Proteolytic processing of human brain alpha spectrin (fodrin): identification of a hypersensitive site

The processing of brain spectrin (fodrin) by calcium-dependent proteases at the postsynaptic membrane has been postulated to be one of the central molecular mechanisms underlying long-term potentiation (LTP). The effect of such processing on the structure and function of brain spectrin, and on spectrin's ability to organize or otherwise regulate receptor function remains unclear. To address these issues, human and bovine brain spectrin were digested under mild conditions with several proteases, and the resulting cleavage fragments analyzed by 2-dimensional chymotryptic 125I peptide mapping. These studies identify an underlying protease-resistant domain structure reminiscent of, yet distinctly different from, human erythroid spectrin. More importantly, fodrin is unusual for the presence of a single, proteolytically hypersensitive site in the center of the alpha subunit, which is the favored site of action by many proteases, including the calcium-dependent neutral proteases. This proteolytically hypersensitive site is a unique feature of alpha nonerythroid spectrin since it is absent from human erythrocyte spectrin and appears to be the site at which the molecule is processed in vivo. In addition, on the basis of gel overlay techniques, it appears that the hypersensitive site is also the site at which calmodulin binds to the alpha-subunit in a calcium-dependent manner. These studies thus establish at the molecular level 2 calcium-dependent mechanisms by which brain spectrin function might be regulated and provide a conceptual and methodological framework for further investigation into the function of this important molecule.     

The effect of synapsin I phosphorylation upon binding of synaptic vesicles to spectrin.

We have previously demonstrated that brain spectrin is attached to small spherical synaptic vesicles via synapsin I. These studies utilized a novel microfiltration assay in which 125I-labelled synaptic vesicles were incubated with brain spectrin which was covalently attached to cellulosic membranes. In these studies purified dephosphosynapsin I was demonstrated to competitively inhibit the binding of the synaptic vesicles to the immobilized brain spectrin with a KI = 45 nM. In the current study we demonstrate that phosphorylation of synapsin I site 1 (0.74 mol Pi/mol synapsin I) with cAMP-dependent protein kinase and sites 2 and 3 (2.0 mol Pi/mol synapsin I) with Ca(2+)-calmodulin kinase II had little effect upon its interaction with brain spectrin. cAMP-dependent protein kinase phosphorylated synapsin I and Ca(2+)-calmodulin kinase II phosphorylated synapsin I both inhibited the binding of 125I-labelled synaptic vesicles to immobilized brain spectrin with a KI of 23 nM and 24 nM respectively. We conclude that phosphorylation of synapsin I does not down-regulate the interaction of synaptic vesicles with brain spectrin.     

Brain Spectrins 240/235 and 240/235E: Differential Expression During Development of Chicken Dorsal Root Ganglia in vivo and in vitro

Brain spectrin, a membrane-related cytoskeletal protein, exists as two isoforms. Brain spectrin 240/235 is localized preferentially in the perikaryon and axon of neuronal cells and brain spectrin 240/235E is found essentially in the neuronal soma and dendrites and in glia (Riederer et al., 1986, J. Cell Biol., 102, 2088 - 2097). The sensory neurons in dorsal root ganglia, devoid of any dendrites, make a good tool to investigate such differential expression of spectrin isoforms. In this study expression and localization of both brain spectrin isoforms were analysed during early chicken dorsal root ganglia development in vivo and in culture. Both isoforms appeared at embryonic day 6. Brain spectrin 240/235 exhibited a transient increase during embryonic development and was first expressed in ventrolateral neurons. In ganglion cells in situ and in culture this spectrin type showed a somato - axonal distribution pattern. In contrast, brain spectrin 240/235E slightly increased between E6 and E15 and remained practically unchanged. It was localized mainly in smaller neurons of the mediodorsal area as punctate staining in the cytoplasm, was restricted exclusively to the ganglion cell perikarya and was absent from axons both in situ and in culture. This study suggests that brain spectrin 240/235 may contribute towards outgrowth, elongation and maintenance of axonal processes and that brain spectrin 240/235E seems to be exclusively involved in the stabilization of the cytoarchitecture of cell bodies in a selected population of ganglion cells.     

