Proline-directed protein kinase (p34cdc2/p58cyclin A) phosphorylates bovine neurofilaments

Phosphatidylinositol phosphates directly bind to neurofilament light chain (NF-L) for the regulation of NF-L self assembly

Phosphatidylinositol phosphates (PtdInsPs) are ubiquitous membrane phospholipids that play diverse roles in cell growth and differentiation. To clarify the regulation mechanism acting on neurofilament light chain (NF-L) self assembly, we examined the effects of various PtdInsPs on this process. We found that PtdInsPs, including PI(4,5)P((2)), directly bind to the positively charged Arg(54) of murine NF-L, and this binding promotes NF-L self assembly in vitro. Mutant NF-L (R53A/R54A) proteins lacking binding affinity to PtdInsPs did not have the same effect, but the mutant NF-L proteins showed greater self assembly than the wild-type in the absence of any PtdInsP. These results collectively suggest that Arg(54) plays a pivotal role in NF-L self assembly by binding with PtdInsPs.

In vitro assay of neurofilament light chain self-assembly using truncated mutants

Neurofilaments (NFs) are heteropolymers composed of light (NF-L), middle (NF-M), and heavy (NF-H) subunits, present in most neurons. NF-L polymerizes on its own to provide a scaffold on which regular NFs form via the cross-bridging of NF-M or NF-H. To clarify the mechanism of regulation of NF-L self-assembly, we developed an assay using truncated mutant NF-L fused to glutathione-S transferase (GST). Western immunoblotting data show that the GST-fused head-rod domains of NF-L are necessary and sufficient for detecting assembled NF-L. The levels of self-assembled NF-L subunits detected using GST fusion proteins were consistent with those detected by electron microscopy and turbidity assay. Our results collectively imply that GST-fused head-rod domains of NF-L are critical tools for analyzing NF-L self-assembly in vitro.

Alpha-ketoisocaproate increases the in vitro 32P incorporation into intermediate filaments in cerebral cortex of rats

In this study we investigated the effects of alpha-ketoisocaproic (KIC), alpha-ketoisovaleric (KIV) and alpha-keto-beta-methylvaleric (KMV) acids on the phosphorylation of intermediate filament (IF) proteins of cerebral cortex of rats. Tissue slices were incubated with [32P] orthophosphate in the presence or absence of the acids. The intermediate filament enriched cytoskeletal fraction was isolated and the radioactivity incorporated into neurofilament subunits, vimentin and glial fibrillary acidic protein was measured. Results demonstrated that KIC significantly increased phosphorylation of these proteins whereas the other acids had no effect. Experiments using protein kinase inhibitors indicated that the effect of KIC was mediated by Ca2+/calmodulin- and cAMP-dependent protein kinases. This study provides evidence that KIC, a key metabolite accumulating in maple syrup urine disease, increases phosphorylation of IF proteins.

Multiple subcellular localizations of PCTAIRE-1 in brain

We developed a selective antibody to a synthetic peptide corresponding to an N-terminal sequence of the PCTAIRE-1 protein. In rodent brain extracts it recognized only the protein doublet characteristic of PCTAIRE-1, and this signal is completely abolished by preincubation of the antibody with the immunopeptide. Immunolabeling experiments done with this PCTAIRE-1-specific antibody reveal that the protein is widely distributed in the rodent brain as are the mRNAs visualized using an antisense riboprobe corresponding to the entire PCTAIRE-1 open reading frame. Two types of PCTAIRE-1 protein localizations were observed: first a diffuse labeling of almost all brain regions, particularly intense in the molecular layer of the cerebellum and the mossy fiber region of the hippocampus, and second a spot-like localization in the nuclei of large neurons such as cerebellar Purkinje cells and pyramidal cells of the hippocampus. Colocalization with the B23 protein allows one to identify these compartments as nucleoli. Our results suggest a nucleolar function of PCTAIRE-1 in large neurons and a role in regions containing important granule cell projections.

