11: i24 kmx upgrade

SAC

Doc No: DLS-SAC-I24KMX Issue: 1 Date: 25 Nov 2020 Page: 1 of 24

1

11: I24 KMX upgrade microfocus and kinetic MX

Prepared for Diamond SAC/DISCO

December 2020

SAC


2

1. Acknowledgements

Working Group

o Mike Hough (University of Essex, Lead/chair)

o Arnaud Basle (University of Newcastle, DUC)

o Andy Dore (Sosei Heptares)

o James Errey (Evotec)

o Helena Kack (Astra Zeneca)

o David Leys (University of Manchester)

o Ivo Tews (University of Southampton)

o Briony Yorke (University of Bradford)

Diamond staff

Robin Owen, Danny Axford, Selina Storm, Sam Horrell, Allen Orville, Dave Hall, Gwyndaf Evans, Martin Walsh, Dave Stuart.

Relevant dates

Working group meetings: 29 July 2020, 29 September 2020, 15 October 2020

Webinar: 3 November 2020

SAC


3

2. Executive Summary

The UK has a rapidly growing community of academic and industrial structural biologists wishing to access dynamic structural information to help explain the mechanisms, binding modes, drugability, and other features of important biological molecules and complexes. Over the last 5 years, through the Diamond XFEL Hub, UK users have been exposed to the potential of serial crystallography for accessing dynamic structural information, and recent developments at I24 have further served to build interest and experience in this field, although it remains a specialist’s technique.

I24 KMX is a flagship upgrade to I24 that will enhance its capability in the measurement of high-quality, low-dose, MX data from cryo-cooled and room temperature 1 – 10 micron-sized crystals with microsecond time resolution. Critically it will build upon I24’s strong track record in delivering high quality microfocus capability by further optimising beam and sample delivery systems to improve signal to noise. In the area of time resolved crystallography I24 KMX will provide well-integrated tools for initiating and tracking reactions in the microsecond to second time domain, thereby complementing the femto- to nano-second time resolution that can be achieved at XFELs. I24 KMX will be designed to deliver, without compromise, a step change in routine cryo-crystallography, whilst also offering serial MX together with in situ validation of intermediate states and appraisal of radiation induced damage through the integration of complementary data collection such as X-ray emission spectroscopy (XES).

The project goal will be to enable dynamic structural studies using microsecond, microbeam serial crystallography techniques within the envelope of a world-class microfocus structural biology beamline. Diamond-II and I24 KMX offer the opportunity to deliver routine, high quality microsecond structural dynamics for the UK, opening this technique up to a broad cross-section of users and adding an extra dimension to the study of biological structure and function.

SAC


4

3. Scientific case

3.1 Introduction

I24 is the highest impact life science beamline at Diamond. For more than a decade it has provided users with state-of-the-art microfocus capability that is high throughput, delivers high quality data and is user friendly. It has been in very high demand from industry and academia alike and has helped expand our understanding, for instance, of a range of therapeutically relevant GPCR structures [1-6] as well as providing insight into foot and mouth disease virus, malaria, Parkinson's disease, cystic fibrosis and antibiotic resistance [7-11]. Its ability to offer room temperature in situ screening and data collection alongside cryo-crystallography was a unique innovation at a microfocus beamline [12]. In recent years the portfolio of capability has been broadened to include fixed target and extruder serial synchrotron crystallography (SSX): capability that provides users with very low dose snapshots of radiation sensitive samples, retaining biological and chemical relevance of the structures, and is beginning to provide time resolved data for some proteins on the millisecond timescale.

The I24 KMX upgrade envisages a significant optical and end-station upgrade to I24 to provide both high quality pin-based microfocus MX at 100K and room temperature pin and serial approaches in a dual end-station environment. There are extremely strong parallels between classical microfocus MX and serial synchrotron crystallography (SSX), and a fundamental aim of I24 KMX is to blur the boundary between the two so both can benefit. Improvements to the X-ray focus and beam intensity realised at the I24 sample position, coupled with optimisation of signal to noise and faster shuttering can, in conjunction with multi-modal data collection (i.e. the collection of secondary forms of data such as X-ray emission spectroscopy (XES) in addition to X-ray diffraction), benefit all experiments.

While serial approaches provide a pathway for MX to tackle one of the Diamond-II science drivers of providing atomic-level insight into the dynamics of macromolecules, implementation for the general user community is far from straightforward. Sample preparation, data collection, driving and following reactions in crystallo all present significant challenges. These experimental considerations combined with challenging data analysis and the need to ensure intermediates are validated form a substantial entry barrier to new users. The cumulative effect of these is illustrated in part by the number of academic and industrial user groups using I24 for serial and standard MX experiments (6 of 34 proposals requested serial beamtime at I24 in AP27). By pursuing both traditional microfocus and serial approaches, I24 KMX aims to address these challenges, lowering the entry barrier to dynamic crystallography and removing the perception that dynamic serial approaches are applicable only to robust systems.

The proposed beamline directly addresses several key aspects of the Diamond-II life science case including structure solution of membrane proteins, dynamic studies of biological systems, structure-based therapeutic discovery and the role of metals in disease.

3.2 Science enabled by project

Through their nature, challenging targets often result in challenging crystals which may be small, badly ordered, or difficult to produce. Dedicated tools and methods are required to tease the best possible data from them, often at the expense of the highest levels of automation or throughput. An exemplar of this is the success of membrane protein crystallography at I24. Membrane proteins make up nearly 30% of known eukaryotic proteins, yet constitute only 3.5% of structures in the Protein Data Bank1. At Diamond MX beamlines excluding I24, membrane proteins represent 1.2% of the deposited structures. In contrast, at I24

1 https://blanco.biomol.uci.edu/index.shtml

SAC


5

the percentage is 6.5%, illustrating the gains that can be made by targeting a particular challenging field. At I24 a well-focused X-ray microbeam coupled with the development of bespoke tools such as the diffraction grid scan [13] and its faster successors ported from fixed target serial crystallography [14], allow invisible samples to be reliably centered in the X-ray beam and well diffracting sub-volumes of crystals to be identified [15-17]. The use and impact of the grid scan and a related tool, the line scan, are illustrated in figure 1. In each case new beamline tools directly benefit the user and aid structure solution and data quality. I24 KMX will seek to continue this marriage of cutting-edge beamline technology and close collaboration with users working on challenging targets. With a focus on protein dynamics and also membrane protein crystallography, we anticipate that tools developed to enable new science in these fields will become part of the standard toolkit for crystallography both at I24 and the other MX beamlines at Diamond.

Figure 1. Top: crystals of the class B human GPCR corticotropin releasing factor [2] in opaque cryo-cooled lipidic cubic phase (LCP) located on I24 using the diffraction grid scan. Bottom: crystals of the mitochondrial ATP-binding cassette (ABC) transporter re-orientated to avoid overlaps (right). Data were subsequently collected using the I24 line scan [18], illustrating translation of a large anisotropic crystal through a small X-ray focal spot to enable structure solution. Figure adapted from Warren et al. (2016) [19].

