16 17 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 1 9 / 2 0 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 1 9 / 2 0 Macromolecular Crystallography Group Beamlines I03, I04, I04-1 and I24 Structure guided antiviral drug design against SARS-COV-2 Related publication: Douangamath A., Fearon D., Gehrtz P., Krojer T., Lukacik P., Owen C. D., Resmick E., Strain-Damerell C., Aimon A., Ábrányi-Balogh P., Brandão-Neto J., Carbery A., Davison G., Dias A., Downes T. D., Dunnett L., Fairhead M., Firth J. D., Jones S. P., Keeley A., Keserü G. M., Klein H. F., Martin M. P., Noble M. E. M., O'Brien P., Powell A., Reddi R. N., Skyner R., Snee M.,Waring M. J.,Wild C., London N., von Delft F. &Walsh M. A. Crystallographic and electrophilic fragment screening of the SARS-CoV-2 main protease. Nat. Commun . 11 , 5047 (2020). DOI: 10.1038/s41467-020-18709-w Publication keywords: COVID-19; SARS-CoV-2 virus; Main protease; X-ray fragment based drug discovery; XChem; Antivirals; Protease inhibitors A t the end of 2019, a new coronavirus was identified as the cause of a disease outbreak in Wuhan, China. This disease now known as coronavirus disease 2019 (COVID-19) is caused by the highly contagious SARS-CoV-2 virus. As yet, we have no drugs against SARS-CoV-2 that have been verified in clinical trials and are limited to treating the disease symptoms. There is a global race to produce a vaccine, but the vaccines in development may not provide long-lasting immunity. Vaccines are not suitable for all patients, including those whose immune systems are weakened by disease (e.g. cancer). It is therefore vital to develop new treatments that target the virus itself. Working towards this goal, researchers at Diamond Light Source explored interactions between drug- like molecules and a part of the virus known to be essential to its reproduction. Researchers carried out the majority of this work on the Macromolecular Crystallography (MX) and XChem beamline (I04-1). Primarily I04-1 was used because it is embedded in the XChem pipeline, capable of handling the large numbers of samples quickly. Speed was vital, as researchers tested thousands of molecules to find promising candidates. Complementary experiments were carried out on other MX beamlines (I03, I04 and I24) to provide additional information. The XChem facility and the state-of-the-art structural biology labs in the Research Complex at Harwell were essential for rapid sample production and turn-around. This rapid research allowed the team to solve the structure of a protein essential to the reproduction of the virus. They have also identified more than 90 compounds worthy of further investigation for development of antivirals for SARS-CoV-2. There is no certainty that the project will succeed, but by working in a fully open source approach to harness expertise across the world, the project aims to do as much as it can to help accelerate the discovery of antiviral therapies effective against COVID-19. A novel coronavirus (CoV) belonging to the β-coronavirus cluster, SARS- CoV-2, has rapidly spread across the globe from an initial outbreak in Wuhan, China, in late 2019 1 .This is the third coronavirus to escape the animal population and cause human disease, following Severe Acute Respiratory Syndrome (SARS) in 2002 and Middle Eastern Respiratory Syndrome (MERS) in 2012 2-3 . SARS- CoV-2 is an enveloped, non- segmented, positive-sense RNA virus and as for other coronaviruses it possesses a large RNA genome of over 30 kb.The genome contains a 5' cap structure along with a 3' poly (A) tail, so it can act as an mRNA for translation of the replicase polyproteins.The replicase gene encodes the non- structural proteins (nsps) and takes up approximately 20 kb of the genome, with the remaining 10 kb encoding the structural and accessory proteins. The SARS-CoV-2 replicase is expressed in the form of two polyproteins pp1a and pp1ab containing the nsps 1-11 and 1-16, respectively. SARS-CoV-2 encodes two cysteine proteases that cleave the replicase polyproteins; a papain- like protease (PL pro ), encodedwithin nsp 3, and a 3C-like or main protease (M pro ), encoded by nsp 5. PL pro cleaves at three sites releasing nsp 1-3, while the M pro is responsible for the remaining 11 cleavage events, releasing nsp 4-16. Several of the nsps then assemble into the replicase-transcriptase complex (RTC) which is responsible for RNA replication and transcription of the viral sub-genomic RNA. To date, there are no antivirals that specifically target SARS-CoV-2. The SARS and MERS outbreaks have generated some research and much of this has been directed to validating a number of suitable antiviral targets, such as the viral proteases, polymerases, and entry proteins (the spike protein). However, significant work remains to develop drugs that target these processes effectively to inhibit viral replication. Furthermore, although vaccine research and development is proceeding at pace, it is not guaranteed that durable high level immunity will be induced. Pursuing novel avenues of innovation leading to chemical entities specifically and rationally designed to target the virus may therefore prove essential. At the start of the SARS-CoV-2 outbreak, many groups in China immediately focused their research efforts on understanding the novel coronavirus in detail. The groups of Zihe Rao and Haitao Yang at Shanghai Tech were able to solve the X-ray structure of M pro in early January 2020 which was an incredible feat, and they were keen to accelerate their work. Having worked with Diamond Light Source in the past, they were keen to apply a high-throughput structure based drug discovery approach to M pro here. In discussion with the Rao andYang groups, we were able to rapidly apply the methods they used to overproduce M pro in sufficient quantities for structural work at Diamond. Using the XChem platform at Diamond, experiments to explore interactions made between M pro and hundreds of low molecular weight organic molecules (‘fragments’) in just tens of hours at beamline I04-1 commenced. The small size and varied functional groups and chemical properties of these fragments allowed us to probe the surface of the enzyme – particularly the enzyme’s active site, the location at which it performs chemistry – for new binding interactions. Then, by merging and growing fragments with useful properties we could iteratively improve the binding of these compounds that can potentially lead to drugs that will bind tightly to the targeted