATERIALS AND METHODS Materials Protein A-Sepharose CL4B was purchased from Pharmacia LKB Biotechnology; [y3*P]ATP and '251-protein A from Du Pont-New England Nuclear; P-cellulose papers from Whatman; the Immuno Pure Immobilization kit from Pierce; and biotinylated CaM from Biomedical Technologies. Other chemicals were of analytical grade. The CaM-kinase I1 substrate peptide syntide-2 was synthesized and prepared as described (Hashimoto and Soderling, 1987). Synthetic peptide inhibitors of cyclic AMP-dependent protein kinase (PKI-tide) and protein kinase C (peptide 19-36) and mastoparan were from Peninsula Laboratories. CaM-kinase 11 purified from rat brain (Hashimoto et al., 1987), rabbit polyclonal antibody [immunoglobulin G (IgG) fraction] to CaM-kinase 11 (Fukunaga et al., 1988), and bovine brain CaM (Gopalakrishna and Anderson, 1982) were kindly provided by Drs. Roger Colbran, Koji Fukunaga, and Debra Brickey, respectively, of our laboratory. Phosphatidylserine and 1,2-diolein were first mixed in chloroform and, after removal of chloroform, were suspended in water as previously described (Hashimoto and Soderling, 1987).
Antibodies against synthetic peptides from CaMkinase I1 Synthetic peptides corresponding to residues 132-1 46 [CaM-kinase (132- 146)] and residues 28 1-309 [CaM-kinase (28 1-309)] were synthesized and purified as previously described (Colbran et al., 1988). Immunization of the rabbits and preparation of the IgG fractions followed the method of Fukunaga et al. (1988). The antibodies were further purified from the IgG fraction by antigen-affinity chromatography. Immobilization of each peptide on Amino Link columns and affinity purification used the procedures outlined in the Pierce Chemical Co. manual. Antibodies were eluted using 5 MNaI and 1 mM sodium thiosulfate solution rather than 0.1 M glycine, pH 2.8. The purified antibodies were collected in 0.1 mg/ml of bovine serum albumin solution, dialyzed against phosphate-buffered saline, concentrated tenfold by Centricon filtration, and kept at -7OOC until use.
1111
Development of ischemia and preparation of brain extracts Adult gerbils (Meriones unguiculatus; 50-80 g of body weight) were subjected to bilateral carotid artery occlusion for 2-10 rnin (Leeet al., 1986; Seubert et al., 1989). At several postsurgical survival times, the gerbils were killed, and samples from the cerebellum, hippocampus, and cortex were taken, frozen on liquid N1, and kept at -70°C until use. One-half of each sample was homogenized using a Teflon pestle-glass tube homogenizer in 1 ml of homogenization buffer [50 mMHEPES (pH 7.9, 0.1% Triton X-100, 4 M EGTA, 10 mMEDTA, 0.32 Msucrose, 1 mMdithiothreito1, 75 p M pepstatin A, 0.1 mM leupeptin, and 0.1 mg/ml of aprotinin] with or without protein phosphatase inhibitors (1 5 mM sodium pyrophosphate, 100 mM 0-glycerophosphate, and 25 mM NaF). The homogenate was centrifuged at 400,000 g a t 4°C for I5 min. The pellet (particulate fraction) was suspended in 1 ml of homogenization buffer.
Kinase assays The standard assay system is essentially the same as that of Fukunaga et al. (1989) containing, in 25 pl, the following constituents: 50 mM HEPES (pH 7.9, I0 mM magnesium (3,000-5,000 cpm/pmol), 1 m d acetate, 0.1 mM [T-~~PIATP ml of bovine serum albumin, 40 p M syntide-2, and 5 p M PKI-tide. The assay for total CaM-kinase I1 contained 1 mM CaCI2and 6 pMCaM, whereas for Ca'+-independent activity it contained 1 mM EGTA. When total protein kinase C was quantified, the assay contained 1 mM CaQ, 3 p A 4 mastoparan, 50 pgml of phosphatidylserine, 5 pg/ml of 1,2-diolein, and no CaM or bovine serum albumin. The assay for cyclic AMP-dependent protein kinase contained 1 mM EGTA, Mcyclic AMP, 67 pMKemptide, and 50 Mpotassium phosphate buffer (pH 6.8) in place of HEPES. We omitted PKI-tide from the reaction mixture. Results are given as mean k SEM values. Tests of the statistical significance of differences were performed by using the t-EASE program (IS1 Software).
