Glycosylation of microtubule-associated protein tau: an abnormal posttranslational modification in Alzheimer's disease.

Alzheimer's disease (AD) is characterized by the presence of numerous neurons with neurofibrillary tangles of paired helical filaments (PHFs). The microtubule-associated protein tau in abnormally hyperphosphorylated form is the major protein subunit of the PHF. We now show that PHF tangles isolated from AD brains are glycosylated, whereas no glycan is detected in normal tau. Deglycosylation of PHF tangles by endoglycosidase F/N-glycosidase F converts them into bundles of straight filaments 2.5 +/- 0.5 nm in diameter, similar to those generated by the interaction of normal tau and abnormally hyperphosphorylated tau (AD P-tau). Deglycosylation plus dephosphorylation, but not deglycosylation alone, of AD P-tau and tau from PHF tangles restores their microtubule polymerization activity. Dephosphorylation of deglycosylated PHF tangles results in increased tau release. Thus, although the abnormal phosphorylation might promote aggregation of tau and inhibition of the assembly of microtubules, glycosylation appears to be responsible for the maintenance of the PHF structure.

Alzheimer's disease hyperphosphorylated tau sequesters normal tau into tangles of filaments and disassembles microtubules.

Microtubule-associated protein tau becomes abnormally hyperphosphorylated in Alzheimer's disease (AD) and accumulates as tangles of paired helical filaments in neurons undergoing degeneration. We now show that in solution normal tau associates with the AD hyperphosphorylated tau (AD P-tau) in a nonsaturable fashion, forming large tangles of filaments 3.3 +/- 0.7 nm in diameter. These tangles, which are not detected in identically treated normal tau or AD P-tau alone, are made up of filaments several microns in length and are labeled with tau antibodies. Dephosphorylation with alkaline phosphatase abolishes the ability of AD P-tau to aggregate with normal tau and prevents tangle formation. AD P-tau disassembles microtubules assembled from normal tau and tubulin. These data provide insight into how the hyperphosphorylation of tau might lead to the formation of the neurofibrillary tangles and the degeneration of the affected neurons in AD.

Cell surface immunoglobulin. VII. Synthesis, shedding, and secretion of immunoglobulin by lymphoid cells of germ-free mice.

Lymphoid cells from the spleen, lymph nodes, and thoracic duct of axenic and control mice were incubated with [(3)H]tyrosine and synthesis and secretion of protein and Ig studied. It was found that only IgM was synthesized by cells from axenic mice whereas cells from control mice also synthesized IgG. Splenocytes from both axenic and control mice had 8S IgM on their surface. Radiolabeled splenocytes from axenic mice were incubated to determine the kinetics of release of (125)I-labeled cell surface IgM and [(3)H]tyrosine-labeled IgM. Cell surface IgM was shed as 8S with an initial half-life of release of 5-8 h whereas [(3)H]tyrosine-labeled Ig was secreted as 19S with an initial half-life of 2-3 h. These findings suggest that two independent pathways are involved. It is suggested that small lymphocytes shed 8S IgM and plasma cells secrete 19S IgM. It was observed that lymphoid cells from axenic mice synthesize a higher proportion of IgM relative to total protein. Electron microscopic examination of splenocytes from such mice revealed a markedly higher proportion of plasma cells and a paucity of lymphoblasts compared to controls. It was suggested, therefore, that axenic mice lack a population of stimulated T cells which can induce a switch from IgM to IgG synthesis and which is capable of suppressing IgM synthesis. Lymphoid cells from axenic mice synthesize and secrete less protein that coprecipitates with antigen-antibody complexes.

Biological markers in Alzheimer's disease.

