The design of therapeutic small molecules as competitive inhibitors with natural ligands or as interrupters of protein:protein interactions requires their specific interaction with the target proteins. Over the past two decades fragment-based methods, involving the elaboration and combination of weak affinity binders < 300 Da, have rapidly developed into an alternative approach to high throughput screening (HTS) for lead discovery. The small size of the fragments allows to sample a larger part of chemical space with a limited number of fragments (1-5 103 vs 106). The generally weak binding affinities (M to mM range) of fragments, however, implies that sensitive biophysical methods have to be used for screening. Screening by serial X-ray crystallography or ligand NMR-based fragment screening form the benchmark methods. Crystallography allows the direct observation of the crystal fragment complexes, but requires that good quality crystals are readily available. NMR-based screening identifies interacting fragments using proteins in solution, and can identify the fragment binding site in case a full assignment of the protein is made or a known competing ligand is used.
At the VIB-VUB Center for Structural Biology we have recently set up an NMR-based fragment screening platform on our 800 MHz spectrometer, using a 1000-fragment Maybridge library. The spectrometer was therefor also equipped with a sample changer. The library was first subjected to quality control by collecting the 1H NMR spectra for all individual fragments. The library was subsequently pooled in pools of 5 based on their chemical compatibility and non-overlapping NMR-spectra, to allow identification of individual binders from spectra collected for the pools. In order to identify hits water-ligand observed via gradient spectroscopy (Water-LOGSY), saturation transfer difference (STD) or Carr-Purcell-Meiboom-Gill (CPMG) spectroscopy can be performed.
So far, we have screened the fragment library for 4 protein targets, from different functional families. The hit rate varied between 1-3%, as expected for fragment screening by NMR. For two of these projects we also performed hit validation experiments by NMR, assessing the binding site of the fragment hits by displacement with a known ligand. For a third project, soakable crystals are available. Hopefully our screening results will be the start of fruitful (structure-based) lead discovery.