Phosphorylation of the high molecular weight neurofilament protein (NF-H) by Cdk5 and p35

The high molecular weight neurofilament protein (NF-H) is highly phosphorylated in the axon. The phosphorylation sites have been identified as KSP (Lys-Ser-Pro) repeats in the tail domain of NF-H. These KSP sequences are present more than 50 times in the NF-H tail, and most of these sites are normally phosphorylated in vivo. These KSP sites can be further divided into two separate consensus sequences, KSPXK and KSPXY (where Y is not K). The extensive phosphorylation of NF-H has been proposed to play a critical role in the determination of axonal diameter. Recent studies have shown that Cdk5, a kinase related to the cell cycle-dependent kinase Cdc2, is expressed in the brain and associates with the cytoskeleton. In vitro phosphorylation studies have shown that Cdk5 in conjunction with its activator, p35, is able to phosphorylate histone H1, dephosphorylated NF-H, as well as a synthetic peptide with the repetitive KSP motif. We have cloned the cDNAs for rat Cdk5 and p35 by reverse transcription-polymerase chain reaction and cDNA library screening and studied the phosphorylation of NF-H both in vivo and in vitro. By transient transfection assays, we have shown that NF-H can only be extensively phosphorylated in the presence of both Cdk5 and p35. This phosphorylation can be inhibited by a Cdk5-dominant negative mutant, an observation which further supports that Cdk5 is a kinase that is able to phosphorylate NF-H. By immunoprecipitating Cdk5 and p35 from the transfected cells, we have been able to show that the KSPXK repeats are the preferred phosphorylation sites for Cdk5, while the KSPXY repeats are not directly phosphorylated by Cdk5 and p35.

A cdc2-like kinase distinct from cdk5 is associated with neurofilaments

An immunoprecipitation assay was used to identify protein kinases which are physically associated with neurofilaments (NF) in mouse brain extracts. Using this approach, we show that a cdc2-related kinase is associated with NF. The cdc2-related kinase was found to be distinct from cdk5 and the authentic cdc2 by a number of criteria. Firstly, it has a molecular mass on SDS-PAGE gels of 34 kDa, similar to that of cdc2, but differing from cdk5 (31 kDa). Secondly, it is not recognized by an antibody specific for cdk5. Thirdly, it is recognized by an antibody raised against the C-terminal region of authentic cdc2, but not by an antibody specific for the PSTAIRE motif. Using immunoblotting, we further show that the cdc2-related kinase copurifies with NF isolated from rat tissues. In vitro kinase assays further demonstrated that immunoprecipitated cdc2-related kinase phosphorylates recombinant NF-H protein. Phosphorylation of NF-H by the cdc2-like activity was not affected by 3 microM olomoucine but was inhibited by 10 microM of this kinase inhibitor. Phosphoamino acid analysis of in vitro phosphorylated NF-H indicates that the immunoprecipitated cdc2-related kinase phosphorylates serine residues.

Cyclin-dependent kinase 5 (Cdk5) and neuron-specific Cdk5 activators

While cyclin-dependent kinase 5 (Cdk5) is widely distributed in mammalian tissues and in cultured cell lines, Cdk5-associated kinase activity has been demonstrated only in mammalian brains. An active form of Cdk5, called neuronal cdc2-like kinase (Nclk) has been purified from mammalian brain and shown to be a heterodimer of Cdk5 and a 25 kDa protein, which is derived proteolytically from a 35 kDa brain and neuron-specific protein. The protein is essential for the kinase activity of Cdk5 and is therefore designated neuronal Cdk5 activator, p25/35Nck5a. Nclk appears to have important neuronal functions. The changes in Cdk5 and Nck5a expression appear to correlate with the terminal differentiation of neurons of the mouse embryonic brain. Transfection of cultured cortical neurons with dominant negative cdk5 mutants or Nck5a antisense DNA may reduce neurite growth, suggesting that Nclk plays an active role in neuron differentiation. A number of cytoskeletal proteins including neurofilament proteins, the neuron-specific microtubule associated protein tau, and the actin binding protein caldesmon are in vitro substrates of Nclk. Although Nck5a has cyclin-like activity, it shows minimal amino acid sequence identity to members of cyclin family proteins. The mechanism of activation of Cdk5 by Nck5a differs from that of cyclin activation of Cdks in that full Cdk5 kinase activity can be achieved in the absence of phosphorylation of Cdk5. An isoform of Nck5a, a 39 kDa protein has been cloned and shown to share extensive amino acid identity and the mechanism of Cdk5 activation with Nck5a. These proteins may represent a subfamily of Cdk activators distinct from cyclins.