The domain of brain beta-spectrin responsible for synaptic vesicle association is essential for synaptic transmission

We have examined the interaction between synapsin I, the major phosphoprotein on the membrane of small synaptic vesicles, and brain spectrin. Using recombinant peptides we have localized the synapsin I attachment site upon the beta-spectrin isoform betaSpIISigmaI to a region of 25 amino acids, residues 211 through 235. This segment is adjacent to the actin binding domain and is within the region of the betaSpIISigmaI that we previously predicted as a candidate synapsin I binding domain based upon sequence homology. We used differential centrifugation techniques to quantitatively assess the interaction of spectrin with synaptic vesicles. Using this assay, high affinity saturable binding of recombinant betaSpIISigmaI proteins was observed with synaptic vesicles. Binding was only observed when the 25 amino acid synapsin I binding site was included on the recombinant peptides. Further, we demonstrate that antibodies directed against 15 amino acids of the synapsin I binding domain specifically blocked synaptic transmission in cultured hippocampal neurons. Thus, the synapsin I attachment site on betaSpIISigmaI spectrin comprises a approximately 25 amino acid segment of the molecule and interaction of these two proteins is an essential step for the process of neurotransmission.     

The domain of brain β-spectrin responsible for synaptic vesicle association is essential for synaptic transmission

We have examined the interaction between synapsin I, the major phosphoprotein on the membrane of small synaptic vesicles, and brain spectrin. Using recombinant peptides we have localized the synapsin I attachment site upon the β-spectrin isoform βSpIIΣI to a region of 25 amino acids, residues 211 through 235. This segment is adjacent to the actin binding domain and is within the region of the βSpIIΣI that we previously predicted as a candidate synapsin I binding domain based upon sequence homology. We used differential centrifugation techniques to quantitatively assess the interaction of spectrin with synaptic vesicles. Using this assay, high affinity saturable binding of recombinant βSpIIΣI proteins was observed with synaptic vesicles. Binding was only observed when the 25 amino acid synapsin I binding site was included on the recombinant peptides. Further, we demonstrate that antibodies directed against 15 amino acids of the synapsin I binding domain specifically blocked synaptic transmission in cultured hippocampal neurons. Thus, the synapsin I attachment site on βSpIIΣI spectrin comprises a 25 amino acid segment of the molecule and interaction of these two proteins is an essential step for the process of neurotransmission.     

Seizure activity-induced changes in polyamine metabolism and neuronal pathology during the postnatal period in rat brain.

Systemic injection of kainic acid (KA) does not cause neuronal pathology in limbic structures in rat brain prior to postnatal day (PND) 21. The present study tested if the development of the pathogenic response is associated with the maturation of a link between seizure activity and polyamine metabolism. Pathology was assessed with histological techniques and with the binding of [3H]Ro5-4864, a ligand for the peripheral type benzodiazepine binding sites (PTBBS), a marker of glial cell proliferation. In agreement with previous results, peripherally administered kainate at doses sufficient to induce intense behavioral seizures produced a loss of Nissl staining in hippocampus after PND 21 but not at earlier ages. The pattern of neuronal damage observed after PND 21 resembled that found in adult animals: extensive losses of Nissl staining in area CA3 of hippocampus and in piriform cortex, more modest effects in CA1 and sparing of the granule cells of the dentate gyrus. Similarly, no increase in [3H]Ro5-4864 binding as a result of KA administration was observed in hippocampus and piriform cortex until PND 21. Ornithine decarboxylase (ODC) activity and putrescine levels were high in the neonatal brain and decreased to reach adult values by PND 21. KA-induced seizure activity did not significantly alter both variables until PND 21. After PND 21, ODC activity and putrescine levels markedly increased 16 h after KA-induced seizure activity in hippocampus and piriform cortex. The magnitude of the effects increased between PND 21 and PND 30, at which point the changes in both parameters were comparable to those found in adults. Polyamines stimulate the activity of the calcium-dependent proteases calpain in brain fractions and may increase calpain-mediated proteolysis in situ. In accord with this, kainate-induced breakdown of spectrin, a preferred substrate of calpain, measured 16 h after KA injection followed a developmental curve parallel to that for kainate-induced increases in putrescine levels. These results indicate that the onset of vulnerability to seizure activity triggered by kainic acid is correlated with the development of an ODC/polyamine response to the seizures and further support a critical role for the ODC/polyamine pathway in neuronal pathology following a variety of insults.     

