Human tau isoforms confer distinct morphological and functional properties to stably transfected fibroblasts

Tau interferes with axonal neurite stabilization and cytoskeletal composition independently of its ability to associate with microtubules

Tau impacts overall axonal transport particularly when overexpressed by interfering with translocation of kinesin along microtubules (MTs) and/or as a cargo of kinesin by outcompeting other kinesin cargo. To discern between which of these mechanisms was more robust during axonal outgrowth, we overexpressed phosphomimetic (E18; which is incapable of MT binding), phospho-null (A18) or wild-type (WT) full-length human tau conjugated to EGFP, the latter two of which bind MTs. Expression of WT and A18 displayed increased acetylated MTs and resistance to colchicine, while expression of E18 did not, indicating that E18 did not contribute to MT stabilization. Expression of all tau constructs reduced overall levels of neurofilaments (NFs) within axonal neurites, and distribution of NFs along neurite lengths. Since NFs are another prominent cargo of kinesin during axonal neurite outgrowth, this finding is consistent with WT, A18 and E18 inhibiting NF transport to the same extent by competing as cargo of kinesin. These findings indicate that tau can impair axonal transport independently of association with MTs in growing axonal neurites.

Alzheimer's disease-like tau neuropathology leads to memory deficits and loss of functional synapses in a novel mutated tau transgenic mouse without any motor deficits

Tau transgenic mice are valuable models to investigate the role of tau protein in Alzheimer's disease and other tauopathies. However, motor dysfunction and dystonic posture interfering with behavioral testing are the most common undesirable effects of tau transgenic mice. Therefore, we have generated a novel mouse model (THY-Tau22) that expresses human 4-repeat tau mutated at sites G272V and P301S under a Thy1.2-promotor, displaying tau pathology in the absence of any motor dysfunction. THY-Tau22 shows hyperphosphorylation of tau on several Alzheimer's disease-relevant tau epitopes (AT8, AT100, AT180, AT270, 12E8, tau-pSer396, and AP422), neurofibrillary tangle-like inclusions (Gallyas and MC1-positive) with rare ghost tangles and PHF-like filaments, as well as mild astrogliosis. These mice also display deficits in hippocampal synaptic transmission and impaired behavior characterized by increased anxiety, delayed learning from 3 months, and reduced spatial memory at 10 months. There are no signs of motor deficits or changes in motor activity at any age investigated. This mouse model therefore displays the main features of tau pathology and several of the pathophysiological disturbances observed during neurofibrillary degeneration. This model will serve as an experimental tool in future studies to investigate mechanisms underlying cognitive deficits during pathogenic tau aggregation.

Muscle-specific microtubule-associated protein 4 is expressed early in myogenesis and is not sufficient to induce microtubule reorganization

The expression of a muscle-specific variant of microtubule-associated protein 4 (mMAP4) has been analyzed during myogenesis of C(2)C(12) cells using an isoform-specific antibody. MMAP4 localizes to microtubules (MTs) and is expressed prior to a very early morphogenetic event, the formation of mononucleate spindle-shaped cells. MMAP4 protein appears at about the same time as titin and coincident with Golgi reorganization, but antedates myosin expression. Misexpression of EGFP-mMAP4 in non-muscle and proliferating C(2)C(12) cells does not induce dramatic changes in MT organization or stability, nor in Golgi organization. Expression of full-length mMAP4 or of a truncated form lacking the MT-binding domain does not disrupt myotube formation or myofibrillogenesis. While previous antisense studies indicated that mMAP4 is necessary for normal myotube formation [Mangan and Olmsted, 1996: Development 122:771-781], these data indicate mMAP4 is not sufficient to induce the reorganization of MTs or the Golgi into patterns typical of muscle cells. Thus, with respect to MT organizing properties, this tissue-specific variant differs from related neuronal MAPs, MAP2, and tau, which induce neural-like changes in MT organization.

Structural and functional differences between 3-repeat and 4-repeat tau isoforms. Implications for normal tau function and the onset of neurodegenetative disease