Immunoprecipitation The purified CaM-kinase I1 (1.2 pg) and the supernatant from the control homogenate (12.6 pg) or the ischemia homogenate (24 ps),which had been homogenized with protein phosphatase inhibitors, were incubated with antibody to CaM-kinase I1 (10 pg of IgG fraction), antibody to CaMkinase (28 1-309) peptide (5 pl), or nonimmune IgG in 0.1 ml of 40 mM HEPES (pH 7.5) and 10 mM 2-mercaptoethanol at 4°C for I h. Fifty microliters of protein A-Sepharose CL-4B suspension (50%vol/vol) was added, and samples were incubated at 4OC for 1 h. The supernatant was collected by centrifugation at 10,000 g at 4°C for 3 min and assayed for CaM-kinase I1 activity.
Other procedures Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed in 10%acrylamide by the method of Laemmli (1970). Molecular masses of proteins were calibrated with standards such as bovine carbonic anhydrase (30,000), ovalbumin (43,000), bovine serum albumin (67,000), and phosphorylase b (94,000). Immunoblotting after sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed by the method of Towbin et al. (1979) using '"I-protein A as described (Fukunaga et al., 1988). CaM overlay was performed by the procedures outlined in the Biomedical Tech-
J. Neurochem., Vol. 58, No. 3, 1992
H. YAMAMOTO ET AL.
1112
nologies manual, using the Vectastain ABC kit. Protein content was determined by the method of Bradford (1976) with bovine serum aibumin as the standard. No significant differences in the amount of control and ischemic proteins were observed ( p > 0.05).
Cortex Supernatant
IM
"1
Cortex Particulate
80
l a j n
RESULTS Effects of ischemia on protein kinase activities Bilateral occlusion of the carotid arteries in gerbils induces ischemia of the forebrain but not the cerebellum, thereby producing a nonischemic internal control. The activities of CaM-kinase I1 were determined in the high-speed supernatant (soluble kinase) and particulate fractions of cortex, hippocampus, and cerebellum. The assay conditions used for CaM-kinase I1 are highly selective for CaM-kinase 11, but not completely specific owing to some contribution by protein kinase C (Hashimoto and Soderling, 1987). As shown in Fig. 1, total CaM-kinase I1 activity (assayed in the presence of Ca2+/CaM)in the supernatants of cortex and hippocampus was almost completely lost after 10 min of ischemia ( p < 0.0005). After 10 min of ischemia, the carotid clamp was released, and reperfusion was continued for 1 2 h. Although there was no recovery of CaM-kinase I1 activity after 30 min, at 2 h of reperfusion a small restoration of CaM-kinase I1 was observed. CaM-kinase I1 in the particulate fraction was more resistant to ischemia. In cortex and hippocampus, there was a gradual loss of CaM-kinase I1 at 5 3 0 min of reperfusion. Loss of particulate CaM-kmase I1 was more pronounced in cortex (25% of control) than in hippocampus (54%). In cerebellum, there was no significant loss of CaM-lunase I1 in either the supernatant or particulate fractions except for the 2-h supernatant. In addition to assaying total kinase activity in the presence of Ca2+/CaM,we also assayed Ca2+-independent kinase activity in the presence of EGTA (Fig. 1). In cortex and hippocampus, the Ca'+-independent kinase activity represented 16-22% of the total kinase activity. It is well established that autophosphorylation of CaM-kinase I1 produces a Ca*'-independent species of the kinase that can be reversed by protein phosphatase treatment. The brain samples assayed in Fig. 1 were homogenized in buffer containing protein phosphatase inhibitors. Triplicate samples of control or ischemic (10 min) brain were homogenized in buffer either with or without phosphatase inhibitors. In the control supernatant samples homogenized without phosphatase inhibitors, the Ca2+-independent activity was only 2.0%, compared with 17% for samples homogenized with phosphatase inhibitors. There was no significant change of total CaM-kinase I1 activity by homogenization without phosphatase inhibitors. This result confirms that the Ca'+-independent activity was due to the autophosphorylated form of CaM-kinase 11. The 10-min ischemic supernatant samples from cortex and hippocampus (Fig. 1) had significantly lower Ca2+independent activity (
Ischemia-Induced Loss of Brain Calcium/CalmodulinDependent Protein Kinase I1 Hideyuki Yamamoto, Koji Fukunaga, *Kevin Lee, and Thomas R. Soderling Department of Molecular Physiology and Biophysics, Vanderbilt University, Nashville, Tennessee, and *Department of Neurosurgery, University of Virginia, Charlottesville, Virginia, U.S.A.
