Phosphate-Dependent Monoclonal Antibodies to Neurofilaments and Alzheimer Neurofibrillary Tangles Recognize a Synthetic Phosphopeptide

Five phosphate-dependent monoclonal antibodies to the neurofilament heavy polypeptide bound strongly to a phosphorylated synthetic peptide, which contains a single Lys-Ser-Pro sequence that occurs in human neurofilaments. Three of the antibodies label Alzheimer's disease neurofibrillary tangles and two do not, suggesting that in tangles an epitope similar to the peptide is available to some but not all of the antibodies. In addition, some antibodies were found to be more affected than others by enzymatic dephosphorylation of the antigen, but because all the antibodies bound the same synthetic phosphopeptide they do not bind to mutually exclusive phosphorylation sites. Instead the more phosphate-dependent antibodies might bind the phosphate group more directly, as suggested by their inhibition by inorganic phosphate and free phosphoserine.     

Epitopes that span the tau molecule are shared with paired helical filaments

Tau protein has been shown to be an integral component of Alzheimer paired helical filaments (PHF). However, the extent to which tau is incorporated into PHF has not been clear because the antibodies used to label PHF generally do not have precisely defined epitopes. Here we define the antigenic sites for five monoclonal antibodies that react with tau and cross-react with SDS-extracted neurofibrillary tangles. The reactive sites were determined by screening a lambda gt11 sublibrary expressing small fragments of the tau sequence. The mapped epitopes were found to span almost the entire length of tau, suggesting that PHF contains tau in its entirety or nearly in its entirety. One antibody was found to cross-react with microtubule-associated protein 2, implying some degree of homology between the two proteins.     

The structure, biochemical properties, and immunogenicity of neurofilament peripheral regions are determined by phosphorylation state

Treatment of freshly isolated, bovine neurofilaments with Escherichia coli alkaline phosphatase removes over 90% of the phosphate groups from serine residues of the Mr 200,000 and 150,000 polypeptide components (NF200 and NF150). Dephosphorylated NF200 and NF150 remain associated with filaments, but migrate in sodium dodecyl sulfate gels with reduced apparent molecular weights. Unusual migration appears to be due to modification at regions of these polypeptides that are peripheral to the neurofilament backbone as defined by limited chymotryptic digestion. Over 90 monoclonal antibodies recognizing epitopes located within the peripheral domain of native NF200 all show reduced affinity for dephosphorylated NF200. A single monoclonal antibody binds within the filament-associated domain of NF200 and its recognition of NF200 is unaffected upon treatment of neurofilaments with phosphatase. Around 50% of our monoclonal antibodies that bind NF150 monospecifically and at epitopes within its peripheral domain have reduced affinities for NF150 from phosphatase-treated filaments, while the remaining 50% bind native and dephosphorylated NF150 equally well. The smallest neurofilament component (NF70) contains few phosphate groups, most of which remain after treatment of neurofilaments with phosphatase. The resulting form of NF70 migrates normally in gels and its recognition by antibodies is unchanged. We conclude that phosphorylation modifies the structure of the two larger neurofilament polypeptides along domains that are peripheral to the filamentous backbone and that these effects are more pronounced for NF200 than for NF150.     

Neurofilament Monoclonal Antibodies RT97 and 8D8 Recognize Different Modified Epitopes in Paired Helical Filament-τ in Alzheimer's Disease

Abstract: Neurofibrillary tangles in Alzheimer's disease have been previously found to be labeled by some neurofilament antibodies that also recognize τ proteins. We have studied the reactivity of two such monoclonal antibodies, RT97 and 8D8, and of an anti-ubiquitin serum with the abnormal paired helical filaments (PHF)-τ (A68) polypeptides known to be the main component of the PHFs constituting the neurofibrillary tangles. 8D8 recognized the three major PHF-τ polypeptides, but RT97 reacted only with the two larger PHF-τ species. PHF-τ polypeptides were labeled by 8D8 and RT97 much more strongly than normal human τ and this labeling was decreased after alkaline phosphatase treatment. Anti-ubiquitin and anti-phosphotyrosine antibodies did not label PHF-τ polypeptides. The immunoreactivity of proteolytic fragments of PHF-τ polypeptides was studied with RT97, 8D8, and a panel of τ antibodies. The epitope for 8D8 on PHF-τ was localized between amino acids 222 and 427 in the carboxyl half of τ. The RT97 epitope on PHF-τ was localized in the amino domain of τ, probably in the 29-amino-acid insertion (insert 1) found towards the amino terminus of some τ isoforms. These results show that the basis for the labeling of neurofibrillary tangles by antibodies 8D8 and RT97 to neurofilament is their ability to react with PHF-τ polypeptides by recognizing sites specifically modified on PHF-τ, including a site specific to some τ isoforms.     

