The grand challenge to structural biology in the 21st century is to somehow obtain high-resolution images of fundamentally dynamic systems. In crystals, dynamics blurs the electron density and blurry density produces no high-resolution spots. Single-molecule diffraction with 4th generation sources promises to provide high-resolution information regardless of motion, but only as one 2D projection per conformer because the molecule is destroyed in the process and diffraction patterns are fundamentally two dimensional. Going to three dimensions by combining many snapshots is unfortunately a form of averaging, equivalent to the problem of blurring in crystals. However, averaging can be avoided if the model itself is augmented to include the dimension of time and validated by comparing the predicted conformations to the available observations. A first step toward this goal is currently tractable: a molecular dynamics simulation of a crystal can “predict” the observed structure factors without being constrained to match them. This is trivial for salts and other small molecules, but currently impossible for large proteins. The intermediate case of a small peptide has proved encouraging as well as instructive on how the simulation could be improved. In fact, two weak density features that were not in the published model were retrospectively verified in the observed electron density maps, showing how the strengths of both techniques combined can produce high-resolution pictures of what would otherwise be a “fuzzy” system.

Structure of disorder

249th National Meeting (2015)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.