Inhibition of calpain-mediated cell death by a novel peptide inhibitor

Calpains are calcium- and thiol-dependent proteases whose overactivation and degradation of various substrates have been implicated in a number of diseases and conditions such as cardiovascular dysfunction and ischemic stroke. With increasing evidence for calpain's role in cellular damage, the development of calpain inhibitors continues to be an important objective. Previously, we identified a highly specific calcium-dependent, calpain interacting peptide L-S-E-A-L, that showed homology to domains A and C of the only known endogenous inhibitor of calpains, calpastatin. This suggested that LSEAL had a calpain inhibitory function and synthetic LSEAL inhibited calpain I and II proteolysis of two calpain substrates, tau and alpha-synuclein. In the present study, we demonstrate that synthetic LSEAL is membrane permeable and is a potent inhibitor in two established models of calpain-mediated cell death using primary rat cortical neurons and SH-SY5Y neuroblastoma cells. In addition, we show that LSEAL inhibits calpain activity towards protein substrates as detected by an antibody to a calpain-specific breakdown product of spectrin. Taken together, these results suggest that LSEAL may represent a novel calpastatin mimetic with the potential for benefit in conditions of increased calpain activity such as stroke, traumatic brain injury or heart attack.     

Specific proteolytic cleavage of the myristoylated alanine-rich C kinase substrate between Asn 147 and Glu 148 also occurs in brain

The myristoylated alanine-rich C kinase substrate (MARCKS) is a major ubiquitous substrate of protein kinase C. The expression of the protein is regulated during cell cycle progression and cell proliferation. Specific proteolytic cleavage of the protein between Asn 147 and Glu 148 was described recently in cultured cells, and the corresponding proteolytic activity was observed in various tissue extracts except for brain. We purified a 40 kDa fragment of MARCKS from bovine brain that we characterized as the C-terminal specific fragment found in other tissues. The identification of the fragment was achieved by in vitro phosphorylation by protein kinase C, calcium-dependent interaction with calmodulin, mass spectrometric analysis, and N-terminal sequencing. These data suggest that specific proteolytic cleavage of MARCKS also occurs in brain and may be a general mechanism of down-regulation of the protein. J. Neurosci. Res. 48:259–263, 1997. © 1997 Wiley-Liss, Inc.     

Amelin and synapsin I are 4.1 related spectrin binding proteins in brain

How do synaptic vesicles move towards the presynaptic plasma membrane, fuse with that membrane, and release their contents during synaptic transmission? The answers to these questions at the molecular level are just beginning to be understood. Synapsin I is a neuron specific phosphoprotein that is associated with the cytoplasmic surface of synaptic vesicles. During synaptic transmission, the translocation of the synaptic vesicles to the presynaptic membrane of the neuron is thought to be mediated through changes in the phosphorylation state of synapsin I. It has been suggested that synapsin I is a spectrin binding protein related to the erythrocyte cytoskeletal protein 4.1, which binds to the terminal ends of the erythrocyte spectrin tetramer. The interaction of synapsin I (through brain spectrin) with the neuronal cytoskeleton may be essential for regulating the movement of synaptic vesicles towards the presynaptic plasma membrane. In addition, we have identified another protein in brain that is immunologically and structurally more closely related to erythrocyte 4.1 than is synapsin I. This protein, termed amelin, is localized in the cell body and dendrites of the neuron, whereas synapsin I is found exclusively in the synaptic terminals, suggesting that there is a family of erythrocyte 4.1 related proteins present in brain with distinct subcellular distribution and functions.     

Proteolysis of NR2B by calpain in the hippocampus of epileptic rats

Overactivation of N-methyl-D-aspartate receptors is known to mediate excitotoxicity due to excessive entry of calcium, leading to the activation of several calcium-dependent enzymes. Calpains are calcium-activated proteases that appear to play a role in excitotoxic neuronal death. Several cellular proteins are substrates for these proteases, particularly the N-methyl-D-aspartate receptor. Recently, cleavage of NR2B subunits has been implicated in excitotoxic neurodegeneration in ischemia. In this work, we investigated the proteolysis by calpains of NR2B subunits of the N-methyl-D-aspartate receptor in the hippocampus of epileptic rats. Our results show that cleaved forms of NR2B subunits are formed after status epilepticus, in the same areas of the hippocampus where calpain activation was detected by immunohistochemical staining of calpain-specific spectrin breakdown products.     

