Our Tau Tales from Normal to Pathological Behavior

INTRODUCTION

Several dementias have in common the formation of intracellular filamentous deposits formed of the microtubule-associated protein tau, in abnormally hyperphosphorylated forms. Through abnormal tau function, they apparently share a common disease mechanism, and are collectively known as tauopathies. This family of diseases includes Alzheimer’s disease (AD), frontotemporal dementia with Parkinsonism linked to chromosome 17 (FTDP-17), amyotrophic lateral sclerosis, cortical basal degeneration, dementia pugilistica, Pick’s disease, progressive supranuclear palsy, and tangle-only dementia. Despite their diverse phenotypic manifestations, brain dysfunction, and degeneration, these tauopathies are linked to the progressive accumulation of filamentous hyperphosphorylated tau inclusions which, in the absence of other disease-specific neuropathological abnormalities, provide circumstantial evidence implicating abnormal tau in the onset and/or progression of neurodegenerative disease.

Our tau-research project started in Khalid and Inge Grundke-Iqbal’s laboratory in 1992. The microtubule associated protein tau, originally described by the Kirschner laboratory in 1975 [1], was identified as the building block from the filamentous aggregates found in the neurons of AD patients. In the abnormal state, tau was hyperphosphorylated. Tau is normally found in cells as a phosphoprotein with ∼3 moles of phosphate per mole of the normal protein. The hyperphosphorylated tau contains a significantly higher phosphate content than the normal tau, up to ∼7–10 moles of phosphate per mole of protein [2] which includes the appearance of new phosphorylation sites. In the central nervous system, tau is a family of six proteins derived from a single gene by alternative splicing of the pro-mRNA [3, 4]. The human brain tau isoforms range from 352 to 441 amino acids, differing in whether they contain three or four tubulin-binding domains/repeats (R) which consist of 31 or 32 amino acids near the C-terminus. At the N-terminus of tau there are two, one, or no inserts of a 29-amino acid repeat (N). Isoform expression and degree of phosphorylation are developmentally regulated. Fetal tau is mainly composed of the 3R0N isoform and is highly phosphorylated normally, but lacks several of the phosphorylation sites seen in paired helical filament (PHF) tau. The degree of phosphorylation of the six isoforms decreases with age, probably because of the activation of phosphatases [5]. All six isoforms have been observed in hyperphosphorylated states in PHFs from AD patients [6–9].

Our interest was in the biological activity of tau as a microtubule associated protein (MAP). We found that hyperphosphorylated tau from AD brains was unable to promote microtubule assembly and that, more importantly, the abnormal protein could inhibit the normal activity of tau and other MAPs. AD P-tau was able to disrupt preformed microtubules and, for the first time, we described a prion-like behavior in tau: the abnormal protein was able to bind normal protein and turn it into an AD P-tau like molecule. We were the first to describe this gain of toxic function in tau in AD patients [10–13]. The toxic behavior was prevalent when AD P-tau was not forming filaments and it was lost upon dephosphorylation.

The majority of axonal proteins are synthesized in the neuronal cell body and transported through the axons along the microtubule tracks. Axonal transport occurs throughout the life of a neuron and is essential to its growth and survival. In vitro, tau promotes the assembly of tubulin into microtubules and stabilizes the assembled ones [1]. In the neurons of patients with AD, the microtubule system is disrupted, interrupting axonal transport, thus preventing vesicles from reaching the synapses. We have shown that hyperphosphorylated tau can disrupt these microtubules by sequestering normal tau through protein-protein interactions [10–13]. As a result, slowly and steadily, the synapses deteriorate by retrograde degeneration.

