Dr Mark Dufton
TG605 THOMAS GRAHAM
Tel : +44 (0)141 548 2440
- Bioinformatics. Software for topological analysis and interpretation, leading to prediction of topomorphic sites in biological macromolecules. Application to discovery of new generation of topomorphic drugs and understanding the process of natural evolution for application to protein engineering. Collaborative partners: Dept of Computer & Information Sciences (Dr John Wilson); Dept of Pharmacy (Prof Simon Mackay).
- Snake venom chemistry, especially the supportive role of venom enzymes in enhancing the efficacy of the key toxins. Collaborative partner: Dr John Parkinson (Pure & Applied Chemistry). Venom action, as a guide for the development of protein-based drugs and strategies for enhancing the speed and efficacy of drug delivery.
- Biomolecular structures and mechanisms, and their evolutionary history.
- Role of biomolecule chain fold topology in determining directionality of structural and functional evolution.
- Drug action through site specific perturbation of chain fold topologies in proteins and nucleic acids.
Industrial drug discovery is presently a very expensive, time consuming and risky business. A great deal of time and money (£millions to £billions) could be saved by the pharma industry with the aid of an automated “first step” that accurately identifies candidate druggable sites in proteins and nucleic acids associated with disease states. Furthermore, if these candidate sites offer a better and safer mode of action for drugs compared to conventionally designed drugs, then the risk of a prolonged investment resulting in an unusable drug is also reduced.
The bioinformatic research we have undertaken over the last 15 years has resulted in the SID technology for the systematic analysis of protein and nucleic acid 3D structures. It establishes the potential for adjustment of these macromolecules by site-specific mutation or small/large molecule binding. The SID technology is a fully tested, versatile, fast and easy to use software suite that is applicable to the 70,000 + biomolecule 3D structures held in the international databases. Presently unpublished uncirculated, and owned entirely by the university, it is being retained in house to protect its substantial revenue-earning potential. Industrial customers can obtain results for particular drug targets by contractual arrangement with the university (RKES).
We have also used the SID technology to create our own results database for all the targets (70,000 +) that can be analysed. This raises the possibility of selling the complete database or “niche” databases relating to one type of drug target.
Snake Venom Research
A century from now, many of the drugs presently in common use will be regarded in the same light as we now view the various remedies available in the Victorian era. They will be seen as unsubtle, too toxic, based on misunderstandings, insufficiently targeted in their action and surprisingly ineffective given the high doses that need to be taken. On the other hand, snake venoms show just what is achievable given an appropriate route of entry, high information content molecules (proteins) and a complimentary selection of enzymes and other factors to allow fast and selective target acquisition. It is inevitable that a full understanding of venom molecule structures and strategies will prompt revolutions in drug design and switch attention away from small molecule “organic” drugs towards multifunctional proteins that can act at very small doses with exquisite speed and selectivity.
The enzyme system currently under investigation is associated with lowering bodily defence mechanisms. When accompanying conventional drugs, there is a clear possibility that these enzymes could allow the drugs to act faster from a smaller dose level. For the drug industry, where new drugs often fail to reach market because the necessary dose levels are too toxic, the results of this research have a clear value.
Mark Dufton is a graduate of the former Department of Chemistry at Essex University.
He completed a BA in Biological Chemistry, followed by a PhD researching snake toxin chemistry and then took up temporary appointments at Essex University as a lecturer in the Department of Chemistry.
After research-only posts at the Universities of Marseilles and Essex concerned with animal venom action, he joined the Organic Section of the Department of Pure and Applied Chemistry at Strathclyde University where he is now Senior Lecturer/Academic Selector.
Mark’s central research interests have always focussed on the theoretical analysis of protein structural data, and this type of research was pioneered long before the term “bioinformatics” came into common use and the subject became fashionable.
He has published highly cited works on animal venom structure/function and the evolutionary behaviour of protein families, and devised original methods for protein data analysis.
His best known publications (as single author) concern an explanation for the enigmatic organisation of the genetic code in terms of minimising the energy cost of proteins, a proposal for the mechanism of serine proteinase action that involves domain movement, and the prediction (recently supported by new fossil evidence), that early mammals were venomous.
Other research output has concerned the discovery and characterisation of an integrated suite of neuropeptide-processing enzymes in cobra venom and the action of venom toxins on membranes.
Having devised new bioinformatic tools for adding value to the rapidly expanding databases of protein sequence and 3D structure, it became clear that the methods had significant potential in the challenging area of drug discovery.
With this aim in mind, the methods were refined technically and automated for application (in collaboration with the Strathclyde Institute for Pharmacy and Biomedical Sciences and the Department of Computer and Information Science) to a broad range of drug discovery programmes taking place at Strathclyde.
Of particular current interest and success is a programme (user interface pictured below) that predicts the locations of possible allosteric control sites in any protein 3D structure.
This is playing a key role as the hitherto elusive “first step” for the automated discovery of protein-targeting drugs that are more selective and subtle in their action than those discovered by traditional means (i.e. drugs that are not directly competitive with the substrates acted upon by target enzymes, or completely abolishing of their action).
Access to these unique bioinformatic tools for commercial interests is available via the Research and Innovation Office at Strathclyde.