Complementing machine learning‐based structure predictions with native mass spectrometry

Abstract

The advent of machine learning-based structure prediction algorithms such as AlphaFold2 (AF2) and RoseTTa Fold have moved the generation of accurate structural models for the entire cellular protein machinery into the reach of the scientific community. However, structure predictions of protein complexes are based on user-provided input and may require experimental validation. Mass spectrometry (MS) is a versatile, time-effective tool that provides information on post-translational modifications, ligand interactions, conformational changes, and higher-order oligomerization. Using three protein systems, we show that native MS experiments can uncover structural features of ligand interactions, homology models, and point mutations that are undetectable by AF2 alone. We conclude that machine learning can be complemented with MS to yield more accurate structural models on a small and large scale.

Mass Spectrometry and Machine Learning Reveal Determinants of Client Recognition by Antiamyloid Chaperones

“…show that the CCS of the heterocomplex is closer to the AF2 rank 1 structure with A bound inside the TTR tetramer, rather than the AF2 rank 3 structure with A bound on the surface of TTR. The ATD of apo-TTR is shown in grey, that of the TTR-A (12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28) complex in blue. (G) Further evidence for binding in the central channel is given by the minimal A binding to TTR after overnight incubation, when the tetramer is stabilized by co-incubation with tafimidis.…”

“…We next performed native MS to experimentally test these predictions. Weak binding of A (12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28) to TTR at a 1:4 stoichiometry was observed, with the intensity of the peaks increasing upon overnight incubation at room temperature (Figure 2C-D). Complex formation was confirmed using MS/MS (Figure 2E).…”

“…It can deliver seemingly plausible models for protein interactions not occurring under J o u r n a l P r e -p r o o f 5 native conditions, and the resulting models require therefore further experimental validation. Mass spectrometry (MS) offers several good analytical strategies complementary to AF2 26,27 :…”

“…It can deliver seemingly plausible models for protein interactions not occurring under native conditions, and the resulting models require therefore further experimental validation. Mass spectrometry (MS) offers several good analytical strategies complementary to AF2 ( 26 , 27 ): The use of physiological buffer conditions and carefully tuned instrument parameters have enabled us to preserve noncovalent interactions during MS analysis. The mass of intact complexes thus reveals protein and ligand interactions.…”

A Perspective on the Prospective Use of AI in Protein Structure Prediction

“…Mass spectrometry (MS) is another experimental approach which well complements structure prediction methods. Indeed, several studies have shown how MS data can be combined with AF2 models. − Specific tools have been developed to include MS data in the modeling process, namely, AlphaLink and its extension to AlphaFold-Multimer, AlphaLink2 …”