Optimization of small molecule drugs binding to highly polar target sites: Lessons from the discovery and development of neuraminidase inhibitors

Binding affinity optimization of small molecules interacting with polar binding sites on target proteins is a formidable, but not uncommon challenge in drug discovery. The challenge relates to the difficulty of integrating favourable and unfavourable polar, non-polar and conformation contributions into overall favourable binding energies. This review describes the surprising breakthrough findings leading to the development of Tamiflu, a clinically efficacious orally bioavailable drug targeting the active site of influenza neuraminidase (NA). The NA active site is highly polar and formed mostly by arginine, aspartate and glutamate residues. This active site structure evolved for efficient interaction with charged sialic acid moieties on glycoproteins and stabilization of an oxocarbonium ion in the transition state of the neuraminidase reaction. The initial strategy of optimizing polar interactions in transition state analogs led to NA inhibitors (NAIs) with sub-nanomolar binding affinities, but such compounds were highly polar and lacked oral bioavailability. The realization of the possibility to achieve high affinity binding in a highly polar active site through optimization of non-polar and van-der-Waals interactions initially appeared counterintuitive and required a few serendipitous findings, but was key to reduce the polarity of drug candidates, avoid large desolvation penalties and achieve oral bioavailability.