Search

Neurodegenerative diseases (NDs) are characterized by the formation of amyloid filaments that adopt disease-specific conformations in the brain. Recently developed small molecules hold promise as diagnostics and possible therapeutics for NDs, but their binding mechanisms to amyloid filaments remain unknown. Here, we used cryo–electron microscopy (cryo-EM) to determine a 2.7 Å structure of Alzheimer’s disease patient-derived tau paired-helical filaments incubated with the GTP-1 PET probe. GTP-1 is bound stoichiometrically along an exposed cleft of each protofilament in a stacked arrangement that matches the fibril’s symmetry. Multiscale modeling revealed favorable pi-pi aromatic stacking interactions between GTP-1 molecules that, together with small molecule–protein contacts, result in high affinity binding. This binding mode offers new insight into designing compounds for diagnosis and treatment of specific NDs.One Sentence SummaryCryo-EM structure reveals a novel stacked arrangement of the GTP-1 PET ligand bound to Alzheimer’s disease tau filaments.

In microscopy-based drug screens, fluorescent markers carry critical information on how compounds affect different biological processes. However, practical considerations, such as the labor and preparation formats needed to produce different image channels, hinders the use of certain fluorescent markers. Consequently, completed screens may lack biologically informative but experimentally impractical markers. Here, we present a deep learning method for overcoming these limitations. We accurately generated predicted fluorescent signals from other related markers and validated this new machine learning (ML) method on two biologically distinct datasets. We used the ML method to improve the selection of biologically active compounds for Alzheimer's disease (AD) from a completed high-content high-throughput screen (HCS) that had only contained the original markers. The ML method identified novel compounds that effectively blocked tau aggregation, which had been missed by traditional screening approaches unguided by ML. The method improved triaging efficiency of compound rankings over conventional rankings by raw image channels. We reproduced this ML pipeline on a biologically independent cancer-based dataset, demonstrating its generalizability. The approach is disease-agnostic and applicable across diverse fluorescence microscopy datasets.