Immunochemical Properties of Ubiquitin Conjugates in the Paired Helical Filaments of Alzheimer Disease

Immunocytochemical and peptide sequencing studies indicate that the regulatory protein ubiquitin (Ub) is incorporated into the paired helical filaments (PHF) of Alzheimer disease. In this study, we showed that some antibodies raised to PHF recognize epitopes of Ub. Analysis of the Ub sequences recognized by the antibodies raised to PHF, along with the known specificity of several monoclonal antibodies raised to artificial Ub conjugates, indicates the immunochemical representation of Ub residues 34-76 in PHF. The Ub epitopes recognized by antibodies raised to PHF are distinct from those recognized by antibodies raised to artificial Ub conjugates in two respects. First, antibodies that are raised to PHF and that recognize Ub react with PHF equally, whether denatured or not, whereas those raised to artificial Ub conjugates show greater reaction after denaturation. Second, mapping of the epitopes recognized by two monoclonal antibodies to PHF onto Ub indicates a distinction in the Ub residues recognized, compared with monoclonal antibodies raised to artificial Ub conjugates. The proximity of their epitopes to the site of conjugation, as well as their affinity for PHF polypeptides, suggests that the PHF antibodies that recognize Ub may be directed specifically to Ub epitopes defined by the protein conjugated to Ub.     

Are Alzheimer neurofibrillary tangles insoluble polymers?

The two methods currently available for the bulk isolation of Alzheimer tangles of paired helical filaments (PHF) are based on a brief treatment of a neuronal-enriched preparation with sodium dodecyl sulfate (SDS) (Method I) and on heating of whole brain homogenate with SDS and beta-mercaptoethanol (Method II). PHF were isolated from the same Alzheimer brain by these two methods, subjected to SDS-polyacrylamide gel electrophoresis and immuno-labelled with monoclonal antibodies to PHF after transferring from the gel to nitrocellulose paper. The PHF isolated by method I revealed the presence of 45 kilodalton to 62 kilodalton PHF polypeptides, whereas the PHF isolated by method II were excluded from the gel. However, PHF isolated by both methods were digested with proteinase-K, though the degradation of PHF of method I was considerably more rapid than that of PHF isolated by method II. These findings should establish that the solubility of PHF might depend on the methods employed for their isolation and that they might not be insoluble polymers of covalently crosslinked polypeptides which accumulate irreversibly in the brain of patients with Alzheimer disease.     

Carbonyl-Related Posttranslational Modification of Neurofilament Protein in the Neurofibrillary Pathology of Alzheimer's Disease

Abstract: We present the first evidence for carbonyl-related posttranslational modifications of neurofilaments in the neurofibrillary pathology of Alzheimer's disease (AD). Two distinct monoclonal antibodies that consistently labeled neurofibrillary tangles (NFTs), neuropil threads, and granulovacuolar degeneration in sections of AD tissue also labeled the neurofilaments within axons of the white matter following modification by reducing sugars, glutaraldehyde, formaldehyde, or malondialdehyde. The epitope recognized by these two antibodies shows a strict dependency for carbonyl modification of the neurofilament heavy subunit. The in vivo occurrence of this neurofilament modification in the neurofibrillary pathology of AD suggests that carbonyl modification is associated with a generalized cytoskeletal abnormality that may be critical in the pathogenesis of neurofibrillary pathology. Furthermore, the data presented here support the idea that extensive posttranslational modifications, including oxidative stress-type mechanisms, through the formation of cross-links, might account for the biochemical properties of NFTs and their resistance to degradation in vivo.     

