Scaffold Hopping

Scaffold Hopping
How to Replace an Undesired Molecular Scaffold

What to do if the most important part of a compound is the one limiting any progress? In its very true definition, the pharmacophore is one crucial part of the molecule that is responsible for its biological activity. But sometimes it is the part with the undesired properties: toxicity, promiscuity, unfavorable physicochemistry, or, even worse, so good that it is patented.
This article will tackle the topic of how to find good replacements for core structures with computational methods.

Almost the Same – But Different

Scaffold hopping is the identification of isofunctional molecular structures with chemically completely different core structures. It is a subset of bioisosteric replacement where the core motif (pharmacophore) is replaced. Important interaction potentials of the molecule are usually maintained.
The beauty of scaffold hopping is to find something that has similar properties as a lead compound but contains a different core motif. An iconic, historical example: Sildenafil (top) and vardenafil (bottom) are depicted in the figure. Vardenafil (developed by Bayer, approved in 2003) contains a different arragement of nitrogens within the ringsystem than sildenafil (marketed by Pfizer, approved in 1998) because this area was not protected by a patent.

Example 1: Virtual Screening

But how to discover replacements for a scaffold?
Virtual screening is a method to predict potential binders for a target of interest. A set of molecules is docked and scored allowing comparison of their predicted binding modes and interaction qualities. After selection of promising candidates, the compounds are aquired (through synthesis or external sources) and pharmacologically assessed. The greatest advantage of this method is that it can discover completely chemically unrelated candidates because it does not directly rely on structural information from a known binder.
Using pharmacophore contraints (like hydrogen bond acceptors/donors, lipophilic or charged groups, aromatic ring systems, etc.) can increase the success rate of the virtual screening as the generated poses will feature important interactions with the target.
Virtual screening can be performed with SeeSAR.

Example 2: Topological Replacement

A part of a molecule can also be replaced for something that can keep the geometrical orientation of the decorations attached to the core. Basically, this method searches for something that has similar 3D coordination of the connection points as a reasonable topological exchange motif for a structure within a ligand.
The ReCore functionality of SeeSAR's Inspirator Mode can be used for this approach. It screens libraries containing fragments of molecules (e.g. ZINC database, PDB) as 3D coordinates and ranks those according to their connecting vector similarity.
Again, pharmacophore constraints can be applied as well to filter for results that feature key interactions with the target.

Example 3: Fuzzy Pharmacophores

This method searches for distant relatives to a compound; Molecules, that share similar pharmacophore properties and functionalities but with a little bit of breathing space.
FTrees (short for Feature Trees) analyzes the overall topology and fuzzy pharmacophore properties of the molecule and translates the data into so-called molecular descriptors. With these descriptors, FTrees swiftly navigates through compound libraries and Chemical Spaces looking for molecules with similar features. FTrees is embedded in our Chemical Space navigation platform infiniSee.
Structure-based methods (e.g. docking) can be used as additional enrichment methods for results.

Example 4: Shape Similarity

Sometimes no binding mode information on a ligand is available. This requires drug hunters to use to ligand-based drug discovery (LBDD) methods.
One of those methods is to screen for compounds that share a similar shape and orientation of functionalities as the query molecule. The Similarity Scanner of SeeSAR can be used to generate molecule superposition based on shape and pharmacophore features. Good results are likely to share the important functionalities in close proximity to the original. Pharmacophore constraints can further be used to fine-tune the results.

Last but not least it should be mentioned that some approach work better than others depending on the individual scenario. Those examples by no means rule out each other but can synergistically be used together.

We hope this introduction to scaffold hopping was helpful. Try out the methods in our drug discovery platforms!