The discovery in 1998 of mutations in the tau gene, which co-segregate with the disease in frontotemporal dementia, provided unequivocal evidence that tau abnormalities alone are enough to cause neurodegenerative disease [14–16]. Three different types of tau mutations have been described: missense, intronic, and one-deletion (ΔLys280). The missense mutations resulted in point mutations that conferred disease progression (i.e., P301L and R406W). The intronic 5' to exon 10 mutations resulted in overexpression of 4R tau proteins, disrupting the balance of 3R/4R proteins in neurons [15, 16]. The exact molecular mechanism of neurodegeneration in the affected patients is not yet understood. Like individuals with AD, FTDP-17 patients show accumulations of hyperphosphorylated tau as neurofibrillary tangles in every case. Hyperphosphorylated tau arises from the emergence of new phosphorylation sites in the protein. All the mutations discovered in tau are dominant, suggesting that the effect of tau mutations result in a gain of toxic function by the protein [17]. The research which was seeded in the Iqbal laboratory focused on understanding the mechanism of tau induced neurodegeneration by first understanding tau’s biological activity as well as the biological activity of hyperphosphorylated tau, using biochemistry, cellular biology, and, most recently, animal models generated to express PH-Tau in neuronal cells. We proposed that mutations on tau associated with disease (FTDP-17) altered its conformation to make it a better substrate for kinases [18]. Using phospho-mimetics, we found that the minimum phospho-sites necessary to acquire such a toxic behavior of tau were at 199, 212, 231, and 262, and tau pseudophosphorylated at those sites behaved as pathological tau [19]. Here is a brief tale of our work on tau.

MICROTUBULES AND TAU IN ALZHEIMER’S DISEASE

When the neurons of patients with AD are studied, a decrease in microtubules is observed with a concurrent increase in the concentration of tau [2]. Three different pools of tau can be observed from the brains of AD patients: AD tau, not hyperphosphorylated and most similar to normal tau; AD P-tau, soluble hyperphosphorylated tau; and PHF-tau, insoluble and hyperphosphorylated tau. Levels of AD tau are decreased by about 60% compared to tau found in normal brain. AD P-tau, as well as normally phosphorylated tau, can be isolated from AD brain in solution [2]. Using brain extracts, we studied the biological activity of tau from AD brains to determine the microtubule-promoting activity in in vitro assembly assays [10]. We found that AD tau has normal microtubule-promoting activity; conversely AD P-tau did not promote microtubule assembly. Even more we found that AD P-tau inhibited the microtubule assembly promoted by normal tau, MAP1A, MAP1B, and MAP2 [20]. After treatment with phosphatases, the microtubule-promoting activity was recovered implicating phosphorylation of tau as the mechanism of microtubule disruption. Interestingly, AD P-tau preincubated with normal tau prior to the addition of tubulin both inhibited the normal microtubule–promoting activity and destroyed microtubules already present. This was probably due to interactions between tau and AD P-tau thereby sequestering it from the tubulin.

AD P-TAU HAS A PRION-LIKE BEHAVIOR

Using both solid phase and solution binding assays, we verified that AD P-tau was able to bind normal tau [11]. Quantitation of the solution binding assay indicated that the AD P-tau binding to normal tau was non-saturable, and visualization by electron microscopy showed us that the products were bundles of filaments [11]. These results suggested that hyperphosphorylation of tau could change the conformation of the protein in such a way that this change could be transferred to normal protein which would seed tau filament self-assembly. This was a new hypothesis in the field of tau biochemistry. The ability of hyperphosphorylated tau to bind normal tau was confirmed by Vandebroek et al. [21] in yeast. They expressed the human largest four-repeat protein (4R2N) and the human largest three-repeat isoform (3R2N), and demonstrated that human tau expressed in yeast acquired pathological phospho-epitopes, assumed a pathological conformation, and formed aggregates. These processes were modulated by yeast kinases Mds1 and Pho85, orthologues of GSK-3β and cdk5 which are kinases known to phosphorylate tau in humans (as reviewed in [22]). They observed that a) tau aggregated more when it was more phosphorylated, b) the mobility in SDS electrophoresis was higher with increased phosphorylation, c) isolated hyperphosphorylated tau was able to assemble into filaments, and d) the isolated hyperphosphorylated tau was able to nucleate the assembly of the normal, non-phosphorylated tau. The authors proposed that hyperphosphorylated tau is the biochemically stable form of tau that is the actual seed or nucleation factor that initiates and promotes the aggregation of tau, as we had proposed for hyperphosphorylated tau isolated from AD brain almost 10 years prior [11]!

The conformational change transfer by AD P-tau to normal tau is a property of a prion protein, and we were the first to describe this property in tau and to show that it was due to hyperphosphorylation. This prion-like activity of AD P-tau was further determined to disrupt the microtubules formed by normal tau or by the other neuronal MAPs, including MAP1b and MAP2 [11, 20]. Furthermore, amorphous aggregates are formed when AD P-tau binds to MAP1b and MAP2 [20].

