A static structure of a protein gives a biologist much insight into its function, but researchers can gain an even deeper understanding by watching the protein in action. Detecting intermediate conformational states that exist only fleetingly, however, is a serious experimental challenge.

Time-resolved crystallography reveals protein dynamics. Image adapted from Nat. Methods 11, 923–926 (2014). Credit: Katie Vicari/Nature Publishing Group

One approach used for studies of protein dynamics is time-resolved X-ray crystallography (TRX). In a TRX study, a laser 'pump' pulse is deployed to initiate a reaction, and then X-ray 'probe' pulses are applied to collect a series of diffraction patterns as the reaction proceeds. In the past, performing such experiments required highly specialized synchrotron sources. These instruments have a 'speed limit' of about 100 picoseconds, precluding experimental visualization of very fast protein conformational changes. But now, recent advances made possible by ultrafast X-ray free-electron lasers (XFELs), offering femtosecond-level time resolution, are allowing researchers to resolve very rapid structural changes that are undetectable by any other method.

In 2014, for example, researchers used XFEL-based TRX to follow the conformational changes that photosystem II undergoes as it catalyzes light-driven water splitting, at 5.5-Å resolution (Nature 513, 261–265, 2014). Later that year, a team used the approach to study the photocycle of photoactive yellow protein, resolving reaction intermediates at an impressive 1.6-Å resolution (Science 346, 1242–1246, 2014). And in 2015, another group took advantage of ultrafast XFEL pulses to follow structural changes in myoglobin resulting from photolysis of the iron–carbon dioxide bond at its active-site heme (Science 350, 445–450, 2015).

Given the intense competition for XFEL beamtime at the very small number of instruments around the world, such experiments are even less generally applicable than a TRX experiment using a specialized synchrotron source. Thus a clever pump-probe pulse-sequence approach that can be used to collect TRX data with any of the standard synchrotron beamlines, which are accessible to a much broader community, is a welcome development (Nat. Methods 11, 1131–1134, 2014). Though the technique's time resolution has not yet been demonstrated to be nearly as good as that of an XFEL, future advances may make this a possibility.

It will certainly be interesting to watch the TRX technique blossom, especially to see how researchers will adapt the approach to study ultrafast reactions in proteins triggered by stimuli other than light.