Structural stability of paired helical filaments requires microtubule-binding domains of tau: a model for self-association.

Highly purified and SDS-soluble paired helical filaments (PHFs) were immunogold labeled and immunoblotted with antibodies to tau: Tau 14 (N-terminal half), AH-1 (microtubule-binding domain), and Tau 46 (C-terminal end). The main component of PHFs was modified tau of 68, 64, and 60 kd, also called A68 or PHF-tau. Trypsin digestion reduced the maximum width of PHFs by 10%-20%, increased aggregation of filaments, and abolished the binding of Tau 14, but had no effect on the binding of AH-1. The smallest tau-reactive tryptic fragments were 13 and 7-8 kd, positive with AH-1, and negative with Tau 46. Our results and the model of Crowther and Wischik suggest that by self-association and anti-parallel arrangement of the microtubule-binding domains, PHF-tau forms the backbone of PHFs.

Ubiquitin-immunoreactive dystrophic neurites in Down's syndrome brains.

Ubiquitin-immunoreactivity was studied in Down's syndrome brains ranging in age from two days to sixty years. Numerous randomly distributed ubiquitin-immunoreactive dot-like structures in the white matter were shown to correspond to granular degeneration of myelin. Granular degeneration of myelin was first detected at age 21 and increased thereafter with age. Other larger and more coarsely granular ubiquitin-immunoreactive structures, most numerous in the middle and upper cortical layers, were consistent with dystrophic neurites. Immunoelectron microscopy demonstrated that the dystrophic neurites contained non-filamentous, membranous, dense bodies. In Down's syndrome, ubiquitin-immunoreactive dystrophic neurites were first detected at age six in the hippocampus, and were consistently more numerous in comparison to age-matched control subjects. In the presence of amyloid, either as diffuse or as compact deposits, ubiquitin-immunoreactive dystrophic neurites frequently formed aggregates consistent with senile plaques. Although apparently independent events, these data suggest that amyloid deposition is associated with local accentuation of ubiquitin-immunoreactive neuritic dystrophy. In addition, since dystrophic neurites appeared substantially earlier in the grey matter in Down's syndrome than in age-matched normals, this may be further evidence that selective aspects of aging are accelerated in Down's syndrome.

Modified Bielschowsky and immunocytochemical studies on cerebellar plaques in Alzheimer's disease.

Senile plaques (SP) in the cerebellum of 23 cases of Alzheimer's disease (AD), three with widespread amyloid angiopathy, were studied with a modified Bielschowsky stain and immunocytochemical methods using antibodies to a beta-amyloid synthetic peptide (beta ASP), phosphorylated neurofilament proteins, ubiquitin, tau protein, and glial fibrillary acidic protein (GFAP). The four subtypes of SP (diffuse plaques, compact plaques, perivascular plaques, and subpial fibrillar deposits) that were observed with the modified Bielschowsky stain were also stained with antibodies to beta ASP. Many cerebellar SP contained ubiquitin-positive granular elements resembling dystrophic neurites. In contrast to neuritic elements in cerebral SP in AD, ubiquitin-positive elements in cerebellar SP were not labeled with antibodies to phosphorylated neurofilament or tau proteins. Various degrees of glial reaction were observed in all subtypes of SP except diffuse plaques. The absence of phosphorylated neurofilament and tau epitopes in neuritic elements in cerebellar SP is not surprising since paired helical filaments have not been seen in the cerebellum. Nevertheless, our results suggest that cerebellar SP are frequently associated with dystrophic neurites.

Mass and physical dimensions of two distinct populations of paired helical filaments.

We studied the ultrastructure of two fractions of paired helical filaments (PHF) from Alzheimer brains separated on sucrose density gradient. Fraction A2 (1M sucrose) contained filaments which were short in length and did not aggregate while those in fraction AL2 (1/1.5 M sucrose interface) were mostly aggregated. By scanning transmission electron microscopy, PHF in fraction A2 had significantly more mass per nm length of filament (107-120 kD/nm) than those in fraction AL2 (79-85 kD/nm), and they were also wider in their maximum and minimum widths but did not differ in their periodicity. Differences in mass and dimensions between two morphologically distinct populations of PHF suggest that a partial proteolysis may be involved in the generation of the aggregated population of PHF. The results suggest that a similar process may be active in the formation of neurofibrillary tangles.

