The previous period of Merit support allowed us to develop novel protocols for immuno-affinity isolation and characterization of very pure protein aggregates. We began with C. elegans models of neurodegenerative diseases, and then applied the same protocols to purify Aβ42- and tau-containing aggregates from affected hippocampal tissue of Alzheimer's patients, seeking aggregate proteins that differentiate them from age-matched controls. We now propose to apply our experience with these models to develop novel therapeutic agents that might delay, prevent, or even reverse Alzheimer's and Parkinson's pathology, through the following Aims: Aim 1. We have access through Dr. Sue Griffin, our collaborator, to a tissue bank of tissue samples from Alzheimer's Disease (caudal hippocampus) and Parkinson's Disease (substantia nigra), from which we will isolate aggregates by immuno-affinity for Aβ42, tau and α-synuclein. We propose to cross-link these aggregates, thoroughly digest them with trypsin, and identify cross-linked peptide pairs that will reveal the protein-protein interactions that mediated their conglomeration. We will use state-of-the-art “click chemistry” reagents to create, tag and recover cross-link sites, and Xlink Identifier software to analyze mass-spectrometry data. To visualize, integrate and interpret the protein-protein interactions thus revealed, we will create protein-interaction networks and analyze them with tools adapted to the nonfunctional interactions that predominate in protein aggregation. Aim 2. Proteomics data from Aim 1 will define critical protein-protein interactions within aggregates containing tau, Aβ42 or α-synuclein, which are highly enriched in Alzheimer or Parkinson brain samples. These protein- protein interactions will be ranked on the basis of their predicted propensity (by ΔG estimation) and stability (by molecular-dynamic simulations). The top 10 protein interfaces from each pulldown will then serve as targets for in silico screening of drug libraries developed to disrupt protein-protein interactions. Such libraries, constructed in diverse ways, include many `PPII' drugs designed to block specific protein-pair interactions, but which have frequently found applications beyond their original targets. For each target, we will perform initial screens of PPII-library drugs in silico, from which the top candidates will be retested in molecular-dynamic simulations. Aim 3. The drugs predicted in silico to most effectively disrupt key protein interactions (Aim 2) will be tested in vivo for reduction of aggregate formation―first in human neuronal-cell cultures expressing APPSwe (forming amyloid aggregates) or split-GFP::tau (fluorescing upon tau oligomerization). For drugs that are protective in either assay, tests will also be conducted in C. elegans aggregation models expressing Aβ42, tau, or α-synuclein. Biotinylated versions of 4–6 top candidates will be assessed for activity, and active biotinylated drugs will be use...