Calpain I activation is specifically related to excitatory amino acid induction of hippocampal damage

Sustained stimulation of receptors for excitatory amino acids leads to both activation of the calcium-dependent cysteine protease calpain I and to the death of receptive neurons. Here, we have examined the relationship between the calpain I activation and neurodegeneration. Calpain I activation was manifested as increased levels of the major proteolytic fragments of the calpain substrate spectrin, detected and quantified by immunoblotting. Intraventricular administration of the excitatory amino acids kainate or N-methyl-D-aspartate (NMDA) produced calpain I-mediated spectrin degradation and hippocampal neuronal loss. The NMDA antagonist 3-(2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid selectively blocked NMDA- but not kainate-induced protease activation and hippocampal damage. Temporally, spectrin degradation preceded the onset of pyramidal cell degeneration monitored by silver-impregnation histochemistry. Only those doses of kainate (0.15-1 microgram) or NMDA (40-80 micrograms) sufficient to cause hippocampal damage markedly increased spectrin breakdown. Both the neuronal damage and calpain I activation induced by kainate occurred primarily in area CA3. Degeneration of hippocampal neurons evoked by colchicine was not accompanied by calpain activation, indicating that proteolysis is not stimulated simply as a secondary response to neuronal destruction. Thus, a close correspondence exists between excitatory amino acid induction of neuronal degeneration and of calpain I-mediated spectrin degradation. The results suggest that calpain I may be an intracellular mediator of excitatory amino acid action, and further, they support the hypothesis that calcium influx and calpain I activation are obligatory events in the initiation of excitatory amino acid neurotoxicity.     

Lesions of entorhinal cortex produce a calpain-mediated degradation of brain spectrin in dentate gyrus. II. Anatomical studies

Lesions of the various afferents to the hippocampus have been widely used to investigate the mechanisms underlying growth and degeneration in adult mammalian CNS. It has been proposed that disturbances in intracellular calcium and activation of calcium-dependent proteases represent key steps in producing come of the consequences of the lesions. In this study, we show that lesions of the entorhinal cortex or of the commissural pathway result in profound changes in the distribution of brain spectrin. At 2 days after lesions of the entorhinal cortex, immunoreactivity to spectrin is markedly increased in the outer molecular layer (OML) of the dentate gyrus; conversely at 2 days after commissural lesions, immunoreactivity to the same antigen is increased in the inner molecular layer. The increase in immunoreactivity to spectrin varies with survival time after lesions of the entorhinal cortex. By 24 h post lesion, the increase is homogeneous across the OML, and becomes more intense by 48 h. Between 1 and 3 weeks the increase is much less than at 48 h and is concentrated at the inner border of the OML. Pretreatment of the animals with the calpain inhibitor leupeptin reduces the increase in spectrin immunoreactivity normally seen 48 h after the lesion of the entorhinal cortex. Changes in the pattern of immunoreactivity to GFAP are very different to that seen with spectrin antibodies and are consistent with the known modifications in astrocytes that follow lesions of hippocampal afferents.(ABSTRACT TRUNCATED AT 250 WORDS)     

Hypothyroidism selectively reduces the rate and amount of transport for specific SCb proteins in the hyt/hyt mouse optic nerve.

Thyroid hormone significantly affects molecular and neuroanatomical properties of the developing nervous system. Altered connectivity in hypothyroidism may reflect reductions in process growth, alterations in process maintenance, or changes in synaptogenesis or synaptic maintenance. These events are dependent on microtubules, neurofilaments, microfilaments, and associated molecular components. Reductions in delivery of microtubules and neurofilaments to the distal axon by slow component a (SCa) of axonal transport may contribute to the neuroanatomical abnormalities of hypothyroidism (Stein et al., J Neurosci Res 28:121-133, 1991). However, hypothyroidism might also affect the axon and synaptic connections by altering slow component b (SCb), which includes actin microfilaments and proteins that contribute to synaptic function, i.e., clathrin, HSC70 (clathrin uncoating ATPase), spectrin, and calmodulin. To determine the effect of hypothyroidism on SCb proteins, slow axonal transport was analyzed in optic nerves of hyt/hyt hypothyroid mice, which have severe primary hypothyroidism, and euthyroid control mice. Clathrin, spectrin, HSC70, and actin showed significant reductions in transport velocity in hyt/hyt optic nerves relative to euthyroid nerves, but the transport rate for calmodulin was less affected. However, the amount of calmodulin was significantly elevated in hyt/hyt nerve over euthyroid nerves. Hypothyroidism selectively reduces transport of SCb proteins, which are thought to play significant roles in synaptic function and in the growth cone. The effects of hypothyroidism on microtubules and neurofilaments combined with actions on SCb suggest that changes in neuronal function associated with reduced thyroid hormone during development and maturity (i.e., alterations in neuronal connectivity, nerve conduction, and synaptic function) may be mediated in part by effects on slow axonal transport.     