TAU SELF-ASSEMBLY AND ‘AD P-TAU-LIKE’ PROTEIN BEHAVIOR IS INDUCED BY HYPERPHOSPHORYLATION

In AD, hyperphosphorylation of tau appears to precede the appearance of the tangles [8]. As described above, tau is a phosphoprotein that in its toxic, hyperphosphorylated state has an increase of 2–4 times the phosphate per mole of protein due to an increase in the number of phospho-sites [2]. Degenerating neurons appear to have tau that has self-assembled into tangles composed of PHFs and short filaments (SFs). AD P-tau was able to self-assemble into these tangles (Fig. 1A) at varying pHs [13]. The PHFs generated by AD P-tau in vitro had similar dimensions to those of AD PHFs extracted from the brain. These filaments all contained a wide part of ∼20 nm, which narrowed to ∼10 nm at every ∼80 nm. Within the bundles of PHFs, some 4-nm protofilaments and SFs of ∼15 nm, similar to the SFs in AD, were also observed. The self-assembly was halted by dephosphorylation of AD P-tau (Fig. 1A) [10] suggesting that hyperphosphorylation of tau is a requirement for its self-assembly into tangles of filaments of varying sizes.

OPEN IN VIEWER

Fig.1

In vitro polymerization of AD P-tau and recombinant tau into tangles of PHF/SF and the effects of dephosphorylation. A) AD P-tau was purified as described in Kopke et al. [2], 0.4 mg/mL (a) without pretreatment and (b) dephosphorylated by alkaline phosphatase was incubated for 90 min and the products of assembly were examined by negative stain electron microscopy. Dephosphorylation completely abolished AD P-tau polymerization. B) Recombinant tau, 0.5 mg/mL, was incubated with rat brain extract as a source of protein kinases in the presence of (a) ATP to induce hyperphosphorylation of tau or (b) non-hydrolyzable ATP, AMP-PNP as a contol. C) Recombinant 2N4R tau with FTDP-17 mutations, 0.5 mg/mL, were incubated with rat brain extract plus ATP to induce hyperphosphorylation then analyzed by negative stain electron microscopy (1-h incubation: a, V337M; b, R406W; 4-h incubation: c, R406W; 6-h incubation: d, P301L; G272V). The research for A and B was originally published in [13]; and the research for C was originally published in [18].

To confirm the role of hyperphosphorylation in the conversion of normal tau into a toxic molecule that has aggregation propensities, the six isoforms of recombinant tau (r-tau) were individually treated with protein kinases present in normal brain extract and followed its ability to bind normal tau and to inhibit its microtubule-promoting activity [12, 13]. Rat brain extract treated r-tau became hyperphosphorylated with the increase to ∼12 moles of phosphate per mole of the protein (phosphorylated tau, P-tau) which is similar to AD P-tau. P-tau also bound to normal tau and was able to self-assemble into tangles of PHFs/SFs in a phosphorylation dependent manner and inhibited the microtubule assembly activity (Fig. 1B) [13]. These results suggested that hyperphosphorylation could convert tau into an AD P-tau-like state.

Several reports have shown that FTDP-17 mutations decrease tau’s ability to promote tubulin assembly into microtubules [23] or increase the ability of tau to self-assemble [24]. We proposed that these mutations may change the conformation of tau making it a better substrate for phosphorylation [25]. Phosphorylation assays using r-tau with FTDP-17 mutations R406W, P301L, V337M, or G272V resulted in faster rate and greater phosphorylation extent (∼16–18 moles versus ∼12 moles of phosphate per mole protein) than normal tau in vitro [18]. This increase in phosphorylation probably correlates to an increased number of sites that become modified based on the higher phosphorylation stoichiometry. We also found that fewer moles of phosphate per mole of protein were required for filament formation in the mutant proteins (Fig. 1C).