Increased phosphorylation of neurofilament subunits in PC12 cells and rat dorsal root ganglion neurons treated with N-Acetyl-Leu-Leu-norleucinal

Treatment of PC12 cells or dorsal root ganglion neurons with the protease inhibitor, N-Acetyl-Leu-Leu-norleucinal, stimulated phosphorylation of the mid-sized and heavy neurofilament subunits and caused the heavy subunit in the perikarya of dorsal root ganglion neurons to become hyperphosphorylated. The closely related inhibitor, N-Acetyl-Leu-Leu-methioninal, did not produce a similar effect. Okadaic acid increased the phosphorylation state of the heavy neurofilament subunit in PC12 cells in a fashion similar to N-Acetyl-Leu-Leu-norleucinal and the effect of both compounds together was greater than for either one alone. There was no increase in cyclin-dependent kinase 5-immunoprecipitable histone H1 kinase activity in PC12 cells treated with N-Acetyl-Leu-Leu-norleucinal despite the presence of enzyme protein. The present study demonstrates that a protease inhibitor can induce the hyperphosphorylation of neurofilament subunits to a level normally seen only in axons. This suggests that perturbations in intracellular proteolysis may lead to the accumulation of phosphorylated neurofilament epitopes in neuronal perikarya in certain pathological states. The results also show that the carboxy-terminal tail domains of the two largest neurofilament subunits are phosphorylated even when cyclin dependent kinase 5 is inactive, indicating that other neuronal kinases are involved in the phosphorylation of Lys-Ser-Pro repeats.

Regulatory properties of neuronal cdc2-like kinase

Neuronal cdc2-like kinase, nclk, is a heterodimer of cyclin dependent protein kinase 5, cdk5, and a 25 kDa subunit derived from a novel, neuron-specific, 35 kDa protein: p35. The characterization and regulation of nclk will be summarized in this minireview. The activity of nclk appears to be governed by highly complex regulatory mechanisms including protein-protein interaction, protein phosphorylation and isoforms. The histone H1 kinase activity of nclk is absolutely dependent of the interaction between the 25 kDa subunit and the catalytic subunit, cdk5. In addition, nclk interacts with other cellular proteins to form macromolecular complexes. The kinase activity of nclk is inhibited in vitro by the phosphorylation reactions of a weel-like protein tyrosine kinase and a protein serine/threonine kinase from bovine thymus. Northern blot analysis has revealed the existence of two populations of p35 mRNA of 2 and 4 kb. A novel cDNA encoding a p35 homologous protein has been obtained from a human hippocampus library.

Phosphorylation and dephosphorylation of distinct isoforms of the heavy neurofilament protein NF-H

