sparkV10 (the new name for FieldStere) is a fast and powerful field based software tool which uses Cresset’s innovative XED technology, together with a database of molecular fragments, to help guide drug discovery projects. In this example sparkV10 is used for chemotype morphing to produce novel, non-intuitive chemical ideas for new BACE inhibitors.
sparkV10 is used for scaffold hopping and making bioisosteric replacement of specified parts of molecules. The latter may be done in order to obtain new intellectual property (IP), to protect existing IP or for gaining improvements in the overall physiochemical properties of a lead molecule, whilst retaining biological activity. sparkV10 can also be used to find a novel chemical linkage between two separate fragments.
Background
BACE is one of the proteases responsible for processing beta amyloid precursor protein (APP) and is thought to be a part of the process that leads to the deposition of amyloid plaques that occur in the brains of Alzheimer’s disease sufferers. BACE is an important aspartyl protease which has received a reasonable amount of attention from the pharmaceutical industry and, as a result, a number of potent small molecule inhibitors are now known. The first group of these inhibitors were developed from analysis of the substrates and were peptidomimetic. Later examples were derived from screening and were thus inherently more drug-like.
Structures
For a number of these inhibitors, for example the hydroxyethylamine inhibitor (figure 1, structure 1) and the cyclic guanidine based inhibitor (figure 1, structure 2), the ligand protein complexes have been obtained in high resolution using X-ray crystallography (PDB 2IQG and 3L59, respectively).
Figure 1: Example BACE inhibitors.
From these complexes the protein appears to be relatively flexible. The catalytic aspartates and adjacent active site residues can be engaged in subtly different ways by the different ligands (figure 2).
Figure 2: Compound 1 (PDB: 2IQG, left) and 2 (PDB: 3L59, right).
To demonstrate how sparkV10 can be used to generate novel ideas, a virtual experiment was performed that effectively morphs both inhibitors and produces example molecules which can exploit the combined patterns of interaction from both inhibitors.
Process
For this experiment compound (1) was used as the chemical seed for the process and compound (2) was used as the reference against which the resultant molecules would be scored. Fragment replacements made in molecule (1) that best matched the fields of compound (2) would thus be ranked highest in the results list.
Figure 3 shows the fragment replacement for compound (1).
Results
Selected examples from the results of the experiment are shown in Figure 4.
Figure 4: The results of the sparkV10 experiment, showing the drug-like molecules produced from the fragment replacement.
Most of the results are more compact relative to compound (1), look reasonably drug-like and are novel. They contain H-bond donor features present in the original seed compound (1) but also have an additional H-bond acceptor feature which is only present in the reference compound (2). Result 18 was particularly interesting and is shown in Figure 5 in the context of the protein.
Figure 5: Result 18 in BACE (PDB: 2IQG).
The methine proton, alpha to the imidazole, mimics the guanidine NH in the ‘field space’.
Conclusions
This chemotype morphing example demonstrates the power of fields and the flexibility of sparkV10 in rapidly producing novel, non-intuitive chemical ideas which are relevant to drug discovery.