Alzheimer's disease (AD) is the most common type of dementia occurring in human population. The disorder is characterized clinically by memory loss and histopathologically by the presence of neurofibrillary tangles and senile plaques in patient's brain. Accuracy of the clinal diagnosis of AD is quite variable (-60 to 95 %), leaving a significant number of AD patients undiagnosed or falsely positively diagnosed. Therefore there is a requirement for biological markers, which would unambiguously discriminate living AD patients from other non-AD individuals. Until now a few diagnostic biomarkers for AD have been identified, which can be divided in two groups: protein markers and genetic markers. The most significant protein biomarkers are levels of tau proteins, ubiquitin and amyloid beta-peptides in cerebrospinal fluid (CSF). Among genetic AD markers, the most relevant are allelic variants of gene for apolipoprotein E and point mutations in genes coding for amyloid precursor protein and presenilin 1 and 2. Nevertheless, neither of recent biomarkers allow the ultimate AD diagnosis, because the disease is multifactorial and heterogenous. Identification of various subgroups of AD will help improvement in diagnoses and development of potent therapeutic drugs (Tab. 2, Fig. 2, Ref. 53).

Lewy bodies contain epitopes both shared and distinct from Alzheimer neurofibrillary tangles.

Most of the identified constituents of the filamentous inclusions characteristic of the neurodegenerative diseases of aging are derived from the cytoskeleton. This study was undertaken to define immunocytochemically the cytoskeletal constituents of the filamentous cytopathologic marker of idiopathic Parkinson disease, the Lewy body (LB). An array of antibodies specific to neurofilaments, tubulin, microtubule associated proteins (tau, MAP1 and MAP2) and Alzheimer neurofibrillary tangles (NFT) were used to immunostain sections containing LB. All the antibodies to tubulin, MAP1 and MAP2 and the majority of the antibodies to neurofilaments and NFT recognized LB. The two monoclonal antibodies to NFT that recognize LB also react with ubiquitin, which has been identified in NFT. The prominent NFT component, tau, is apparently not incorporated into LB. These findings suggest that the presence of tau might not be a prerequisite to the formation of abnormal filaments. Therefore, although LB contain elements of neurofilaments, microtubules and ubiquitin, as do other abnormal neuronal filaments, they are distinct in composition. These distinctive and shared features may provide useful insights regarding the mechanisms underlying the formation of filaments in LB as well as those of other neuronal inclusions.

Distribution, levels, and activity of glycogen synthase kinase-3 in the Alzheimer disease brain.

A number of studies have implicated a proline-directed protein kinase, glycogen synthase kinase-3 (GSK-3) in the hyperphosphorylation of tau in Alzheimer's disease (AD). Toward understanding the role of GSK-3 in the abnormal hyperphosphorylation of tau in AD we have found that GSK-3 is prominently present in neuronal cell bodies and their processes and co-localizes with neurofibrillary changes in AD brain. Furthermore, the levels of GSK-3 as determined by indirect ELISA are approximately 50% increased in the postsynaptosomal supernatant from AD brains as compared to the controls. However, no increase in GSK-3 enzyme activity was detected. In AD brain, with its reduced phosphatase activity, even normal levels of GSK-3 activity might be sufficient for the hyperphosphorylation of tau.

An antigenic profile of Lewy bodies: immunocytochemical indication for protein phosphorylation and ubiquitination.

An antigenic profile of subcortical and cortical Lewy bodies was determined in the presence or absence of neurofibrillary tangles in the same brain using antisera and monoclonal antibodies to various cytoskeletal elements as well as to determinants not present in the normal cytoskeleton. The cores of many Lewy bodies were strongly reactive with a monoclonal antibody to paired helical filaments which has been shown to recognize ubiquitin. This antibody also stained Marinesco bodies in the same tissue sections. Two monoclonal antibodies to phosphorylated epitopes of neurofilament proteins (SM I 31, SM I 34) stained the peripheries of about 40% of all discernable Lewy bodies on untreated paraffin sections. Reactivity with a monoclonal antibody to neurofilaments (SM I 33) appeared only after pretreatment of the sections with phosphatase. Lewy bodies did not bind antibodies to tau protein. Our results show that, as previously shown for neurofibrillary tangles, Lewy bodies also contain ubiquitin. The uncovering of neurofilament epitopes by treatment with phosphatase indicates that abnormal phosphorylation of cytoskeletal elements may play a role in the pathogenesis of the Lewy body.