Dynamic crystallography Time resolved macromolecular crystallography has enjoyed a renaissance in recent years, due in no small part to the promise of femtosecond, and slower, time resolution delivered by X-ray free electron lasers using serial femtosecond crystallography (SFX) [20, 21]. Together, advances in sources and sample delivery have opened the door to following irreversible processes in crystallo. More recently these serial sample delivery techniques have been exploited at, and optimised for, data collection at synchrotrons – so-called serial synchrotron crystallography (SSX) [22] - a field in which I24 and the XFEL Hub are world leaders. Time-resolved SSX at I24 can already reach the millisecond timescale and the combination of Diamond II and KMX upgrades will facilitate a push into the microsecond time domain, enabling capture at room temperature of short-lived and radiation sensitive intermediates that are inaccessible by more traditional freeze-trapping methods. Even for long-lived intermediates when freeze-trapping could be employed, room temperature approaches are preferable as a 100 K structure may not be representative of the physiologically relevant structure of that state [23].

Routine dynamic crystallography is a new capability for Diamond-II, representing a key link in understanding how life operates on the atomic scale. The proposed aim of probing and resolving timescales of

SAC


6

microseconds and longer should not be seen as a limitation: this is much faster than the average 70 millisecond turnover time of enzymes in solution [24] and, as figure 2 illustrates, encompasses a huge range of processes in biology. Furthermore, catalytic turnover is typically slowed in crystallo which will bring more enzyme mechanisms within reach [25].

Figure 2. Timescales where I24 KMX will play a role, highlighting biological processes that occur in the microsecond to second time domain. At all sources, longer time scales are accessible, though at XFELs there is a natural bias towards the shorter timescales that more closely approximate pulse length. Adapted from the Diamond-II science case; figure originally adapted from Cell Biology by Numbers by Ron Milo, Rob Phillip (2015).

Room temperature data collection The value of ambient (room) temperature structures of proteins is becoming increasingly recognized [26], as the use of cryoprotectants and cryo-cooling to ~100K limits protein dynamics and may lead to non-physiological conformations being observed [27]. Room temperature structures, where protein conformational dynamics may be more representative of those in vivo, are therefore highly desirable, even if they are more challenging to obtain. In particular, non-cryogenic samples are important for time-resolved crystallography in order to allow e.g. enzyme reactions, or receptor/sensor protein conformational rearrangements to be observed.

Validated diffraction data Conformational rearrangements or differences between intermediates along reaction pathways may be subtle and difficult to resolve using only electron density maps from X-ray diffraction. An example of this is the changes in manganese oxidation state in the oxygen evolving complex of photosystem II during photosynthesis [28]. Interpretation of data is particularly challenging when the resolution of diffraction data is limited, as is often the case for membrane proteins. In order to address this, I24 KMX will enable simultaneous collection of both diffraction and complementary spectroscopic data (e.g. X-ray emission (XES) and/or electronic absorption, fluorescence, or Raman spectroscopies). By acting as a dual data source, I24 KMX will provide the experimenter with both atomic and electronic information, reducing model ambiguity and significantly increasing functional insight of catalytic intermediates. While coordinated data collections of this type are currently possible in tour de force experiments, I24 KMX will aim to make multi-dimensional data a routine part of MX diffraction data collection.

SAC


7

Watching catalysis in crystallo Reactions and catalysis can be triggered in crystals by several means, the fastest being laser illumination. Such laser excitation allows ultra-fast (ps-ns) biological processes such as carbon monoxide binding in myoglobin [29] and isomerization of photoactive proteins [30] to be probed in crystals. The use of photocages can expand this laser pump-probe approach to make it applicable to enzymes that are not light dependent [31]. For slower processes, such as changes in conformational modes or ligand binding, substrate can be mixed with [32] or projected onto crystals [33]. For substrate binding, and larger conformational changes, a key premise is that data should be collected from microcrystals (less than ~5 microns in size). This is because microcrystals allow efficient soaking of substrates into crystals, meaning that diffusion times become small and the distribution of states becomes more homogenous over the crystal volume [33, 34]. Crucially, serial approaches, photocages and mixing methods allow irreversible reactions / processes to be analysed via time-resolved crystallography.

It is clear that variations in crystal volume, substrate concentration and changes in other small experimental factors will alter the rate of reaction between samples and limit the achievable time resolution. I24 KMX will address this through on-line tracking of in crystallo reactions and cross-validation of where a crystal is on the catalytic pathway, which will aid in merging, processing, and interpretation of diffraction data to form validated time-resolved structures.

The insight added to X-ray diffraction by spectroscopic approaches is illustrated by the use of X-ray emission spectroscopy (XES) to characterise metal oxidation states in the oxygen-evolving complex in photosystem II [28]. XES spectra provided key information to complement limited resolution (5.0 Å) diffraction data confirming the initial electronic state of the manganese cluster and how it changed throughout the dynamic experiment. In addition, other forms of spectroscopy such as UV-Vis absorption and fluorescence can also provide information on (changes to) electronic structure [35] and will also be investigated for incorporation into the I24 KMX sample environment. In a new capability for Diamond, I24 KMX will enable diffraction data to be ‘tagged’ and selectively processed using simultaneously collected spectroscopic data as a filter. While aimed at time-resolved experiments, this facility will also benefit conventional microfocus experiments where only thin wedges of data can be collected from crystals.

Antibiotic resistance A case study of where such a dynamic approach to MX could have a significant impact is in understanding the sequential acylation and diacylation of the ß-lactam ring that occurs when ß-lactams interact with penicillin binding proteins and with ß-lactamase enzymes. ß-lactamases provide resistance to antibiotics – a significant risk to human health – and determination of the atomic structures of their catalytic intermediates can provide insight into compounds that inhibit their function [36]. Some of the first experiments at the European XFEL determined the structure of a ß-lactamase providing a first step to revealing transition states during inhibitor binding [37].

Membrane proteins and drug discovery A key area in structural biology concerns membrane proteins such as GPCRs. Remarkably, 60% of drugs target membrane proteins yet they comprise less than 4% of structures in the Protein Data Bank. X-ray crystallography has the power to visualise unusual binding modes in GPCRs [38] by providing a snapshot of a ground, or bound, state. GPCRs are dynamic molecules where ligand binding produces a conformational change that ultimately triggers a cellular response. KMX will build on the first steps of serial [39, 40] and microfocus high throughput GPCR crystallography [41] to provide visualisation of conformational changes in challenging samples such as GPCRs. This is of real interest for GPCRs as, despite the insight snapshots of the active and inactive sites provide, the transitional mechanisms between these end points remain unclear. In particular, the allosteric nature of coupled conformational motions and ligand response in GPCRs [42, 43] may mean that critical transient sites (e.g. druggable pockets) only become apparent through use of a dynamic time resolved approach where the kinetic context is critical. Key GPCR dynamics are believed to occur on the microsecond to millisecond timescale [44, 45] matching the targeted time regime of I24 KMX.

SAC


8

Metalloproteins and radiation damage Around half of all proteins contain a metal co-factor, with approximately one quarter to a third requiring a metal to carry out their function [46]. Metal centres are, however, exquisitely prone to X-ray induced radiation damage [47] and the extremely high flux densities that will be realised at I24 KMX will provide two challenges for the experimenter. First, crystals will have an extremely short lifetime and it will not be possible to obtain a complete dataset from a single crystal due to the effects of global damage (loss of diffracting power, cell expansion, increasing non-isomorphism). This challenge can be addressed in part through monitoring scaling statistics and implementation of multi-dataset merging software such as xia2.multiplex [48]. Second, local damage (disulphide reduction, metal centre electronic state changes, electron driven reactivity) will occur on much shorter dose (time) scales and is more insidious. Proteins containing redox centres, such as transition metals or flavin cofactors, are highly susceptible to electronic state and structural changes as a consequence of X-ray dose (figure 3). This is particularly evident for high valent metal centres such as the Fe(IV) ferryl heme centres used for catalysis in peroxidases and cytochromes P450, as well as in non-heme Fe(IV) centres used in isopenicillin N synthase (IPNS) and the 2-oxoglutarate (2OG) dependent DNA demethylase AlkB. Importantly, these site-specific changes occur at much lower doses than loss of diffraction (figure 3) [49, 50], providing a strong case for the incorporation of complementary methods to identify and track X-ray induced changes.