Identification and characterization of specific binding sites of [3H]spermidine in synaptic membranes of rat brain.

Synaptic membranes of rat brain contained specific binding sites of [3H]spermidine (SPD) that exhibited an inverse temperature dependency, structure selectivity, reversibility, saturability, low affinity and high density with an uneven distribution profile. The affinities were not significantly different from each other in the rodent brain, while the highest density was found in the medulla-pons among the central structures examined with progressively lower densities in the midbrain, striatum, cerebellum, hypothalamus, hippocampus and cerebral cortex. The binding was insensitive to digestion by various proteases and glycosidases but sensitive to potentiation by phospholipases. A clear correlation was seen between the abilities of several natural and synthetic polyamines to displace [3H]SPD binding and to potentiate [3H] (+)-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine binding to open cation channels associated with an N-methyl-D-aspartate (NMDA)-sensitive subclass of brain excitatory amino acid receptors. Treatment of brain membranes with deoxycholic acid resulted in a significant solubilization of [3H]SPD binding sites. Furthermore, [3H]SPD markedly associated with the acidic phospholipid phosphatidylserine irrespective of the presence of synaptic membranes in a manner sensitive to inhibition by a variety of calmodulin antagonists. These results suggest that endogenous polyamines may play a stimulatory role in neuronal responses mediated by the NMDA receptor ionophore complex through an interaction between their positive charges and negative charges of membranous phosphatidylserine in rat brain.     

Immunoelectron-microscopic localization of the 180 kD component of the neural cell adhesion molecule N-CAM in postsynaptic membranes

In order to investigate the expression of cell adhesion molecules in synapses, we have studied the localization of the neural cell adhesion molecule N-CAM in the cerebellum and hippocampus of adult mice by immunocytological and immunochemical methods. Of the three molecular components of N-CAM with relative molecular masses (Mr) of 120, 140, and 180 kD, N-CAM 120 is not detectable in synaptosomal membranes, whereas N-CAM 140 is expressed on both pre- and postsynaptic membranes and N-CAM 180 is restricted to postsynaptic sites, with localization of the N-CAM 180-specific epitope in postsynaptic densities. Specificity of immunoreactivity is indicated by the observation that antibodies to the neural cell adhesion molecule L1 do not label synaptic membranes, whereas antibodies to two major components of postsynaptic densities, actin and erythrocyte spectrin, react with synaptic structures. Interestingly, N-CAM 180 is only detectable in subpopulations of synapses in the intact tissue. Isolated synaptosomes, opened for unimpeded accessibility of antibody by hypoosmotic treatment, also reveal a partial expression of N-CAM 180 in that 67% are labeled by antibodies to N-CAM 180, while antibodies to actin and erythrocyte spectrin react with 95% and 88% of all synaptosomes, respectively. N-CAM 180 does not appear to be differentially expressed in synapses of a particular morphological type, but is detectable in all types of synapses in the cerebellum and hippocampus, except for mossy fiber synapses and synapses between basket and Purkinje cells, which are generally N-CAM 180-negative. Since N-CAM 180 has been shown to be characteristic of stabilized or stabilizing cell contacts, possibly by its association with the cytoskeleton-membrane linker protein spectrin (Pollerberg et al.: J. Cell Biol. 101:1921-1929, '85; Nature 324:462-465, '86; Cell Tissue Res. 250:227-236, '87), we would like to suggest N-CAM 180 plays an important role in determining the stability of contacts between pre- and postsynaptic membranes and state of synaptic activity.     

