Two types of brain calmodulin-dependent protein kinase II: morphological, biochemical and immunochemical properties

"Memory bytes" - molecular match for CaMKII phosphorylation encoding of microtubule lattices

Learning, memory and long-term potentiation (LTP) are supported by factors including post-synaptic calcium ion flux activating and transforming the hexagonal calcium-calmodulin kinase II (CaMKII) holoenzyme. Upon calcium-induced activation, up to six kinase domains extend upward, and up to six kinase domains extend downward from the CaMKII association domain, the fully activated holoenzyme resembling a robotic insect 20 nanometers in length. Each extended kinase domain can be phosphorylated, and able to phosphorylate other proteins, thus potentially further encoding synaptic information at intraneuronal molecular sites for memory storage, processing and distribution. Candidate sites for phosphorylation-encoded molecular memory include microtubules, cylindrical lattice polymers of the protein tubulin. Using molecular modeling, we find spatial dimensions and geometry of the six extended CaMKII kinase domains can precisely match those of microtubule hexagonal lattice neighborhoods (both A- and B-lattices), and show two feasible phosphorylation mechanisms. In one, phosphorylation sites (e.g., valine 208) on a CaMKII extended kinase domain interact with serine 444 on a C-terminal "tail" of tubulin. In the second, the CaMKII kinase domain unfurls, enabling phosphorylation sites to contact threonine and serine sites on the tubulin surface. We suggest sets of six CaMKII kinase domains phosphorylate hexagonal microtubule lattice neighborhoods collectively, e.g., conveying synaptic information as ordered arrays of six "bits", and thus a "byte", with (minimally) 2⁶ (64) possible bit states per CaMKII-microtubule interaction. We model two levels of interaction between CaMKII and microtubules, suggesting a testable framework for molecular memory encoding.

Calcium/calmodulin-dependent protein kinase II expression in motor neurons: effect of axotomy

Although Ca2+/calmodulin-dependent (CaM) protein kinase II isoforms are present in the nervous system in high amounts, many aspects of in vivo expression, localization, and function remain unexplored. During development, CaM kinase IIalpha and IIbeta are differentially expressed. Here, we examined CaM kinase II isoforms in Sprague-Dawley rat sciatic motor neurons before and after axotomy. We cut the L4-5 spinal nerves unilaterally and exposed the proximal nerve stumps to a fluoroprobe, to retrogradely label the neurons of origin. Anti-CaM kinase IIbeta antibody showed immunoreactivity in motor neurons, which decreased to low levels by 4 days after axotomy. We found a similar response by in situ hybridization with riboprobes. The decrease in expression of mRNA and protein was confined to fluorescent motor neurons. For CaM kinase IIalpha, in situ hybridization showed that the mRNA was in sciatic motor neurons, with a density unaffected by axotomy. However, these neurons were also enlarged, suggesting an up-regulation of expression. Northern blots confirmed an mRNA increase. We were unable to find CaM kinase IIalpha immunoreactivity before or after axotomy in sciatic motor neuron cell bodies, suggesting that CaM kinase IIalpha is in the axons or dendrites, or otherwise unavailable to the antibody. Using rats with crush lesions, we radiolabeled axonal proteins being synthesized in the cell body and used two-dimensional polyacrylamide gel electrophoresis with Western blots to identify CaM kinase IIalpha as a component of slow axonal transport. This differential regulation and expression of kinase isoforms suggests separate and unique intracellular roles. Because we find CaM kinase IIbeta down-regulates during axonal regrowth, its role in these neurons may be related to synaptic transmission. CaM kinase IIalpha appears to support axonal regrowth.

An intact N terminus of the gamma subunit is required for the Gbetagamma stimulation of rhodopsin phosphorylation by human beta-adrenergic receptor kinase-1 but not for kinase binding

Cleavage after lysine 32 in the Ggamma2 subtype and after lysine 36 in the Ggamma3 subtype of purified mixed brain Gbetagamma by endoproteinase Lys-C blocks Gbetagamma-mediated stimulation of phosphorylation of rhodopsin in urea-extracted rod outer segments by recombinant human beta-adrenergic receptor kinase (hbetaARK1) holoenzyme while hbetaARK1 binding to rod outer segments is partially affected. This treatment does not attenuate the binding of the treated Gbetagamma to C-terminal fragments of hbetaARK1 containing the pleckstrin homology domain. Lys-C proteolysis also does not alter the association of the Gbetagamma with phospholipids, its ability to support pertussis toxin-catalyzed Galphao/Galphai ADP-ribosylation, or its ability to inhibit forskolin-stimulated platelet adenylate cyclase. The Gbeta subunit remains noncovalently associated with the cleaved Ggamma fragments. Thus, in addition to recruiting hbetaARK1 to its receptor substrate, Ggamma contributes secondary and/or tertiary structural features to activate the kinase.