1. Previous immunohistochemical studies led to the suggestion that distinctly phosphorylated neurofilament isoforms exist in different types of neurons. We have recently examined this hypothesis by direct biochemical experiments, which revealed that the heavy neurofilament protein NF-H of bovine ventral root cholinergic neurons is more acidic and markedly more phosphorylated than that of bovine dorsal root neurons. 2. In the present study we employed this system to study the degree to which distinctly phosphorylated NF-H isoforms differ in the extents to which they can be phosphorylated and dephosphorylated in vitro. This was performed utilizing alkaline phosphatase and protein kinase PK40ERK, which is specific to serines of Lys-Ser-Pro (KSP) repeats. The results obtained reveal that: 3. The more extensively phosphorylated ventral root NF-H is dephosphorylated more rapidly than dorsal root NF-H. 4. Ventral root NF-H and dorsal root NF-H in their native form are both poor substrates of PK40ERK. 5. Following dephosphorylation, ventral root and dorsal root NF-H are phosphorylated extensively and differentially by this kinase. Under these conditions, PK40ERK catalyzes the incorporation of, respectively, 4.2 +/- 1.3 and 2.8 +/- 0.6 mol of phosphate per molecule of ventral root NF-H and dorsal root NF-H. The ratio of phosphates incorporated into ventral root NF-H to those incorporated into dorsal root NF-H is 1.46 +/- 0.17. 6. These findings support the hypothesis that different classes of neurons contain distinctly phosphorylated neurofilaments and show that ventral root and dorsal root neurons are a useful model system for studying the distinct characteristics of neurofilament phosphorylation in different types of neurons.

Neuronal cdc2-like kinase

Neurofilament proteins and the neuron-specific microtubule-associated protein tau are phosphorylated in vivo at sites conforming to the phosphorylation consensus motif of the cell-cycle-control protein kinase, p34cdc2-cyclin. Abnormalities in the phosphorylation of these proteins are associated with neurodegenerative disorders, such as amylotrophic lateral sclerosis and Alzheimer's disease. A cdc2-like kinase composed of cyclin-dependent kinase 5 (cdk5) and a brain-specific regulatory subunit is proposed to be responsible for the cdc2-like phosphorylation of these neuronal proteins.

Porcine brain neurofilament-H tail domain kinase: its identification as cdk5/p26 complex and comparison with cdc2/cyclin B kinase

Using dephosphorylated neurofilament (NF) proteins as substrates, the kinase with a higher activity for the dephosphorylated NF-H than the phosphorylated form of NF-H was searched for in the porcine brain extract. Most NF-H kinase activity in the brain extract pelleted with microtubules. The NF-H kinase purified from a high salt extract of the microtubule pellets was composed of cdk5 and a 26 kDa protein, a fragment of the 35 kDa regulatory subunit of cdk5. In contrast to the association of the active kinase with microtubules, each of uncomplexed cdk5 and the 35 kDa regulatory subunit was differently distributed in the supernatant fraction and the pellet, respectively, by ultracentrifugation of the brain extract. Dephosphorylated forms of NF-H and NF-M became reactive to antibodies recognizing in vivo phosphorylation sites (SMI31, 34, and 36, JJ31 and 51) by phosphorylation with cdk5/p26. cdk5/p26 showed similar enzymatic properties to p34cdc2/cyclin B kinase; the substrate specificity and inhibition by a p34cdc2 kinase specific inhibitor, butyrolactone I. However, p34cdc2/cyclin B kinase was distinguished from cdk5/p26 by its binding to p13suc1 protein and by its reactivity to anti-p34cdc2 antibodies. In spite of similar enzymatic properties of cdk5/p26 and p34cdc2/cyclin B kinase, cdk5/p26 did not display M-phase promoting activity when assayed with a cell-free system of Xenopus egg extract.

A brain-specific activator of cyclin-dependent kinase 5

Phosphorylation of the neurofilament proteins of high and medium relative molecular mass, as well as of the Alzheimer's tau protein, is thought to be catalysed by a protein kinase with Cdc2-like substrate specificity. We have purified a novel Cdc2-like kinase from bovine brain capable of phosphorylating both the neurofilament proteins and tau. The purified enzyme is a heterodimer of cyclin-dependent kinase 5 (Cdk5) and a novel regulatory subunit, p25 (ref. 8). When overexpressed and purified from Escherichia coli, p25 can activate Cdk5 in vitro. Unlike Cdk5, which is ubiquitously expressed in human tissue, the p25 transcript is expressed only in brain. A full-length complementary DNA clone showed that p25 is a truncated form of a larger protein precursor, p35, which seems to be the predominant form of the protein in crude brain extract. Cdk5/p35 is the first example of a Cdc2-like kinase with neuronal function.

