For decades, the protein tau has been cast as a villain in the story of Alzheimer’s disease. It forms the neurofibrillary tangles that are a hallmark of the condition; its abnormal hyperphosphorylation is a defining pathological feature; and anti-tau therapies are among the most actively pursued strategies for treating dementia. But a new study from Flinders University in Adelaide, Australia, published in Nature Communications, reveals that tau has a secret life — one that is essential for healthy memory function.
The study, led by Arne Ittner of the Flinders Health and Medical Research Institute, shows that a precise, controlled chemical modification of tau — phosphorylation at a specific site called threonine-205 (T205) — is required for the formation of long-lasting memories in mice. Without it, memories are either not stored properly or become inaccessible to natural recall.
“Tau is not merely a problem protein that appears in disease,” Ittner said. “It plays a fundamental role in how the brain encodes information. This has direct implications for the development of anti-tau therapies, because removing tau or blocking its function could harm normal cognition.”
Engram organiser and noise reducer
The researchers used a combination of behavioural tests, genetic manipulations, optogenetics, and mass spectrometry to dissect tau’s role in memory formation. They employed three independent memory paradigms — cued fear conditioning, touchscreen pairwise discrimination, and the Morris water maze — and tracked memory at different time points ranging from 24 hours to four months.
Three distinct functions emerged.
First, tau acts as an engram cell organiser. During the encoding of a memory, a specific subset of neurons — the engram ensemble — is selected to store the trace. The team found that tau’s phosphorylation at T205, triggered by the learning event itself, is required for the correct selection of which neurons become part of the engram. Without it, the engram was poorly defined.
Second, tau reduces background neural noise. By analysing c-Fos expression — a marker of neuronal activity — the researchers found that in mice lacking tau or carrying a mutation that prevents T205 phosphorylation, non-engram neurons in the vicinity fired inappropriately during recall. The memory signal was buried in excess activity, like a conversation held in a crowded room. EEG confirmed that overall hippocampal network activity was normal; the defect was specifically local dysregulation, not global hyperactivity.
Third, tau couples sensory cues to the engram. In tau-deficient mice, direct optogenetic stimulation of the engram cells — bypassing natural sensory input — successfully retrieved the memory, proving that the trace was still stored. The problem was access: without tau, natural cues could not trigger recall. “The memory is still there,” Ittner explained. “It’s just not reachable through the normal channels.”
A subtle switch
The critical modification is phosphorylation at T205. Using CRISPR gene editing, the team created mice in which T205 could not be phosphorylated (T205A knock-in). These mice showed the same remote-memory impairment as full tau knockouts. Mass spectrometry confirmed that T205 was the most abundant phosphorylation site induced by memory encoding. A different set of experiments showed that p38γ — a kinase that phosphorylates T205 — was the enzyme responsible.
The study also explored the relationship between tau’s healthy function and its pathological role. When the researchers expressed a disease-associated mutant form of tau (P301S) specifically in engram cells, it produced anterograde amnesia (impaired new memory formation) when present during encoding, and retrograde amnesia (loss of existing memories) when present during recall. This provides a mechanistic bridge between tau’s physiological role in organising engrams and its pathological role in dismantling them.
Caution for drug developers
The findings carry an immediate cautionary message for the development of anti-tau therapies. Several approaches — including tau-lowering antisense oligonucleotides, tau-targeting antibodies, and kinase inhibitors — are in clinical trials for Alzheimer’s disease. If tau is essential for normal memory function, depleting it or blocking its phosphorylation could cause cognitive side effects.
“Future research will hopefully be able to confirm concepts developed in our study in human memory,” the authors wrote. The study was conducted entirely in mice, and the specific phosphorylation site (T205) and its regulatory kinase (p38γ) have been identified as potential targets for more subtle therapeutic interventions — rather than removing tau wholesale, one might correct only the pathological modifications while preserving its physiological functions.
The study was funded by the Australian National Health and Medical Research Council, the Australian Research Council, the BrightFocus Foundation, and Dementia Australia Research Foundation. Ittner and co-author Lars Ittner are inventors on patents related to targeting p38γ and Thr-205 tau, licensed to Celosia Therapeutics.
Source: Kosonen, R., Stefanoska, K., Lin, Y. et al. “Tau T205 phosphorylation modulates engram cell recruitment and remote memory in mice.” Nature Communications (2026). DOI: 10.1038/s41467-026-73207-9