Ischemia-induced translocation of Ca2+/calmodulin-dependent protein kinase II: potential role in neuronal damage

The activities of Ca2+/calmodulin (CaM)-dependent, Ca2+/phospholipid-dependent, and cyclic AMP-dependent protein kinases (CaM-KII, PKC, and PKA, respectively) were determined in rat brains after global ischemia. Both CaM-KII and PKC activities were significantly depressed in both hippocampal and cerebral cortical regions of ischemic animals, whereas no change was detected in PKA activity. The loss of CaM-KII activity was more dramatic and more sustained than the loss of PKC activity and correlated with the duration of ischemia. These decreases in enzyme activity were found in both supernatant and pellet fractions from crude homogenates. When the supernatant and pellet were analyzed for the amount of CaM-KII 50-kDa protein, a significant decrease was detected in supernatant fractions that paralleled a gain in the amount of CaM-KII in the pellet. Thus, the loss of CaM-KII activity in the supernatant can be explained by translocation of the enzyme to the pellet. Whether inactivation of CaM-KII occurs during or after the enzyme translocates from the supernatant to the pellet is unknown. Our results indicate that loss in CaM-KII activity parallels neuronal damage associated with ischemia; down-regulation of CaM-KII activity coincided with translocation of the enzyme to the particulate fraction, and it is proposed that this may be, in fact, a mechanism for controlling excessive CaM-KII phosphorylation.

Developmental changes in phosphorylation of MAP-2 and synapsin I in cytosol and taxol polymerised microtubules from chicken brain

In cytosol, cyclic AMP stimulated phosphorylation of microtubule associated protein-2 (MAP-2) increased from 2 days to adult in proportion to the increase in the concentration of MAP-2. By contrast, the calmodulin stimulated phosphorylation of MAP-2 decreased in proportion to the decrease in the concentration of calmodulin stimulated protein kinase II (CMK II). Similarly, the cAMP stimulated phosphorylation of the site on synapsin I labeled by the cAMP stimulated protein kinase (PKA) changed little during development whereas the calcium/calmodulin stimulated phosphorylation of the CMK II site decreased dramatically in proportion to the decrease in the concentration of CMK II. The decrease in the concentration of CMK II which occurs in cytosol during synapse maturation was also observed in taxol polymerised microtubules and the effects of the change in the relative concentrations of CMK II and PKA on the phosphorylation of MAP-2 and synapsin I in this fraction were similar to that observed in the cytosol. These results are consistent with the hypothesis that the developmental changes in phosphorylation of endogenous substrates by PKA is controlled largely by changes in the concentration of those substrates, whereas the concentration of CMK II is limiting so that the developmental changes in the phosphorylation of endogenous substrates by CMK II are a function of the concentration of CMK II itself as well as the concentration of endogenous substrates. Some possible functional consequences of this during synapse maturation are discussed.

Developmental regulation of type II calcium/calmodulin-dependent kinase isoforms in rat cerebellum

Two distinct isoforms of a Type II calcium/calmodulin-dependent protein kinase were separated from high-speed supernates (cytosol) of rat neonatal [postnatal day 10 (P10)] and adult [postnatal day 40 (P40)] cerebellum using cation-exchange chromatography. The isoenzymes contained variable amounts of three subunits of apparent Mr's of 50 kDa (alpha), 58 kDa (beta'), and 60 kDa (beta). The specific activity of calmodulin-dependent kinase (CaM kinase II) in crude homogenates increased sixfold between P10 and P40 using exogenous MAP 2 as substrate. Cytosol from cerebellum at P40 contained a predominant isoform (approximately 40% of total cytosolic activity) with a 1:5 molar ratio of alpha:beta',beta subunits that eluted with 150 mM NaCl (designated 150) and a less abundant isoform (approximately 20% of total cytosolic activity) containing a 1:8 molar ratio of alpha:beta',beta subunits that eluted with 350 mM NaCl (designated 350). In neonatal cerebellum at P10, the relative abundance of the two isoforms was reversed such that approximately 50% of the cytosolic calmodulin-dependent kinase activity was recovered in the 350 isoform, whereas only 20% of the total cytosolic kinase activity was recovered in the 150 isoform. Previous studies indicate that cerebellar granule cells may contain an all beta',beta isoform of CaM kinase II that lacks alpha subunit. Thus, to assess the cell-specific localization of kinase isoforms within cerebellum, cytosol prepared from primary cultures of rat cerebellar granule cells was applied to cation-exchange chromatography and analyzed for calmodulin-dependent kinase activity. The cells contained both isoforms of the kinase that were present in fresh tissue suggesting that granule cell-enriched cultures express all three kinase subunits. The data demonstrate that rat cerebellum contains unique mixtures of CaM kinase II isoenzymes and that their expression is developmentally regulated.