The pyruvate dehydrogenase complex during aging.

Acetylcholine synthesis and pyruvate oxidation decline with age. To determine the role of the pyruvate dehydrogenase complex in these age-related deficits, its activity and activation state were measured in vivo and in vitro in the brains of 3-, 10- and 30-month-old mice. Aging did not alter the active form of pyruvate dehydrogenase complex in vivo, although the total complex was 17% lower at 30 than at 3 months of age. In vitro, total or active forms of pyruvate dehydrogenase complex did not change with age. The results suggest that neither changes in total activity nor in the activation state of the pyruvate dehydrogenase complex account for the age-related deficits in oxidative or acetylcholine metabolism.

Phosphorylation of tau at THR212 and SER214 in human neuronal and glial cultures: the role of AKT.

We have reported recently that the microtubule-associated protein tau is phosphorylated in vitro by Akt, an important kinase in anti-apoptotic signaling regulated by insulin and growth factors. We also established that Akt phosphorylates tau separately at T212 and S214, two sites previously shown to be phosphorylated by glycogen synthase kinase 3beta (GSK3beta) and protein kinase A (PKA), respectively. In the present studies, we examined the relationship between Akt and T212/S214 in primary cultures of human neurons and astrocytes, and evaluated the contribution of two other kinases. In intact cells, we found a very low content of active (phospho-S473) form of Akt. We also found a low content of phospho-S214 but not phospho-T212 of tau, suggesting that only phospho-S212 may depend on Akt activity in situ. We upregulated Akt activity using two experimental models: treatment with a protein phosphatase inhibitor, okadaic acid, and transfection with a constitutively active Akt gene construct (c-Akt). Under these conditions, phosphorylation of tau at T212 and S214 was regulated independently, with little change or downregulation of phospho-T212 and dynamic upregulation of phospho-S214. Our studies revealed that Akt may influence the phospho-S214 content in a meaningful manner. They also revealed that PKA may only partially contribute to the phosphorylation of S214. In comparison, okadaic acid treatment severely depleted the content of GSK3beta and downregulated the remaining GSK3beta activity by Akt-dependent inhibition, consistent with minimal changes in phospho-T212. In summary, these results strongly suggest that in primary cultures, Akt selectively phosphorylates tau at S214 rather than T212. Our studies raise the possibility that tau S214 may participate in Akt-mediated anti-apoptotic signaling.

Conformation of paired helical filaments blocks dephosphorylation of epitopes shared with fetal tau except Ser199/202 and Ser202/Thr205.

To determine if the high phosphate content of paired helical filaments (PHFs) in Alzheimer's disease (AD) is a result of limited access to filament phosphorylation sites, we studied in vitro dephosphorylation of intact PHFs, PHFs with filamentous structure abolished by formic acid treatment (PHF(FA)) and fetal human tau protein. Samples were treated with alkaline phosphatase for up to 24 h at 37 degrees C and then immunoblotted with eight well characterized tau antibodies, that recognize two phosphorylation-insensitive sites and six phosphorylation-sensitive epitopes at Thr181, Ser199/202, Ser202/Thr205, Thr231, Ser262/356 and Ser396/404. Intact PHFs were effectively dephosphorylated only at the two N-terminal epitopes Ser199/202 and Ser202/Thr205, with little change in electrophoretic mobility. In contrast, PHF(FA) were dephosphorylated at all epitopes, with particular effectiveness at those in the C-terminus and with significant increase in electrophoretic mobility. The fetal tau epitopes were effectively dephosphorylated except at Thr181 and Thr231 with marked increase in mobility. The extent of dephosphorylation of PHF(FA) was equal or more effective than in fetal tau, except for Thr181 that was minimally dephosphorylated in both proteins. The results indicate that intact PHFs, but not PHF(FA) or fetal tau display differential dephosphorylation of the N- and C-terminal epitopes. The results confirm that the filamentous conformation may significantly contribute to hyperphosphorylation of PHFs in the C-terminus. The filamentous conformation, however, does not limit access to two N-terminal epitopes Ser199/202 and Ser202/Thr205. The access to these sites in AD may be limited by other factors, e.g., inhibition of phosphatase binding.