The electronic structure of a metal centre is an essential aspect of reactivity. The availability of in situ spectroscopic methods, fully integrated into the beamline to characterise radiation damage processes and verify the state of any particular structure is essential to ensure that valid biological information is obtained and provides a secondary motivation for the collection of data complementing diffraction.

Figure 3. Comparison of the heme environment in the iron containing dye-type heme peroxidase DtpAa as determined by SFX and SSX. Even when the absorbed dose is low (10s of kGy compared to the ~600 kGy required to halve the diffracting power of protein crystals at room temperature [50], X-ray induced site-specific changes in metalloproteins can be significant. Data from Ebrahim et al IUCrJ (2019) [49].

Maximising signal, minimising noise The inevitable onset of radiation damage may be slowed in part by the use of higher photon energies. Recent, and ongoing, work at I24 using a CdTe Eiger2 detector has shown that, provided a detector with an appropriate sensor is used, significantly more diffraction data can be collected per unit absorbed dose; with the proviso that X-ray energies of around 22-26 keV are used rather than the more typical 12.4 keV. For a given absorbed dose data collected at higher energies are also of

SAC


9

higher resolution. Exploitation of the energy dependence of radiation damage will benefit both traditional and serial crystallography at I24 KMX.

Routine operation at high X-ray energies will be complemented by an improved focus, top-hat profile for larger beam sizes and increased flux density for probing weakly diffracting samples (details in §4 and §5.2). The upgrade provides the opportunity for a large co-ordinated upgrade to less valued yet critical beamline components such as slits, apertures, scatter guards and beam-stop which have a significant impact on signal to noise levels and hence on the quality of data that can be obtained from small crystals [51].

Sample delivery and automation The standard sample mount for MX data collection at Diamond and many other synchrotron sites is the SPINE pin. This will continue to be a dominant mode of sample mount at I24 KMX which will continue to integrate and exploit the highly successful automation developments led by I03. Automated cryo-MX capability will ensure I24 KMX continues to also offer high throughput microfocus MX, balancing the load across the other beamlines, retaining Diamond’s overall capacity and meeting the UK’s demand for synchrotron-based protein crystallography. To benefit pin-based microfocus MX, existing hardware will be developed or replaced to allow still faster rastering over samples and routine kappa goniometry to allow better access to all regions of reciprocal space.

Serial sample delivery continues to be a rapidly evolving area. I24 has helped lead the way in fixed target SSX [52], and extruder data collection at I24 is also routine (eg PDB entry 6GUY, manuscript in preparation), though both are incorporated into I24 in an ad hoc informal manner. The XFEL Hub has contributed extensive experience to the development of the acoustic drop tape drive [53] used at LCLS, PAL and SACLA though not, as yet, at Diamond. I24 KMX will aim to provide these as a minimum, but a key aspect will be the flexibility to allow established methods to evolve and incorporate new and emerging techniques. A recent example of the impact an evolving sample delivery and sample environment can have is illustrated by the implementation of anaerobic fixed target SSX at I24 [54]. Enzyme-catalysed reactions using atmospheric dioxygen are manifested in most kingdoms of life and have fundamental roles in many aspects of biology, with the importance of oxygenases and oxidases reflected by the award of the 2019 Nobel Prize for Medicine in this field. The provision of a room temperature low-dose method for probing oxygenases, and other oxygen sensitive enzymes such as the nitrogen-fixation enzyme nitrogenase which requires anaerobicity, at synchrotrons will be a unique tool provided by I24 KMX. End-station schematics and sample delivery possibilities are detailed further in §5.3.

Bridging and blurring of the gap between serial and conventional pin-based approaches is a key aim of I24 KMX. An exemplar is mesh-based rotation crystallography, where each microcrystal is rotated by a few degrees during data collection and several diffraction images are collected. This incurs a higher dose than serial crystallography (single image per crystal) but allows for improved data quality and structure solution from fewer crystals. For the user an incremental progression to room temperature data collection in terms of methods, beamline tools, and number of crystals required, provides a clear pathway to observing in crystallo changes and dynamics while the pipelines developed for tracking changes provide validation for static structure determination. Essential to the success of this is a single, flexible and well-supported user interface for all forms of data collection (see §5.5).

Reducing sample consumption A critical focus for the beamline will be on ensuring the serial approaches offered are sample efficient, since this currently limits the impact of the method. Currently data from many hundreds or thousands of crystals are merged to obtain an SSX dataset: hugely daunting numbers for the majority of crystallographers and even more so for projects when producing each single crystal is a challenge. The use of a wide bandpass multilayer helps address this challenge, significantly reducing the number of crystals required to form a dataset [55]. The gain in flux provided by the multilayer also provides access to the microsecond time domain by reducing the time required to obtain an interpretable diffraction pattern.

SAC


10

Developments to serial approaches to reduce sample consumption will also be pursued, building on new techniques such as the use of acoustic drop ejection to load fixed targets [56] and offline characterisation of sample holders prior to collection [57].

A flexible future A key feature of I24 has been the ability to be flexible. Within the envelope of a world class microfocus beamline, new capabilities and approaches such as in situ data collection, grid scans, serial crystallography, acoustic sample delivery, and data merging software have been developed to address challenging problems in structural biology. Several of these now form the bedrock of data collection at the MX beamlines. I24 KMX seeks to continue this approach, providing a flexible instrument for MX at Diamond that can drive science by allowing novel experiments. Key capabilities such as cryo-pin data collection, fixed-target serial crystallography, photoexcitation and the availability of complementary data will provide the foundation for new experiments as MX evolves at Diamond-II.

3.3 Diamond-II portfolio

The capability of I24 KMX to probe enzymes on the atomic scale at room temperature provides a fundamental pillar in the Diamond-II vision for integrated structural biology. No single method can provide a molecular description of a biological system, and MX comprises one part of a suite of complementary tools incorporating small angle scattering, X-ray microscopy, and electron microscopy. Although MX remains the leading tool for atomic visualisation of enzymes, development in all of these fields is pivotal to Diamond-II (X4SCM (flagship proposal 9); K04 XCHEM (proposal 12); nano-bioimaging (proposal 13); eBIC). MX provides a rapid means of determining the shape of enzymes in detail and reveals how they interact both with each other and with small molecules such as drug fragments. The ability to collect a series of structures of an enzyme along a catalytic reaction coordinate on timescales ranging from microseconds to seconds will enable molecular stop-motion movies to be produced, providing direct evidence of how protein structure relates to function.

At Diamond II, I24 will develop further its role in the MX suite. Its highly optimised microfocus capability with high flux at high energies straddles the capabilities between the “workhorse” ultra-high throughput beamlines I03, I04 and K04, is able to step-up to meet demand for (micro)-cryo-crystallography as required and bridges the gap to the difficult experiments with the smallest crystals that will be investigated at VMXm.

