Back

Large-Scale Mutagenesis Identifies Sequence Changes That Bias Tau Toward CTE-like Aggregation

Dashnaw, C. M.; Rastogi, S.; Joachimiak, L. A.; Memon, I.; Jaczynska, K.; Nguyen, B. A.; Saelices, L.; White, C. L. A.; Shahmoradian, S.

2026-10-09 bioengineering
10.64898/2026.10.08.757173 bioRxiv
Show abstract

Tau amyloid fibrils adopt disease-specific structures across tauopathies, yet the sequence determinants that favor one pathological fold over another remain poorly defined. Here we combine computational saturation mutagenesis and energetic profiling, in vitro aggregation, cryo-EM and cellular seeding assays to identify sequence changes that alter tau fold compatibility. Across six tauopathy fibril structures, approximately 12,000 substitutions revealed that mutational sensitivity is concentrated within interaction networks formed by amyloidogenic motifs rather than uniformly distributed across the fibril core. Comparison of Alzheimer's disease (AD) and chronic traumatic encephalopathy (CTE) folds identified residues within the 335-363 region that differentially influence fold compatibility. The S341L/S352I double mutant accelerated aggregation and remodeled packing within the PAM4-containing amyloidogenic region, producing an alternative amyloid architecture in a minimal tau fragment. S341L/S352I biosensors preferentially responded to CTE-derived over Alzheimer's disease-derived seeds, while aggregates generated from recombinant S341L/S352I tau exhibited CTE-like cellular templating behavior. Although these assemblies are not established to reproduce the atomic CTE fold, the results demonstrate that targeted sequence changes can bias tau aggregation and propagation toward a CTE-compatible state. This framework provides a strategy for defining sequence-structure relationships in polymorphic amyloids and developing conformation-selective models and biosensors.

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.