Abstract: Forebrain ischemia in gerbils, produced by brief bilateral carotid occlusion, induced the dramatic loss of Ca2+/ calmodulin-dependent protein kinase I1 (CaM-kinase 11) as determined by both kinase activity assays and western blot analysis. In cortex and hippocampus, cytosolic CaM-kinase I1 was completely lost within 2-5 min of ischemia. Particulate CaM-kinase I1 was more stable and decreased in level -40% after 10 min of ischemia followed by 2 h of reperfusion. CaMkinase I1 in cerebellum, which does not become ischemic, was not affected. The rapid loss of CaM-kinase I1 within 25 min was quite specific because cytosolic cyclic AMP kinase
and protein kinase C in hippocampus were not affected. These data indicate that cytosolic CaM-kinase I1 is one of the most rapidly degraded proteins after brief ischemia. Because the multifunctional CaM-kinase I1 has been implicated in the regulation of numerous neuronal functions, its loss may destine the neuronal cell for death. Key Words: Protein kinaseIschemia-Neuronal cell death-Hippocampus. Yamamoto H. et al. Ischemia-induced loss of brain calcium/calmodulindependent protein kinase 11. J. Neurochem. 58, 11 10-1 117 ( 1992).
Transient ischemic events such as stroke result in delayed neuronal cell death occumng over several days, and certain regions of the brain such as hippocampus appear particularly vulnerable (Kirino, I 982; Pulsinelli et al., 1982; Johansen et al., 1983; Suzuki et al., 1983). Several studies have suggested that prolonged elevation of intracellular calcium may initiate neuronal cell death, perhaps by activation of the Ca*+-dependent protease calpain (Choi and Rothman, 1990). Elevation of intracellular calcium levels could occur by the following multiple mechanisms: (a) inhibition of the Ca2+ATPase extrusion pump due to ischemic decreases in ATP levels (Kass and Lipton, 1986), (b) influx of extracellular calcium through voltage-gated or receptoroperated channels (Choi and Rothman, 1990),and/or (c) release of sequestered intracellular calcium. Benveniste et al. (1984) have reported that the extracellular content of glutamate is increased after ischemia, and the glutamate-gated N-methyl-paspartate ion channel has high permeability for calcium (Dingledine, 1983). This hypothesis is strengthened by the observations that
protease inhibitors ameliorate postischemic cell death after transient forebrain ischemia (Lee et al., 1991) and improve the recovery of synaptic transmission after hypoxia in hippocampal slices (Arai et al., 1990). Furthermore, levels of substrate proteins of calpain, such as spectrin (Seubert et al., 1989) and microtubule-associated protein 2 (Kitagawa et al., 1989),are decreased by ischemia in vivo, and spectrin levels are reduced by hypoxia in vitro (Arai et al., 1991). We wanted to investigate the effects of ischemia on the proteolysis of Ca2+/calmodulin (CaM)-dependent protein kinase I1 (CaM-kinase 11) for several reasons. First, CaM-lunase I1 is a substrate for calpain in vitro (Kwiatowski and King, 1989; Rich et al., 1990). Limited proteolysis by calpain generates a fully activated catalytic 30-kDa fragment of CaM-kinase I1 that, on prolonged treatment, is further degraded to inactive fragments (Rich et al., 1990). Second, CaM-kinase I1 is the most abundant protein kinase in brain, and immunohistochemical studies have demonstrated its concentration in hippocampal pyramidal neurons
Received May 14, 199I ; revised manuscript received August 13, 1991; accepted August 17, 1991. Address correspondence and reprint requests to Dr. T. R. Soderling at his present address: Vollum Institute for Advanced Biomedical Research, Oregon Health Sciences University, Portland, OR 97201, U.S.A. The present address of Dr. H. Yamamoto is Vollum Institute for
Advanced Biomedical Research, Oregon Health Sciences University, Portland, OR 9720 1, U.S.A. The present address of Dr. K. Fukunaga is Department of Pharmacology, Kumamoto University Medical School, Kumamoto 860, Japan. Abbreviations used: CaM, calmodulin; CaM-kinase 11, Ca2+/calmodulin-dependent protein kinase 11; IgG, immunoglobulin G.