Tau, MAP2, and MAP4 are members of a microtubule-associated protein (MAP) family that are each expressed as "3-repeat" and "4-repeat" isoforms. These isoforms arise from tightly controlled tissue-specific and/or developmentally regulated alternative splicing of a 31-amino acid long "inter-repeat:repeat module," raising the possibility that different MAP isoforms may possess some distinct functional capabilities. Consistent with this hypothesis, regulatory mutations in the human tau gene that disrupt the normal balance between 3-repeat and 4-repeat tau isoform expression lead to a collection of neurodegenerative diseases known as FTDP-17 (fronto-temporal dementias and Parkinsonism linked to chromosome 17), which are characterized by the formation of pathological tau filaments and neuronal cell death. Unfortunately, very little is known regarding structural and functional differences between the isoforms. In our previous analyses, we focused on 4-repeat tau structure and function. Here, we investigate 3-repeat tau, generating a series of truncations, amino acid substitutions, and internal deletions and examining the functional consequences. 3-Repeat tau possesses a "core microtubule binding domain" composed of its first two repeats and the intervening inter-repeat. This observation is in marked contrast to the widely held notion that tau possesses multiple independent tubulin-binding sites aligned in sequence along the length of the protein. In addition, we observed that the carboxyl-terminal sequences downstream of the repeat region make a strong but indirect contribution to microtubule binding activity in 3-repeat tau, which is in contrast to the negligible effect of these same sequences in 4-repeat tau. Taken together with previous work, these data suggest that 3-repeat and 4-repeat tau assume complex and distinct structures that are regulated differentially, which in turn suggests that they may possess isoform-specific functional capabilities. The relevance of isoform-specific structure and function to normal tau action and the onset of neurodegenerative disease are discussed.

Tubulin immunoreactive neuronal intranuclear inclusions in the human brain

Intranuclear filamentous and crystalline inclusion bodies have been described in the nuclei of a variety of cells in both normal and pathological states. The functional significance of these structures remains to be elucidated. Moreover, although the proteinaceous nature of these inclusions has been inferred in some histochemical studies, the identity of their constituent proteins remains to be determined. In the present study, immunohistochemistry was used to investigate the presence of intranuclear inclusions in neurones of the human brain which are intensely immunoreactive for the neuronal cytoskeletal protein class III beta tubulin. The ability to label these structures immunohistochemically was exploited to investigate the topographic pattern of distribution of these inclusions in the human brain. Intranuclear inclusions were rod-shaped, polygonal, or irregular in shape. They were present in neurones and ependymal cells. Intranuclear inclusion-bearing neurones were distributed in an anatomically heterogeneous pattern in the brain. Areas exhibiting relatively high densities of inclusions included the substantia inominata and anterior olfactory nucleus, dentate gyrus, substantia nigra, inferior olivary nucleus, and dentate nucleus of the cerebellum. In addition, intranuclear inclusions were prevalent in neurones in layers II, V, and VI of the cerebral cortex. They were particularly prevalent in the mesial basal temporal neocortex. The relationship of these structures to the intranuclear rods and sheets of the classical microscopists is uncertain. The demonstration that they are composed, at least in part, of tubulin, a major cytoskeletal protein, provides important clues regarding the mechanisms underlying their formation and provides a springboard for developing hypotheses regarding their functional significance. Furthermore, the ability to demonstrate these inclusions immunohistochemically provides an avenue for further studies directed at elucidating the potential involvement of these inclusions in various pathological settings.

Phylogenetic diversity of the expression of the microtubule-associated protein tau: implications for neurodegenerative disorders

The microtubule-associated protein tau regulates the dynamic stability of the neuronal cytoskeleton by interacting with microtubules. It is encoded by a single gene, but expressed in a variety of isoforms due to differential RNA splicing. Six isoforms can be found in the human central nervous system. These isoforms differ in their ability to promote the assembly of microtubules as well as in their capacity to stabilize existing microtubule structures. Furthermore, some of the isoforms of tau are specifically involved in the pathogenesis of neurodegenerative disorders. Thus, splicing of tau might critically influence the physiological functions of tau protein as well as the pathogenesis of neurodegenerative diseases with tauopathy. The present study addresses the differential expression of the six isoforms of tau in the central nervous system of 12 mammalian species including Homo sapiens. The occurrence of each of the six tau isoforms was highly variable. However, species that were phylogenetically related expressed a similar pattern of tau isoforms. These results suggest a phylogenetic descent of splicing paradigms, which can be matched with known phylogenetic concepts based on morphological and molecular genetical studies. Especially, the unique expression pattern of tau isoforms in the human central nervous system implicates a possible link to the particular vulnerability of humans to neurodegenerative disorders with tauopathy, namely Alzheimer's disease, frontotemporal dementia and Pick's disease.

