Recent, rapid technical advances to microscopes, detectors and image processing have substantially improved the resolution of cryo-electron microscopy (cryo-EM), causing the broader biology community to sit up and take notice of this once niche technique. An increasing number of near-atomic-resolution structures of biologically interesting protein complexes solved by cryo-EM are being reported in high-profile journals. But these advances notwithstanding, the cryo-EM community has been unable to penetrate the 3-Å resolution barrier, despite predictions showing that there is no theoretical limit to reaching atomic (∼2-Å) resolution.

X-ray crystallography is routinely used to solve protein structures below 2.3-Å resolution, which allows visualization of fine details such as hydrogen bonding, salt bridges and ordered water molecules. The ability to attain such resolution with cryo-EM—which uses samples frozen in a thin layer of ice rather than crystallization and is particularly suitable for studying large protein complexes—is certain to open new doors in structural biology.

In recent work, Sriram Subramaniam of the US National Cancer Institute and colleagues reported the highest-resolution structure solved by cryo-EM to date, of a complex between Escherichia coli β-galactosidase and an inhibitor (phenylethyl β-d-thiogalactopyranoside) (Bartesaghi et al., 2015)1. The reported resolution: 2.2 Å.

Just last year, Subramaniam's group reported a 3.2-Å structure of E. coli β-galactosidase, a fairly ordinary enzyme of about 460 kDa whose structure had been solved by crystallography, allowing the researchers to vet the cryo-EM structure. Reaching 3.2 Å was commendable, but Subramaniam was eager to explore what his group might do to break through the 3-Å barrier. “There are so many steps to go from purified protein to final structure,” he notes. “I just felt we needed to look at each step carefully and make sure we're doing the best we could.”

For their most recent work, Subramaniam's team used the same equipment and methods as in 2014, but performed an exhaustive study of different conditions for specimen preparation, imaging and data processing steps. The outcome of their work is a potentially generalizable set of imaging condition and image processing workflows. Once sample preparation is also optimized, these developments offer the hope of routine structure determination at ∼2-Å resolution, at least for structurally homogeneous proteins. They obtained the best resolution when the specimen ice thickness was just right: thin enough to obtain strong signals but thick enough to obtain particles in different orientations (important for image processing). Despite the great recent cryo-EM advances, “we have not paid a lot of attention to specimen preparation,” says Subramaniam. “Nothing has changed in almost three decades of doing it the same way.” He hopes that his group's work will inspire the development of new methods for preparing better cryo-EM specimens.

Subramaniam's study comes on the heels of another recent 3-Å barrier-breaking report from Holger Stark and colleagues at the Max Planck Institute for Biophysical Chemistry. Stark's team reported a cryo-EM structure of the E. coli 70S ribosome in complex with the elongation factor Tu at a resolution of 2.65–2.9 Å using a microscope equipped with a spherical aberration corrector (Fischer et al., 2015)1. These two studies together solidly demonstrate that resolutions that were formerly the realm of X-ray crystallography are now possible with cryo-EM.