Tau proteins form reversible clusters

Tau is the protein that helps stabilize the internal scaffolding of neurons. In Alzheimer’s, tau is famous for something else — the twisted fibrils and tangles that show up after neurons are already in trouble. But a 2025 paper in the *Journal of the American Chemical Society* argues that tau does not go straight from harmless to tangled. Under the right conditions, phosphorylated tau can first gather into soft, reversible droplets. ### What’s the new idea here? Basically, the paper says early tau assembly can be dynamic rather than permanent. The team compared ordinary tau with phosphorylated tau — tau carrying phosphate groups, a modification heavily linked to Alzheimer’s biology. Only the phosphorylated version spontaneously formed liquid-like condensates without extra crowding agents. And those condensates could disappear again when the temperature dropped. (pubs.acs.org) ### Why does phosphorylation matter so much? Phosphorylation changes tau’s shape and charge. In this work, native mass spectrometry picked up just 2 to 9 phosphate groups per molecule, but that was enough to loosen tau’s structure and reduce long-range internal contacts. That looser, more flexible form seems to make self-condensation possible. So the phosphate tags are not just labels of disease — they may actively reshape tau into a state that likes to cluster. (pubs.acs.org) ### What does “reversible droplets” actually mean? Think less “hard plaque” and more “protein dew.” When the sample warmed from 4°C to room temperature, phosphorylated tau separated into micrometer-scale droplets through liquid–liquid phase separation. When the sample cooled, the droplets dissolved. That reversibility is the point. It means at least one early tau state behaves like a switchable condensed phase, not a one-way march into insoluble junk. (pubs.acs.org) ### Does that mean tangles are reversible too? No — and this is the catch. The paper does not show that mature Alzheimer’s tangles melt away. It shows that an earlier phosphorylated tau state can reversibly condense before fibrils form. In the same experiments, neither tau nor phosphorylated tau made amyloid fibrils on their own. Fibrils needed extra inducers such as dextran sulfate or RNA. So the reversible step seems to sit upstream of the more familiar, harder-to-reverse aggregates. (phys.org) ### Where does RNA fit in? RNA also pushed tau into droplets, but in a different way. The paper separates two routes: phosphorylation-driven self-coacervation and RNA-assisted complex coacervation. The first was reversible with temperature. The RNA-driven route was not, and it unfolded on a different timescale. That matters because “tau droplets” are not one single thing — the path into condensation changes the biology. (phys.org) ### Why are people connecting this to Alzheimer’s? Because the cell experiments showed something odd and suggestive. Reporter cells exposed to phosphorylated tau in its liquid-droplet state developed aggregates at the nuclear envelope, and that pattern resembles nuclear-envelope tau changes described in Alzheimer’s brain tissue. That does not prove these droplets are the disease-driving species in humans. But it gives researchers a plausible intermediate state to study instead of treating tau aggregation as a jump straight to tangles. (phys.org) ### Is this the whole story? Not quite. Another 2026 preprint points the other way on a different tau fragment: phosphorylation at Ser352 kept Tau-R4 in liquid condensates at neutral pH, but falling pH pushed those condensates irreversibly into amyloid filaments. So the broader picture is not “tau clusters are reversible, full stop.” It’s that tau can pass through reversible and irreversible condensed states depending on the phosphorylation pattern, the tau region, and the chemical environment. (pubs.acs.org) ### Bottom line? The useful shift is conceptual. Tau aggregation may be less like a sudden crash into tangles and more like a sequence of state changes, with a soft, reversible stage early on. If that holds up in neurons and in vivo, therapies may need to target the conditions that let tau condense in the first place — before the protein hardens into the pathology everyone already knows. (pubs.acs.org) (biorxiv.org)