Tau immunoreactivity and SDS solubility of two populations of paired helical filaments that differ in morphology.

To further understand the processes that lead to the formation of neurofibrillary tangles from paired helical filaments (PHF) in Alzheimer brains, we studied two morphologically distinct fractions of PHF separated on sucrose density gradient. In a fraction with mostly short and non-aggregated PHF, the majority of filaments could be solubilized in SDS. In a fraction containing primarily PHF aggregated into clusters or bundles, sometimes resembling neurofibrillary tangles, filaments were less soluble in SDS. Immunogold labelling with a panel of tau-immunoreactive antibodies demonstrated that N-terminal epitopes of tau were preserved in the short filaments, but were reduced or absent in aggregated filaments. In contrast, C-terminal epitopes were present in both fractions. Furthermore, the accessibility of the microtubule-binding domain to immunolabelling was markedly impaired in short and non-aggregated filaments compared to aggregated filaments. These results are consistent with proteolytic degradation of the N-terminal epitopes and preservation of the C-terminal epitopes and the microtubule-binding domain of tau in the aggregated filaments. Partial proteolysis may be involved in the generation of aggregated PHF in neurofibrillary tangles.

Phosphate analysis and dephosphorylation of modified tau associated with paired helical filaments.

We performed phosphate analysis of tau proteins isolated from normal human brain, tau proteins associated with paired helical filaments (PHF-tau), and Alzheimer tau not associated with PHF. These tau fractions were of high purity. Normal and Alzheimer tau were purified by heat treatment, acid extraction and calmodulin-affinity chromatography with or without HPLC. Fractions containing primarily PHF-tau polypeptides of 60, 64 and 68 kDa and their degraded fragments were purified either on a sucrose density gradient as filaments (PHF) or by heat treatment and acid extraction as amorphous proteins (PHF-tau). PHF and PHF-tau were found to contain 6-8 mol phosphate/mol protein while normal and Alzheimer tau proteins contained 1.9 and 2.6 mol phosphate/mol protein, respectively. Upon 2-h incubation with alkaline phosphatase, PHF lost two of the phosphate groups without apparent changes in the stability and morphology of PHF. The released phosphate originated from the N-terminal half of PHF-tau as determined by immunoblotting with antibodies to epitopes blocked by phosphorylation. Tau-1 and E-2, and by a prominent shift in the electrophoretic mobility of some fragments of PHF-tau. The shift in mobility was not observed with the C-terminal fragments of 25-26 kDa, which retained the epitope to Tau 46. The results suggest that the phosphorylation sites not affected by phosphatase may be located in the 25-26 kDa C-terminal region of PHF-tau and may play a role in structural stability of PHF.

Binding of Alz 50 depends on Phe8 in tau synthetic peptides and varies between native and denatured tau proteins.

