Ab initio prediction of protein 3-D structures as a road to 'Terra Incognita' of their conformational landscapes

Is the vast conformational space of oligopeptides and proteins amenable to ab initio structural predictions? We will propose multidisciplinary research that combines state-of-the-art computational chemistry and accurate machine-learned oligopeptide potentials with applications in structural and evolutionary biology. First, we will develop a new computational tool, denoted as QMLFold. This would allow us to enter ‘terra incognita’ of protein structure predictions, hitherto inaccessible to AlphaFold2/3 and its competitors. This underexplored territory encompasses oligopeptides, intrinsically disordered proteins, proteins consisting of non-proteinogenic (non-coded, xeno) amino acids, or smaller de novo designed (metallo)-proteins; all these biomolecules have a great biological and industrial potential.

QMLFold will view protein folding from a conceptually different perspective, as a ‘universal solvation problem’ (i.e., as a covalently linked ensemble of amino acid side chains, solvated in themselves). This is in stark contrast to the standard atomistic approach, adopted in force-field simulations. We will formulate a new and non-trivial extension of the efficient COSMO-RS solvation model to three dimensions (3D-COSMO-RS) and couple it to machine-learned (ML) oligopeptide potentials. The latter were proven to provide conformational free energies of longer peptides at the “QM-accuracy”, at a fraction of second of CPU time.

The experimental calibration and validation of QMLFold will involve the synthesis and structural characterization (NMR, VCD, ECD) of selected oligopeptides with the pre-designed secondary structure.

Two areas of potential applications will be then explored:

• Biocatalysis: for de novo design of catalytic metallopeptides or metalloproteins beyond the known peptide scaffolds
• Xeno-biology: to understand whether differences in conformational complexity of coded and non-coded amino acids played a role in the evolution of proteins as we know them today.

Field of study: Modelling of chemical properties on nano- and biostructures

References:

• Unke, O. T.; Chmiela, S.; Sauceda, H. E.; Gastegger, M.; Poltavsky, I.; Schütt, K. T.; Tkatchenko, A.; Müller, K. Machine Learning Force Fields. Chem. Rev. 2021, 121, 10142-10186. doi:10.1021/acs.chemrev.0c01111
• Wang, Y.; Wang, T.; Li, S.; He, X.; Li, M.; Wang, Z.; Zheng, N.; Shao, B.; Liu, T. Enhancing geometric representations for molecules with equivariant vector-scalar interactive message passing. Nat. Commun. 2024, 15, 313. doi:10.1038/s41467-023-43720-2
• Frank, J. T.; Unke, O. T.; Müller, K.; Chmiela, S. A Euclidean transformer for fast and stable machine learned force fields. Nat. Commun. 2024, 15, 6539. doi:10.1038/s41467-024-50620-6
• Anstine, D.; Zubatyuk, R.; Isayev, O. AIMNet2: A Neural Network Potential to Meet your Neutral, Charged, Organic, and Elemental-Organic Needs. ChemRxiv 2024. doi:10.26434/chemrxiv-2023-296ch-v3
• Zhang, S.; Makoś, M. Z.; Jadrich, R. B.; Kraka, E.; Barros, K.; Nebgen, B. T.; Tretiak, S.; Isayev, O.; Lubbers, N.; Messerly, R. A.; Smith, J. S. Exploring the frontiers of condensed-phase chemistry with a general reactive machine learning potential. Nat. Chem. 2024, 16, 727-734. doi:10.1038/s41557-023-01427-3