Proteolytic fragments of Alzheimer's paired helical filaments

We found that at least a part of Alzheimer's paired helical filaments (PHF) was cleaved by proteases to release three major polypeptides, the apparent molecular weights of which were 10K, 26K, and 36K daltons. These proteolytic fragments were strongly labeled with the antibodies specific to PHF. Absorption with highly purified PHF eliminated this labeling. The monospecific antibodies bound to the 10K daltons protein, the most intensely immunolabeled one, stained isolated neurofibrillary tangles composed of PHF. From these observations, we conclude that these polypeptides released by proteases, especially the 10K daltons protein, are derived from PHF themselves.     

Neurofilament Proteins in Cultured Chromaffin Cells

Antibodies were raised against the 200-kd, 145-kd, and 68-kd subunits of a rat neurofilament preparation. Immunoblots showed that each antibody was specific for its antigen and that it did not cross-react with any of the two other neurofilament polypeptides. Use of the three antibody preparations to stain bovine chromaffin cells in culture by the indirect immunofluorescence technique indicated that the three neurofilament polypeptides are present in chromaffin cells maintained in culture for 3 or 7 days. The three anti-neurofilament antibodies labelled the cells in a similar pattern: very thin filaments specifically localized around the nucleus were observed whereas neurites and growth cones, developed by cultured chromaffin cells, were generally not stained. Some fibroblasts were present in our cultures but they were never stained by any of the neurofilament antibodies. This indicated that the antibodies used do not react with vimentin, the major intermediate filament protein found in fibroblasts. The three neurofilament antibodies were also used to immunoprecipitate specifically three proteins of molecular weights 210 kd, 160 kd, 70 kd from solubilized extracts of cultured chromaffin cells that were radiolabelled with [35S]methionine. These proteins correspond in molecular weight to the neurofilament triplet found in bovine brain. Finally, the presence of neurofilaments in freshly isolated chromaffin cells was tested by immunoblotting using the 68-kd antibody. A 70-kd protein was specifically stained by this antibody, suggesting that neurofilaments are not only present in cultured chromaffin cells but also in the adrenal gland in vivo. It is concluded from these results that chromaffin cells contain completely assembled neurofilaments. This additional neuronal property again illustrates that chromaffin cells are closely related to neurons and therefore represent an attractive model system for the study of functional aspects of adrenergic neurons.     

Phosphorylation on carboxyl terminus domains of neurofilament proteins in retinal ganglion cell neurons in vivo: influences on regional neurofilament accumulation, interneurofilament spacing, and axon caliber

The high molecular weight subunits of neurofilaments, NF-H and NF-M, have distinctively long carboxyl-terminal domains that become highly phosphorylated after newly formed neurofilaments enter the axon. We have investigated the functions of this process in normal, unperturbed retinal ganglion cell neurons of mature mice. Using in vivo pulse labeling with [35S]methionine or [32P]orthophosphate and immunocytochemistry with monoclonal antibodies to phosphorylation- dependent neurofilament epitopes, we showed that NF-H and NF-M subunits of transported neurofilaments begin to attain a mature state of phosphorylation within a discrete, very proximal region along optic axons starting 150 microns from the eye. Ultrastructural morphometry of 1,700-2,500 optic axons at each of seven levels proximal or distal to this transition zone demonstrated a threefold expansion of axon caliber at the 150-microns level, which then remained constant distally. The numbers of neurofilaments nearly doubled between the 100- and 150- microns level and further increased a total of threefold by the 1,200- microns level. Microtubule numbers rose only 30-35%. The minimum spacing between neurofilaments also nearly doubled and the average spacing increased from 30 nm to 55 nm. These results show that carboxyl- terminal phosphorylation expands axon caliber by initiating the local accumulation of neurofilaments within axons as well as by increasing the obligatory lateral spacing between neurofilaments. Myelination, which also began at the 150-microns level, may be an important influence on these events because no local neurofilament accumulation or caliber expansion occurred along unmyelinated optic axons. These findings provide evidence that carboxyl-terminal phosphorylation triggers the radial extension of neurofilament sidearms and is a key regulatory influence on neurofilament transport and on the local formation of a stationary but dynamic axonal cytoskeletal network.     