Upon excess phosphorylation, tau will acquire the ability to bind normal tau. This occurs maximally after the incorporation of ∼4 moles of phosphate per mole of protein [18] and polymerizes into filaments after ∼10 moles of phosphate per mole of protein [13, 18]. These results suggest that at least two different conformational states of tau are induced by phosphorylation: one in which the hyperphosphorylated tau is able to bind normal tau, and one in which it is able to self-assemble into filaments. These mechanisms may be regulated by changes in phosphorylation mediating the neutralization of the charged regions on the protein. As previously shown, the N-terminal inserts of tau can neutralize the positive charge of the flanking regions of tau and induce self-assembly of unmodified protein. In agreement with our model, oxidation of tau by the addition of carbonyls to Lys, which also neutralizes the charge, increases tau filament formation [26]. Different mechanisms can lead to the conformational change to acquire tau toxic conformation, that as we have shown, can be transferred to the normal, unmodified protein.

As discussed above, hyperphosphorylation confers upon tau a toxic property in which microtubule stability is decreased because of its ability to bind normal tau and MAPs. It is possible that this toxic property is lost upon increased self-assembly as PHF-Tau do not bind normal MAPs and do not inhibit microtubule assembly [27]. Our hypothesis is that tau gets hyperphosphorylated, binds normal MAPs, disrupts microtubules, and interrupts axoplasmic transport, with the consequent degeneration of the synapse. If tau self-assembles into PHF/SF, then it cannot bind normal MAPs, and the microtubules can be still functional.

WHAT IS HYPERPHOSPHORYLATED TAU? HOW CAN WE STUDY THE GAIN OF TOXIC FUNCTION?

Hyperphosphorylated tau is understood to be a protein in which there is an increase in moles of phosphate per mole of protein. However, there is much discussion as to whether hyperphosphorylation actually relates to a general increase in this ratio or increased phosphorylation at specific sites within the molecule. One method to mimic the negative charge of the phosphate group and length of the side chain is pseudophosphorylation where the codons for Ser or Thr residues are replaced with that for Glu. This is a widely accepted approach to mimic phosphorylation [19, 28–32]. A mouse model was developed to study hyperphosphorylated tau using a tau protein with 10 pseudophosphorylation sites [33]. This mouse did not appear to have any of the hallmark traits of dementia-related neurodegeneration indicating that it is more likely phosphorylation at specific sites than overall phosphate per molecule.

This understanding led us to more closely examine the role of protein conformation as the structure of tau, and other intrinsically disordered proteins, may be determined by long-range interactions which can be modulated by phosphorylation and other post-translational modifications [34]. Intermolecular association of tau has been linked to interactions through the microtubule binding domain (MTBD) while self-assembly appears to be inhibited by the flanking regions of this domain [18, 35] (Fig. 2). The presence of the two N-terminal inserts of tau, which are highly negative, can induce tau self-assembly potentially by neutralizing the charge of the flanking region, as we have shown that non-modified full length tau is able to self-assemble in short filaments [13]. As a disordered protein, tau has little defined secondary structure. Nevertheless, the study of tau structure in PHF/SF from AD brains showed that the structure of tau is important in tau self-assembly [36], confirming our observations on tau self-assembly from the whole molecule. Regions of tau have a strong basic charge (pI >  9) and are separated from other domains by Pro residues, which can induce a bend in the amino acid chain. These very basic regions that are N-terminal to the microtubule binding domains can mask the intermolecular attraction of the MTBD. Three residues in this region, Thr212, Thr231, and Ser262, appear to be 50% phosphorylated when tau begins to polymerize [18] thus decreasing their theoretical pI and increasing the probability of tau self-assembly. On the C-terminal side of the MTBD there is a basic region up to Pro397 that is followed by an acidic segment. Phosphorylation at Ser396 and/or Ser404 may open up this segment and increase intermolecular interactions thereby increasing tau self-assembly.