Alz 50 is a monoclonal antibody that in Western blotting analysis recognizes both normal tau as well as hyperphosphorylated tau proteins associated with paired helical filaments (PHF-tau) in Alzheimer disease (AD). Within tissue sections of AD brain, however, Alz 50 immunolabels only PHF, which suggests that the antibody recognizes a conformational epitope. Using competitive enzyme-linked immunosorbent assay, we demonstrate that Alz 50 binds to tau synthetic peptides with low affinity (KD between 0.27 to 2.7 x 10(-5) M) and that the binding is specific for the RQEF sequence corresponding to N-terminal residues 5-8 of tau. The Alz 50 epitope appears to be largely dependent on Phe8, a strongly hydrophobic amino acid residue, since the substitution of Phe8 with Ala8 in the synthetic peptide abolishes Alz 50 binding. The effects of tau conformation on Alz 50 binding were studied with various normal tau proteins with either low or high phosphate content (adult vs. fetal) and PHF-tau proteins. The normal tau fractions were isolated from both adult and fetal human brains using affinity chromatography (native form) and heat/perchloric acid treatments (denatured form). PHF-tau was isolated as Sarcosyl-insoluble fraction. With competitive ELISA, the denatured form of normal tau (fetal and adult) bound Alz 50 with the same high affinity as did PHF-tau (KD between 1.3 to 1.8 x 10(-7) M). In contrast, the native form of tau from either brain was unable to fully compete for Alz 50 and at most only 50% of the Alz 50 binding sites in native tau were occupied. These results suggest that native tau may exist either in complexes with other proteins or in a form of dimers/oligomers, in which only some N-termini are available for binding (e.g. head-to-tail assembly). The results also suggest that denaturation rather than phosphorylation of tau has more significant effect on interactions of tau with Alz 50.

Assembled tau filaments differ from native paired helical filaments as determined by scanning transmission electron microscopy (STEM).

Paired helical filaments (PHF) are abnormal, approximately 20-25-nm wide periodically twisted filaments, which accumulate in Alzheimer's disease (AD) brain and other neurodegenerative disorders, including corticobasal degeneration (CBD). PHF are primarily composed of highly phosphorylated tau protein. However, both phosphorylated and non-phosphorylated forms of tau are able to assemble in vitro into filaments similar in the ultrastructural appearance to PHF. In the present study, filaments were assembled in vitro from unmodified recombinant human tau and the physical mass per unit length of filaments and the mass density were determined using scanning transmission electron microscopy (STEM). Two general types of filaments were observed. One type was composed of 11.4 nm-wide, 10-75 nm long, frequently twisted and PHF-like filaments, with a mass per unit length (44 kDa/nm) approximately one third of that observed in isolated AD filaments. The other were straight filaments, approximately 6.8-nm wide and 0.2-2 microm long, which often formed parallel clusters of two or more filaments. Triple clusters were 19. 2-nm wide and had a mass per unit length (70 kDa/nm) approximately two thirds of that seen in isolated AD filaments. Despite different morphology, both twisted and straight filaments had mass densities between 0.48-0.55 kDa/nm3. These values are significantly higher than those reported for PHF found either in AD (0.40 kDa/nm3) or CBD (0.33 kDa/nm3). These results suggest that the packing of tau differs in vivo from that observed in vitro and that specific tau isoform content, elongation of tau molecules by phosphorylation or other factors may be required to reproduce pathological assembly. Therefore mass density determinations appear to be an important criterion in comparing various filaments.

Paired helical filaments in corticobasal degeneration: the fine fibrillary structure with NanoVan.

Paired helical filaments (PHF) composed of hyperphosphorylated tau proteins are characteristic findings in neurodegenerative disorders, including Alzheimer's disease (AD) and corticobasal degeneration (CBD). The filaments in CBD differ from those in AD by a reduced number of tau isoforms and less stable ultrastructure. To further compare the ultrastructure of both filaments, we employed a novel staining reagent, NanoVan, as well as aurothioglucose and uranyl acetate. With commonly used uranyl acetate, both kinds of filaments appeared as twisted ribbons 15-20-nm and 21-23-nm wide, respectively, without significant internal substructure. With application of aurothioglucose, only few structural details were apparent. With NanoVan, AD filaments showed similar structure to that with uranyl acetate but CBD filaments displayed a highly heterogeneous appearance consistent with the dissociation of the 20-25-nm-wide filaments along two longitudinal axes. This was evident by the presence of thinner, 12-13-nm-wide filaments and filaments that splayed into two 20-25-nm-wide components at one or both ends. Moreover, detection of a prominent, 7-8-nm-wide axial region distinguished up to four protofilaments per one filament. Each protofilament appeared to contain two 3-5-nm-wide fibrils separated by an approximately 1-nm-wide axial region. The results suggest that 3-5-nm fibrils are the smallest structural subunits of filaments in CBD and that NanoVan may be an unique reagent in detecting eight-fibril organization in these less stable filaments.