Distribution of active glycogen synthase kinase 3beta (GSK-3beta) in brains staged for Alzheimer disease neurofibrillary changes.

Accumulation of paired helical filaments (PHFs) in neurofibrillary tangles, neuropil threads, and dystrophic neurites is one of the major neuropathological hallmarks of Alzheimer disease (AD). The principal protein subunit of PHFs is the abnormally hyperphosphorylated tau. Glycogen synthase kinase 3beta (GSK-3beta) is one of the candidate kinases involved in PHF-tau formation. To play a role in PHF-tau formation, it would be expected that GSK-3beta is active in tangle bearing neurons. In the present study, we investigated the regional and intracellular distributions of active and inactive forms of GSK-3beta in brains staged for neurofibrillary changes. We found that neurons with tangle-like inclusions positive for active, but not inactive, GSK-3beta appear initially in the Pre-alpha layer of the entorhinal cortex and extend to other brain regions, coincident with the sequence of the development of neurofibrillary changes. Active, but not inactive, GSK-3beta was found to initially accumulate in the cytoplasm of pretangle neurons. These data provide direct in situ evidence that is consistent with the involvement of GSK-3beta in PHF-tau formation.

A silver impregnation method for labeling both Alzheimer paired helical filaments and their polypeptides separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

The Gallyas silver impregnation which is specific to neurofibrillary changes of paired helical filaments (PHF) and 15 nm straight filaments, was adapted to stain polypeptides separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Both PHF and tau polypeptides were readily and consistently stained by the Gallyas stain. This technique stained PHF greater than tau greater than high-molecular-weight microtubule-associated polypeptides (MAPS). Tubulin was stained only weakly. Neurofilament triplet, ubiquitin, bovine serum albumin and histones were unstained. The staining of PHF and tau polypeptides by Gallyas silver stain is consistent with the presence of tau in PHF.

Immunoreactivity of neuronal lipofuscin with monoclonal antibodies to the amyloid beta-protein.

Monoclonal antibodies generated against a synthetic peptide corresponding to amino acids 1 to 24 of cerebrovascular amyloid beta-protein do not only stain amyloidotic blood vessels and the amyloid deposits of the (senile) neuritic plaques, but also the neuronal pigment lipofuscin. Staining of lipofuscin is observed in both cerebral and cerebellar cortices, subcortical nuclei as well as the brain stem, and is identical in Alzheimer and normal control brain. Western blots of a lipofuscin enriched fraction show an anti-beta-protein reactive polypeptide migrating at approximately 31 kDa position on SDS-polyacrylamide gel electrophoresis. These results suggest that this polypeptide is associated with lipofuscin and is most likely derived from the predicted amyloid precursor protein. This implicates that, unlike in Alzheimer's disease where this protein is also processed extraneuronally in a manner to release an amyloid fiber forming fragment, the end point of its processing in the nerve cell seems to accumulate on a lipopigment characteristic for normal aging.

Ubiquitination and abnormal phosphorylation of paired helical filaments in Alzheimer's disease.

The most characteristic cellular change in Alzheimer's disease is the accumulation of aberrant filaments, the paired helical filaments (PHF), in the affected neurons. There is growing evidence from a number of laboratories that dementia correlates better with the accumulation of PHF than of the extracellular amyloid, the second major lesion of Alzheimer's disease. PHF are both morphologically and biochemically unlike any of the normal neurofibrils. The major polypeptides in isolated PHF are microtubule-associated protein tau. Tau in PHF is phosphorylated differently from tau in microtubules. This abnormal phosphorylation of tau in PHF occurs at several sites. The accumulation of abnormally phosphorylated tau in the affected neurons in Alzheimer's disease brain precedes both the formation and the ubiquitination of the neurofibrillary tangles. In Alzheimer's disease brain, tubulin is assembly competent, but the in vitro assembly of microtubules is not observed. In vitro, the phosphate groups in PHF are less accessible than those of tau to alkaline phosphatase. The in vitro dephosphorylated PHF polypeptides stimulate microtubule assembly from bovine tubulin. It is hypothesized that a defect in the protein phosphorylation/dephosphorylation system is one of the earliest events in the cytoskeletal pathology in Alzheimer's disease. Production of nonfunctional tau by its phosphorylation and its polymerization into PHF most probably contributes to a microtubule assembly defect, and consequently, to a compromise in both axoplasmic flow and neuronal function. Index Entries: Alzheimer's disease; mechanisms of neuronal degeneration; neurofibrillary changes; paired helical filaments: biochemistry; microtubule-associated protein tau; abnormal phosphorylation; ubiquitination; microtubule assembly; axoplasmic flow; protein phosphorylation/dephosphorylation.