Posthoc phosphorylation of proteins derived from ischemic rat hippocampus, striatum and neocortex

Disruption of the brain's protein phosphorylation system by ischemia may cause irreversible metabolic and structural alterations leading eventually to cell death. To examine the effect of ischemia on the phosphorylation state of brain proteins, tissue homogenates derived from the hippocampus, striatum and neocortex of normal rats and rats subjected to severe forebrain ischemia were phosphorylated with [gamma-32P]ATP. The phosphorylated proteins were separated by two-dimensional polyacrylamide gel electrophoresis and changes were assessed by autoradiography. Cerebral ischemia caused marked alterations of the phosphorylation state of many brain proteins; phosphorylation of some proteins was increased while phosphorylation of others was decreased. Despite differences in the sensitivity of the hippocampus, striatum and neocortex to ischemic injury the direction and approximate magnitude of protein phosphorylation changes caused by ischemia were similar in all three regions. Since the pattern of protein phosphorylation in the ischemia-vulnerable hippocampus was identical to that in the ischemia-resistant paramedian neocortex we conclude that abnormalities of protein phosphorylation may be necessary for ischemic injury to neurons but none are sufficient to explain the selective vulnerability of certain brain regions to ischemic damage.

Separation of isozymic forms of type II calcium/calmodulin-dependent protein kinase using cation-exchange chromatography

Three distinct isozymes of a type II calcium/calmodulin-dependent protein kinase (CaM kinase II) from rat forebrain cytosol were separated using S-Sepharose cation-exchange resin. About 90% of the applied kinase activity was recovered in three protein peaks. Each isozymic form of the kinase was purified 200-300 fold by chromatography on S-Sepharose, calmodulin-affinity and gel filtration resins. All 3 forms of CaM kinase II had apparent molecular masses of 650-700 kDa, but contained variable proportions of 50 kDa and 58-60 kDa subunits. The molar ratios of the 50 kDa/58-60 kDa kinase subunits in each holoenzyme were determined by protein staining and [125I]calmodulin-binding studies and were approximately: 6/1, 3/1 and 1/1. The isozyme containing a 3/1 ratio of subunits corresponds to the predominant form of CaM kinase II in forebrain representing 70-80% of the total activity in cytosol, whereas the other forms each represent 5-10% of the total cytosolic activity. The substrate specificities and time courses of substrate phosphorylation for the isozymes were comparable. These studies provide a basis to examine regional, subcellular, and developmental differences in the isozymic forms of CaM kinase II which may subserve different neuronal functions.

Calmodulin selectively modulates the guanylate cyclase activity by repressing the lipid phase separation temperature in the inner half of the bilayer of rat brain synaptosomal plasma membranes

The association of [125I-]calmodulin with rat brain synaptosomal plasma membranes, when incubated for 1 h at 25 degrees in the presence or in absence of 20 microM Ca2+, follows a sigmoid path with a Hill coefficient h = 1.79 +/- 0.12 and h = 1.72 +/- 0.11, respectively. The total association of calmodulin with the membrane increased approx. 60%-80% at all the range of calmodulin concentrations used in the presence of 20 microM Ca2+. A three fold increase of guanylate cyclase activity was shown in the presence of low concentrations of calmodulin (up to 10 nM); higher concentrations (up to 40 nM) however, led to a progressive inhibition of the enzyme activity with respect to maximal stimulation. Calmodulin increased the lipid fluidity of synaptosomal plasma membranes labeled with 1,6-diphenyl-1,3,5-hexatriene (DPH), as indicated by the steady-state fluorescence anisotropy [(ro/r)-1]-1. Arrhenius-type plots of [(ro/r)-1]-1 indicated that the lipid separation of the membrane at 22.7 +/- 1.2 degrees was perturbed by calmodulin such that the temperature was reduced to 16.3 +/- 0.9 degrees and 15.5 +/- 0.8 degrees in the absence or in the presence of 20 microM Ca2+. Arrhenius plots of guanylate cyclase and acetylcholinesterase activities exhibited break points at 26.7 +/- 1.4 degrees and 22.3 +/- 1.0 degrees in control synaptosomal plasma membranes, respectively. The break point for the guanylate cyclase was reduced to 16.3 +/- 0.9 degrees in calmodulin treated synaptosomal plasma membranes whereas that of acetylcholinesterase remained unaffected (21.1 +/- 0.9 degrees).(ABSTRACT TRUNCATED AT 250 WORDS)