Role of Hypertension in Aggravating Abeta Neuropathology of AD Type and Tau-Mediated Motor Impairment.

Epidemiological evidence suggests that hypertension may accelerate the onset and progression of Alzheimer's disease (AD). In this study, we explored the role of hypertension in the neurodegenerative changes associated with Abeta and tau aggregation. We induced hypertension in APP(swe) Tg2576 and P301L-tauTg mouse models. In Tg2576 mice, experimental hypertension was associated with a significant increase of the accumulation of Amyloid-beta (Abeta) peptides in brain tissue and a significant reduction of Abeta peptides in serum (P < .05). These results indicate that hypertension may promote AD-type Abeta neuropathology in Tg2576. In P301L-tauTg mice we found that the presence of hypertension was significantly associated with aggravated motor function assessed by hindlimb extension test (P = .01). These results suggest that hypertension may play a role in accelerating the progression of motor dysfunction associated with tau-related alterations. Our studies suggest that the management of blood pressure (BP) may alleviate AD-type Abeta neuropathology and neurological disorders associated with abnormal tau metabolism.

Differential changes in phosphorylation of tau at PHF-1 and 12E8 epitopes during brain ischemia and reperfusion in gerbils.

Cortical neurons are vulnerable to ischemic insult, which may cause cytoskeletal changes and neurodegeneration. Tau is a microtubule-associated protein expressed in neuronal and glial cells. We examined the phosphorylation status of tau protein in the gerbil brain cortex during 5 min ischemia induced by bilateral common carotid artery occlusion followed by reperfusion for 20 min to 7 days. Control brain homogenates contained 63, 65 and 68 kD polypeptides of tau immunoreactive with Alz 50, Tau 14 and Tau 46 antibodies raised against non-phosphorylated tau epitopes. Gerbil tau was also immunoreactive with some (PHF-1 and 12E8) but not all (AT8, AT100, AT180 and AT270) antibodies raised against phosphorylated tau epitopes. PHF-1 recognized a single 68 kD polypeptide and 12E8 bound the 63 kD polypeptide. During 5 min ischemia, PHF-1 immunoreactivity declined to 6%, then recovered to control levels after 20 min of blood recirculation and subsequently increased above control values 3 and 7 days later. In contrast, 12E8 immunoreactivity remained stable during ischemia and reperfusion. Our results suggest that the two phosphorylated epitopes of tau are regulated by different mechanisms and may play different roles in microtubule dynamics. They may also define various pools of neuronal/glial cells vulnerable to ischemia.

Inhibition of cyclooxygenase as potential novel therapeutic strategy in N141I presenilin-2 familial Alzheimer's disease.

The present study was designed to further explore the potential cause/effect relationship between the expression of both the N141I presenilin (PS)2 mutant familial Alzheimer's disease (FAD) gene and cyclooxgenase (COX) in respect to the mechanism associated with programmed cell death in Alzheimer's disease (AD). We found that expression of mutant N141I PS2 resulting in apoptotic cell death in H4 neuronal cells coincided with >4-fold induction in the expression of the inducible form of COX-2, but not the constitutive COX-1. Moreover, we found that the expression of the N141I PS2 FAD gene strongly promoted (>2-fold) glycogen synthase kinase (GSK)-3beta activity coincidental with a reduction in the level of beta-catenin translocated from the cytoplasmic to the nuclear compartment. Most interestingly, we found that inhibition of COX-2-mediated generation of prostaglandin (PG)-E2 in H4 neuronal cells with the preferential COX-2 inhibitor nimesulide protects against N141I PS2-mediated apoptotic cell death coincidental with an inhibition of GSK-3beta activity and subsequent normalization of beta-catenin cellular distribution. The clinical relevance of this finding was confirmed by the evidence that COX-2 protein and PG-E2 concentrations were selectively increased >2-fold in the cerebral cortex of subjects harboring the N141I PS2 FAD mutation relative to wild-type PS2 AD cases. This study demonstrates for the first time that COX-2 may be a downstream effector of mutant N141I PS2-mediated apoptotic cell death and that inhibition of COX-2 may neuroprotect in AD through modulation of a GSK-3beta-beta-catenin-mediated response. The study provides support for the potential pharmacogenomic identification of N141I PS2 FAD cases that might preferentially benefit from inhibition of COX-2 during the progression of clinical dementia.