Room temperature crystallography has seen a resurgence in recent years and the development of in-situ diffraction experiments at I24 and I03 played a key role in this at Diamond with successes reflected ultimately by one branch of the redeveloped I02 being dedicated to in situ crystallography in the form of VMXi. Through use of a double multilayer monochromator VMXi is able to provide the highest flux available at an MX beamline at Diamond which, in combination with a sub ten micron beamsize, allows screening and data collection from microcrystals within crystallisation media. The VMXi beam properties are also well suited to serial crystallography and work with I24 and the XFEL Hub to incorporate extruder, drop-on-demand and fixed target data collection is ongoing. The XFEL Hub is currently leading development of an on-demand system that can be used at both VMXi and the SwissFEL Cristallina beamline which will benefit the user communities of both sources. This complementarity and diversity in serial approaches across VMXi and I24 KMX can, together with the local expertise of the XFEL Hub, provide for and sustain the build-up of the community for serial experiments at Diamond-II while providing flexibility for both beamlines to adapt to demand for cryo (I24) and in situ (VMXi) experiments.

As a coordinated suite of instruments, Diamond MX beamline developments are deployed across the group and refined as exemplified by, for example, robotic developments from the I03 automation team through to the fast raster scanning developed by the I24 for SFX experiments which now sits at the heart of highly efficient X-ray centering across the MX beamlines – investments in individual MX beamlines for Diamond II

SAC


11

will inevitably be recouped by exploitation across MX and beyond (robotics and fast raster scanning have been deployed further across Diamond for example).

Coordinated access to MX using conventional and SSX methods, SAXS and imaging beamlines in Diamond-II, and eBIC will transform the rate of information gathering, and its depth, covering both spatial and temporal domains to truly understand structure, dynamics and function in the life sciences. This will allow users to build on joint experiments such as those on the oncoprotein smoothened GPCR which used crystal structures determined at I24 in conjunction with SAXS data collected on B21 to shed light on the structural mechanism by which the activity of a GPCR is controlled by ligand mediated interactions [58].

Solution and single molecule dynamic techniques, for example X-ray foot printing mass spectrometry (XFMS) [59]and diffracted X-ray tracking (DXT) [60] on the soft condensed matter beamlines will also provide useful complementary data to KMX studies on membrane proteins and other systems. The development of dynamic SAXS, XFMS and DXT together with time-resolved MX at Diamond-II will allow identification of both large conformational changes and visualization of key protein-ligand interactions providing a structural basing of such allosteric processes.

I24 KMX can provide crucial and unique tools within the Diamond-II life sciences portfolio for probing dynamics, including high spatial resolution, room temperature data collection, access to a wide range of time domains and high throughput. For I24 KMX Diamond-II provides increased flux, in particular at higher X-ray energies, and a very low emittance source that can be easily focused to the sub ten micron beam size required at the sample position.

3.4 Academic user community and beneficiaries

The strength of Diamond in the field of structural biology is reflected in part by the recent milestone of 10,000 Protein Data Bank depositions passed by the Diamond MX beamlines. With depositions from groups based in the UK, Europe, Middle East, USA, Africa, and Australia from both academia and industry, the magnitude of this milestone, and the short time in which it was reached, illustrates the breadth, diversity, and size of Diamond’s MX user community. The extent to which this MX community rely on I24 is highlighted by many of the statements of support for I24 KMX (section §7) and also highlight how I24 is used for successful data collection from their most challenging projects or when data collection has failed elsewhere due to limitations imposed by the size or quality of samples.

A major I24 upgrade will improve the delivery of high impact science through microfocus cryo-data collection capability by combining a high intensity well-focused X-ray beam with tools such as grid scans, multi-crystal data collection and in situ validation. The beamline will be extremely well suited to data collection from a wide range of challenging crystals. I24 KMX will provide routine access to low dose datasets from challenging radiation sensitive protein samples where the interpretation of biological function can be compromised by X-ray damage. I24 KMX will enable Diamond to build upon its competitive position in SSX by developing dynamic crystallography at Diamond to create an exciting and highly relevant asset for the already vibrant XFEL-Hub community. Far from merely providing a pathway towards XFEL experiments however, KMX will be able to probe key biological timescales from microseconds to seconds, provide physiologically relevant low-dose room temperature structures and improve upon its high-throughput cryo-MX capability for both academic and industrial users.

While pin-based microfocus MX is a well-established technique, tools and approaches for efficiently obtaining the best possible data from the most challenging samples are continually evolving. New tools will be demonstrated to users through regular training sessions such as BAG training. For less widely used techniques such as serial and dynamic crystallography, dedicated training workshops will be co-organised by I24 and the XFEL-Hub. The first of these is already scheduled for May 2021. Through workshops such as

SAC


12

this and BAG training, I24 KMX and the XFEL-Hub will together promote the capability of Diamond in dynamic MX and foster growth in the community. 3.5 Industrial user community and beneficiaries - impact on UK PLC

Structure-based drug design (SBDD) is a key tool for the pharmaceutical industry. X-ray crystallography provides a high throughput means of realising SBDD allowing many tens of potential ligands to be screened and visualised at atomic resolution within a single visit. I24 has been used by 26 different companies and is regularly the beamline most used by industrial clients at Diamond. I24 has generated over 15% of all industrial revenue since the start of the industry service in 2008. Industrial use of I24 is also reflected through publications: 47 (6%) of the publications deriving from I24 include an industrial co-author. Industrial use of I24 reflects the quality of the beam provided, the development and availability of tools such as grid scans and automated processing, and, critically, the drive from the beamline to advance tools based on feedback from (all) users. Grid scans are an essential part of structure-based GPCR drug discovery for example, and since the introduction of grid scans on I24 speed has increased by an order of magnitude with automated scoring able to keep up.

I24 KMX will aim to provide structures faster and from smaller samples than is currently routinely possible. A focus will continue structure solution from challenging crystals; for example, those grown in LCP or larger but highly non-isomorphous crystals such as those often obtained for GPCRs and other membrane proteins. Automation and throughput are essential aspects of structure-based drug discovery and I24 KMX will deliver these. Multiple structures and views of binding are critical to understand key ligand docking modes, especially in GPCRs [17]. This will be accommodated at I24 KMX through high throughput and the advent of dynamic crystallography; which will blur the boundaries between multi-crystal microfocus and serial MX to facilitate challenging samples.

Industrial use of serial techniques is in its infancy. Due to the potential of room temperature SSX to provide a complete assessment of protein-drug kinetics, identification of transient binding modes, and enable optimisation of drugs targeted to transition states, industrial use of serial techniques can be expected to increase. This has already begun with ongoing collaborative experiments between I24 and multiple industrial groups. Application of key synchrotron attributes such as throughput, exploitability by non-expert users, and automated processing pipelines will drive uptake by both the academic and industrial user communities provided that the beamline continues to evolve based on user feedback and experience(s). We anticipate that this approach will maximise usage opportunities for both academia and industry and not exclude experimenters who feel they may only occasionally have need for micro-focus. Maintaining a good cross-section of users doing “standard MX” is perhaps also the best way to bring in new dynamics projects as and when they emerge.

I24 KMX is expected to provide fundamental understanding in the following science areas.

