Growing Glycans in Rosetta: Accurate de novo glycan modeling, density fitting, and rational sequon design
Jared Adolf-Bryfogle,Jason W. Labonte,John C. Kraft,Maxim Shapovalov,Sebastian Raemisch,Thomas Lütteke,Frank DiMaio,Christopher D. Bahl,Jesper Pallesen,Neil P. King,Jeffrey J. Gray,Daniel W. Kulp,William R. Schief
DOI: https://doi.org/10.1371/journal.pcbi.1011895
2024-06-25
PLoS Computational Biology
Abstract:Carbohydrates and glycoproteins modulate key biological functions. However, experimental structure determination of sugar polymers is notoriously difficult. Computational approaches can aid in carbohydrate structure prediction, structure determination, and design. In this work, we developed a glycan-modeling algorithm, GlycanTreeModeler , that computationally builds glycans layer-by-layer, using adaptive kernel density estimates (KDE) of common glycan conformations derived from data in the Protein Data Bank (PDB) and from quantum mechanics (QM) calculations. GlycanTreeModeler was benchmarked on a test set of glycan structures of varying lengths, or "trees". Structures predicted by GlycanTreeModeler agreed with native structures at high accuracy for both de novo modeling and experimental density-guided building. We employed these tools to design de novo glycan trees into a protein nanoparticle vaccine to shield regions of the scaffold from antibody recognition, and experimentally verified shielding. This work will inform glycoprotein model prediction, glycan masking, and further aid computational methods in experimental structure determination and refinement. Many biological proteins are chemically modified to induce specific structure and function. Carbohydrates (glycans) are one such modification that play an important role in signaling, stability, solubility, aggregation, and the immune system. In this work, we have developed and extensively benchmarked a computational protocol for predicting the structures of these glycans, while improving the Rosetta Software Suite through the development of new general analysis frameworks, such as the SimpleMetric system and extensive glycan tools for modeling and design. We describe the benchmarking, optimization, and use of a novel computational method for three dimensional modeling of glycans and glyco-conjugates called the GlycanTreeModeler . This method is unique in that it grows glycans layer-by-layer using extensive data-driven methods. We thoroughly benchmark the method and detail iterative improvement of the GlycanTreeModeler through scoring and kinematic improvements, and show how these methods can be useful in glycan-related computational tasks and glycan masking of a novel vaccine scaffold in-vitro and in-vivo.
biochemical research methods,mathematical & computational biology