Rationalising glycomics with GPU-accelerated equilibrium simulations: a novel route to 3D-structure biological function and molecular design

Carbohydrates, amino acids, nucleic acids and lipids constitute the fundamental building blocks of life. Although carbohydrates were the first of these to be extracted from living organisms and characterised, our understanding of protein and nucleic acid biological function is far advanced, which is, in part, due to the excellent progress made in the 1950s on probing the structure of crystallised biological matter with x-rays. By investigating the geometries adopted by amino acids and nucleic acids, Watson, Crick, Pauling, Phillips and others theorised that biological function is a manifestation of microscopic shape, which drove a revolution in Biochemistry: the realisation that proteins comprise a string of amino acids folded into specific functional shapes and work as micro-machines. It also led to discovery of the DNA-double helix, which carries our genetic information and is the progenitor of Molecular Biology and Genetics. Unfortunately, a similar revolution has not occurred in our understanding and harnessing of carbohydrates, for example, although the heparin carbohydrate has been routinely used in surgical procedures for almost a century, the mechanism of action has only recently begun to be understood. The problem is that carbohydrates do not crystallise readily and other techniques to investigate their microscopic shape have not been developed. Consequently, the relationship between carbohydrate composition and function is tenuous. Carbohydrates, therefore, represent a major unexplored frontier in Biochemistry and due to their fundamental industrial importance (e.g., food, paper, wood, pharmaceuticals, biomaterials) any breakthrough in understanding this relationship would certainly lead to another revolution in biotechnology. This research is driving toward providing this important link between carbohydrate composition and function by developing new techniques for investigating their microscopic shape. Rather than using x-rays, we are using precise and extensive computer simulations, advanced methods for refining pure carbohydrates and a molecular microscope based on magnetic resonance (similar to MRI-scanners found in hospitals) to achieve this goal. To date our research has focused on unravelling some of the mysteries surrounding the large polymeric carbohydrate molecules called glycosaminoglycans (GAGs), which fill the space between cells, bonding them together, conferring strength to organs, joints and skin, while allowing our bodies to grow and change. Microscopically, carbohydrate molecules are composed of chemical rings, joined together to form polymeric chains. We found that to accurately describe the shape of these molecules we need to understand both what is happening around the connection joints between these rings and also how the rings flex dynamically. Computer simulations of carbohydrates, which take into account thousands of interactions with solvent water, have so far managed to investigate the joints, but up until now limitations in computer hardware has not permitted an investigation into ring flexing because it happens on timescales that are hundreds of times longer. We overcame this problem using graphics processors (hardware used for computer gaming) to dramatically speed-up simulations and provide insight into the GAG heparan sulphate, an anticoagulant similar to heparin that lines blood vessels. It has revealed that ring flexing is central to carbohydrate function and that absence of ring flexing leads to regions that are stiffened, which consequently interact with other molecules in the body. We now plan to research and test this hypothesis further and determine whether this ring flexing behaviour is central to the function of other GAGs, such as those found in cartilage and skin, whether it is important in other human and plant carbohydrates and whether it can be used for understanding how to design novel pharmaceuticals and biomaterials based on carbohydrates.

Carbohydrate puckering occurs on microsecond timescales, which was previously inaccessible to accurate computer simulations. Recently, we used graphics processing units (GPUs) to overcome this limitation and to hypothesise that puckering depends on both monosaccharide type and chemical environment. We showed that this major determiner of 3D-shape resolves an otherwise unexplainable observation: that the protein-binding domains in the glycosaminoglycan (GAG) heparan sulphate are more rigid than the rest of the polymer. We now aim to perform essential research into this fundamental hypothesis, which will transform our understanding of basic biology. Specifically we will: Confirm the universal importance of puckering in the GAG glycome by performing GPU-accelerated microsecond simulations on chondroitin and dermatan sulphate. We will quantify monosaccharide puckering, create free energy landscapes for key sulphation patterns, and make polymeric predictions via mesoscale simulations; predictions will be tested against NMR measurements on pure oligosaccharides and scattering experiments on polymers. Investigate puckering in the wider mammalian and plant glycomes using microsecond simulations of high-mannose branched oligosaccharides (from protein post-translational modifications), fucosylated blood-group epitopes, cell-adhesion carbohydrates (Lewis-x) and amylopectin; novel observations will be validated using ultra-high-field NMR experiments. Rationalise carbohydrate binding and mimetic design using microsecond simulations of bioactive antithrombin-III heparan sulphate oligosaccharides and therapeutics (as a model system). After validation by NMR, a novel puckering-based quantitative structure-activity relationship will be developed to inform biology and drive virtual screening for new bioactive molecules. Bioassay testing will be used to validate that our approach can rationalise discovery of novel carbohydrate-based foods, drugs and biotechnological devices.

The four main impacts (in order) resulting from our research will be: (1) world-leading peer-reviewed papers, (2) patentable intellectual property and know-how that can be commercialised, and (3) knowledge-transfer to industry and (4) engagement with the public on the importance of carbohydrates to health and disease; in other words, realising pioneering bioscience, pursuing the development of novel biotechnologies, driving commerce, and disseminating the outcomes to a wider audience. The PI has a track-record of delivering high-quality publications, has filed several patents and has taken research from the lab-bench to commercial reality via the spin-out company Conformetrix (www.conformetrix.com), which was based on research conducted under BBSRC research and translational funding (now employing five PhD-level scientists in the North-West of England). It is also the applicants' view that commercial, transferable outputs from University research are as important as basic scientific publications and, while keeping an eye on producing quality publications, they will remain aware of commercial opportunities and liaise with the technology transfer office when relevant. Clearly, the general science community will benefit from this basic research into carbohydrates and it has the potential to impact many research fields from plant and food sciences to the pharmaceutical sciences. Furthermore, the basic principle of puckering equilibria, once integrated into the theory of how carbohydrates function at the molecular level is so fundamental that it has the potential to change the way undergraduate textbooks report carbohydrates and the way we view carbohydrates and their importance in general science and medicine. The proposed non-peer reviewed dissemination of the research outcomes (website, popular science publication and press releases) will engage and benefit the public by providing a non-specialist window into the transformational capacity of UK-bioscience. The pharmaceutical and biotechnology sectors will also benefit from our research, and in particular industrial researchers (with a remit to find novel anti-thrombotic and anti-inflammatory drugs, vaccines and wound healing agents) will benefit from the knock-on improvement in carbohydrate medicinal chemistry methods, such as quantitative structure-activity relationships (QSAR). Furthermore, the structural information that we will provide can be directly used to inform development of chemical mimetics. Our unique understanding of the pericellullar environment will have a profound impact on those interested in the physiological and toxicological effects of pharmaceuticals and drug delivery to cells (particular cancerous and infected tissue). It is also anticipated that the results from our research could impact those trying the grow cells in vitro (e.g., in fermentors), and in particular, cells that are not surrounded by an intact matrix environment, are therefore in an alien environment and subjected to cellular stress. Engagement with industry, for example in the proposed industrial partnership visits and postdoctoral research assisstant (co-investigator) secondment, will forge much needed new relationships between academia and industry. This will open up new communication channels and lay the foundations for future collaborations and innovations with our group and other BBSRC researchers.