Mechanism of neurofibrillary degeneration in Alzheimer's disease.

Neurofibrillary degeneration associated with the formation of intraneuronal neurofibrillary tangles of paired helical filaments (PHF) and 2.1 nm tau filaments is one of the most characteristic brain lesions of Alzheimer's disease. The major polypeptides of PHF are the microtubule associated protein tau. tau in PHF is present in abnormally phosphorylated forms. In addition to the PHF, the abnormal tau is present in soluble non-PHF form in the Alzheimer's disease brain. The level of tau in Alzheimer's disease neocortex is severalfold higher than in aged control brain, and this increase is in the form of the abnormally phosphorylated protein. The abnormally phosphorylated tau does not promote the assembly of tubulin into microtubules in vitro, and it inhibits the normal tau-stimulated microtubule assembly. After in vitro dephosphorylation both PHF and non-PHF abnormal tau stimulate the assembly of tubulin into microtubules. The activities of phosphoseryl/phosphothreonyl protein phosphatase 2A and nonreceptor phosphotyrosyl phosphatase(s) are decreased in AD brain. It is suggested that 1. A defect(s) in the protein phosphorylation/dephosphorylation system is one of the early events in the neurofibrillary pathology in AD; 2. A decrease in protein phosphatase activities, at least in part, allows the hyperphosphorylation of tau; and 3. Abnormal phosphorylation and polymerization of tau into PHF most probably lead to a breakdown of the microtubule system and consequently to neuronal degeneration.

Hyperphosphorylated tau in SY5Y cells: similarities and dissimilarities to abnormally hyperphosphorylated tau from Alzheimer disease brain.

Unlike normal tau, abnormally hyperphosphorylated tau (AD P-tau) from Alzheimer disease (AD) does not promote but instead inhibits microtubule assembly and disrupts already formed microtubules. Tau in the human neuroblastoma cell line SH-SY5Y is hyperphosphorylated at several of the same sites as AD P-tau, and accumulates in the cell body without any association to the cellular microtubule network. The aim of the present study was to elucidate why the SY5Y tau does not affect the viability of the cells. We found that, like AD P-tau, SY5Y tau because of hyperphosphorylation does not bind to microtubules and inhibits the tau-promoted assembly of microtubules. However, the tau/HMW MAP ratio is about 10 times less in SY5Y cells than in AD brain. These findings suggest that the hyperphosphorylated tau from SY5Y cells has similar biological characteristics as AD P-tau from AD brain, but is not lethal to the SY5Y cells because of its low tau/HMW MAP ratio.

Levels of normal and abnormally phosphorylated tau in different cellular and regional compartments of Alzheimer disease and control brains.