Transglutaminase and the Neuronal Cytoskeleton in Alzheimer''s Disease

Transglutaminase [EC 2.3.2.13, (R)-glutaminyl-peptide:amine gamma-glutamyltransferase], an enzyme that catalyzes the introduction of glutamine-lysine cross-links into proteins, was purified. Neurofilament and microtubule proteins were substrates for this enzyme but the insoluble neurofibrillary tangles (NFT) isolated from Alzheimer's disease brain were not substrates. In vitro cross-linking of neurofilaments and microtubules by the enzyme did not produce paired helical filaments (PHF), which are the major ultrastructural component of NFT. These results make it unlikely that PHF are formed by the straightforward cross-linking of neurofilaments or microtubules.     

Studies on the biosynthesis of neurofilament proteins

To determine whether the triplet polypeptides of neurofilaments arise by degradation of precursor, we studied the biosynthesis of neurofilament polypeptides both in vivo and in cell-free systems. Neurofilament-enriched fractions and polyribosomes were prepared from the same rabbit spinal cord homogenates. At 1 h after intracisternal administration of [34S]methionine, radiolabeled neurofilament proteins were detected in spinal cord homogenates as well as in isolated filaments. When polyribosomes from rabbit spinal cord were allowed to incorporate [35S]methionine into protein, triplet polypeptides were among the proteins labeled. Addition of spinal cord polyribosomes to rabbit reticulocyte lysates led to several cycles of translation of the spinal cord mRNA; the three neurofilament polypeptides were among the proteins synthesized in this system. The results demonstrate that the triplet polypeptides of neurofilaments are synthesized as such in the course of individual translational events and do not arise from degradation of P200 or a larger precursor.     

Matters Arising

Abstract: The possibility that neurofilaments could be involved in the transduction of chemical and mechanical energy in axons led us to investigate whether neurofilament proteins can hydrolyze ATP. We fractionated neurofilaments from rabbit spinal cord and found that preparations highly enriched for neurofilament proteins hydrolyzed ATP at a substantial rate (as high as 0.4 μmol/min/mg protein). However, the ATPase activity was neither inhibited by anti-neurofilament antibody, nor was it precipitated by the antibody under circumstances that precipitated most of the neurofilament polypeptides. We conclude that neurofilament proteins do not hydrolyze ATP at a significant rate under the conditions of our assay; if hydrolysis of ATP is a physiological function of neurofilaments, additional factors are required.     

Multiple phosphorylated variants of the high molecular mass subunit of neurofilaments in axons of retinal cell neurons: characterization and evidence for their differential association with stationary and moving neurofilaments

The 200-kD subunit of neurofilaments (NF-H) functions as a cross-bridge between neurofilaments and the neuronal cytoskeleton. In this study, four phosphorylated NF-H variants were identified as major constituents of axons from a single neuron type, the retinal ganglion cell, and were shown to have characteristics with different functional implications. We resolved four major Coomassie Blue-stained proteins with apparent molecular masses of 197, 200, 205, and 210 kD on high resolution one- dimensional SDS-polyacrylamide gels of mouse optic axons (optic nerve and optic tract). Proteins with the same electrophoretic mobilities were radiolabeled within retinal ganglion cells in vivo after injecting mice intravitreally with [35S]methionine or [3H]proline. Extraction of the radiolabeled protein fraction with 1% Triton X-100 distinguished four insoluble polypeptides (P197, P200, P205, P210) with expected characteristics of NF-H from two soluble neuronal polypeptides (S197, S200) with few properties of neurofilament proteins. The four Triton- insoluble polypeptides displayed greater than 90% structural homology by two-dimensional alpha-chymotryptic iodopeptide map analysis and cross-reacted with four different monoclonal and polyclonal antibodies to NF-H by immunoblot analysis. Each of these four polypeptides advanced along axons primarily in the Group V (SCa) phase of axoplasmic transport. By contrast, the two Triton-soluble polypeptides displayed only a minor degree of alpha-chymotryptic peptide homology with the Triton-insoluble NF-H forms, did not cross-react with NF-H antibodies, and moved primarily in the Group IV (SCb) wave of axoplasmic transport. The four NF-H variants were generated by phosphorylation of a single polypeptide. Each of these polypeptides incorporated 32P when retinal ganglion cells were radiolabeled in vivo with [32P]orthophosphate and each cross-reacted with monoclonal antibodies specifically directed against phosphorylated epitopes on NF-H. When dephosphorylated in vitro with alkaline phosphatase, the four variants disappeared, giving rise to a single polypeptide with the same apparent molecular mass (160 kD) as newly synthesized, unmodified NF-H. The NF-H variants distributed differently along optic axons. P197 predominated at proximal axonal levels; P200 displayed a relatively uniform distribution; and P205 and P210 became increasingly prominent at more distal axonal levels, paralleling the distribution of the stationary neurofilament network.(ABSTRACT TRUNCATED AT 400 WORDS)     

