Protein shape sampled by ion mobility mass spectrometry consistently improves protein structure prediction

Abstract

Ion mobility (IM) mass spectrometry provides structural information about protein shape and size in the form of an orientationally-averaged collision cross-section (CCSIM). While IM data have been used with various computational methods, they have not yet been utilized to predict monomeric protein structure from sequence. Here, we show that IM data can significantly improve protein structure determination using the modelling suite Rosetta. We develop the Rosetta Projection Approximation using Rough Circular Shapes (PARCS) algorithm that allows for fast and accurate prediction of CCSIM from structure. Following successful testing of the PARCS algorithm, we use an integrative modelling approach to utilize IM data for protein structure prediction. Additionally, we propose a confidence metric that identifies near native models in the absence of a known structure. The results of this study demonstrate the ability of IM data to consistently improve protein structure prediction.

Investigation of Charge-State-Dependent Compaction of Protein Ions with Native Ion Mobility–Mass Spectrometry and Theory

“…Because linear regression analysis of the individual structural features as a function of charge did not yield any new significant trends, we sought to determine whether some combination of these features might be more meaningful in explaining the observed trends in CCS than each feature alone. Other IM–MS researchers have used sophisticated machine learning and PA CCS computations on condensed-phase structures to analyze ion mobility results for very large data sets. , Here, we use principal component analysis (PCA) and TM computations for MD-compacted structures, owing to the relatively small data set, for which more advanced machine learning approaches might be inappropriate or extremely time-consuming. We selected one metric from each of the seven different categories of features (hydrogen bonds, surface hydrogen bonds, surface residues, polar contacts involving charged side chains, salt bridges, and α-helical and β-strand secondary structure content), expressed as a percent change relative to the original unsimulated structure so that all features would be on the same protein-normalized scale.…”

Protein complex prediction using Rosetta, AlphaFold, and mass spectrometry covalent labeling

“…Information obtained from hydroxyl radical footprinting (HRF), HDX, and DEPC labeling experiments have been shown to improve tertiary structure prediction with Rosetta 36 – 43 by using calculated solvent exposure metrics for models to select for experimentally accurate predictions. Similarly, protein shape and size information obtained through collisional cross-section data from IM experiments has also improved Rosetta structure prediction 44 . A method iSPOT 45 , which uses a combination of multiple biophysical methods (integration of shape information from small-angle X-ray scattering and protection factors probed by hydroxyl radicals), has been shown as a powerful approach for integrated modeling of multiprotein complexes.…”

Toward an Integrated Machine Learning Model of a Proteomics Experiment

“…In cross-linking experiments, CCS values can be used to distinguish cross-linked peptides from unlinked peptides of similar mass-to-charge ratio . Comparing measured and predicted CCS values has also been used to interrogate protein structure and dynamics, including protein complexes. , …”