• Chemistry of complex molecular processes that govern life and disease • Aid discovery of advanced drugs and new antibiotics • Structure discovery of membrane proteins, in particular, GPCRs • Structure determination of metalloproteins at extremely low X-ray doses • Methods for time-resolved structural biology

SAC


13

3.6 Comparison to other facilities, current and planned

Table 1 summarises comparable synchrotron beamlines: MicroMAX 2 (MAX-IV, Lund); ID29 3 (ESRF, Grenoble); FMX 4 (NSLS-II, Brookhaven); and P14-EH2 5 (PETRA III, Hamburg). Nanofocus and more specialised beamlines such as VMXm at Diamond are not included though are discussed briefly below. Unique to KMX is routine parallel readout of multiple detectors so both atomic and electronic structural data can be simultaneously obtained from crystals and a focus on high energy operation. A complementary focus on room temperature structure determination from modest numbers of crystals, exploiting showers of crystals often obtained during crystallisation. Serial techniques are also key, and this will help to bridge the gap to serial experiments for non-expert users, who comprise the vast majority of the UK’s structural biologists. This would complement the massively automated capabilities of VMXi to collect room temperature data from 1000s of crystals per day in crystallisation media. There is a synergy in building a room temperature community around I24, VMXi, and the XFEL-Hub to develop current and emerging methods. Together, and working with the XChem facility, a range of opportunities to provide complementary information at near physiological temperatures will be enabled, particularly the area of drug discovery.

Table 1. Vital statistics of I24, VMXi and KMX with comparable existing and upcoming beamlines at other sources. Note that these beamlines are in varying states of design and commissioning so numbers may be estimates, hardware subject to change, and capabilities somewhat aspirational. I24 KMX would also compare to XFEL beamlines in terms of its scientific focus although, due to the nature of the source, it primarily targets slower timescales (microsecond or slower) and cannot provide true ‘zero dose’ structures. For a wide range of targets this is not a limitation, matching well to the turnover time of many enzymes as discussed above. Serial X-ray diffraction MX methods are also complementary to the emerging fields of serial electron diffraction [61, 62] and time resolved cryoEM [63]. Both of these approaches suffer from limitations, such as a restriction on crystal thickness to a few hundred nanometres

2 https://www.maxiv.lu.se/accelerators-beamlines/beamlines/micromax/ 3 https://www.esrf.eu/cms/live/live/en/sites/www/home/UsersAndScience/Experiments/MX/About_our_beamlines/id29.html 4 https://www.bnl.gov/ps/beamlines/beamline.php?r=17-ID-2 5 https://www.embl-hamburg.de/services/mx/P14_EH2/

SAC


14

or the need to cryo-cool. While both will certainly evolve in the coming years, serial X-ray diffraction sits in a sweet spot, with a common theme between all fields being sample delivery: approaches developed for KMX can have a reach well beyond the beamline and SSX.

The evolution of I24 to I24 KMX will also address the current high demand for I24 beamtime and the desire to target the most challenging targets and experiments. With fully incorporated core SSX techniques, automated switching between experimental approaches will become possible, allowing users to make full use of allocated beamtime and a change to more routine experiments for evening, night or weekend shifts.

3.7 Combined impact of project and added value in relation to activities on the Harwell campus and

beyond

I24 KMX is already a collaborative effort: the I24 team works closely with the XFEL Hub to develop methods for serial crystallography and many of these have been incorporated into the proposal. Additionally, the processing of serial data has advanced through collaboration with the DIALS team. Both of these will continue with, for example, the ongoing incorporation of a disc drive for droplet delivery and the DIALS target of improving analysis of serial data (§5.5). As detailed above (§3.3) there is strong synergy between I24 KMX and several of the other proposed Diamond-II beamlines including K04, X4SCM and hard X-ray bioimaging. I24 KMX will also complement ongoing Diamond developments such as the nano/microfocus MX beamline VMXm, and HeXi electron diffraction instrument that will be embedded there, through transfer of sample mounting, delivery and data collection techniques.

Cross-site collaborations already exist with the Central laser facility, evidenced by the current I24 Octopus proposal (currently on hold due to COVID restrictions). This collaboration will continue as use of the PORTO and other laser systems is essential for dynamic crystallography. The aims of I24 correlate highly with those of Structural Biology at the Rosalind Franklin Institute, with both aiming to target sample delivery and structure solution for in situ, real time molecular research.

SAC


15

4. Beamline performance specification and requirements

• The Diamond-II lattice together with CPMU and a DMM will offer high flux (>1014 ph s-1) at both ~12.4 keV and 25 keV plus.

o Flux of monochromatic beam factor 2 increase on I24 o Flux of DMM beam factor of 20 increase on I24

• Variable focus of ~2 - 30 micron (though would anticipate predominately operating at the smaller end of this scale).

o Beamsize in vertical approx. same as nominal I24 o Beamsize in horizontal approx. factor of 2 smaller than I24

• Highly flexible sample environment based around cryo-pin and serial capabilities (fixed target, extruder, drop-on-demand) with ability to accommodate novel approaches.

o Current serial capabilities currently incorporated into endstation on an ad hoc basis. Represents new capability.

• Fast shuttering. o Current open/close time of shutter is tens of milliseconds. Fast shutter or chopper for accessing

fast time points that is fully synchronised with sample delivery represents new capability o Complex shuttering, exploiting for example Hadamard encoding, represents new capability

• Light delivery to the sample position should be built into the initial design o Laser capability is currently provided by the PORTO laser system which is shared between

several beamlines. An additional laser (with more modest specifications) would reduce dependence on scheduling needs of other beamlines

o Light delivery to the sample position will also facilitate routine optical spectroscopy representing new capability.

• Multiple detectors built into initial design. o Parallel, tagged, readout and subsequent processing represents new capability o Spectroscopic data represents new capability.

• Integrating detector. o Essential for sub millisecond exposures. New capability

• State-of-the-art automated sample exchange for both cryogenic and room temperature samples allowing high throughput cryo-MX and remote SSX for less complex experiments.

o Automation of SSX represents new capability

4.1 Additional developments required

For serial experiments two areas are seen as bottlenecks: sample preparation and data processing. Offline sample preparation and characterisation will be developed with the XFEL Hub (§5.4). Serial and multi-crystal data processing will be developed in collaboration with DIALS (§5.5). Sufficient support is also required to allow data acquisition of all data types through a single user interface. The interface must be flexible enough and sufficiently supported to allow new or modified experiments to be incorporated over time.

SAC


16

5. Schematic outline of beamline or project

5.1 Source

I24 KMX will make use of the existing 17.6mm period CPMU. In order to achieve extremely high fluxes both at ~12.4 keV and at high energies (> 20 keV), a minimum gap of 4 mm is required.

5.2 Optics

Optical design will be based on the existing I24 layout but improvements to mirrors will be required to allow the best possible focus to be obtained: bimorphs degrade overtime so replacement or repolishing will be required. To facilitate the delivery of a top-hat profile for larger beam sizes, wave-front modifiers will be used [64]. The new I24 DCM (due to be installed August 2021) includes space for incorporation of a multilayer crystal and this is required for I24 KMX. Faster shuttering, possibly with a chopper or other ultrafast shutter, at the secondary source is also a requirement.

5.3 End-station(s)

Figure 4. (a) current I24 end-station with existing horizontal goniometer removed and replaced with breadboard. (b) same space four forms of sample delivery shown; pin (blue), fixed target (orange), tape drive (yellow) and XES mirrors & detector shown in place (green). Forms of data that could be recorded shown in pink (diffraction), green (XES) and cyan (XRF).