Organization of mammalian neurofilament polypeptides within the neuronal cytoskeleton

Neurofilaments in the axons of mammalian spinal cord neurons are extensively cross-linked; consequently, the filaments and their cross- bridges compose a three-dimensional lattice. We have used antibody decoration in situ combined with tissue preparation by the quick- freeze, deep-etch technique to locate three neurofilament polypeptides (195, 145, and 73 Kd) within this lattice. When antibodies against each polypeptide were incubated with detergent-extracted, formaldehyde-fixed samples of rabbit spinal cord, each antibody assumed a characteristic distribution: anti-73-Kd decorated the neurofilament core uniformly, but not the cross-bridges; anti-145-Kd also decorated the core, but less uniformly; sometimes the anti-145-Kd antibodies were located over the bases of cross-bridges. In contrast, anti-195-Kd primarily decorated the cross-bridges between the neurofilaments. These observations show that the 73-Kd polypeptide is a component of the central core of neurofilaments, and that the 195-Kd polypeptide is a component of the inter-neurofilamentous cross-bridges. It is consistent with this conclusion that we found few cross-bridges between neurofilaments in the optic nerves of neonatal rabbits during a developmental period when the ratio of 195 to 73 or 145-Kd polypeptides is much lower than in adults. The ratio of 195-Kd polypeptide to the other two neurofilament polypeptides also appeared much lower in the cell bodies and dendrites than in axons of adult spinal cord neurons, when the dispositions of the three polypeptides were studied by immunofluorescence experiments. The cell bodies apparently contain neurofilaments composed primarily of 145- and 73-Kd polypeptides, because we observed antibody decoration of individual neurofilaments in the cell bodies with anti-73- and -145-Kd, but not with anti-195-Kd. We conclude that the 195-Kd polypeptide participates in a cross-linking function, and that this function is, at least in certain neurons, most prevalent in the mature axon.     

Neuron-like derivatives of cultured F9 embryonal carcinoma cells express characteristics of parietal endoderm cells

Murine F9 embryonal carcinoma cells exposed to retinoic acid and dibutyryl cyclic AMP gradually arborize and acquire a neuron-like morphology in monolayer culture. Whether F9 cells can be induced to differentiate into cells with features specific to neural cells is controversial. We analyzed the intermediate filament content and pericellular matrix proteins of F9 cells after exposing them to retinoic acid, dibutyryl cyclic AMP, and nerve growth factor. In long-term cultures, a great majority of the cells appeared neuron-like, but showed intra- and pericellular laminin and type IV collagen, and frequently cytokeratin filaments as well. Several monoclonal antibodies to neurofilaments did not react with these cells in immunofluorescence or immunoblotting, though they recognize either all or individual mouse neurofilament triplet proteins. Polyclonal antibodies to neurofilament proteins gave a diffuse, nonfibrillar, vinblastine-resistant fluorescence in the morphologically neuron-like cells, but in immunoblotting failed to reveal polypeptides compatible with neurofilament triplet proteins. In long-term cultures, most of the cells appeared to have partially or totally lost the intermediate filaments. This was confirmed with anti-IFA antibodies which normally react with all intermediate filament proteins. The F9-derived cells did not respond to nerve growth factor in any detectable way. We conclude that the morphologically neuron-like derivatives of F9 cells display characteristics of modified parietal endoderm-like cells and do not show unequivocal features of neural cells.     

The polypeptide composition of axoplasm and of neurofilaments from the marine worm Myxicola infundibulum.