cdc2-like kinase from rat spinal cord specifically phosphorylates KSPXK motifs in neurofilament proteins: isolation and characterization

A protein kinase that phosphorylates a specific KSP sequence [K(S/T)PXK], which is abundant in high molecular weight neurofilament (NF) proteins, was identified and isolated from rat spinal cord. Characterization of this enzyme activity revealed a close relationship with p34cdc2 kinase with respect to its molecular mass (32.5 kDa by SDS/PAGE) and substrate specificities. It could phosphorylate a synthetic peptide analog of the simian virus 40 large tumor antigen, reportedly a specific substrate for p34cdc2 kinase. Histone (H1) and peptide analogs of the KSP sequence present in the C-terminal end of rat and mouse neurofilament proteins were phosphorylated. This kinase did not phosphorylate alpha-casein and peptide substrates of other known second messenger-dependent or -independent kinases. Dephosphorylated rat NF protein NF-H was strongly phosphorylated by the purified enzyme; NF proteins NF-M and native NF-H, but not NF-L, were slightly phosphorylated. Studies on synthetic peptide analogs of KSP repeats with substitution of specific residues, known to be present in the C-terminal regions of NF-H, revealed a consensus sequence of X(S/T)PXK, characteristic of the p34cdc2 kinase substrate. On Western blots, the enzyme was immunoreactive with antibody against the C-terminal end of cdc2 kinase (mouse) and neuronal cdc2-like kinase from rat but not with an antibody against the conserved PSTAIRE region of the p34cdc2 kinase. The antibody against the C-terminal end of cdc2 kinase could immunoprecipitate (immunodeplete) the purified kinase activity. Since the adult nervous system is composed primarily of postmitotic cells, the present observations indicate a nonmitotic role for this cdc2-like kinase activity. The effective phosphorylation of NF-H by this kinase suggests a function in axonal structure.

Phosphorylation of tau by proline-directed protein kinase (p34cdc2/p58cyclin A) decreases tau-induced microtubule assembly and antibody SMI33 reactivity

Tau protein was evaluated as a substrate for a proline-directed protein kinase (p34cdc2/p58cyclin A) which recognizes the phosphorylation site motif X-Ser/Thr-Pro-X. The shortest human tau isoform, expressed as a recombinant protein, was phosphorylated to a stoichiometry of 2 mol phosphate/mol tau. Phosphoamino acid analysis revealed phosphorylation of both serine and threonine residues. Phosphorylation of recombinant tau resulted in a decreased ability to induce microtubule assembly but had no effect on the final extent of microtubule formation or on the rate of cold-induced microtubule disassembly. Phosphorylation of tau by the proline-directed protein kinase completely blocked immunoreactivity with antibody SMI33. Phosphorylation did not create the epitopes for the phosphate-dependent antibodies SMI31 or SMI34. Antibody SMI33 recognizes neurofibrillary tangles after treatment with alkaline phosphatase, suggesting that the proline-directed protein kinase may phosphorylate tau at sites that are phosphorylated in Alzheimer's disease.

Dynamics of the neuronal intermediate filaments

We have analyzed the dynamics of neuronal intermediate filaments in living neurons by using the method of photobleaching of fluorescently-labeled neurofilament L protein and immunoelectron microscopy of incorporation sites of biotinylated neurofilament L protein. Low-light-level imaging and photobleaching of growing axons of mouse sensory neurons did not affect the rate of either axonal growth or the addition of intermediate filament structures at the axon terminal, suggesting that any perturbations caused by these optical methods would be minimal. After laser photobleaching, recovery of fluorescence did occur slowly with a recovery half-time of 40 min. Furthermore, we observed a more rapid fluorescence recovery in growing axons than in quiescent ones, indicating a growth-dependent regulation of the turnover rate. Incorporation sites of biotin-labeled neurofilament L protein were localized as numerous discrete sites along the axon, and they slowly elongated to become continuous arrays 24 h after injection. Collectively, these results indicate that neuronal intermediate filaments in growing axons turn over within the small area of the axoplasm possibly by the mechanism of lateral and segmental incorporation of new subunits.