Space around the sample is at a premium at all synchrotron and XFEL beamlines and the ability to empty this space to incorporate new sample delivery modes or detectors is a major asset. Figure 4 shows a concept design for a possible end-station arrangement for I24 KMX. Removal of the existing horizontal goniometer means that plate-based in situ experiments will no longer be possible at I24 and at Diamond-II this capability will be provided efficiently by VMXi. With the high-throughput capabilities of VMXi, this will not impact Diamond’s capacity in this area. The removal of the horizontal goniometer does however free up a large volume of space adjacent to the sample position. It may also be possible to gain space around the sample position by pushing it downstream through modified on-axis viewing system optics.

A highly flexible sample environment is envisaged. This will build on the existing model developed at I24 where pin micro-crystallography and serial approaches make use of the same sample position, and make used of identical core instrumentation such as sample visualisation and automation. Key functionality is

o High precision retractable goniometer equipped with cold nitrogen stream. o Full integration of established ‘core’ serial approaches such as fixed targets, drop-on-demand,

and viscous media extruder.

SAC


17

o Ability to accommodate new developmental approaches. SSX is a rapidly evolving field so the ability to ‘empty’ the sample environment to accommodate novel approaches is extremely valuable and, crucially, does not compromise core functionality if built into the initial design.

Existing instrumentation will be assessed against current-state-of-the-art to establish where improvements can be made within the existing space envelope. Currently, for example, the Smaract sample stages have a maximum velocity of 20 mm s-1, limiting the maximum speed of grid scans and fixed target experiments. The in-house kappa goniometer is underused due to a lack of robustness and software integration and so these will be addressed. Replacement of all sample stages will benefit both pin and serial experiments.

Multiple detectors and parallel (tagged) readout so different data types can be cross-correlated should be built into initial design. Data types include.

o Diffraction (high energy, large area, high frame rate, integrating, forward direction) o XES (Multi-crystal wavelength dispersive X-ray spectrometer in von Hamos geometry); can

double as a standard X-ray fluorescence detector o in situ spectroscopies such as Raman, UV-Vis absorption, fluorescence.

In addition to co-ordinated triggering, a significant demand here is on good meta data being routinely recorded to facilitate subsequent data analysis.

Fast shuttering is required for both conventional and serial experiments. The existing Cedrat piezo shutter has a rise time of 2ms which is sufficient for many-frame rotation experiments and serial experiments when the shutter opens and closes once per experiment. For single-shot and pump probe experiments fast shuttering is required and this may be required to be synchronous with a laser shutter. Possibilities for fast X-ray shutter are

o Chopper for accessing fast time points that is fully synchronised with sample delivery and detectors.

o Sequential synchronised piezo type shutters for shorter opening times. Custom choppers with, for example, a novel arrangement of apertures, will enable complex shuttering. This will facilitate strategies such as Hadamard encoding which offers access to faster time domains through selective merging of multiple time points [65]. It may also be possible to achieve Hadamard encoding by electronic detector gating or use of a detector such as the Timepix.

A strength of Diamond MX is automation and bringing together of sample tracking, data collections and the results of automated processing and analysis through ISPyB. The BART robot developed at I03 has been incorporated into I24 and is fast and reliable. Currently it can only be used for cryo-cooled samples held in the 37-puck capacity Dewar. A key development for I24 KMX will be the development of automation for serial approaches reflecting the preferred user mode of remote access. While it will always be beneficial for experimenters to be at the beamline for complex experiments to work closely with instrumentation and beamline staff, serial experiments are not unlike challenging pin-based experiments where a significant fraction of beamtime can be dedicated to screening crystals or optimising the experiment. Automation through use of, for example, a small additional hotel or ’Dewar’ for fixed target mounts will make beamtime more efficient, increase throughput, and make serial experiments more accessible to new users.

5.4 Sample preparation facilities

Sample preparation and offline characterisation is key for I24 KMX. Generating the thousands of crystals required for a serial experiment requires a step change from the nanolitre vapour diffusion methods developed for MX [66]. Tools for assessing crystals and slurries obtained are also not widespread and form a key part of complementary lab space. To make more efficient use of beamtime, offline spectroscopic tools should also be available to allow the timescales of reactions in crystallo to be experimentally determined and verified ahead of any beamtime.

SAC


18

Key tools are • Slurry characterisation to allow size distributions and concentration of microcrystal slurries to be

measured. • Offline sample preparation in close proximity to the beamline is essential. This should include

crystallisation (available in the RCaH), microcrystal visualisation, crystal slurry concentration/dilution, ligand soaking, anaerobic glove box, a darkroom for light sensitive proteins or photocage. The ability to test sample delivery offline for challenging samples ahead of beamtime will be important.

• Offline laser and spectroscopy capability for sample characterisation, verifying progression of reactions and timescales of reactions in crystals prior to X-ray experiments.

• Protocols for a permanently dark or selective wavelength lighting of the sample environment (including during hutch search) should be incorporated.

It is important to note that almost all of these facilities are not specific to I24 KMX but need overlaps heavily with the XFEL Hub, the rest of the MX group and other proposed beamlines such as X4SCM.

5.5 Computing infrastructure and support

I24 KMX will produce large volumes of data that need to be processed on timescales comparable to the experiment. A large fraction of processing will be automated but users may also wish to process data once on disk or selectively reprocess automatically based on meta-data/data from other detectors. For pin-based data collections, immediate feedback is required for grid scans as these are used to position crystals in the X-ray beam and results are a prerequisite for the next experimental step. For data collections, per-image analysis provides useful information on how well crystals are positioned in the X-ray beam and if the data collection parameters used are appropriate. Integration and scaling of data gives the first indication of if the experiment is successful or it should be repeated. These represent the ‘standard’ automated processing run on the Diamond cluster by all the MX beamlines though increased load might be expected from automated and user-initiated merging of multiple datasets.

For serial data fast per-image analysis is required to show if crystals are present and diffract, and if so, in what numbers. This should keep up with data collection. Data indexing and integration can be slower but should still be performed on comparable time scales with data collection while merging and scaling of reflection intensities for structure solution and the generation of electron density maps represents the final step and can be slower still. For all experiments, even slow automated processing provides input to subsequent experimental data collections however and ‘slower still’ implies results within tens of minutes. For both pin-based and serial data collections a key aspect will be the merging of data from multiple samples over a visit and possibly between visits.

Serial MX data formed from a series of discrete ‘still’ diffraction images are typically of poorer quality than rotation experiments. Due to this reduction in quality and as SSX is a relatively new and emerging area the data collected are seen by the user community as significantly more challenging to process. Key to addressing this will be close collaboration between I24 KMX and DIALS (whose next Wellcome funded grant is entitled ‘DIALS: making serial crystallography data analysis accessible for biomedical researchers ’) pairing beamline developments with software advances from the start.

Due to the large volumes of data collected and the challenging nature of the experiments it is likely that data analysis will extend beyond the current period of 28 days before spinning disc access is removed.

Support above the current level is required from data acquisition to allow collection of rotation and serial data by users through a common interface, the integration of new tools for pin and serial experiments and also novel experiments which will come up as the beamline and sample delivery and triggering techniques evolve.

SAC


19

6. References

1. Doré, A. S., et al., Structure of the Adenosine A2A Receptor in Complex with ZM241385 and the Xanthines XAC and Caffeine. Structure (London, England : 1993), 2011. 19(9): p. 1283-1293.

