A deep learning model for predicting optimal distance range in crosslinking mass spectrometry data.

Macromolecular assemblies play an important role in all cellular processes. While there has recently been significant progress in protein structure prediction based on deep learning, large protein complexes cannot be predicted with these approaches. The integrative structure modeling approach characterizes multi-subunit complexes by computational integration of data from fast and accessible experimental techniques. Crosslinking mass spectrometry is one such technique that provides spatial information about the proximity of crosslinked residues. One of the challenges in interpreting crosslinking datasets is designing a scoring function that, given a structure, can quantify how well it fits the data. Most approaches set an upper bound on the distance between Cα atoms of crosslinked residues and calculate a fraction of satisfied crosslinks. However, the distance spanned by the crosslinker greatly depends on the neighborhood of the crosslinked residues. Here, we design a deep learning model for predicting the optimal distance range for a crosslinked residue pair based on the structures of their neighborhoods. We find that our model can predict the distance range with the area under the receiver-operator curve of 0.86 and 0.7 for intra- and inter-protein crosslinks, respectively. Our deep scoring function can be used in a range of structure modeling applications.

Facilitating In Situ Cross-Linking and Mass Spectrometry by Antibody-Based Protein Enrichment.

Cross-linking of living cells followed by mass spectrometry identification of cross-linked peptides (in situ CLMS) is an emerging technology to study protein structures in their native environment. One of the inherent difficulties of this technology is the high complexity of the samples following cell lysis. Currently, this difficulty largely limits the identification of cross-links to the more abundant proteins in the cell. Here, we describe a targeted approach in which an antibody is used to purify a specific protein-of-interest out of the cell lysate. Mass spectrometry analysis of the protein material that binds to the antibody can then identify considerably more cross-links on the target protein. By using an antibody against the CCT chaperonin, we identified over 200 cross-links that provide in situ evidence for the subunit arrangement of the CCT particle and its interactions with prefoldin. Similar targeting with an antibody against tubulin provided in situ evidence for the structure of the microtubule. Finally, the approach was also successful in identifying cross-links within a protein that expresses at a low level. These results demonstrate the general utility of antibody-based sample simplification for in situ CLMS and greatly expand the scope of protein systems that are amenable to in situ structural studies.

Mass spectrometry reveals the chemistry of formaldehyde cross-linking in structured proteins.

Whole-cell cross-linking coupled to mass spectrometry is one of the few tools that can probe protein-protein interactions in intact cells. A very attractive reagent for this purpose is formaldehyde, a small molecule which is known to rapidly penetrate into all cellular compartments and to preserve the protein structure. In light of these benefits, it is surprising that identification of formaldehyde cross-links by mass spectrometry has so far been unsuccessful. Here we report mass spectrometry data that reveal formaldehyde cross-links to be the dimerization product of two formaldehyde-induced amino acid modifications. By integrating the revised mechanism into a customized search algorithm, we identify hundreds of cross-links from in situ formaldehyde fixation of human cells. Interestingly, many of the cross-links could not be mapped onto known atomic structures, and thus provide new structural insights. These findings enhance the use of formaldehyde cross-linking and mass spectrometry for structural studies.