Phosphorylation of tau by glycogen synthase kinase-3 beta in intact mammalian cells: the effects on the organization and stability of microtubules

The phosphorylation state of tau changes during neurodevelopment and highly phosphorylated tau accumulates in the paired helical filaments found in Alzheimer's disease. In non-neuronal mammalian cells transiently expressed tau is predominantly not phosphorylated at sites known to be phosphorylated in paired helical filaments. However this pattern of phosphorylation is induced by both glycogen synthase kinase-3 alpha and -3 beta and here we show that this results in a change in the intracellular properties of tau. Within cells tau is bound to cytoskeletal structures and causes changes in cellular cytoarchitecture with the induction of thick and stable microtubule bundles. This morphology is lost when tau is co-expressed with glycogen synthase kinase-3 beta; microtubules become less stable and are not bound by tau. Independently of any direct or indirect effects on tau, glycogen synthase kinase-3 beta induces some but relatively slight changes in microtubule organization with the loss of a prominent centrosomal microtubular origin. The cytoskeleton is critical to cell function and within post-mitotic neurons has a highly specialized structure induced, in part, by the neuronal-specific microtubule-associated proteins such as tau. In vitro studies have suggested that the properties of tau are regulated by phosphorylation as highly phosphorylated tau does not promote tubulin polymer assembly. We have demonstrated, in intact cells, that tau highly phosphorylated in the presence of glycogen synthase kinase-3 beta loses the properties of microtubule binding and stabilization, suggesting that regulation of tau phosphorylation by this enzyme might be an important mechanism whereby cytoskeletal function is modulated during neurodevelopment and lost in neurodegeneration.

Domains of tau protein, differential phosphorylation, and dynamic instability of microtubules

The dynamic instability of microtubules is thought to be regulated by MAPs and phosphorylation. Here we describe the effect of the neuronal microtubule-associated protein tau by observing the dynamics of single microtubules by video microscopy. We used recombinant tau isoforms and tau mutants, and we phosphorylated tau by the neuronal kinases MARK (affecting the KXGS motifs within tau's repeat domain) and cdk5 (phosphorylating Ser-Pro motifs in the regions flanking the repeats). The variants of tau can be broadly classified into three categories, depending on their potency to affect microtubule dynamics. "Strong" tau variants have four repeats and both flanking regions. "Medium" variants have one to three repeats and both flanking regions. "Weak" variants lack one or both of the flanking regions, or have no repeats; with such constructs, microtubule dynamics is not significantly different from that of pure tubulin. N- or C-terminal tails of tau have no influence on dynamic instability. The two ends of microtubules (plus and minus) showed different activities but analogous behavior. These results are consistent with the "jaws" model of tau where the flanking regions are considered as targeting domains whereas the addition of repeats makes them catalytically active in terms of microtubule stabilization. The dominant changes in the parameters of dynamic instability induced by tau are those in the dissociation rate and in the catastrophe rate (up to 30-fold). Other rates change only moderately or not at all (association rate increased up to twofold, rates of rescue or rapid shrinkage decreased up to approximately twofold). The order of repeats has little influence on microtubule dynamics (i.e., repeats can be re-arranged or interchanged), arguing in favor of the "distributed weak binding" model proposed by Butner and Kirschner (1991); however, we confirmed the presence of a "hotspot" of binding potential involving Lys274 and Lys281 observed by Goode and Feinstein, 1994. Phosphorylation of Ser-Pro motifs by cdk5 (mainly Ser 202, 235, and 404) in the flanking regions had a moderate effect on microtubule dynamics while phosphorylation at the "Alzheimer"-site Ser262 MARK eliminated tau's interactions with microtubules. In both cases the predominant effects of phosphorylation are on the rates of tubulin dissociation and catastrophe whereas the effects on the rates of association or rescue are comparatively small.

Cell cycle-dependent phosphorylation and microtubule binding of tau protein stably transfected into Chinese hamster ovary cells

Tau protein, a neuronal microtubule-associated protein, is phosphorylated in situ and hyperphosphorylated when aggregated into the paired helical filaments of Alzheimer's disease. To study the phosphorylation of tau protein in vivo, we have stably transfected htau40, the largest human tau isoform, into Chinese hamster ovary cells. The distribution and phosphorylation of tau was monitored by gel shift, autoradiography, immunofluorescence, and immunoblotting, using the antibodies Tau-1, AT8, AT180, and PHF-1, which are sensitive to the phosphorylation of Ser202, Thr205, Thr231, Ser235, Ser396, and Ser404 and are used in the diagnosis of Alzheimer tau. In interphase cells, tau becomes phosphorylated to some extent, partly at these sites; most of the tau is associated with microtubules. In mitosis, the above Ser/Thr-Pro sites become almost completely phosphorylated, causing a pronounced shift in M(r) and an antibody reactivity similar to that of Alzheimer tau. Moreover, a substantial fraction of tau is found in the cytoplasm detached from microtubules. Autoradiographs of metabolically labeled Chinese hamster ovary cells in interphase and mitosis confirmed that tau protein is more highly phosphorylated during mitosis. The understanding of tau phosphorylation under physiological conditions might help elucidate possible mechanisms for the hyperphosphorylation in Alzheimer's disease.