2. Hollenstein, K., et al., Structure of class B GPCR corticotropin-releasing factor receptor 1. Nature, 2013. 499: p. 438-443.

3. Dore, A.S., et al., Structure of class C GPCR metabotropic glutamate receptor 5 transmembrane domain. Nature, 2014. 511: p. 557-562.

4. Cheng, R.K.Y., et al., Structural insight into allosteric modulation of protease-activated receptor 2. Nature, 2017. 545(7652): p. 112-115.

5. Jazayeri, A., et al., Crystal structure of the GLP-1 receptor bound to a peptide agonist. Nature, 2017. 546(7657): p. 254-258.

6. Robertson, N., et al., Structure of the complement C5a receptor bound to the extra-helical antagonist NDT9513727. Nature, 2018. 553: p. 111.

7. Porta, C., et al., Rational Engineering of Recombinant Picornavirus Capsids to Produce Safe, Protective Vaccine Antigen. PLoS Pathog, 2013. 9(3): p. e1003255.

8. Baragaña, B., et al., Lysyl-tRNA synthetase as a drug target in malaria and cryptosporidiosis. Proceedings of the National Academy of Sciences, 2019. 116(14): p. 7015-7020.

9. Gladkova, C., et al., Mechanism of parkin activation by PINK1. Nature, 2018. 559(7714): p. 410-414.

10. Sigoillot, M., et al., Domain-interface dynamics of CFTR revealed by stabilizing nanobodies. Nature Communications, 2019. 10(1): p. 2636.

11. Dong, H., et al., Structural basis for outer membrane lipopolysaccharide insertion. Nature, 2014. 511(7507): p. 52-56.

12. Axford, D., et al., In-situ macromolecular crystallography using microbeams. Acta Cryst, 2012. D68: p. 592-600.

13. Aishima, J., et al., High-speed crystal detection and characterization using a fast-readout detector. Acta Cryst, 2010. D66(9): p. 1032-1035.

14. Sherrell, D.A., et al., A modular and compact portable mini-endstation for high precision, high speed fixed target serial crystallography at FELs and synchrotron sources Journal of Synchrotron Radiation, 2015. 22: p. 1372-1378.

15. Hino, T., et al., G-protein-coupled receptor inactivation by an allosteric inverse-agonist antibody. Nature, 2012. 482(7384): p. 237-240.

16. Evans, G., et al., Macromolecular Microcrystallography. Crystallography Reviews, 2011. 17(2): p. 105-142.

17. Rappas, M., et al., Comparison of Orexin 1 and Orexin 2 Ligand Binding Modes Using X-ray Crystallography and Computational Analysis. Journal of Medicinal Chemistry, 2020. 63(4): p. 1528-1543.

18. Shintre, C.A., et al., Structures of ABCB10, a human ATP-binding cassette transporter in apo- and nucleotide-bound states. Proceedings of the National Academy of Sciences, 2013. 110(24): p. 9710-9715.

19. Warren, A.J., et al., Exploiting Microbeams for Membrane Protein Structure Determination, in The Next Generation in Membrane Protein Structure Determination, I. Moraes, Editor. 2016, Springer International Publishing: Cham. p. 105-117.

20. Schlichting, I., Serial femtosecond crystallography: the first five years. IUCrJ, 2015. 2(2): p. 246-255.

SAC


20

21. Orville, A.M., Recent results in time resolved serial femtosecond crystallography at XFELs. Current Opinion in Structural Biology, 2020. 65: p. 193-208.

22. Diederichs, K. and M. Wang, Serial Synchrotron X-Ray Crystallography (SSX), in Protein Crystallography: Methods and Protocols, A. Wlodawer, Z. Dauter, and M. Jaskolski, Editors. 2017, Springer New York: New York, NY. p. 239-272.

23. Keedy, D.A., et al., An expanded allosteric network in PTP1B by multitemperature crystallography, fragment screening, and covalent tethering. eLife, 2018. 7: p. e36307.

24. Bar-Even, A., et al., The Moderately Efficient Enzyme: Evolutionary and Physicochemical Trends Shaping Enzyme Parameters. Biochemistry, 2011. 50(21): p. 4402-4410.

25. Makinen, M.W. and A.L. Fink, Reactivity and Cryoenzymology of Enzymes in the Crystalline State. Annual Review of Biophysics and Bioengineering, 1977. 6(1): p. 301-343.

26. Helliwell, J., What is the structural chemistry of the living organism at its temperature and pressure? Acta Crystallographica Section D, 2020. 76(2): p. 87-93.

27. Keedy, D.A., et al., Mapping the conformational landscape of a dynamic enzyme by multitemperature and XFEL crystallography. eLife, 2015. 4: p. e07574.

28. Kern, J., et al., Simultaneous Femtosecond X-ray Spectroscopy and Diffraction of Photosystem II at Room Temperature. Science, 2013. 340(6131): p. 491.

29. Schotte, F., et al., Watching a signaling protein function in real time via 100-ps time-resolved Laue crystallography. Proceedings of the National Academy of Sciences, 2012. 109(47): p. 19256-19261.

30. Tenboer, J., et al., Time-resolved serial crystallography captures high-resolution intermediates of photoactive yellow protein. Science, 2014. 346(6214): p. 1242.

31. Tosha, T., et al., Capturing an initial intermediate during the P450nor enzymatic reaction using time-resolved XFEL crystallography and caged-substrate. Nature Communications, 2017. 8(1): p. 1585.

32. Stagno, J.R., et al., Structures of riboswitch RNA reaction states by mix-and-inject XFEL serial crystallography. Nature, 2016. 541: p. 242.

33. Mehrabi, P., et al., Liquid application method for time-resolved analyses by serial synchrotron crystallography. Nature Methods, 2019. 16(10): p. 979-982.

34. Schmidt, M., Reaction Initiation in Enzyme Crystals by Diffusion of Substrate. Crystals, 2020. 10. 35. Yi, J., et al., Synchrotron X-ray-Induced Photoreduction of Ferric Myoglobin Nitrite Crystals Gives

the Ferrous Derivative with Retention of the O-Bonded Nitrite Ligand. Biochemistry, 2010. 49(29): p. 5969-5971.

36. Rabe, P., et al., Roles of 2-oxoglutarate oxygenases and isopenicillin N synthase in β-lactam biosynthesis. Natural product reports, 2018. 35(8): p. 735-756.

37. Wiedorn, M.O., et al., Megahertz serial crystallography. Nature Communications, 2018. 9(1): p. 4025.

38. Oswald, C., et al., Intracellular allosteric antagonism of the CCR9 receptor. Nature, 2016. 540(7633): p. 462-465.

39. Weinert, T., et al., Serial millisecond crystallography for routine room-temperature structure determination at synchrotrons. Nature Communications, 2017. 8(1): p. 542.

40. Ishchenko, A., et al., Toward G protein-coupled receptor structure-based drug design using X-ray lasers. IUCrJ, 2019. 6(6): p. 1106-1119.

41. Rucktooa, P., et al., Towards high throughput GPCR crystallography: In Meso soaking of Adenosine A2A Receptor crystals. Scientific Reports, 2018. 8(1): p. 41.

42. Weis, W.I. and B.K. Kobilka, The Molecular Basis of G Protein–Coupled Receptor Activation. Annual Review of Biochemistry, 2014. 87(1): p. 897-919.