Two monoclonal antibodies recognize Alzheimer's neurofibrillary tangles, neurofilament, and microtubule-associated proteins.

Two monoclonal antibodies that recognize Alzheimer's neurofibrillary tangles (ANTs), AD10 and AB18, have been characterized by immunoblotting against human and calf spinal cord neurofilament (NF) and calf brain microtubule preparations. Both antibodies bind to the 200-kilodalton (kd) (NF-H) and 160-kd (NF-M) but not to the 68-kd (NF-L) NF triplet proteins. They also bind to high-molecular-weight microtubule-associated proteins (MAPs) and tau. AD10 immunostains MAP2 and MAP1 families, whereas AB18 stains mainly MAP1 bands. Preincubation of intact filament preparation or nitrocellulose strips containing electroblotted NF proteins with Escherichia coli alkaline phosphatase completely blocks AD10 binding and partially blocks binding of AB18. These results suggest that the determinants recognized by these antibodies are phosphorylated. Immunoblotting of peptide fragments generated by limited proteolysis of NF proteins with alpha-chymotrypsin and Staphylococcus aureus V8 protease shows that the localization of the antigenic determinants to AD10 and AB18 in NF-H is approximately 100 and 60 kd, respectively, away from the carboxy terminal, a region previously shown to form the NF projection side arm. In NF-M, the antigenic determinants to both antibodies are located also in the projection side arm, in a 60-kd polypeptide adjacent to the alpha-helical filament core. The results show that ANTs contain at least two phosphorylated antigenic sites that are present in NF and MAPs, a finding suggesting that ANTs may be composed of proteins or their fragments with epitopes shared by cytoskeletal proteins.

Phosphatase and carbocyanine dye binding define different types of phosphate groups in mammalian neurofilaments.

The phosphorylation state of human and bovine spinal cord neurofilaments (NF) was studied by direct phosphate analysis and carbocyanine dye ("Stains-all") binding to NF polypeptides resolved on SDS-polyacrylamide gels. Electrophoretically purified NF-H (200 kDa), NF-M (160 kDa), and NF-L (68 kDa) of human origin contained 24, 18, and 4 mol phosphate/mol protein, whereas bovine NF contained 53, 23, and 5 mol phosphate/mol protein, respectively. Incubation of NF preparations with E. coli alkaline phosphatase removed about 55% of the phosphate from NF-H, about 30% of the phosphate from both human and bovine NF-M, but did not change the phosphate content of NF-L. This treatment also inhibited or substantially reduced the binding of electroblotted NF-H and NF-M to 2 anti-NF monoclonal antibodies known to recognize phosphorylated sites on projection side arms. "Stains-all" was found to be a very sensitive probe for detection of phosphorylated cytoskeletal proteins. Without the phosphatase treatment, NF and other phosphoproteins, MAP1, MAP2, tubulin, and tau, all bound the carbocyanine dye on SDS gels, forming blue dye-protein complexes. Measured densitometrically at 615 nm, the staining intensity (relative units/mol protein) was 9, 9, and 3 for human and 10, 13, and 6 for bovine NF-H, NF-M, and NF-L, respectively. NF-H bound the dye less efficiently than was expected from its phosphate content. After phosphatase treatment, NF-H, with half of its phosphate residues remaining, no longer formed blue complex with "Stains-all," the staining intensity of NF-M decreased by 20-40%, and the staining of NF-L was not changed.(ABSTRACT TRUNCATED AT 250 WORDS)

Calpain-induced proteolysis of normal human tau and tau associated with paired helical filaments.