Identification of an amelin isoform located in axons

A new axonal isoform of amelin, an analogue of the erythrocyte spectrin binding protein termed protein 4.1 has been identified in mouse brain. This new isoform has a molecular weight of 93 kDa, and migrates to a more acidic pH (pH 7.5-8.0) than the previously described amelin E (pH 8.5) on two dimensional NEPHGE-SDS PAGE. The 93 kDa protein looks nearly identical to amelin E on two dimensional chymotryptic iodopeptide mapping, and both share partial homology with rbc protein 4.1. The new isoform is located in axons, and the soma of neurons in mouse cerebellum, while amelin E is located in neuronal soma and dendrites. The axonal amelin antibody detects a 97 kDa protein in embryonic tissue which diminishes during development; and a 93 kDa protein which is first seen at postnatal day 1 of mouse brain ontogeny, increasing constantly to its adult concentrations. This time course of expression is quite different than amelin E, which is present at embryonic day 15 and diminishes constantly reaching its lowest concentration in the adult brain. We hypothesize that axonal amelin and amelin E may play important roles in the interaction of brain spectrin(240/235) and brain spectrin(240/235E) with f-actin and neuronal membranes.     

Distribution of calpains I and II in rat brain

Calpains I and II are calcium-dependent proteases that have been implicated in several aspects of brain function, including neurofilament turnover, Wallerian degeneration, and excitatory synaptic transmission. In this study, specific affinity-purified antibodies against each of the enzymes were used to determine their cellular distribution in rat brain. Differences between the two were found throughout the brain, with calpain I being located primarily in neurons, whereas calpain II was more prominent in glial cells. In myelinated axons, calpain II was present at low levels but calpain I was not detectable. In all brain areas, both enzymes were concentrated in cell bodies, with lesser amounts in neuronal and glial processes. Calpain I was only detectable proximally in dendrites and was not found in spiny branchlets of either pyramidal or Purkinje cells. These results suggest that calpain II is the likely form of the enzyme involved in calcium-activated proteolytic phenomena in axons. They do not support the existence of a role for calpain at excitatory axospinous synapses.     

Synaptotagmin mutants Y311N and K326/327A alter the calcium dependence of neurotransmission

Synaptotagmin I, a calcium-binding synaptic vesicle protein, is thought to act as the calcium sensor for fast neurotransmission, but what synaptotagmin I does, upon binding calcium, to trigger exocytosis is still unknown. To begin to examine the role of synaptotagmin I's interactions with calcium-dependent binding partners, three mutant versions of synaptotagmin I reported to affect calcium-dependent self-oligomerization (Y311N, K327A, and K326/327A) were expressed in cultured mouse hippocampal neurons lacking endogenous synaptotagmin I, and effects on neurotransmission were evaluated by comparison with transmission rescued by wild-type synaptotagmin I. All three mutants reduced transmitter release. To separate effects on calcium binding from effects on calcium-dependent oligomerization, we measured the calcium dependence of exocytosis for two of the mutants. Both showed apparent calcium affinity much lower than wild-type, a reduction sufficient to account for the neurotransmission defects. We conclude that self-oligomerization is unlikely to play any significant role in triggering synaptic vesicle exocytosis.     

Temporal relation of calcium-calmodulin binding and neuronal damage after global ischemia in rats.

BACKGROUND AND PURPOSE: This study explores the temporal relation of the severity of ischemia and calcium-calmodulin binding in vulnerable and resistant brain regions in a commonly used model of global ischemia. METHODS: Immunohistochemical assay of free calmodulin unbound to calcium and light microscopic histological damage were measured in rats after 5, 10, or 20 minutes of global ischemia. RESULTS: After 24 hours of reperfusion, decreased calmodulin staining, representing increased calcium influx and calcium-calmodulin binding, correlated with increasing durations of ischemia across all brain regions. Based on a 4-point scale (4, extensive stain; 0, no staining), calmodulin staining after 5 minutes versus 10 minutes of ischemia was 3.2 versus 1.9, respectively (p less than 0.05) and after 10 minutes versus 20 minutes of ischemia was 1.9 versus 1.0, respectively (p less than 0.01). The CA1 region displayed the greatest sensitivity to ischemia. Similar but less dramatic results were seen after 2 hours of reperfusion. After 72 hours of reperfusion, histological damage closely correlated with calcium-calmodulin binding after variable durations of ischemia. A threshold of 10 minutes of ischemia was required to cause calcium-calmodulin binding and irreversible neuronal damage. Surviving neuronal populations showed recovery of calmodulin staining 7 days after ischemia, representing a return of free calmodulin and normal calcium homeostasis. CONCLUSIONS: These correlations between calcium-calmodulin binding, histological damage, and duration of ischemia support the causal role of calcium influx in global ischemic injury and suggest the need for very rapid intervention after ischemia if calcium-mediated damage is to be prevented.