Ice in biomolecular cryocrystallography.

David W Moreau,Hakan Atakisi,Robert E Thorne

doi:10.1107/s2059798321001170

David W Moreau, Hakan Atakisi + Show 1 more

Open Access

https://doi.org/10.1107/s2059798321001170

Copy DOI

Abstract

Diffraction data acquired from cryocooled protein crystals often include diffraction from ice. Analysis of ice diffraction from crystals of three proteins shows that the ice formed within solvent cavities during rapid cooling is comprised of a stacking-disordered mixture of hexagonal and cubic planes, with the cubic plane fraction increasing with increasing cryoprotectant concentration and increasing cooling rate. Building on the work of Thorn and coworkers [Thorn et al. (2017), Acta Cryst. D73, 729-727], a revised metric is defined for detecting ice from deposited protein structure-factor data, and this metric is validated using full-frame diffraction data from the Integrated Resource for Reproducibility in Macromolecular Crystallography. Using this revised metric and improved algorithms, an analysis of structure-factor data from a random sample of 89 827 PDB entries collected at cryogenic temperatures indicates that roughly 16% show evidence of ice contamination, and that this fraction increases with increasing solvent content and maximum solvent-cavity size. By examining the ice diffraction-peak positions at which structure-factor perturbations are observed, it is found that roughly 25% of crystals exhibit ice with primarily hexagonal character, indicating that inadequate cooling rates and/or cryoprotectant concentrations were used, while the remaining 75% show ice with a stacking-disordered or cubic character.

Highlights

Ice diffraction frequently contaminates diffraction data collected from biomolecular crystals at cryogenic temperatures (Rupp, 2009; Pflugrath, 2015)
Ice diffraction in protein crystallography can arise from ice in the internal crystal solvent, from ice formed in residual cryoprotectant-containing solvent on the crystal surface and from frost
The ice diffraction obtained by cooling to 180 K is isotropic, indicating a small grain size, and its resolution dependence is consistent with stacking-disordered ice with a substantial cubic fraction

Summary

Introduction

Ice diffraction frequently contaminates diffraction data collected from biomolecular crystals at cryogenic temperatures (Rupp, 2009; Pflugrath, 2015). Ice may appear as contaminating frost on the sample or sample-holder surface, due to exposure to moist ambient air during handling or data collection, or from accumulated frost in the liquid nitrogen used to initially cool and to store the crystals (Pflugrath, 2004). Ice that forms from solvent confined to the solvent cavities of the crystal, from bulk-like solvent containing substantial cryoprotectant or from bulk-like solvent that is rapidly cooled is typically highly polycrystalline, producing continuous and largely isotropic ice rings. Ice that forms from bulk-like solvent containing little cryoprotectant or that is cooled slowly tends to be comprised of fewer, larger crystals, producing ‘lumpy’, anisotropic, quasi-continuous diffraction rings. Frost is typically comprised of an even smaller number of larger dendritic crystals

Methods

Results

Conclusion