An automated platform for structural analysis of membrane proteins through serial crystallography

Abstract

Membrane proteins are central to many pathophysiological processes yet remain very difficult to analyze at a structural level. Moreover, high-throughput structure-based drug discovery has not yet been exploited for membrane proteins due to lack of automation. Here, we present a facile and versatile platform for in meso membrane protein crystallization, enabling rapid atomic structure determination at both cryogenic and room temperature and in a single support. We apply this approach to two human integral membrane proteins, which allowed us to capture different conformational states of intramembrane enzyme-product complexes and analyze the structural dynamics of the ADIPOR2 integral membrane protein. Finally, we demonstrate an automated pipeline combining high-throughput microcrystal soaking, automated laser-based harvesting and serial crystallography enabling screening of small molecule libraries with membrane protein crystals grown in meso. This approach brings badly needed automation for this important class of drug targets and enables high-throughput structure-based ligand discovery with membrane proteins.

Elucidating polymorphs of crystal structures with intensity-based hierarchical clustering analysis on multiple diffraction datasets

“…In the case of serial femtosecond crystallography (SFX) (Barends et al ., 2022) or serial synchrotron rotation crystallography (SSROX) (Gati et al ., 2014; Hasegawa et al ., 2017), a much larger number of datasets is required because every single frame covers only a small portion of reciprocal space, posing challenges in data collection and analysis. Here, the automation of data collection and analysis is crucial and has provided opportunities for expanding the target and achieving structure determination for diverse protein samples (Healey et al ., 2021).…”