Diamond Annual Review 2021/22

D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 2 1 / 2 2 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 2 1 / 2 2 17 16 Macromolecular Crystallography Group Dave Hall, Science Group Leader M acromolecular crystallography (MX) exploits the hard energy, high flux X-rays created at Diamond Light Source to enable our international academic and industrial user community to investigate the structure and function of biological macromolecules at atomistic resolution and up to millisecond timescales. This provides deep insight into the details of biological activity key to our understanding of the processes of life, exemplified by the selection of outstanding scientific highlights in this section. Diamond provides access to a suite of seven MX beamlines 1 to a large international academic and industrial user community. The beamlines cover a very broad range of capabilities from high throughput, micro and nano-focus beams, extremely long wavelengths, room temperature in situ collection from crystallisation plates, (time resolved) serial synchrotron crystallography (SSX), a fragment-based screening platform (XChem) and the Membrane Protein Laboratory. The COVID-19 pandemic continued to have impact on the research carried out on the beamlines as well as on the user programme and how we operated the suite of MX beamlines. For a significant period of this report the MX team continued to operate the beamlines whilst users were not able to come to site due to (inter)national restrictions. User access continued to be supported either via remote access, automated collection or with on-site staff support to enable experiments requiring sample handling, preparation or more complex data collections such as crystallisation for VMXi, XChem fragment screening, long wavelength experiments on I23 or serial synchrotron crystallography and room temperature collection at I24 and VMXi. As restrictions have eased over time opportunities for site-based access for users have been provided as soon as possible. Early in 2020 a special rapid access user call opened for COVID-19 research with, to date, 69 proposals received from groups from across the world studying a range of SARS-CoV-2 targets via MX, cryo-EM, bioSAXS and SR-CD. Access was provided to the high throughput beamlines I03 and I04 for rapid turnaround of high-resolution structures, room temperature studies on I24 and VMXi, drug repurposing studies across the MX beamlines and the XChem I04-1 facility undertook large scale fragment screening where over this period the team improved the duty cycle per sample by 33%, in part following the installation of an Eiger2 XE detector capable of operating at 500 frames per second, bringing it in line with earlier installations on I03 and I04. To date there have been nine successful fragment screening campaigns supported by the I04-1 XChem facility with data collected from more than 22,000 crystals (Figure 1) . One such screen on the Main protease triggered the COVID Moonshot initiative 2 to design an antiviral drug in a fully open manner which already has lead candidates. In parallel automated unattended data collection (UDC) has been used for several SARS-CoV-2 projects including a concerted campaign coordinated through University of Oxford to study the antigenic landscape of variants with a focus on Spike domains complexing with Fab and ACE2. To date over 59 structures have been deposited in the PDB (public protein data bank repository) from this campaign. Overall, more than 25,000 data sets have been collected on Diamond MX beamlines related to COVID-19 research since February 2020 with over 20%of the SARS-CoV-2 structures in the PDB deposited as a result of work at Diamond. See the Supporting international COVID-19 research section for more details. UDC has also undergone continuous improvements over the year to broaden the capability across beamlines I03, I04 and I04-1 and provide more options for our users with this access mode now becoming increasingly popular, in particular with industrial users. To handle the variable beam size and flux characteristics across the energy range of beamline I04, the use of radiation dose in calculations for UDC has been implemented. To date more than 60,000 data sets have been collected across these beamlines via UDC. Collecting automatically with rapid turnaround frees up significant user time and provides flagship upgrade for Diamond-II. VMXi has achieved routine room temperature structure determination from crystals as small as 10 µm in situ within crystallisation plates, including of COVID-19 targets. User operation is fully remote with rapid automated feedback when crystals have been identified within plates. Typical room temperature datasets comprise data from1-12 individual crystals and resolutions better than 1.8 Å are regularly achieved. In partnership with the XFEL Hub at Diamond, time resolved SSX is being developed for VMXi. The micro/nanofocus MX beamline VMXm will enable data collection from crystals smaller than 1 µm. Work has continued throughout this year to further improve beamdelivery across the beamline energy range, sample handling and data acquisition and analysis such that it can currently cater for crystals in the range 2-10 µm in size mounted on TEM grids and special holders. The beamline has recently announced a call to the user community for applications for use of commissioning beamtime. Excellent progress continues to be made on the electron diffraction HeXI project with successful external scientific reviews completed and a full beamline-grade Conceptual Design Review for phase 1 recently finalised. The commissioning of offline electron microscopes, detectors, and software continues to inform conceptual design and technical decisions in collaboration with Diamond support groups. Productive external collaborations are also in place with user groups, institutes, and industry to further facilitate hardware, workflow, and processing solutions for rapid, routine electron diffraction. Throughout the last year work has continued for the Diamond-II upgrade programme. Muchwork has been done to identify critical components requiring upgrading across the suite as well as look at future competitive upgrade paths, including work on the conceptual design reviews for K04 (flagship beamline to replace and transform I04-1 XChem) and KMX, a flagship project to upgrade I24 to further improve the serial synchrotron crystallography offer alongside upgrading its microfocus capabilities. 1. https://www.diamond.ac.uk/Instruments/Mx.html 2. https://postera.ai/moonshot 3. Storm, S. L. S. et al. Experimental evidence for the benefits of higher X-ray energies for macromolecular crystallography . IUCrJ 8 , (2021). DOI: 10.1107/ S2052252521008423 rapid, consistent feedback to their research aims. No one cannot fail to be aware of the impressive results made in the last year of AI driven automated protein structure prediction by AlphaFold2 (DeepMind) and Robetta (Baker, University of Washington). These have had a profound impact on the field of structural biology and are being actively used to aid experimental structure determination. To this end we have implemented AlphaFold2 into our automated pipelines so that users can provide their target sequence and we generate predicted protein structures that are used in downstream automated software to aid experimental structure solution more readily than previously possible. Despite the success of AlphaFold2 in producing models for molecular replacement for successful phasing, several projects have still required the long- wavelength beamline I23 for structure solution. An app is now available to help evaluate the potential success for phasing for native sulphur SAD phasing. In addition, an increasing number of projects are using the extended wavelength range of the beamline to determine the nature of elements bound to proteins, nucleic acids and their complexes to gain further insight into their chemistry and function. I04 has recently introduced the option to dial a target dose instead of an exposure time. This is calculated using Raddose3D assuming a standard protein sample. Developments are ongoing to include additional sample information that can be provided by the user. This dose-based mechanism is implemented in UDC as well as the interactive GUI and is available over the entire beamline energy (6 – 18 keV) and beam size (8 x 5 to 110 x 100 µm 2 (h x v)) range. This option provides the user with a much better guide for different experimental aims without the risk of under- or overexposing samples. I24 continues to offer tuneable microfocus MX with extremely high flux densities at all energies. The I24 monochromator was replaced this year, upgrading one of the oldest components on the beamline resulting in significantly increased beam stability at the sample position. A cadmium telluride Eiger detector is now available allowing efficient high energy (>20 keV) data collection. The use of high energies exploits the energy dependence of howX-rays interact with crystals and allowsmore data to be collected fromeach crystal, especially when they are of a limited size 3 . The ease of use and breadth of SSX available at I24 has continued to develop: both static and light/substrate driven dynamic SSX is available as well as tools such as anaerobic data collection and offline slurry characterisation. Developments will feed into the I24 KMX Figure 1. XChem campaigns on SARS-CoV-2 targets. VMXm beamline at Diamond