Molecular modelling and structure-based inhibitor design for the angiotensin converting enzyme 2 and its homologues

The work presented in this thesis focuses on the M2 family of metalloproteases, in particular A<;E2 and·ACE3, two recently discovered homologues of the Angiotensin Converting Enzyme (ACE). ACE is an important, well validated . cardiovascular disease target and the key enzyme in the Renin-Angiotensin System. ACE2 increases the complexity of this system by counter regulating ACE and is currently under intense biological investigation. Another homologue, ACE3, was first identified in mouse. Here, we further investigate its . . biological role. A range of computational methods are used to investigate these enzymes from a structural and functional perspective. This involves close collaboration with experimental researchers throughout these studies. .'\. We investigate ACE3 in several organisms and compare it to other ACE family members. In addition, homology modelling is undertaken based on the AC.E crystal structure to analyse the structural conservation of the active site. The mutation of Glu386 to GIn in the catalytic HEXXH zinc binding motif, as well as a number of other active site mutations are found to be critical i~ disrupting catalytic activity in ACE3. Comparative genomics reveals that the ACE3 gene is present in several mammalian species (Le. rat, mouse, dog and cow) and located downstream in close proximity of the ACE gene. The inferred amino acid sequences are st~ongly conserved amongst all putative ACE3 orthologues in te~s of sequence identity (up to 81%) and in terms of characteristic features such as the HQMGH motif and two transmembrane domains. In human, it is found to be a pseudogene due to deletions and insertions in the gene. Reconstruction of the ancestral ACE3 amino acid sequence shows a Close relationship with ACE and the other ACE3 sequences. The characteristic ACE3 HQMGH motif is further identified in ten other mammalian species but is lacking in non marru:nalian vertebrates and invertebrates which have either ACE or ACE2. ACE2 is then investigated by applying a range of computational Structure-based Drug Design methods to develop potential new inhibitors. A lead modification study of a potent'; co-crystallised ACE2 inhibitor (MLN-4760) guided by chemical synthesis is undertaken'. The aim is to synthesise and test the novel compounds in a bioassay and to relate the measured inhibitory activity back to computational predictions. To this end, the Rl'-group is replaced by a small . , library of twelve fragments on the peptidomimetic scaffold of MLN-4760. The new compounds are evaluated and ranked by docking with GOLD and the lead optimisation tool LeadOpt. GOLD is able to redock MLN-4760 with a low RMSD when compared to the crystal structure and several scoring functions are evaluated for their pose selection abilities. The ranking of the docked . 'Y compounds with GOLD contrasts with those of LeadOpt and issues arising from this comparison are further discussed. In addition, other aspects are investigated by docking such as the esterification of the essential zinc binding group and stereoisomerism. Ultimately, only three compounds from the library could be synthesised and tested experimentally, the best compound being a '159 J1M inhibitor ofACE2. Following on from the above study an alternative approach, pharmacophore modelling and virtual screening, is applied to identify novel ACE2 inhibitors. A set of pharmacophore models is generated by exploiting structural information of the MLN-4760 inhibitor-bound ACE2 structure. A similar pharmacophore model is created for three, inhibitor-bound, ACE structures. Protein-ligand interactions are studied in ACE2 and compared to ACE. The models are then validated by screening an internal ACE inhibitor data set as well as the Derwent World Drug Index. A large library of about 2.5 million unique compounds from 26 commercial databases is screened with Catalyst. The most promising compounds are experimentally tested, resulting in six novel, micromolar ACE2 inhibitors, the best compound being a 60 J1M inhibitor. Retrospective docking with GOLD reveals binding modes similar to those predicted by the pharrnacophore modelling and this is used to further analyse and compare the newly discovered inhibitors. 'Finally, selectivity is evaluated by docking into ACE and the results are compared to experimental findings.EThOS - Electronic Theses Online ServiceGBUnited Kingdo