The role of tau phosphorylation in transfected COS-1 cells

Tau cDNAs from each of the six human isoforms were transfected into COS-1 cells and, in every case, more than one peptide was observed. The diversity of expressed isoforms was due to different levels of tau phosphorylation. Tau phosphorylation results in a decrease of the protein electrophoretic mobility. The major contribution to this mobility shift is due to the phosphorylation at the at the C-terminus of the molecule, as inferred from the expression of tau fragments. Phosphorylation takes place in some of the sites modified in neural cells and in the basis of AD patients. Copolymerization studies indicate that the level of phosphorylation, as well as the localization of the modified residues, may affect the binding of the protein to microtubules. These results indicate that phosphorylation regulates tau function inside the cell.

Stimulation of MAP kinase by v-raf transformation of fibroblasts fails to induce hyperphosphorylation of transfected tau

A proportion of the microtubule-associated protein, tau, is in an elevated state of phosphorylation in foetal and adult brain whereas all of the tau in paired helical filaments, which are characteristic of Alzheimer's disease is hyperphosphorylated; it is important therefore to elucidate the mechanisms that regulate tau phosphorylation. Here we describe results that show that although MAP kinase can hyperphosphorylate tau in vitro, activation of MAP kinase in transformed fibroblasts does not result in hyperphosphorylation of transfected tau, whereas glycogen synthase kinase-3 beta (GSK-3 beta) when co-transfected with tau does result in tau hyperphosphorylation. The findings imply that GSK-3 beta may be a stronger candidate than MAP kinase for inducing tau hyperphosphorylation in vivo.

Influence of phosphorylation on isoform composition and function of a microtubule-associated protein from developing Artemia

A novel 49 kDa protein, which exhibits nucleotide-dependent cross-linking of microtubules in vitro and localizes to ordered microtubule arrays by immunofluorescent staining, has been purified to apparent homogeneity from the brine shrimp, Artemia. Electrophoretic analysis involving isoelectric focusing and two-dimensional gels, supplemented by staining of Western blots with affinity-purified antibody, revealed that the 49 kDa protein consists of five isoforms with pI values of 6.0-6.2. The amount of 49 kDa protein increased slightly, but its isoform composition did not change significantly, during development of Artemia gastrula to third-instar larvae. Treatment with alkaline phosphatase caused the 49 kDa protein to undergo a mobility shift on gel electrophoresis, and, by use of an antibody to phosphoserine, at least two isoforms of the protein were shown to be phosphorylated. The serine phosphate, presumably added by a post-translational mechanism, did not influence binding of the 49 kDa protein to microtubules. Under conditions in which microtubules were cross-linked, the 49 kDa protein failed to interact with actin filaments. Our results demonstrate that the 49 kDa protein, like other structural microtubule-associated proteins such as tau and MAP2, is composed of several isoforms, some of which are phosphorylated. This protein has the potential to regulate the spatial distribution of microtubules within cells but does not link microfilaments to one another or to microtubules.

Phosphorylation and accumulation of tau without any concomitant increase in tubulin levels in Chinese hamster ovary cells stably transfected with human tau441

Eucaryotic expression vectors bearing a 1.4 kb cDNA encoding the 4 repeat isoform of human tau, tau441, in either the sense or anti-sense orientation with respect to a cytomegalovirus (CMV) promoter were constructed. The resulting constructs were used to transiently express tau in Chinese Hamster Ovary cells and to generate non-neuronal stable cell lines. Immunocytochemical studies of these cells show that tau is expressed in the sense but not the anti-sense or vector containing lines. Some of the cells expressing tau showed fine elongated processes which were stained by tau antibodies. The general tau immunostaining pattern appeared diffuse and punctuate. The expressed tau was seen both unbound and bound to microtubules. In some cells labeling with antibodies that specifically recognize hyperphosphorylation of tau was observed. The size of this population increased with increasing numbers of cell passages. However, no increase in steady-state tubulin level was observed following tau441 expression. These studies show that tau can accumulate in the cells without a concomitant increase in tubulin.

Microtubules and microtubule-associated proteins

Microtubule research is becoming increasingly diverse, reflecting the many isoforms and modifications of tubulin and the many proteins with which microtubules interact. Recent advances are particularly visible in four areas: microtubule motor proteins (their structures, stepping modes, and forces); microtubule nucleation (the roles of centrosomes and gamma-tubulin); tubulin folding (mediated by cytoplasmic chaperones); and the expanding list of microtubule-associated proteins, knowledge of their phosphorylation states, and information on their effects on microtubule dynamics.

