The development of inorganic nanoparticles (NPs) for biological applications such as nanoscale adjuvants, enzyme mimics, contrast agents, amyloid fibrillation inhibitors requires a better understanding of their interactions with proteins. The abundant knowledge about protein-protein interactions (PPIs) can potentially serve as a guide for designing protein-NP assemblies, but the chemical and biological inputs used in computational packages for PPI are not applicable to inorganic NPs. Analyzing chemical, geometrical and graph-theoretical descriptors for protein complexes, we found that geometrical and graph-theoretical descriptors can predict interaction sites in protein pairs with accuracy &gt;80% and classification probability ~90%. Furthermore, these descriptors are uniformly applicable to biological and inorganic nanostructures. We extended the protein-trained machine learning algorithms to carbon-based nanoparticles and found a nearly exact match between experimental and predicted interaction sites in their complexes with multiple proteins. These findings can be extended to other inorganic NPs to predict their bioinspired nanoassemblies with chemically dissimilar structures.

Universal structural descriptors for intermolecular interactions: From protein-protein to protein-nanoparticle

ACS Spring 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.