Chillin' in Flatland - Development of HCIE (Heterocycle Isostere Explorer)

点击翻译按钮获取中文摘要

The 2009 publication "Escape from flatland" shows a correlation between a success of potential leads in clinical trials and the presence of sp3 hybridised carbons, suggesting that more 3D molecules have a better chance of becoming drugs, possibly due to their higher solubility and better fit with the 3D protein cavities. However, since then it has been shown that flat core scaffolds are equally good starting points for accessing 3D shape space as 3D core scaffolds. This suggest that the 3D shape desired for increased success in drug discovery could be achieved through a choice of substitutes and functional groups, while still keeping the aromatic, flat scaffold cores. That is a desirable prospect, as the chemistry required to introduce sp3 hybridised carbons is more challenging and time-consuming. Moreover, the chemistry of small, aromatic heterocycles is still largely unexplored as shown by the creation of VEHICLe - virtual exploratory heterocyclic library. VEHICLe is a database of theoretically possible heteroaromatic ring systems, consisting of around 25000 molecules. At the time of its creation, in 2008, only 1701 of those systems (around 7%) have been synthesized and reported in literature. However, the authors' analysis has shown that more than 3000 additional molecules in the database are predicted to be synthetically tractable. This shows a potential for new chemical space exploration and introduction of new chemical structures in design of chemical probes and drugs with unique molecular architectures, while using the already known and simpler chemistry of heterocycles. The aim of this project would be to develop a computational tool, the HeteroCycle Isostere Explorer (HCIE) to discover new heterocyclic cores for compound optimisation in chemical probe development and drug discovery. The main function of HCIE would be to search the VEHICLe database for heterocycle bioisosteres using shape, electrostatic potential and vector similarity. This method of suggesting isosteres distinguishes HCIE from other isostere search tools, like SwissBioisostere, in that it doesn't require any previous knowledge of bioisosteric replacements to make its predictions and thus will to lead to novel isostere suggestions. Additionally, the VEHICLe database would be analysed to search for areas of "flatland" that are currently underexplored and could present novel physicochemical property spaces. Lastly, software to analyse the synthetic tractability and predict synthesis routes for heterocycles in VEHICLe would be developed and incorporated into HCIE, helping the user to incorporate novel heterocycles into their target molecules and de novo design software. Those predictions would then be validated by attempting to synthesize some of the not previously synthesized heterocycles in VEHICLe. This project falls within the following EPSRC research areas: Computational and theoretical chemistry, Chemical biology and biological chemistry and Synthetic organic chemistry. It would be supervised by Professors Paul Brennan and Fernanda Duarte and conducted in collaboration with Exscientia.