Distribution of the phosphorylated microtubule-associated protein tau in developing cortical neurons

During brain development, the microtubule-associated protein tau presents a transient state of high phosphorylation. We have investigated the developmental distribution of the phosphorylated fetal-type tau in the developing rat cortex and in cultures of embryonic cortical neurons, using antibodies which react with tau in a phosphorylation-dependent manner. The phosphorylated fetal-type tau was present in the developing cortex at 20 days but not at 18 days of embryonic life and was not detected before four to five days in neuronal culture. The cyclin-dependent kinase p34cdc2 was expressed only in germinal layers in the embryonic brain and was not co-localized with phosphorylated tau. After 10 days of postnatal life, the phosphorylated tau progressively disappeared from cortical neurons, disappearing first from the deepest cortical layers where neurons are ontogenetically the oldest. Phosphorylated tau was found in axons and dendrites of cortical neurons at all developmental stages whereas unphosphorylated tau tended to disappear from dendrites during development. The timing of appearance of phosphorylated tau in the cortex, by comparison with the expression of other developmental markers, indicates that phosphorylated tau is present at a high level only during the period of intense neuritic outgrowth and that it disappears during the period of neurite stabilization and synaptogenesis, concomitantly to the expression of adult tau isoforms. In control cultures and in cultures treated with colchicine, the phosphorylated tau was not associated to cold-stable and to colchicine-resistant microtubules. These in vivo results suggest that the high expression of phosphorylated tau species is correlated with the presence of a dynamic microtubule network during a period of high plasticity in the developing brain.

Aggregation of tau protein by aluminum

Aluminum has been detected in Alzheimer neurofibrillary tangles, but the significance of its presence is unknown. The principal component of tangles is the paired helical filament (PHF), comprised of tau protein. We investigated whether aluminum could induce tau protein to form filaments or aggregate. When 10 microM bovine tau or non-phosphorylated recombinant human tau was combined with 400 microM or more aluminum, tau protein appeared to aggregate, observed as a dose-dependent decrease in electrophoretic mobility on SDS-PAGE. Tau appeared as a smear above the region of the expected tau bands and, at higher aluminum doses, failed to enter the gel. A tau fragment encompassing the microtubule binding domains did not show decreased mobility in the presence of aluminum, but did form aggregates that failed to electrophorese. However no fibrillar structures were observed in the aluminum-treated tau samples when observed by electron microscopy. The effect of aluminum on tau mobility was reversed by incubating with 1 mM deferoxamine. In contrast, the morphology of PHF fibrils was unaffected by deferoxamine treatment and the characteristic abnormal mobility of PHF-tau was not reduced by deferoxamine. This suggests that aluminum is not, by itself, a significant factor in maintaining the assembly of PHF-tau as fibrils or in its abnormal mobility on SDS gels. Aluminum treatment of 3T3 fibroblasts transfected with human tau resulted in toxicity, but did not change tau expression levels or induce tau aggregation. In conclusion, aluminum appears to induce isolated tau protein to aggregate in a phosphate-independent way, without the formation of fibrils.(ABSTRACT TRUNCATED AT 250 WORDS)

Phosphorylation of tau protein in tau-transfected 3T3 cells

The tau protein of Alzheimer paired helical filaments (PHFs) is aberrantly phosphorylated, as evidenced by its reactivity with several phosphate-dependent antibodies. We sought to identify whether this unusual phosphorylation state exists in tau expressed by transfected NIH 3T3 fibroblasts. Immunoblot analysis of cell clones transfected with constructs for either the 3-repeat or 4-repeat isoforms of tau revealed two tau bands, with the lower band migrating with unmodified tau in each case. Antibodies T3P and tau-1 were used to probe these bands, as they also react with PHF-tau in a phosphate-dependent manner. The epitopes for both antibodies were phosphorylated in both tau isoforms. Only the upper band was phosphorylated at the T3P site whereas phosphorylation at the tau-1 site was not always associated with a shift of tau mobility on gels. Tau in both bands was soluble, in contrast to PHF-tau, and was competent to bind to exogenously added bovine microtubules. Colchicine treatment of the cells resulted in an inhibition of phosphorylation at both sites, through an unknown mechanism. In conclusion human tau expressed in 3T3 cells was phosphorylated at the T3P and tau-1 sites as is PHF-tau, although no PHFs formed and the phosphorylated tau was competent to bind to microtubules.