Integrating multi-temperature sample preparation with particle rebalancing addresses preferred orientation

Single-particle cryo-electron microscopy (cryo-EM) often suffers from preferred orientation, which incurs particle misalignments and compromises reconstruction quality. While the use of detergents, modified grids, tilted data collection, and algorithmic corrections can partially mitigate this issue, these approaches are usually time-consuming and case-dependent. Here, we present a facile and effective approach that modulates particle orientation by varying cryo-EM sample preparation temperatures. By merging thermophilic 70S ribosome datasets collected at different temperatures and applying a three-dimensional (3D) reconstruction-based particle rebalancing strategy, we achieved improved reconstruction, as evidenced by smoother directional Fourier shell correlation (FSC) curves, higher 3D FSC sphericity, and narrower anisotropy transition zone in Fourier shell occupancy. We demonstrated the efficacy of this approach by determining the structure of γ-glutamylmethylamide synthetase, a protein complex prone to severe preferred orientation, ultimately obtaining a high-quality reconstruction. In all, our work integrates high-temperature sample preparation and particle rebalancing to address preferred orientation, holding potential for broad applicability in single-particle cryo-EM.