Turning symmetric protein scaffolds into robust enablers of structural biology

Images that show relevant biological molecules, their fine details and their interactions, help biochemical and pharmaceutical researchers in their experimental programmes. Structural knowledge of biological molecules has historically transformed our view of what they do and how they do it. Today, such knowledge is increasingly being used to drive the discovery of drugs and vaccines to combat diseases that range from bacterial infection to cancer.

Two experimental techniques are key to producing these images. Macromolecular X-ray crystallography (MX) is the most mature method for determining structures of the molecules (usually proteins) that govern the operation of biological systems. However, as recognized by the award of the 2017 Nobel Prize in Chemistry, electron microscopy (EM) has recently come of age as a viable alternative for atomic resolution visualization of large macromolecules.

Despite significant technical advances in recent years for both MX and EM, these two techniques do not yet provide a complete answer to observing all biological molecules in a timely fashion. EM is typically only used to study relatively large biological molecules and complexes. X-ray crystallography is less demanding in respect of size, but, requires the protein of interest to form crystals - billions of identical copies arrayed in a near-faultless three-dimensional grid. This can only be accomplished by trial and error, by testing a protein in thousands of different solutions to see if it can be persuaded to crystallise. Alas, this fails more often than not, and for many of the most therapeutically interesting, it is difficult if not impossible.

Our approach to addressing this issue is to prepare nano-sized 3D scaffolds ("crysalins") on to which a protein of choice can be attached in a crystalline array. Whilst we have successfully produced the nano-scaffolds and attached examplar targets onto them, we remain one step away from this becoming a universal platform for structure determination.

Currently the target proteins are attached only loosely to the scaffold and, because X-rays show you an average picture of the repeating unit of the array, any looseness "smears out" the image of the target protein. This programme of work will firm up the connection to ensure that each target molecule is attached to the scaffold in exactly the same position and orientation.

An added benefit of developing rigid target-scaffold connections is that the building blocks that make up our scaffolds are large and symmetric and therefore perfectly suited to imaging on an electron microscope. Even small targets (of a size usually inaccessible to EM) that are rigidly and uniformly attached to such building blocks have been shown to be straightforwardly imaged by EM. We are therefore producing a building block that can either be used in isolation to facilitate electron microscopy or assembled into a lattice to enable crystallography.

An area in which MX still reigns supreme is high-throughput structure determination, especially where one is looking at a single protein target binding to a number of different small chemicals. If suitable crystals of a protein are available, they can be soaked in solutions of these chemicals and imaged. Indeed, one can screen a large library of chemicals and trawl through the structural images to find candidates that might be refined as therapeutic drugs. The whole process of refinement can be done rationally, seeing how variants of the original chemical bind and how their shape and properties might be changed in order to improve them to enhance their effectiveness. This process is known as structure-based ligand design, SBLD. We will demonstrate the potential of crysalin technology for SBLD in conjuction with the "XChem" chemical screening platform and apply it to proteins that are potential therapeutic targets under study in the laboratories of the PIs.

We have developed a system of self-assembling 1D, 2D, and 3D nano-scaffolds that are built using high-symmetry protein oligomers; and we have shown that the 2D and 3D scaffolds provide crystal lattices that are capable of hosting target macromolecules. This grant aims to develop a reliable method to incorporate target molecules in an ordered fashion, which will be transformative in three areas of structural biology. Firstly, any target, even if not intrinsically crystallisable, can be effectively crystallized for high-resolution X-ray structure determination. Secondly, proteins can be tethered to high-symmetry scaffolds particles to provide a vehicle for robust structure determination by cryoEM, incluging targets far smaller than 100 kDa. Thirdly, it provides a reliable route to crystal systems sufficiently robust for the high-throughput crystallography required in structure-based drug design, including crystal-based fragment screening.

The research will focus on conferring ordered attachment by engineered introduction of multiple interactions between the surfaces of lattice and target; work to date had naively fused target and lattice by flexible linkers. We will incorporate both general binding motifs and specific binder scaffolds into the lattice, and show that they are both ordered and induce order to bound targets. We will explore many (hundreds) adaptor possibilities using new miniaturised, parallelised techniques for high-throughput cloning, expression, purification, crystallization, diffraction and structure solution. Effective adaptor combinations will be further evaluated in CryoEM; and suitability for drug discovery will be shown by XChem fragment screening of a clinically relevant target.

The work will be open access and patent free but will also aim to achieve widespread use through early engagement with vendors interested in achieving a commercial product.

This research has the potential to transform the effectiveness (reliability) and efficiency (throughput) of structural biology. As such it could be transformative for basic research into mechanistic cell biology, and have equal impact on industries (such as biotechnology, the pharmaceutical industry and agrichemicals) that rely heavily on structural biology. Since these sectors underpin the discovery of novel vaccines and therapies, there is scope for substantial impact on national health. Because they also contribute substantially to Gross Value Added in the British economy, impact in these sectors can, further, be anticipated to have broader societal impact through wealth generation.

The ultimate goal of mechanistic cell biology is to derive an atomic resolution model for the action and interactions of the biological macromolecules that sustain life. X-ray crystallography, and cryoEM can populate such a model, but their impact remains limited by challenges of sample preparation that leave many molecules outside their scope of applicability. The research in this program will provide a set of recipes that can be applied to most macromolecular targets to allow them to be imaged at atomic resolution, either through crysalin-assisted X-ray crystallography or crysalin-assisted cryoEM. While there will remain certain targets that cannot readily be imaged (e.g. short-lived macromolecular complexes), a large number of biologically important targets will be brought within scope, to the benefit of the broadest imaginable subset of biomedical researchers

As a result of readier access to atomic structures, a further impact will be in respect of genomic annotation. From the many sequenced genomes, there remain a vast array of proteins for which functional annotation based on sequence homology has not been possible. For some of these proteins, knowledge of their 3D structure may reveal distant evolutionary relationships that, in turn, suggest functional roles: 3D structure and function can be conserved at levels of sequence conservation that are scarcely detectable. As such, this research could provide further benefit by helping to realise the promise of the genomic revolution.

Maybe the key impact of this research, however, will be in providing platforms for structure determination of chosen macromolecules that will open them up as targets for structure-based drug discovery and structure-based vaccine design. For macromolecular targets for which suitably robust experimental systems can be put in place, structural biology can play a central role at all stages of drug discovery - from confirming the tractability of a target, through finding chemical hit matter, to supporting iterative medicinal chemistry. The best possible outcome of this research will be to make many more targets amenable to providing such robust experimental systems. If this potential is realised, targets for drug discovery will, in future, be chosen for reasons of clear disease linkage, rather than expeditiously in consideration of experimental accessibility. This change could lead to the testing of more clinical hypotheses through clinical trials, and to a lower failure rate in clinical trials, in turn leading to improved patient outcomes with lower cost to the nation. Where this opportunity is taken up by UK industries, the result could be additional tax revenue to combine virtuously with lower costs of health care.

We have chosen one exemplar target that is of particular ODA relevance, namely cyclin T. This target might enable the development of curative anti-HIV therapies that could address challenges for developing countries which are disproportionately disadvantaged by the financial costs of endemic AIDS. We further anticipate that the platforms we are generating could help to make techniques of structural biology affordable in developing countries, hence serving an additional ODA goal.