The end of the era of crystallography has been long predicted also on this blog, but has it come now with the recent report of an atomic resolution protein structure solved by cryo–electron microscopy (cryo-EM)?
Yip et al. report a 1.25Å resolution structure of apoferritin which made me check again if cryo-EM is already at the stage where it will generate structural data, relevant for my current research, in a high-throughput manner. I am interested in data that displays the conformational ensemble of a protein or structural changes due to allosteric binding events.
Cryo-EM naturally captures a high number of images, and thus conformations, of a protein in solution and yields a 3D structure model by subsequent averaging of many of these particle images through image analysis. Due to the nature of acquisition that no protein crystals are required, cryo-EM can depict assemblies of large protein complexes as a whole, e.g. the mitochondrial respiratory complex or Corona virus double membrane pores. Thus, the era of cryo-EM has long begun but when (and I think many agree that it is a ‘when’ and not an ‘if’) the technique is able to generate protein structure models at a quality level comparable to X-ray crystallography, cryo-EM will spread even quicker as before.
As I usually only analyse structural data and I am not involved in generating it, so a lot of the technical details don’t mean much to me but I wanted to share some numbers of their study that got me excited because they fit into my world of statistics and PDB structures: the setup reached a pixel size of 0.482Å which is roughly a third of a covalent bond and the total electron dose per image was 50 e– per Å2. For an intermediate model of the target protein with a resolution of 1.5Å they required 22,000 particle images which took about 2 hours of acquisition time. In contrast, for their final model with 1.25Å resolution over 1,000,000 particle images were necessary.
Figure 1 shows three example residues of their study at 1.25Å resolution: red mesh in the first row depicts electron density at a high density threshold, grey mesh in the second row at a low density threshold and the third row both in one picture, zoomed-in. Whereas in the first row a clear separation between individual atoms is visible (= atomic resolution), the second threshold points out the location of hydrogen atoms. The combination of both illustrates the location of all individual atoms of the three example residues.
Yip et al. account their study’s achievements in resolution and model quality to the improvements in microscopy hardware, namely a Titan Krios G3 electron microscope equipped with a monochromator (making sure emitted electrons are of a very similar energy/speed) and a second-generation spherical aberration corrector (reducing optical aberrations and in my understanding of the paper, in combination with the microscope, the key improvement).
It is very cool to see cryo-EM being able to resolve individual atoms and even hydrogens as another example of this technique’s exciting progress. The authors state that these improvements will mainly show with targets that have a very high biochemical quality and stability (here: apoferritin) because for the “vast majority of macromolecular complexes” the bottlenecks are currently more sample preparation and image analysis that can handle dynamic, moving proteins well. Therefore, I will have to wait a bit more until cryo-EM labs yield models (at crystallographic resolution) of conformational ensembles in larger quantities.
Reference
Yip, K.M., Fischer, N., Paknia, E. et al. Atomic-resolution protein structure determination by cryo-EM. Nature (2020). https://doi.org/10.1038/s41586-020-2833-4