Application of synthetic phospho- and unphospho- peptides to identify phosphorylation sites in a subregion of the tau molecule, which is modified in Alzheimer's disease

Phospho- and unphospho- peptides were used to define the essential sequence for a tau epitope, which is recognized by Tau-1 antibody and phosphorylated in Alzheimer's disease (AD). The epitope was mapped within the amino acid residues 192-199 of tau and was phosphorylated by the p34cdc2/p58cyclin A proline directed kinase (PDPK), but not by purified mitogen activated protein kinase (p42mapk). Addition of phosphate to the last serine of the epitope was the most effective in abolishing the reactivity of the epitope to Tau-1 antibody. Our results suggest that one and possibly more members of the PDPK family may play a role in the pathogenesis of AD.

Neuronal cdc2-like kinase: a cdc2-related protein kinase with predominantly neuronal expression

Recent studies have shown that there exists a family of protein kinases structurally and functionally related to the yeast cell cycle regulatory kinase cdc2 [Meyerson, M., Faha, B., Su, L.-K., Harlow, E. & Tsai, L.-H. (1991) Cold Spring Harbor Symp. Quant. Biol. 56, 177-186 and Meyerson, M., Enders, G. H., Wu, C.-L., Su, L.-K., Gorka, C., Nelson, C., Harlow, E. & Tsai, L.-H. (1992) EMBO J. 11, 2909-2917]. Two members of cdc2 family, p34cdc2 (also named cdk1) and cdk2, have been identified in mammalian cells. cdk1 kinase regulates the progression from G2 to M phase, and cdk2 kinase has been proposed to regulate the progression from G1 to S phase. In this work, we have cloned and structurally characterized a third member of the cdc2 kinase family with 58% amino acid sequence identity to mouse cdk1 and 61% identity to human cdk2. We call this kinase neuronal cdc2-like kinase (nclk) because, in contrast to either cdk1 or cdk2, nclk is expressed at high levels in terminally differentiated neurons no longer in the cell cycle. Previous studies have shown [Hisanaga, S., Kusubata, M., Okumura, E. & Kishimoto, T. (1991) J. Biol. Chem. 266, 21798-21803 and Guan, R. J., Hall, F. L. & Cohlberg, J. A. (1992) J. Neurochem. 58, 1365-1371] that cdk1 kinase, but not other structurally defined protein kinases, could phosphorylate the repeated Lys-Ser-Pro (KSP) motifs found in mammalian high and middle molecular mass neurofilament subunits in vitro, but the precise molecular nature of the endogenous neuronal KSP kinase has remained undefined. The structural similarity of nclk to cdk1 kinase and its high level of expression in terminally differentiated neurons suggest that nclk may play a role in the phosphorylation of the neurofilament KSP repeats in vivo, a function distinct from cell cycle regulation.

Okadaic acid induces the rapid and reversible disruption of the neurofilament network in rat dorsal root ganglion neurons

Treatment of 15-17 day old dissociated cultures of rat dorsal root ganglia with 1 microM okadaic acid caused a reduction in the mobilities of neurofilament subunits on SDS-polyacrylamide gels, signifying an increase in their phosphorylation levels. When cultures were exposed to okadaic acid for 0.5 hrs and harvested in buffer containing Triton X-100, NF-H was nearly completely redistributed to the detergent- soluble fraction while NF-M and NF-L required a longer exposure to the drug before undergoing a similar shift. This redistribution of subunits corresponded with striking changes in the immunofluorescence staining pattern for neurofilaments. Upon removal of okadaic acid from the culture medium following a 0.5 hr treatment, NF-L and NF-M returned to the Triton X-100 insoluble fraction within 2 hrs while NF-H required 10 hrs for recovery.