The development of Megahertz (MHz) capable XFEL facilities has allowed us to investigate the structure and function of biological samples such as proteins via the technique known as Serial Femtosecond Crystallography (SFX). Data collection occurs at very fast rates of up to 2,700 data points in 0.6 milliseconds. Furthermore, due to the highly focused and intense X-ray pulses, once a sample has been exposed to a single X-ray pulse, radiation damage will occur, leading to destruction of the sample. However, undamaged data can be collected from the sample before this damage can progress, and this is referred to as “diffraction before destruction”.

The development of Megahertz (MHz) capable X-ray Free Electron Laser (XFEL) facilities has allowed us to investigate the structure and function of biological samples such as proteins via the technique known as Serial Femtosecond Crystallography (SFX). Data collection occurs at very fast rates of up to 2,700 data points in 0.6 milliseconds. Furthermore, due to the highly focused and intense X-ray pulses, once a sample has been exposed to a single X-ray pulse, radiation damage will occur, leading to destruction of the sample. However, undamaged data can be collected from the sample before this damage can progress, and this is referred to as “diffraction before destruction”.

“But is this really the case? Can we really only get one data point from a sample before it is destroyed? And if we can get more than just the one, can we somehow exploit this, or are incidences of multiple hits problematic from a data quality perspective?” - These are some of the questions we asked ourselves during the research period that led to the paper “Megahertz pulse trains enable multi-hit serial crystallography experiments at X-ray Free Electron Lasers” by Holmes et al published in Nature Communication, 2022.

 While it is true that when a sample is exposed to a highly intense and focused X-ray pulse it becomes damaged and is subsequently destroyed. This is not the case for larger X-ray beam diameters where the radiation dose on the sample area is much lower (if the sample is smaller than the diameter of the beam). Furthermore, if the flux is not evenly distributed throughout the beam area, and therefore if the sample interacts with the X-ray pulse towards the edges of the beam, the radiation dose will be even more significantly diminished. Therefore, the size and shape of the beam affects how much radiation dose your sample will receive and if it will be destroyed. 

 Also, one must consider how the sample is delivered to the beam. At these facilities the crystal exists in a liquid, and therefore injectors are used to form a liquid stream which contains the crystals. The injector is designed so that it can produce a stream that aims to achieve a one-by-one delivery of crystals to the X-ray interaction area. The flow rate of the sample can be adjusted in different ways to adjust the  speed the crystal travels through the X-ray beam. The slower the crystal is moving through the beam, the more likely that the crystal will be hit twice before either exiting the beam or being destroyed. Generally, at an experiment the flow rate is optimized so that the crystal is only hit once in the central part of the beam.

 In Holmes et al. we have demonstrated that a crystal can be hit more than once and survive. In fact, we have thrice-successfully solved the structure of Lysozyme, separating the diffraction patterns obtained at the European XFEL into three discrete datasets based on whether the crystal identified in the diffraction pattern was hit once or twice, and whether it was the first or second hit for those hit twice. From these three discrete datasets, high resolution structures were successfully obtained and radiation damage was not observed in any of the structures. No significant differences in the protein backbone were found, and upon examination of the disulfide bonds, areas that are highly prone to radiation damage, no signs of bond braking was observed.

 Curious as to exactly why we were able to obtain a second high resolution diffraction pattern from single crystals, radiation dose calculations revealed that estimated doses were below 0.2 MGy which means that the sample received a dose below the expected radiation damage limit for protein crystals as a result of the beam size and shape.  

 But what does this paper actually mean for the advancement of SFX? Now that we have shown that it is possible to get multiple hits from a single crystal using a MHz XFEL source, we can take advantage of this multi hit “phenomenon” by controlling the pulse interval, therefore allowing us to capture protein movements within the same crystal that will help us to understand how proteins work when exposed to different situations. By controlling the pulse intervals, we can investigate the movements of these proteins at different time scales,  such as investigating how proteins react to certain stimuli (whether it is a drug or light source), and watching as it changes and restructures itself. In effect we will be able to generate molecular movies of proteins from the same crystal.