1. Axoplasm from Myxicola contains two major polypeptides associated with neurofilaments, together with actin, tubulin and many minor polypeptide components. 2. Some of the minor polypeptides with molecular weights between 140,000 and 50,000 purify with neurofilaments under a variety of conditions and they appear to represent an integral part of the filament structure. 3. Peptide fingerprinting shows that the two major neurofilament polypeptides are almost identical. The fingerprint patterns from these major polypeptides share features with those obtained from the minor components. 4. Peptide fingerprinting has enabled us to propose a scheme for the main sites at which papain cleaves the major neurofilament polypeptides. In addition fingerprinting indicates how the minor components are related to the major polypeptides. 5. It is suggested that many of the minor neurofilament polypeptides could arise by proteolysis in vivo.     

Mitotic mechanisms in Alzheimer's disease?

The mechanism(s) leading to widespread hyper-phosphorylation of proteins in Alzheimer's disease (AD) are unknown. We have characterized seven new monoclonal antibodies recognizing independent phospho- epitopes in the paired helical filament proteins (PHF) found in AD brain. These antibodies show pronounced immunoreactivity with cultured human neuroblastoma cells that are in the M phase of cell division, but have no discernible reactivity with interphase cells. Immunoreactivity with these antibodies does not localize to the microtubule spindles or chromosomes in M phase, but is confined to the surrounding cytoplasm. Similar staining in M phase is observed with cultured cells of various tissue types and species. Cells arrested in M phase with the microtubule depolymerizing agent, nocodazole, show marked increases in immunoreactivity with the antibodies by immunofluorescence staining, ELISA, and immunoblotting. In neuroblastoma cells, the appearance of the TG/MC phospho-epitopes coincides with activation of mitotic protein kinases, but not with the activity of the neuronal specific cyclin- dependent kinase, cdk5. These data suggest that the TG/MC epitopes are conserved mitotic phospho-epitopes produced as a result of increased mitotic kinase activity. To investigate this possibility in AD, we examined the staining of human brain tissue with MPM-2, a marker antibody for mitotic phospho-epitopes. It was found that MPM-2 reacts strongly with neurofibrillary tangles, neuritic processes, and neurons in AD but has no staining in normal human brain. Our data suggest that accumulation of phospho-epitopes in AD may result from activation of mitotic posttranslational mechanisms which do not normally operate in mature neurons of brain.     

Immunochemical evidence that fragments of phosphorylated MAP5 (MAP1B) are bound to neurofibrillary tangles in Alzheimer's disease

To test the hypothesis that cortical neurons undergo massive sprouting in Alzheimer's disease brain, we investigated whether neurofibrillary tangles contain fetal antigens. Two monoclonal antibodies to tangles specifically labeled an approximately 300 kd protein in the neonatal brain homogenate, which was subsequently identified as MAP5 (MAP1B). Conversely, two monoclonal antibodies to MAP5 were found to stain tangles. All four reacted with only a phosphorylated species of MAP5. By careful immunochemical analysis, at least three independent phosphorylated epitopes that should have distinct conformations were shown to be shared by tangles and MAP5. However, several monoclonal antibodies to nonphosphorylated MAP5 did not stain tangles. From these observations, we conclude that fragments of phosphorylated MAP5 are bound to tangles. Since MAP5, in particular, a phosphorylated species, is known to be involved in neurite outgrowth, this result supports the sprouting hypothesis.     

The location of phosphorylation sites and Ca2+-dependent proteolytic cleavage sites on the major neurofilament polypeptides from Myxicola infundibulum.

1. When axoplasm is incubated with [32P]Pi the main phosphorylated components are the neurofilament polypeptides. 2. Activation with Ca2+ of the proteinase present in axoplasm causes degradation of these neurofilaments and the peptides produced by this reaction have been analysed by fingerprinting. 3. Fingerprinting shows that initially the Ca2+-activated proteinase cleaves the neurofilament polypeptides at three major sites producing polypeptides with mol.wts. 70,000, 50,000 and 47,000. 4. These polypeptides sediment with filaments, originate from the tail-region of the molecule and contain a little radioactive label. 5. As these polypeptides are produced, other polypeptides that come from the head-region of the molecule are liberated as soluble products that contain the bulk of the radioactivity. 6. Fingerprinting therefore shows that at least two regions on the molecule are phosphorylated and that the major one is located towards the head-end of the polypeptides.