The major components of neurofibrillary tangles (NFT) in Alzheimer's disease are bundles of paired helical filaments (PHF) which are primarily composed of highly phosphorylated tau proteins (PHF-tau). To further understand the mechanism of PHF accumulation in NFT, we examined the calpain-induced proteolysis of highly purified and primarily non-aggregated PHF and normal tau proteins with various contents of phosphate isolated from either fetal (F-tau) or adult human brain (N-tau). The extent of proteolysis was determined by decreases in tau immunoreactivity using Western-blot analysis and a panel of site-specific tau antibodies (Alz 50, Tau-2, Tau 14, Tau-1, AT8, E-11, AH-1 and PHF-1). We found that full-size polypeptides of N-tau and F-tau were similarly and rapidly proteolyzed in vitro by calpain (calpain II, 3.3 units/mg protein) during a 10-min incubation at 30 degrees C, and that their half lives (t1/2) were 1.5 min and 1.8 min, respectively. Analysis of immunoblots suggests that full-length polypeptides of tau are first degraded into large fragments similar in size to that generated endogenously, then into smaller fragments. Since both endogenous and in-vitro-generated tau fragments retained N-terminal epitopes, the results suggest that most of the calpain-sensitive sites may be located in the C-terminal half of the tau molecule. In contrast, PHF were extremely resistant to degradation and only a fivefold higher concentration of calpain (16.7 units/mg protein) induced partial proteolysis of PHF. A major calpain-generated fragment was a 45-kDa polypeptide derived from the C-terminal region of PHF-tau, which forms a core of filaments. The results suggest that the inaccessibility of potential calpain-digestion sites in the filament core could contribute to the resistance of PHF to calpain and subsequently lead to the accumulation of PHF in Alzheimer's disease. The results also suggest that hyperphosphorylation of tau may be marginally involved in the resistance of PHF to degradation by calpain. Ultrastructural examination revealed that, in contrast to previous studies with trypsin, calpain did not alter the morphologic appearance of filaments; after incubation with calpain, the majority of PHF remained short and disperse and the number of PHF aggregated into NFT-like clusters was not significantly increased. The results suggest that the role of calpain in promoting the aggregation and clustering of filaments is limited.

Ubiquitin immunoreactivity of paired helical filaments differs in Alzheimer's disease and corticobasal degeneration.

To establish whether there is a relationship between ubiquitination and ultrastructural appearance of filaments, we compared the ubiquitin immunoreactivity of paired helical filaments (PHFs) in Alzheimer's disease (AD) and corticobasal degeneration (CBD). PHFs in these disorders share a limited similarity since filaments in CBD are wider and twisted at longer intervals than those in AD, and also display less ultrastructural stability. Preparations enriched in SDS-soluble filaments were isolated from AD and CBD brains and subjected to tau and ubiquitin immunogold labeling. Both preparations contained mostly dispersed individual PHFs, which labeled for the amino and carboxyl termini of tau. Immunolabeling of ubiquitin was variable, however, being more intense in AD than CBD samples. SDS-insoluble filaments were prepared from PHFs by boiling in the presence of SDS and 2-mercaptoethanol and collected by sedimentation. In both disorders, the pellets contained highly aggregated and bundled filaments, which were devoid of the amino but not the carboxyl terminal region of tau. Again, ubiquitin labeling was more intense in AD than CBD filaments. The present results suggest that ubiquitination has limited influence on SDS solubility, aggregation and bundling of PHFs; however, it may be one of the factors responsible for the ultrastructural variability and/or stability of filaments.