Fragment screening is a method routinely applied in pharmaceutical drug development by now. Here, small molecules (<300 Da) that are identified during the screening process are developed into lead compounds with high affinity. While high-throughput screening usually requires large compound libraries, fragment libraries can be much smaller in size, containing only several hundred to thousand molecules. We developed a small fragment library composed of 361 compounds and validated it on the aspartyl protease endothiapepsin for which we obtained a high hit rate during crystallographic screening.[1,2]
We used this library to screen protein kinase A (PKA) and PIM1-kinase. For pre-screening we applied a thermal shift assay (TSA) to identify suitable fragments for crystallographic screening. In the TSA 31 fragments were identified for PKA and 52 for PIM1. These fragments were then subjected to crystallographic screening where we obtained 15 complex structures for PKA and 13 for PIM1, indicating also here a high crystallographic hit rate. In comparison we tested a random collection of fragments for crystallographic screening with PKA resulting in a much lower hit rate (21%). Deviating hit lists in the TSA assay and only one common fragment observed in both kinase structures suggest that fragments might be selective binders. TSA screening results, observed binding motifs and structural differences in the ATP-binding pockets of both kinases will be discussed in detail.