Phosphorylation of Ser262 strongly reduces binding of tau to microtubules: distinction between PHF-like immunoreactivity and microtubule binding.

Tau protein, a component of Alzheimer paired helical filaments, can be phosphorylated by several kinases. Of particular interest is the phosphorylation at Ser/Thr-Pro motifs because the resulting state of tau is similar to that found in Alzheimer's disease, as judged by its immunoreactivity. This state can be mimicked by a brain extract kinase activity and by MAP kinase. We have now studied the effect of these modes of phosphorylation on the interaction between tau and microtubules. Although MAP kinase efficiently phosphorylates many Ser/Thr-Pro motifs of tau, its effect on microtubule binding is only moderate. By contrast, phosphorylation of a single residue, Ser262, has a major effect on binding. Ser262 is not phosphorylated by MAP kinase or other proline-directed kinases, but is phosphorylated by a 35/41 kd kinase in brain, whose purification is described.

Tau domains, phosphorylation, and interactions with microtubules.

We consider the interactions of tau protein with microtubules from two points of view, phosphorylation and domain structure. Tau can be phosphorylated at many sites and by several kinases, notably by proline-directed kinases (MAPK, GSK-3, cdk5) which generate Alzheimer-like antibody epitopes. Other kinases phosphorylate Ser 262, a site that has a particularly pronounced influence on the affinity of tau for microtubules. All of these sites can be cleared by phosphatases PP-2a and calcineurin. The site Ser262 lies within the repeat domain of tau. However, when probing the domains of tau for their effects on microtubule binding, nucleation, assembly, or bundling, the repeat domain has only a weak influence. Whereas the repeat domain of tau binds to microtubules with low affinity, repeat-less tau binds strongly yet unproductively in terms of microtubule assembly. Productive binding of tau to microtubules depends on the combination of (some) repeats with the flanking regions, as if the flanking regions acted as "jaws" for the proper positioning of tau on the microtubule surface.

The Alzheimer-like phosphorylation of tau protein reduces microtubule binding and involves Ser-Pro and Thr-Pro motifs.

Tau protein can be transformed into an Alzheimer-like state by phosphorylation with a kinase activity from brain [Biernat et al. (1992) EMBO J. 11, 1593-1597]. Here we show that the phosphorylation at Ser-Pro motifs strongly decreases tau's affinity for microtubules. The major reduction occurs during the first of the three main stages of phosphorylation. The data explain the lower stability of microtubules resulting from the pathological tau phosphorylation.

Glycogen synthase kinase-3 and the Alzheimer-like state of microtubule-associated protein tau.

The Alzheimer-like state of tau protein includes phosphorylation by a proline-directed Ser/Thr kinase present in normal or pathological human brain. Extending earlier results on MAP kinase, we show here that the proline-directed kinase, GSK3, can induce an Alzheimer-like immune response involving several distinct and phosphorylatable epitopes at Ser-Pro motifs, as well as a gel mobility shift, similar to MAP kinase. Both kinases behave like microtubule-associated proteins in that they co-purify through cycles of assembly and disassembly, and both kinases are directly associated with paired helical filaments.

Structure, microtubule interactions, and phosphorylation of tau protein.

This paper summarizes recent structural and functional studies on tau protein, its interactions with microtubules, its self-assembly into paired helical filaments (PHF)-like fibers, and its modification by phosphorylation. The structure of tau in solution resembles that of a random coil. Both tau and Alzheimer PHFs have very little secondary structure, making it improbable that the assembly of tau into PHFs is based on interacting beta sheets. Tau's binding to microtubules can be described by a "jaws" effect. The domain containing the repeats binds very weakly, while the flanking regions (jaws) bind strongly, even without the repeats. However, only the combination of flanking regions and repeats makes binding productive in terms of microtubule nucleation and assembly. Although the majority of Alzheimer-like phosphorylation sites are outside the repeats they have only a weak influence on binding, whereas the phosphorylation at Ser262 inside the repeats inhibits binding and makes microtubules dynamically unstable. This site can be phosphorylated by kinases present in brain tissue, and it is uniquely phosphorylated in Alzheimer brain.

Microtubule-associated protein tau, paired helical filaments, and phosphorylation.

This paper summarizes our recent studies on microtubule-associated protein tau and its pathological state resembling that of the paired helical filaments of Alzheimer's disease. The Alzheimer-like state of tau protein can be identified and analyzed in terms of certain phosphorylation sites and phosphorylation-dependent antibody epitopes. It can be induced by protein kinases which tend to phosphorylate serine or threonine residues followed by a proline; this includes mitogen-activated protein kinase (MAPK) and glycogen-synthase kinase 3 (GSK-3). Both of these are tightly associated with microtubules as well as with paired helical filaments. Structurally, tau appears as a rod-like molecule; it tends to self-associate into dimers whose monomers are antiparallel. Constructs of truncated tau made up of antiparallel dimers of the microtubule binding domain can be assembled into paired helical filaments in vitro.

