Fragment-based drug discovery (FBDD) is based on the idea that using small (< 300 Da), highly soluble compounds to screen against a target will give higher hit rates and sample chemical space more efficiently compared to screens using larger, drug-like compounds.  

Because fragments bind weakly, detecting their binding experimentally is challenging and requires high fragment concentrations and very sensitive biophysical methods. The smaller the fragments, the more of a problem this is. But the idea of going smaller is attractive, as this creates smaller libraries that sample chemical space more efficiently [1].

While the binding of these very small fragments is difficult to measure experimentally, there have been attempts to simulate it, and various computational binding site mapping and druggability estimating methods are based on simulating the binding of small molecular probes on the surface of the target protein [2]. These have been used both to detect binding sites, as well as to highlight individual interactions that contribute to ligand binding. These smaller probes are generally seen as tools to probe the landscape of the binding pocket, rather than as chemical starting points themselves.  

Of the available biophysical methods, X-ray crystallography is most readily suited to observing the binding of these very small fragments, with the added benefit that it reveals the binding mode. A notable early such method is Multiple Solvent Crystal Structures (MSCS), in which crystal structures of the target were solved with organic solvents [3]. This was useful for detecting druggable sites on proteins, but the solvents themselves were considered unlikely starting points for a medicinal chemistry campaign. In addition, crystallographic poses do not give information about how tightly a compound is bound.

Recently, there have been a number of publications about using very small fragments (VSFs) to experimentally probe binding sites with crystallography[4, 5]. These fragments showed high hit rates and raised the possibility of being used as a ‘pre-screen’, in order to determine what larger fragments to screen next. Elliot has written about this idea on the OPIG blog previously here.

In a similar vein, I recently presented this paper by Draxler et al. at group meeting. The authors used a combination of computational screening using the SILCS-Hotspots method, and biophysical screening using crystallography and NMR. They also obtained binding affinity information for some of the very small fragments and were able to calculate ligand efficiencies for a subset of these very small  fragments. It’s a cool paper that combines experimental and computational approaches on very small fragments. I expect we will see more publications in this exciting field in the future. 

[1] Draxler, S. W. et al. Hybrid Screening Approach for Very Small Fragments: X-ray and Computational Screening on FKBP51. J. Med. Chem. 63, 5856–5864 (2020).

[2] Rathi, P. C. et al. Predicting ‘Hot’ and ‘Warm’ Spots for Fragment Binding. J. Med. Chem. 60, 4036–4046 (2017).

[3] Mattos, C. et al. Multiple solvent crystal structures: Probing binding sites, plasticity and hydration. J. Mol. Biol. 357, 1471–1482 (2006).

[4] Wood, D. J. et al. FragLites—Minimal, Halogenated Fragments Displaying Pharmacophore Doublets. An Efficient Approach to Druggability Assessment and Hit Generation. J. Med. Chem. 62, 3741–3752 (2019).

[5] O’Reilly, M. et al. Crystallographic screening using ultra-low-molecular-weight ligands to guide drug design. Drug Discov. Today 24, 1081–1086 (2019).