Fragment based drug discovery (FBDD) is a powerful and versatile approach that is widely used in both industry and academia to find new small-molecule therapeutics. The initial “hits” in FBDD are very small molecules, which due to their size form few interactions with the target protein and therefore have weak binding affinity. Subsequently, significant medicinal chemistry and biophysical characterisation effort is required to elaborate validated fragment hits into higher affinity lead compounds. Inevitably, the synthesis - evaluation cycle becomes a bottleneck for a drug discovery campaign.
In attempt to accelerate the fragment-to-lead process, we tested a new screening approach of soaking protein crystals with minimally purified compounds generated via microscale parallel chemistry to identify those that made additional interactions in the binding site. It was anticipated that screening by X-Ray crystallography in this way could provide key structural information and allow selection of only the most promising analogues from the library. The best compounds would then be synthesised on batch scale and characterised fully.
Herein, we report an initial proof-of-principle study that we performed on an established FBDD target with tractable crystallography - Escherichia coli DsbA (EcDsbA). EcDsbA is a disulfide oxidoreductase, which is responsible for the formation of disulfide bonds in diverse substrate proteins. EcDsbA has a critical role in the folding of many proteins that contribute to bacterial virulence, and is therefore a target for the development of novel antibacterial drugs. Due to its flat hydrophobic binding site that displays little selectivity for protein substrates, developing potent and selective EcDsbA inhibitors has proven to be extremely challenging.
In the pilot study, we successfully screened a library of 93 unpurified reaction mixtures. As a result we managed to identify 5 new compounds that showed clear density in the X-Ray data. These were resynthesised and characterised via a range of biophysical techniques and were found to have equal or higher affinity to the starting compound. Analysis of parallel reaction libraries by X-ray crystallography can therefore be used to accelerate fragment-based drug design and in the discovery of high affinity lead series.