1110
LOSS OF CaM-KINASE II IN ISCHEMIA (Fukunaga et al., 1988), where it constitutes 52% of protein (Erondu and Kennedy, 1985). Last, because CaM-kinase I1 phosphorylates multiple proteins and constitutes the major postsynaptic density protein at excitatory synapses in forebrain (reviewed by Nairn et al., 1985; Colbran et al., 1989), its destruction could be lethal for the cell. Taft et al. (1988) reported the reduction of Ca2+-dependentphosphorylation of some proteins after ischemia. These proteins could be substrates for or subunits of CaM-kinase 11. The studies reported in this article establish that CaM-kinase I1 in the cytosolic fraction from cortex and hippocampus is completely lost by periods of ischemia of 1 2 min. CaM-kinase I1 in the particulate fraction disappears more slowly, with a half-life of > l o min or -2 h after 10 rnin of occlusion. These data strongly suggest that very brief ischemia causes immediate proteolysis of soluble CaM-kinase 11, which would be deleterious to neuronal cell function. MATERIALS AND METHODS Materials Protein A-Sepharose CL4B was purchased from Pharmacia LKB Biotechnology; [y3*P]ATP and '251-protein A from Du Pont-New England Nuclear; P-cellulose papers from Whatman; the Immuno Pure Immobilization kit from Pierce; and biotinylated CaM from Biomedical Technologies. Other chemicals were of analytical grade. The CaM-kinase I1 substrate peptide syntide-2 was synthesized and prepared as described (Hashimoto and Soderling, 1987). Synthetic peptide inhibitors of cyclic AMP-dependent protein kinase (PKI-tide) and protein kinase C (peptide 19-36) and mastoparan were from Peninsula Laboratories. CaM-kinase 11 purified from rat brain (Hashimoto et al., 1987), rabbit polyclonal antibody [immunoglobulin G (IgG) fraction] to CaM-kinase 11 (Fukunaga et al., 1988), and bovine brain CaM (Gopalakrishna and Anderson, 1982) were kindly provided by Drs. Roger Colbran, Koji Fukunaga, and Debra Brickey, respectively, of our laboratory. Phosphatidylserine and 1,2-diolein were first mixed in chloroform and, after removal of chloroform, were suspended in water as previously described (Hashimoto and Soderling, 1987).
Antibodies against synthetic peptides from CaMkinase I1 Synthetic peptides corresponding to residues 132-1 46 [CaM-kinase (132- 146)] and residues 28 1-309 [CaM-kinase (28 1-309)] were synthesized and purified as previously described (Colbran et al., 1988). Immunization of the rabbits and preparation of the IgG fractions followed the method of Fukunaga et al. (1988). The antibodies were further purified from the IgG fraction by antigen-affinity chromatography. Immobilization of each peptide on Amino Link columns and affinity purification used the procedures outlined in the Pierce Chemical Co. manual. Antibodies were eluted using 5 MNaI and 1 mM sodium thiosulfate solution rather than 0.1 M glycine, pH 2.8. The purified antibodies were collected in 0.1 mg/ml of bovine serum albumin solution, dialyzed against phosphate-buffered saline, concentrated tenfold by Centricon filtration, and kept at -7OOC until use.
1111
Development of ischemia and preparation of brain extracts Adult gerbils (Meriones unguiculatus; 50-80 g of body weight) were subjected to bilateral carotid artery occlusion for 2-10 rnin (Leeet al., 1986; Seubert et al., 1989). At several postsurgical survival times, the gerbils were killed, and samples from the cerebellum, hippocampus, and cortex were taken, frozen on liquid N1, and kept at -70°C until use. One-half of each sample was homogenized using a Teflon pestle-glass tube homogenizer in 1 ml of homogenization buffer [50 mMHEPES (pH 7.9, 0.1% Triton X-100, 4 M EGTA, 10 mMEDTA, 0.32 Msucrose, 1 mMdithiothreito1, 75 p M pepstatin A, 0.1 mM leupeptin, and 0.1 mg/ml of aprotinin] with or without protein phosphatase inhibitors (1 5 mM sodium pyrophosphate, 100 mM 0-glycerophosphate, and 25 mM NaF). The homogenate was centrifuged at 400,000 g a t 4°C for I5 min. The pellet (particulate fraction) was suspended in 1 ml of homogenization buffer.
Kinase assays The standard assay system is essentially the same as that of Fukunaga et al. (1989) containing, in 25 pl, the following constituents: 50 mM HEPES (pH 7.9, I0 mM magnesium (3,000-5,000 cpm/pmol), 1 m d acetate, 0.1 mM [T-~~PIATP ml of bovine serum albumin, 40 p M syntide-2, and 5 p M PKI-tide. The assay for total CaM-kinase I1 contained 1 mM CaCI2and 6 pMCaM, whereas for Ca'+-independent activity it contained 1 mM EGTA. When total protein kinase C was quantified, the assay contained 1 mM CaQ, 3 p A 4 mastoparan, 50 pgml of phosphatidylserine, 5 pg/ml of 1,2-diolein, and no CaM or bovine serum albumin. The assay for cyclic AMP-dependent protein kinase contained 1 mM EGTA, Mcyclic AMP, 67 pMKemptide, and 50 Mpotassium phosphate buffer (pH 6.8) in place of HEPES. We omitted PKI-tide from the reaction mixture. Results are given as mean k SEM values. Tests of the statistical significance of differences were performed by using the t-EASE program (IS1 Software).