Microtubule associated protein tau is abnormally phosphorylated in Alzheimer disease (AD) brain. In the present study we investigated (i) whether tau is axonal or both axonal and somatodendritic, (ii) whether tau is a marker of Alzheimer neurofibrillary pathology, and (iii) whether the levels of tau in the cytosol (100,000 x g supernate) from AD brain are altered. Frozen autopsied tissue from 20 AD, 17 normal aged and 15 neurological control cases obtained 3-8 h postmortem were analyzed. Levels of normal, total, and abnormally phosphorylated tau were determined by a radioimmunoslot-blot assay using mAb Tau-1 as the primary antibody. Both frontal gray matter homogenate and cytosol from normal brains had 30-45% higher levels of normal tau than the corresponding fractions from the white matter. In AD frontal and temporal cortices, the total tau levels were 6- to 7-fold higher than in cerebellar cortex (P < 0.01 and P < 0.02). Furthermore, tau levels of cerebellar cortex, an area of the brain unaffected with Alzheimer neurofibrillary changes, were indistinguishable between AD and control groups. The levels of normal tau in cytosol from both frontal gray and white matters in AD were reduced by approximately 40% (P < 0.05). The levels of total tau in AD frontal and temporal cortex were 4- to 5-fold higher than in the corresponding tissue from control cases (P < 0.01) and this increase was in the form of abnormally phosphorylated tau. These studies suggest (i) that there is probably at least as much tau in the somatodendritic compartment as in the axonal compartment, (ii) that the abnormally phosphorylated tau is a biochemical marker of the neurofibrillary pathology in AD, and (iii) that the levels of normal tau are significantly reduced in the 100,000 x g brain supernate from AD cases.

Microtubule-associated protein tau. Identification of a novel peptide from bovine brain.

The microtubule-associated protein tau isolated from bovine brain was cleaved with CNBr and the 3 largest peptides of approx. 21, 19 and 18 kDa were obtained. Dephosphorylation of the CNBr digest of tau with alkaline phosphatase changed the electrophoretic mobility of these peptides to 19, 18 and 17 kDa. Amino acid sequencing of the total CNBr digest of tau revealed at least 3 sequences, two of which were highly homologous to previously published mouse and human tau sequences derived from cDNAs. A third amino acid sequence of 17 residues with heterogeneity at position 11 showed no homology with the cDNA-derived tau sequences. These studies suggest that the amino acid sequences of mammalian tau predicted from their cDNAs might be incomplete.

Rapid Alzheimer-like phosphorylation of tau by the synergistic actions of non-proline-dependent protein kinases and GSK-3.

Tau protein from Alzheimer disease (AD) brain is phosphorylated at eleven Ser/Thr-Pro and nine Ser/Thr-X sites. The former sites are phosphorylated by proline-dependent protein kinases (PDPKs), the latter by non-PDPKs. The identities of both the PDPKs and non-PDPKs involved in AD tau hyperphosphorylation are still to be established. In this study we have analyzed the interactions between a PDPK (GSK-3) and several non-PDPKs (A-kinase, C-kinase, CK-1, CaM kinase II) in the phosphorylation of one isoform (tau 39) of human tau. We found that the rate of phosphorylation of tau 39 by GSK-3 was increased several-fold if tau were first prephosphorylated by the non-PDPKs. Further, several Alzheimer-like epitopes in tau can be induced only slowly after phosphorylation of tau by GSK-3 alone. After a prephosphorylation of tau by the non-PDPKs, however, the rate of induction of these epitopes by GSK-3 is increased several-fold. These results suggest that one role of non-PDPK-catalyzed phosphorylation is the modulation of PDPK-catalyzed phosphorylation of tau in AD brain.

The regulation of phosphorylation of tau in SY5Y neuroblastoma cells: the role of protein phosphatases.

In Alzheimer disease brain the microtubule associated protein (MAP) tau is abnormally hyperphosphorylated. The role of protein phosphatases (PP) in the regulation of phosphorylation of tau was studied in undifferentiated SY5Y cells. In cells treated with 10 nM okadaic acid (OA), a PP-2A/PP-1 inhibitor, the PP-1 and -2A activities decreased by 60% and 100% respectively and the activities of MAPKs, cdc2 kinase and cdk5, but not of GSK-3, increased. OA increased the phosphorylation of tau at Thr-231/Ser-235 and Ser-3961404, but not at Ser-262/356 or Ser-199/202. An increase in tyrosinated/detyrosinated tubulin ratio, a decrease in the microtubule binding activities of tau, MAP1b and MAP2, and cell death were observed. Treatment with 1 microm taxol partially inhibited the cell death. These data suggest (1) that OA induced hyperphosphorylation of tau is probably the result of activated MAPK and cdks in addition to decreased PP-2A and PP-1 activities and (2) that in SY5Y cells the OA induced cell death is associated with a decrease in stable microtubules.