OPEN IN VIEWER

Fig.2

A hypothetical scheme of the phosphorylation-induced self-assembly of wild-type and FTDP-17 mutated tau proteins. Tau self-assembles mainly through the microtubule binding domain/repeat R3 in 3R tau proteins and through R3 and R2 in 4R tau proteins (R2 and R3 have β-structure). Regions of tau molecule both N-terminal and C-terminal to the repeats are inhibitory. Hyperphosphorylation of tau neutralizes these basic inhibitory domains, enabling tau-tau interaction. In the case of the C-terminal region beyond Pro397 (398–441), a highly acidic segment masks the repeats. Phosphorylation (red Ps) of tau at Ser396 and/or 404 opens this segment, allowing tau-tau interaction through the repeats. FTDP-17 mutations make tau a more favorable substrate for phosphorylation than the wild-type tau. The mutated tau proteins achieve the conformation required to self-assemble at a lower level of incorporated phosphate. Although the FTDP-17 mutant tau proteins have conformations that are more prone to polymerize, in the absence of hyperphosphorylation, the highly basic segments and the C-terminus interfere with polymerization. Phosphorylation sites are indicated by red Ps at Ser/Thr positions in tau (left panel): 199, 202, 205, 212, 231, 235, 262, 396, 404, and 422; and in FTDP-17 mutant tau (right panel): 199, 212, 231, 262, and 396, respectively. This figure was reproduced with permission from Alzheimer’s & Dementia [22].

Using this information, pseudophosphorylated sites were studied in the presence and absence of mutations related to FTDP-17, since it was shown to increase the phosphorylation effect. To determine which residues to change to Glu, recombinant tau was phosphorylated in vitro and the phosphorylated sites were determined at the point that self-assembly occurred, about 5 moles of phosphate incorporated per mole of protein by about 2 hours of incubation. Upon analysis, nine sites were found to be phosphorylated about 50% : Ser199, Ser202, Ser205, Thr212, Thr231, Ser235, Ser262, Ser396, and Ser404. From these results, we studied the tau gene (MAPT) mutated at each site to Ala (non-phosphorylatable) or Glu (pseudophosphorylated) in the normal tau or R406W background. Upon transfection into PC-12 cells, the vectors containing Ala mutations acted similarly to non-mutated tau at each of the sites tested. Mutations to Glu, in most cases, resulted in tau dissociation from tubulin but complete microtubule disruption was not observed [19]. This indicated to us that a single phosphorylation event was not enough to convert tau into an AD P-tau like toxic molecule.

After multiple combinations containing two or three pseudophosphorylation sites, it was determined that the strongest effect was observed with the triple mutant tauT212E/S235E/S262E which bound weakly to microtubules in CHO cells and decreased tubulin staining. This pseudophosphorylated tau appeared to be aggregated in both the cytoplasm and nuclear space and was able to sequester normal tau in a manner similar to that of tau isolated from AD brain [19]. When compared to wildtype tau, we found that Ser199 in the pseudophosphorylated tau was very highly phosphorylated. This suggests that phosphorylation at these four sites is able to convert tau into a toxic species which was enhanced by the FTDP-17 mutation R406W. We decided that tau hyperphosphorylation was due to phosphorylation at specific sites within the molecule and we generated phospho-sites at these four residues with the R406W mutation and we named it Pathological Human Tau (PH-Tau).

TOXIC GAIN OF FUNCTION OBSERVED IN A TAUOPATHY MODELS

We generated tau-transgenic flies to study PH-Tau effects in vivo. We found in Drosophilia that PH-Tau expressed in a pan-neuronal fashion has a marked effect on the olfactory learning [37]. We have recently developed and characterized a new mouse model in which PH-Tau is expressed in neuronal cells under the control of the CaMKII promoter [38]. This model expressed the protein at two different levels: PH-Tau_low (4% of normal tau when the promoter is repressed) and PH-Tau_high (14% of normal tau when the promoter is induced). These levels may be correlated with the different levels of in vitro phosphorylation that change the tau binding abilities described above. Substantial differences in cognitive abilities, synaptic morphology, and neuronal loss were observed between PH-Tau_low and PH-Tau_high [38]. Low levels of PH-Tau resulted in cognitive deficits and reduced CA1 synapse number, synaptic protein levels were reduced and PH-Tau appeared to be present in the neuronal body and nuclei. With high levels of PH-Tau there was neuronal death primarily in CA3 as well as astrocytosis in certain brain regions with no apparent effect on CA1 synapses and the processes of the neurons had disappeared [38]. Interestingly, PH-Tau had distinct biochemical properties when expressed at low and high levels that could account for the different phenotypes. At low PH-Tau, we observed a high molecular weight tau species (∼100 kD) that was significantly reduced when high levels of PH-Tau were induced. Furthermore, PH-Tau in the induced animals was truncated at 421. Preliminary work in this mouse model indicates disruptions in mitochondrial morphology in the CA1 and CA3 regions of the hippocampus of mice expressing PH-Tau. These changes may be due to mitochondrial dysfunction that has been shown to play an increasing role in AD [39–46].

