UC Berkeley

While cryo-EM imaging technology and software suites have led to a “resolution revolution”, the poor reproducibility and inefficiency offered by current specimen preparation techniques remain challenges largely unexamined and under-prized. Regardless of camera or software proficiency, a poor-quality specimen will always produce an equivalently substandard 3D reconstruction. To further advance the high-resolution EM pipeline and establish quality control for grid specimens, we tested three techniques to assess optimum conditions for vitrification of small (<500kDa), multi-subunit biological macromolecular complexes.

We performed specimen optimization for a 300kDa, five-component, macromolecular complex, PRC2. PRC2 is a key regulator of gene silencing in eukaryotes and mutations in its catalytic subunit, Ezh2, are linked to a number of human cancers and degenerative diseases. To date, no high-resolution structure of the complete complex has been solved. PRC2 presents a number of unique challenges for the structural biologist, including its small size, highly flexible regions, and conformation heterogeneity. A high-resolution structure of PRC2 would provide unparalleled insight into the biochemical mechanisms that mediate gene silencing by methylation of histone tails. For these reasons, PRC2 is an ideal candidate for optimizing cryo-EM grid preparation techniques translatable to similar complexes and for future research.

Previous EM studies of PRC2 revealed a distinct preferred orientation when bound to carbon substrates, necessitating the collection of tens of thousands of images to generate a complete structure containing high-resolution features. To overcome this impediment, we experimented with two other grid specimen preparation techniques. First, we vitrified solution-suspended PRC2 in open holes on carbon mesh support. Despite modifications of chemical and physical buffer and substrate parameters to stabilize the complex, denaturation upon surface binding at the air-water interface and dissociation of the complex during vitrification proved insurmountable. We next tested a method of affinity binding to a streptavidin monolayer substrate covering the holes of a carbon mesh support. We found that chemically biotinylated PRC2 was intact, monodisperse, and bound the grid in random orientations, all factors critical to high-resolution cryo-EM structure determination. Negative stain EM analysis revealed a more even distribution of views for 3D reconstruction. From these results, we conclude that streptavidin monolayer substrates are an option for overcoming the obstacle of preferred orientations for small complexes and provide an easily reproducible protocol for grid preparation.