Structural biology and drug discovery of difficult targets: the limits of ligandability

Over the past decade, researchers in the pharmaceutical industry and academia have made retrospective analyses of successful drug campaigns in order to establish "rules" to guide the selection of new target proteins. They have identified features that are considered undesirable and some that make targets "unligandable." This review focuses on the factors that make targets difficult: featureless binding sites, the lack of hydrogen-bond donors and acceptors, the presence of metal ions, the need for adaptive changes in conformation, and the lipophilicity of residues at the protein-ligand interface. Protein-protein interfaces of multiprotein assemblies share many of these undesirable features, although those that involve concerted binding and folding in their assembly have better defined pockets or grooves, and these can provide opportunities for identifying hits and for lead optimization. In some protein-protein interfaces conformational changes-often involving rearrangement of large side chains such as those of tyrosine, tryptophan, or arginine-are required to configure an appropriate binding site, and this may require tethering of the ligands until higher affinity is achieved. In many enzymes, larger conformational rearrangements are required to form the binding site, and these can make fragment-based approaches particularly difficult.