Immunoprecipitation The purified CaM-kinase I1 (1.2 pg) and the supernatant from the control homogenate (12.6 pg) or the ischemia homogenate (24 ps),which had been homogenized with protein phosphatase inhibitors, were incubated with antibody to CaM-kinase I1 (10 pg of IgG fraction), antibody to CaMkinase (28 1-309) peptide (5 pl), or nonimmune IgG in 0.1 ml of 40 mM HEPES (pH 7.5) and 10 mM 2-mercaptoethanol at 4°C for I h. Fifty microliters of protein A-Sepharose CL-4B suspension (50%vol/vol) was added, and samples were incubated at 4OC for 1 h. The supernatant was collected by centrifugation at 10,000 g at 4°C for 3 min and assayed for CaM-kinase I1 activity.
Other procedures Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed in 10%acrylamide by the method of Laemmli (1970). Molecular masses of proteins were calibrated with standards such as bovine carbonic anhydrase (30,000), ovalbumin (43,000), bovine serum albumin (67,000), and phosphorylase b (94,000). Immunoblotting after sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed by the method of Towbin et al. (1979) using '"I-protein A as described (Fukunaga et al., 1988). CaM overlay was performed by the procedures outlined in the Biomedical Tech-
J. Neurochem., Vol. 58, No. 3, 1992
H. YAMAMOTO ET AL.
1112
nologies manual, using the Vectastain ABC kit. Protein content was determined by the method of Bradford (1976) with bovine serum aibumin as the standard. No significant differences in the amount of control and ischemic proteins were observed ( p > 0.05).
Cortex Supernatant
IM
"1
Cortex Particulate
80
l a j n
RESULTS Effects of ischemia on protein kinase activities Bilateral occlusion of the carotid arteries in gerbils induces ischemia of the forebrain but not the cerebellum, thereby producing a nonischemic internal control. The activities of CaM-kinase I1 were determined in the high-speed supernatant (soluble kinase) and particulate fractions of cortex, hippocampus, and cerebellum. The assay conditions used for CaM-kinase I1 are highly selective for CaM-kinase 11, but not completely specific owing to some contribution by protein kinase C (Hashimoto and Soderling, 1987). As shown in Fig. 1, total CaM-kinase I1 activity (assayed in the presence of Ca2+/CaM)in the supernatants of cortex and hippocampus was almost completely lost after 10 min of ischemia ( p < 0.0005). After 10 min of ischemia, the carotid clamp was released, and reperfusion was continued for 1 2 h. Although there was no recovery of CaM-kinase I1 activity after 30 min, at 2 h of reperfusion a small restoration of CaM-kinase I1 was observed. CaM-kinase I1 in the particulate fraction was more resistant to ischemia. In cortex and hippocampus, there was a gradual loss of CaM-kinase I1 at 5 3 0 min of reperfusion. Loss of particulate CaM-kmase I1 was more pronounced in cortex (25% of control) than in hippocampus (54%). In cerebellum, there was no significant loss of CaM-lunase I1 in either the supernatant or particulate fractions except for the 2-h supernatant. In addition to assaying total kinase activity in the presence of Ca2+/CaM,we also assayed Ca2+-independent kinase activity in the presence of EGTA (Fig. 1). In cortex and hippocampus, the Ca'+-independent kinase activity represented 16-22% of the total kinase activity. It is well established that autophosphorylation of CaM-kinase I1 produces a Ca*'-independent species of the kinase that can be reversed by protein phosphatase treatment. The brain samples assayed in Fig. 1 were homogenized in buffer containing protein phosphatase inhibitors. Triplicate samples of control or ischemic (10 min) brain were homogenized in buffer either with or without phosphatase inhibitors. In the control supernatant samples homogenized without phosphatase inhibitors, the Ca2+-independent activity was only 2.0%, compared with 17% for samples homogenized with phosphatase inhibitors. There was no significant change of total CaM-kinase I1 activity by homogenization without phosphatase inhibitors. This result confirms that the Ca'+-independent activity was due to the autophosphorylated form of CaM-kinase 11. The 10-min ischemic supernatant samples from cortex and hippocampus (Fig. 1) had significantly lower Ca2+independent activity (
