Abstract: In the rapidly evolving landscape of peptide and protein design, the emergence of Artificial Intelligence (AI) has propelled the field into new realms of possibility. Leveraging rational and computational methodologies, researchers can now engineer an extensive array of stable assemblies from scratch. Yet, amidst these advances, the enigma of complete comprehension of protein folding persists. Another challenge that looms before us is transcending the confines of naturally occurring protein structures to venture into the 'dark matter of protein structure space'. In the first part of this presentation, I will provide a comprehensive overview of my research while delving into the design of the 310-helicalbundle. Unlike the ubiquitous α-helix, 310-helices occupy a more modest presence in protein structures. Distinct 310-helical parameters yield energetically less favourable internal energies, rendering them less prone to conform within defined higher-order structures. In the natural context, 310-helices seldom exceed six residues and do not typically engage in regular super-secondary, tertiary, or quaternary interactions. Remarkably, my work demonstrates that, despite their scarcity in nature, synthetic protein-like assemblies can be meticulously constructed from 310-helices. The presentation provides an insightful exploration of the rational design process, unveils solution-phase characterizations, and showcases X-ray crystal structures of water-soluble bundles formed by 310-helices. Furthermore, I will also expand on the synergistic combination of our design principles and AI-guided protein design, culminating in the design of a singular, single-chain parallel seven helix barrel. In the closing stages, I will discuss about my scientific ambitions to design novel protein folds for new functions.

Meeting ID: 977 0880 7191
Passcode: 277891