protein, stopping it working. Finding a molecule with useful properties is a low probability event, but the high throughput capabilities of the XChem platform and the I04-1 beamline were leveraged to allowM pro to be screened against thousands of different molecules over several days. Additionally, in collaboration with the London research group at the Weizmann Institute of Science in Rehovot, Israel, mechanistic inhibitors have been designed specifically to target and form a covalent bond with a cysteine residue in the active site and these were also assessed.These molecules form a covalent bond with the enzyme and were first identified by incubating the protein and ligand then performing intact protein mass spectrometry (MS). If the mass of the protein increased then it indicated that the molecule had bound. Large numbers of potential covalent inhibitors could be screened rapidly by MS and those found bound to M pro could be followed up crystallographically at I04-1 to visualise how and where they bound to M pro . Escherichia coli bacteria were used to produce milligram quantities of M pro protein that was then purified to crystallisation standard. The construct for M pro protein expression has been shared with other groups in the UK and around the world, allowing multiple inhibitor studies to be carried out by many methods including nuclear magnetic resonance, surface plasmon resonance and mass spectrometry, to accelerate progress. The sample was crystallised using the crystallisation facility at Harwell, which is a joint venture between Diamond, the Research Complex at Harwell and the Rosalind Franklin Institute. The quality of the initial crystals was poor but they appeared over a matter of hours. Using this to our advantage, multiple cycles of optimisation were applied to improve the crystal quality in a rapid fashion. The crystal structure of M pro was solved to high resolution (1.25 Å) on the Microfocus MX beamline (I04) at Diamond providing a high precision ground state structure for comparison with liganded structures. Crystal structures of the protein were solved for more than one thousand fragment binding experiments, and over one hundred binding molecules have been discovered. The average resolution of all datasets was 2.1 Å, allowing mechanisms of binding and changes in protein conformation to be described with confidence. The majority of fragments bind in the active site of the protein, and additional high value fragments have been found that bind at the dimerisation interface of the protein (Fig. 1). As dimerisation is essential for M pro activity, this site is worthy of particular attention. As of the end of March 2020, 68 binding events have been structurally characterised, and based on these hits, screening for improved fragments is ongoing. Based on the screens, one series of promising mechanistic covalent binding inhibitors has been progressed to find improved binders with greater potency. It remains a challenge to transform weakly binding fragments into clinical candidates. Due to the time-sensitive nature of the current COVID-19 outbreak, and to accelerate progress in drug development, the data have been made immediately available to the wider community through regular structure depositions in the protein data bank via the fragalysis platform at Diamond and through a crowd sourced drug design programme, Covid-Moonshot. However, time is of the essence to combat the pandemic and developing antivirals or vaccines typically takes a decade or more of research. A potential strategy for accelerating the discovery of antivirals for SARS-CoV-2 is to repurpose existing drugs approved for other diseases. So, in parallel to starting from scratch using the X-ray fragment-based approach (albeit in a high-throughput way never before realised through the Covid-Moonshot approach), we have established a joint initiative with Exscientia Ltd, an AI drug discovery start-up based in Oxford Science Park to screen almost every known approved and investigational drug – 15,000 clinical molecules – against the SARS-CoV-2 proteases. At the time of writing these screens are near to completion. The hope is that by being able to start with clinically-approved drug molecules we can move more rapidly to clinical trials and potentially provide treatment for patients. This is not your normal Diamond Annual Review highlight, but a snapshot of the rapidly moving open-source approach to accelerated drug discovery in a time of crisis. Science is typically a competitive environment and achieving academic recognition through the publication of work is a well-defined process. However, this has led in modern science to a research culture that cares exclusively about what is achieved and not about how it is achieved. The completely open approach being pursued here, and by other scientific groups who have dropped everything to focus on COVID-19 to work selflessly and openly, releasing results and sharing them with the world as soon as they are produced, will help us respond toWellcome’s call 4 for how we can all reimagine how we conduct research. The work in progress highlighted here has been a truly open and collaborative effort involving the Walsh and von Delft teams at Diamond and collaborators at the Weizmann (London group), Exscientia Ltd, the University of Oxford (Schofield and Vakonnakis groups), Rosalind Franklin Institute (Owens group) and the University of Newcastle (Kawamura, Noble and Waring groups). References: 1. Zhu N. et al. A Novel Coronavirus from Patients with Pneumonia in China, 2019. N. Engl. J. Med. 382 , 727–733 (2020). DOI: 10.1056/NEJMoa2001017 2. KuikenT. et al. Newly discovered coronavirus as the primary cause of severe acute respiratory syndrome. Lancet 362 , 263–270 (2003). DOI: 10.1016/S0140-6736(03)13967-0 3. Bermingham A. et al. Severe respiratory illness caused by a novel coronavirus, in a patient transferred to the United Kingdom from the Middle East, September 2012. Euro Surveill. 17(40) , 20290 (2012) 4. Farrar J.Why we need to reimagine howwe do research. Wellcome (2019) https://wellcome.ac.uk/news/why-we-need-reimagine-how-we-do- research (accessed March 2020) Funding acknowledgement: This work has been made possible by repurposing of funding from Diamond Light Source,Weizmann Institute, Rosalind Frankin Institute, Universities of Oxford & Newcastle. Corresponding author: Dr MartinWalsh, Diamond Light Source,
[email protected] Figure 1: The dimar protease showing fragments bound at active site and other target areas.