43. Wootten, D., et al., Mechanisms of signalling and biased agonism in G protein-coupled receptors. Nature Reviews Molecular Cell Biology, 2018. 19(10): p. 638-653.

SAC


21

44. Dror, R.O., et al., Activation mechanism of the <em>β</em><sub>2</sub>-adrenergic receptor. Proceedings of the National Academy of Sciences, 2011. 108(46): p. 18684.

45. Klein Herenbrink, C., et al., The role of kinetic context in apparent biased agonism at GPCRs. Nature Communications, 2016. 7(1): p. 10842.

46. Waldron, K.J. and N.J. Robinson, How do bacterial cells ensure that metalloproteins get the correct metal? Nature Reviews Microbiology, 2009. 7(1): p. 25-35.

47. Holton, J., A beginner's guide to radiation damage. J. Synch. Rad., 2009. 16(2): p. 133-142. 48. Gildea, R.J. and G. Winter, Determination of Patterson group symmetry from sparse multi-crystal

data sets in the presence of an indexing ambiguity. Acta Crystallographica Section D, 2018. 74(5): p. 405-410.

49. Ebrahim, A., et al., Dose-resolved serial synchrotron and XFEL structures of radiation-sensitive metalloproteins. IUCrJ, 2019. 6(4): p. 543-551.

50. de la Mora, E., et al., Radiation damage and dose limits in serial synchrotron crystallography at cryo- and room temperatures. Proceedings of the National Academy of Sciences, 2020. 117(8): p. 4142.

51. Evans, G., D. Axford, and R.L. Owen, The design of macromolecular crystallography diffraction experiments. Acta Cryst, 2011. D67(4): p. 261-270.

52. Owen, R.L., et al., Low-dose fixed-target serial synchrotron crystallography. Acta Crystallographica Section D, 2017. 73(4): p. 373-378.

53. Fuller, F.D., et al., Drop-on-demand sample delivery for studying biocatalysts in action at X-ray free-electron lasers. Nature Methods, 2017. 14: p. 443.

54. Rabe, P., et al., Anaerobic fixed-target serial crystallography. IUCrJ, 2020. 7(5): p. 901-912. 55. Meents, A., et al., Pink-beam serial crystallography. Nature Communications, 2017. 8(1): p. 1281. 56. Davy, B., et al., Reducing sample consumption for serial crystallography using acoustic drop

ejection. Journal of Synchrotron Radiation, 2019. 26(5): p. 1820-1825. 57. Oghbaey, S., et al., Fixed target combined with spectral mapping: approaching 100% hit rates for

serial crystallography. Acta Crystallographica Section D, 2016. 72(8): p. 944-955. 58. Byrne, E.F.X., et al., Structural basis of Smoothened regulation by its extracellular domains. Nature,

2016. 535(7613): p. 517-522. 59. Gupta, S., et al., Recent Advances and Applications in Synchrotron X-Ray Protein Footprinting for

Protein Structure and Dynamics Elucidation. Protein & Peptide Letters, 2016. 23(3): p. 309-322. 60. Sekiguchi, H. and Y.C. Sasaki, Dynamic 3D visualization of active proteinʹs motion using diffracted

X-ray tracking. Japanese Journal of Applied Physics, 2019. 58(12): p. 120501. 61. Bücker, R., et al., Serial protein crystallography in an electron microscope. Nature

Communications, 2020. 11(1): p. 996. 62. Wolff, A.M., et al., Comparing serial X-ray crystallography and microcrystal electron diffraction

(MicroED) as methods for routine structure determination from small macromolecular crystals. IUCrJ, 2020. 7(2): p. 306-323.

63. Kaledhonkar, S., et al., Late steps in bacterial translation initiation visualized using time-resolved cryo-EM. Nature, 2019. 570(7761): p. 400-404.

64. Dhamgaye, V., et al., Correction of the X-ray wavefront from compound refractive lenses using 3D printed refractive structures. Journal of Synchrotron Radiation, 2020. 27(6): p. 1518-1527.

65. Yorke, B.A., et al., Time-resolved crystallography using the Hadamard transform. Nat Meth, 2014. 11(11): p. 1131-1134.

66. Beale, J.H., et al., Successful sample preparation for serial crystallography experiments. Journal of Applied Crystallography, 2019. 52(6): p. 1385-1396.

SAC


22

7. Expressions of interest & support from the community

103 letters of support for the I24 KMX proposal were received including 30 on behalf submitted of institutions and companies. Statements are included as appendix A in this document and a tabulated summary of statements of support is included below.

A common theme from the letters is the dependence of supporter’s research on I24 and the direct impact a beamline can have on what can be achieved. Examples of this given include the ability to collect data of sufficient quality for structure solution from small, challenging crystals that would otherwise be discarded or require significant laboratory time and resource to optimise. 62% of respondents highlighted the need for a beamline able to tackle small, difficult samples with many further highlighting it’s applicability and role in membrane protein structure determination (30%) and to drug discovery (26%) and metalloproteins (17%). Several respondents highlighted I24 as the European beamline that consistently provides the best diffraction data from challenging samples: the aim of the I24 KMX upgrade is to ensure this remains true in years to come at Diamond-II.

Support for new and developing techniques is also evident with over half (59%) highlighting dynamic crystallography as a current or near future focus. This was also reflected by the letters of support received for the Diamond-II Wellcome Trust application with a significant number supporting serial crystallography opportunities and several making explicit statements in support of I24 and/or dynamics relevant to their research. These point to a community that is interested in dynamics in several areas of research from catalysis through to drug discovery and they foresee the proposed I24 KMX upgrade making these experiments accessible. The development of the beamline from its current state through to I24 KMX at Diamond II in conjunction with groups such as the XFEL Hub will only continue to raise the profile of time resolved serial crystallography, facilitating new science and a user community around it.

Synchrotron diffraction data cannot provide full insight into the structure and function of biomolecules. Reflecting this complementary, i.e. spectroscopic, forms of data collection have been recognised as important in many letters of support (highlighted in >20% of letters) but I24 KMX also sits alongside other techniques such as SAXS, microelectron diffraction and cryo electron microscopy. 24% of letters highlight the importance of dynamic serial approaches at a synchrotron complementing those used at XFEL sources. An important role of I24 KMX raised in the letters is that the beamline can make the most challenging experiments accessible as I24 has done in the past.

The user community clearly see I24 KMX as a positive evolution of a beamline already essential to their research and are incredibly supportive of the proposal. Evolution is the key word here. I24 has evolved to encompass new techniques not envisaged when the beamline was first designed. To continue to be world leading it must continue to do so, allowing users to perform innovative experiments and driving future science. This is what a novel instrument is about.

SAC


23

Statements of support summary

Total number of submissions: 103

Engagement webinar summary

Date of webinar: Tuesday 3rd November 2020 Number of attendees: 99

Attendee location Percentage of attendees

UK 71% International 29%

Key for statements

in Appendix A

Respondent's primary field

of research

Percentage of

respondents

Life Sciences & Biotech 95.1% Chemistry 4.9%

Respondent location Percentage of respondents

UK 79.4% International 20.6%

Type of organisation supporting Percentage of respondents

Academic 90.2% Government 2.9% Industry 5.9% Other 1.0%

Diamond user status Percentage of respondents

Not currently a user at Diamond 7.8% Currently a user at Diamond 92.2%

11: i24 kmx upgrade

Documents