Dephosphorylation of microtubule-associated protein tau by protein phosphatase-1 and -2C and its implication in Alzheimer disease.

Microtubule-associated protein tau is abnormally hyperphosphorylated and forms the major protein subunit of paired helical filaments (PHF) in Alzheimer disease brains. The abnormally phosphorylated sites Ser-199, Ser-202, Ser-396 and Ser-404 but not Ser-46 and Ser-235 of Alzheimer tau were found to be dephosphorylated by protein phosphatase-1 and this dephosphorylation was activated by Mn2+. In contrast, protein phosphatase-2C did not dephosphorylate any of these sites. Both protein phosphatase-1 and -2C had high activities towards [32P]tau phosphorylated by cAMP-dependent protein kinase. These results suggest that both protein phosphatase-1 and -2C might be associated with normal phosphorylation state of tau, but only the former and not the latter phosphatase is involved in its abnormal phosphorylation in Alzheimer disease.

Alzheimer paired helical filaments. Restoration of the biological activity by dephosphorylation.

In a normal mature neuron, microtubule associated protein tau promotes the assembly of tubulin into microtubules and maintains the structure of microtubules. In Alzheimer disease brain, tau is abnormally hyperphosphorylated and is the major protein subunit of paired helical filaments (PHF). In the present study, the biological activity of tau in PHF and the effect of dephosphorylation on this activity were examined. PHF were isolated from Alzheimer disease brains and tau from the untreated or alkaline phosphatase-treated PHF was extracted by ultrasonication in microtubule assembly buffer. Tubulin was isolated by phosphocellulose chromatography of three cycled microtubules from bovine brain. PHF-tau did not promote assembly of bovine tubulin into microtubules whereas tau from the dephosphorylated PHF produced a robust microtubule assembly. These studies suggest (i) that in Alzheimer disease tau in PHF is functionally inactive because of abnormal phosphorylation and (ii) that the abnormally phosphorylated site(s) in PHF that inactivates PHF-tau is accessible to enzymatic dephosphorylation in vitro.

Role of protein phosphatase-2A and -1 in the regulation of GSK-3, cdk5 and cdc2 and the phosphorylation of tau in rat forebrain.

In Alzheimer disease brain the activities of protein phosphatase (PP)-2A and PP-1 are decreased and the microtubule-associated protein tau is abnormally hyperphosphorylated at several sites at serine/threonine. Employing rat forebrain slices kept metabolically active in oxygenated artificial CSF as a model system, we investigated the role of PP-2A/PP-1 in the regulation of some of the major abnormally hyperphosphorylated sites of tau and the protein kinases involved. Treatment of the brain slices with 1.0 microM okadaic acid inhibited approximately 65% of PP-2A and produced hyperphosphorylation of tau at Ser 198/199/202, Ser 396/404 and Ser 422. No significant changes in the activities of glycogen synthase kinase-3 (GSK-3) and cyclin dependent protein kinases cdk5 and cdc2 were observed. Calyculin A (0.1 microM) inhibited approximately 50% PP-1, approximately 20% PP-2A, 50% GSK-3 and approximately 30% cdk5 but neither inhibited the activity of cyclin AMP dependent protein kinase A (PKA) nor resulted in the hyperphosphorylation of tau at any of the above sites. Treatment of brain slices with 1 microM okadaic acid plus 0.1 microM calyculin A inhibited approximately 100% of both PP-2A and PP-1, approximately 80% of GSK-3, approximately 50% of cdk5 and approximately 30% of cdc2 but neither inhibited PKA nor resulted in the hyperphosphorylation of tau at any of the above sites. These studies suggest (i) that PP-1 upregulates the phosphorylation of tau at Ser 198/199/202 and Ser 396/404 indirectly by regulating the activities of GSK-3, cdk5 and cdc2 whereas PP-2A regulates the phosphorylation of tau directly by dephosphorylation at the above sites, and (ii) that a decrease in the PP-2A activity leads to abnormal hyperphosphorylation of tau at Ser 198/199/202, Ser 396/404 and Ser 422.