Domains of tau protein and interactions with microtubules.

The role of the neuronal microtubule-associated protein tau has been studied by generating a series of tau constructs differing in one or several of its subdomains: length and composition of the repeat domains, extensions of the repeats in the N- or C-terminal direction, constructs without repeats, assembly vs projection domain, and number of N-terminal inserts. The interaction of the mutant tau proteins with microtubules was judged by several independent methods. (i) Direct binding assays between tau and taxol-stabilized microtubules yield dissociation constants and stoichiometries. (ii) Light scattering and X-ray scattering of assembling microtubule solutions reflect the capacity of tau to promote microtubule nucleation, elongation, and bundling in bulk solution. (iii) Dark field microscopy of assembling microtubules allows one to assess the efficiency of nucleation and bundling separately. The repeat region alone, the N-terminal domains alone, or the C-terminal tail alone binds only weakly to microtubules. However, binding is strongly enhanced by combinations such as the repeat region plus one or both of the flanking regions which could be viewed as "jaws" for tau on the microtubule surface (the proline-rich domain P upstream of the repeats and the "fifth" repeat R' downstream). Such combinations make tau's binding productive in terms of microtubule assembly and stabilization, while the combination of the flanking regions without repeats binds only unproductively. Efficient nucleation parallels strong binding in most cases, i.e., when a construct binds tightly to microtubules, it also nucleates them efficiently and vice versa. In addition, the proline-rich domain P in combination with the repeats R or the flanking domain R' causes pronounced bundling. This effect disappears when the N-terminal domains (acidic or basic) are added on, suggesting that the tau isoforms are not "bundling proteins" in the proper sense. In spite of the wide range of binding strength and nucleation efficiency, the stoichiometries of binding are rather reproducible (around 0.5 tau/tubulin dimer); this is in remarkable contrast to the effect of certain types of phosphorylation which can strongly reduce the stoichiometry.

Phosphorylation-dependent epitopes of neurofilament antibodies on tau protein and relationship with Alzheimer tau.

We have studied the phosphorylation of tau protein from Alzheimer paired helical filaments, of tau from normal human brain, and of recombinant tau isoforms. As a tool we used monoclonal antibodies against neurofilament protein [Sternberger, N., Sternberger, L. & Ulrich, J. (1985) Proc. Natl. Acad. Sci. USA 82, 4274-4276] that crossreact with tau in a phosphorylation-dependent manner. This allowed us to deduce the state of phosphorylation in normal and pathological tau, as well as antibody epitopes. The epitope of antibody SMI33 is at the first Lys-Ser-Pro sequence motif (residues 234-236) and requires an unphosphorylated Ser-235. Antibody SMI31 binds between Ser-396 (in the second Lys-Ser-Pro motif) and Ser-404, both of which must be phosphorylated. SMI34 has a conformational epitope that depends on the interaction between regions on either side of the microtubule-binding region; it also requires phosphorylation. The phosphorylatable serines detected by the SMI antibodies are part of Ser-Pro motifs and can be phosphorylated by a protein kinase activity that can be used to induce a paired helical filament-like state in human brain tau in vitro. The phosphates are incorporated in several stages that can be identified by antibody reactivity and gel shift. This suggests a role for the phosphorylation sites in Alzheimer disease, as well as the involvement of a Ser-Pro-directed protein kinase.

Phosphorylation of microtubule-associated protein tau: identification of the site for Ca2(+)-calmodulin dependent kinase and relationship with tau phosphorylation in Alzheimer tangles.

The microtubule array in neuronal cells undergoes extensive growth, dynamics and rearrangements during neurite outgrowth. While little is known about how these changes are regulated, microtubule-associated proteins (MAPs) including tau protein are likely to perform an important role. Tau is one of the MAPs in mammalian brain. When isolated it is usually a mixture of several isoforms containing between 341 and 441 residues that arise from alternative splicing. Tau can be phosphorylated by several protein kinases. Phosphorylation at certain sites results in major structural and functional changes, as seen by changes in electrophoretic mobility, interaction with microtubules, molecular length and elasticity. Here we show that the sites of phosphorylation by four kinases (PKA, PKC, CK and CaMK) all lie in the C-terminal microtubule-binding half of tau, but only the phosphorylation by CaM kinase shows the pronounced shift in electrophoretic mobility characteristic for tau from Alzheimer neurofibrillary tangles. By using a combination of limited proteolysis, protein sequencing and protein engineering we show that a single phosphorylation site is responsible for this shift, located at Ser 405 in the C-terminal tail of the protein outside the region of internal repeats. Phosphorylation at this site not only reduces the electrophoretic mobility of tau, it also makes the protein long and stiff, as shown earlier. The site is likely to be phosphorylated in tau from Alzheimer neurofibrillary tangles.