CONCLUSIONS AND VISIONS

Twenty years ago, our work showed that hyperphosphorylated tau sequesters healthy tau protein and causes healthy tau to become pathological though the mechanism of neuronal death was unclear [11]. Through the years, and through much hard work by many researchers in the field, a clearer picture is being drawn. Our studies, including biochemical data using tau from AD brains and recombinant tau as well as the studies using pseudophosphorylated tau, allow us to better understand the modifications of tau that can modulate different events at the cellular levels with important consequences for its physiology (Fig. 3). We have observed tau translocation into the cell nucleus [19]. Presence in the nucleus can cause hyperphosphorylated tau to alter the interaction with DNA [47] and may influence protein expression, in turn affecting cellular function. It is known that hyperphosphorylated tau, especially when it has other mutations, causes not only a destabilization of the microtubules (see above), but also the actin microfilaments [48]. Disruption of the microfilaments in cells can lead to zeiosis of the cell membrane. We have observed in cell culture that as the membrane pinches off during exocytosis, there is the release of hyperphosphorylated tau-containing membrane vesicles throughout the surrounding cellular environment (data not shown). We propose that these vesicles drift toward, and interact with, surrounding cells and that the contents are taken up by endocytosis. As the pathological protein moves from cell to cell it can sequester more healthy tau, propagating its prion-like behavior from neuron to neuron, causing a disruption of all cytoskeleton components, destabilizing the organelles, disrupting protein synthesis, and eventually inducing zeiosis and continuing disease transmission (Fig. 3A).

OPEN IN VIEWER

Fig.3

Proposed mechanisms of neurodegeneration. A) PH-tau induces not only microtubule disruption, but it is also translocated in the nucleus, causes intracellular degeneration, protein aggregation, and vacuole formation. The presence of tau in the nucleus might be involved in alterations of protein expression. As a result of cell death or cell altered metabolism tau can be released from the cells, it is possible that the released conformationally altered tau molecule can propagate the disease to neighboring cells. This figure was reproduced with permission from Alzhiemer’s & Dementia [22]. B) (Left) Low level of PH-tau expression results in translocation to the nucleus, synaptic dysfunction, and mitochondrial disruption. The presence of tau in the nucleus might be involved in alterations of protein expression. B) (Right) High levels of PH-Tau expression results in protein aggregation, microtubule disruption, and loss of synapses. As a result of cell death or cell altered metabolism tau can be released from the cells, it is possible that the released conformationally altered tau molecule can propagate the disease to neighboring cells. This figure was reproduced with permission from the editors of Protein Folding Disorders in the Central Nervous System [50].

We could picture different scenarios where the levels of hyperphosphorylated tau start appearing in the cell because of kinase overactivity, phosphatase deficiency, changes in the substrate conformation, failure in the clearance system, or a combination of them. At the beginning of the diseases, the conformationally modified tau might move in the cell, translocating in the nucleus, locating in synapses, interfering with mitochondria homeostasis (Fig. 3B left). As a consequence, cognitive impairment without significant structural changes might be observed [38]. As the pathological tau increases in the neurons, the toxic effect on the cytoskeleton and the retrograde neurodegeneration appears (Fig. 3B right). Our results reinforce the key role of tau in the development of pathology, in AD and other tauopathies. Despite the different mechanisms, it appears that reduction in the levels of hyperphosphorylated tau remains a key target for tauopathies, in combination with therapies to prevent cytoskeleton disruption [49]. Understanding the mechanism of transmission will allow us to design blockers of tau secretion and/or uptake, or regulators of microglia or other mechanisms to reduce extracellular pathological tau, which will help us halt the progression of the disease. From our models, it is apparent that low levels of conformationally altered tau is enough to trigger pathological effects. To understand these mechanisms triggered by tau but seemingly unrelated to microtubule structure will point to new therapeutic target development.