AI-designed high-resolution 14N solid-state NMR for characterization of biological molecules.

An atomic-level description of biomolecular 3D structure is vital for understanding disease mechanisms and designing new drugs. Solid-state nuclear magnetic resonance (NMR) is a powerful method for analyzing the molecular structure and dynamics of biological solids, such as fibrils, viral capsids, and membrane proteins, which are associated with the progression of various diseases. However, obtaining high-quality structures requires high-resolution NMR spectra, which traditionally demand expensive and often impractical isotopic labeling of samples. This project will develop high-sensitivity, high-resolution solid-state NMR methods based on nuclei (1H and 14N) with nearly 100% natural abundance, enabling the study of label-free biomolecules. Until now, conventional NMR method development has failed to provide spectroscopic solutions for achieving high-resolution 14N spectra due to significant quadrupolar broadening. This project will leverage recent breakthroughs in artificial intelligence (AI) to aid in the design of NMR pulses capable of orientation-selective excitation, resulting in high-resolution 14N spectra. Success in this project will redefine biomolecular analysis, providing important insights into biomolecular properties and paving the way for the investigation of previously elusive biological systems.