My final Structure of Christmas may look like an unremarkable enzyme, but it heralded the arrival of a game-changing method in structural biology.

My ninth (and final) Structure of Christmas is beta-galactosidase: a pretty run-of-the-mill enzyme that turns compound sugars into monosaccharides. When you put a special dye on it, it turns the dye blue (whee!). It’s a mainstay of molecular biology and millions of students have used it in countless experiments, both fascinating and mundane. It doesn’t have much of a ‘wow’ factor – it’s a solid member of a respectable family of sugar-cleaving enzymes.

What is so special about it is the way its structure was determined.

X rays: not for everyone

All the other structures in this series were determined by X-ray crystallography. In fact, most of what we know about biological molecules at the atomic level is based on proteins that can be crystallised and withstand an onslaught of X rays.

But many proteins do not crystallise well. If X-ray crystallography were the only method available for determining structures, we’d still have only a rough idea of their shapes. Fortunately, cryo-Electron Microscopy (cryo-EM) came along to provide a viable alternative.

Cryo-EM is really an entirely different technique. When it was used to resolve the structure of beta-galactosidase at a resolution just as good as many X-ray structures, it opened up a completely new approach to structural biology. Cryo-EM was the first game-changing methodological development to come along in decades, offering fresh opportunities to determine the structures of unwieldy proteins and large complexes at very high resolution.

This has all come about in the past five years.

Finer than light

Cryo-EM is a technique for looking at things under magnification. But rather than shining a light (visible or other) on an object to see how it either reflects or absorbs that light (which is what microscopes and X-ray crystallography do), cryo-EM uses beams of electrons.

Electrons can have up to 100,000-fold shorter wavelengths than light, so they can reveal an extremely high level of detail. But they are also very energetic, so in the early days of EM you had to ‘stain’ a sample to ensure it interacted sensibly with the electrons and didn’t get fried.

At first, EM was considered to be X-ray crystallography’s poor cousin, showing only blob-like structures. These early EM structures didn’t give nearly enough information for anyone to tell how biologically relevant the structures were. There were just too many manipulations needed to get the samples to work in the electron microscope, and the resulting images were a bit of a mess. The stain acted like a blanket of snow on your garden: you could see the shape of objects beneath it, but couldn’t always be sure what they were, and certainly you wouldn’t see any detail.

Putting the “cryo” in cryo-EM

Slowly but steadily, EM improved. The first major breakthrough was a technique whereby biological samples suspended in water could be flash-frozen to a glass-like state, such that the water did not make crystals. This got around the need for staining, and allowed one to keep extremely cold, biologically relevant samples.

This feat, first achieved at EMBL Heidelberg, put the “cryo” into cryo-EM.

The second breakthrough was actually a series of incremental developments, with detectors getting better and better. Early instruments sensed 1 in 50 of the electrons –current instrumentation picks up nearly every electron. That means we can detect quite a lot of signal before the samples become damaged by the electron beam.

Protecting your image

While all this has been going on, image processing has become substantially more sophisticated. That was another big factor in making cryo-EM a success.

If one assumes that a large set of particles all have the same structure but are presented in different orientations, it is possible to find, merge and average across multiple (two-dimensional) instances computationally. This averaging makes it possible to see the main amino-acid backbone and key residues – and so to model the 3D structure.

This is incredible. It is the closest we get to actually seeing the individual atoms of these tiny machines that power us in their ‘natural’ (or natural-ish) environments.

For structural biologists, being able to determine the structures of proteins that don’t crystallise opens new doors. There are many such proteins, and nearly every large university department or structural biology institute is tooling up with one of the new powerful electron microscopes.

But it doesn’t end there.

Tilt and shoot

Cryo-EM sample types have few limitations (biological material in glass-like frozen water) and the fundamental analysis is computational. This provides a lot of flexibility. You don’t need to isolate a ‘single’ protein and coax it into a crystal. Indeed, it is possible to capture different states, configurations or complexes for a series of structures.

Furthermore, structural biologists can now bridge the grey area between molecules and the cells in which they are normally found. To study larger objects, such as intact cells, you can tilt them in small steps and capture images for all orientations, then combine the images into a 3D map (i.e. a tomogram).

Tomograms are interesting in and of themselves, but it is also possible to obtain a high-resolution picture of the protein or complex in its natural cellular environment. To do this you capture multiple copies of a protein or complex and average their maps (this is called sub-tomogram averaging).

Amazing work from the Briggs lab in EMBL Heidelberg involved doing precisely this with HIV-capsids, caught in the act of budding out of cells. (These structures will be in the PDB and EMDB in a few days.)

A century of understanding chemistry

X-ray crystallography has given us almost a century of understanding the chemistry of life at the atomic level. It will continue to give us insights, now joined by another, orthogonal process to understand life at a similar scale.

Expect at least another 100 years of remarkable discoveries.

Reference

Bartesaghi A, Merk A, Banerjee S, Matthies D, Wu X, Milne JL and Subramaniam S (2015) 2.2 Å resolution cryo-EM structure of β-galactosidase in complex with a cell-permeant inhibitor. Science 348:1147-1151.