Fragment screening purely with protein crystallography.

We screen for fragments using X-ray crystallography as the primary screen. There are several unique features in our screening methodology. As a result of using X-ray diffraction as our primary screen, we do not use affinity data to bias our data collection or design in progressing hits toward a lead. Another difference in our methodology is that we choose to group our compounds as shape-similar groups. We also screen in a first pass mode without recollecting failed diffraction experiments. This method of screening results in an average loss of 5-10% of the data sets for the primary screen. The remaining data sets offer enough information to successfully advance three to five scaffolds into the secondary library design. We do not deconvolute the wells which show evidence of fragment binding by repeating the soaks with single compounds. Instead, evaluation of the possible fragments is done by refinement and examination of the resulting electron density difference maps. These methods allow us to complete the initial screen of a primary library of fragments in less than 3 months. A secondary library of fragments is designed using the base structures with electron density envelopes from the successful fragment hits of the primary library. Chemistry is chosen to probe interactions with the target and push the observed binding pocket limits in order to more clearly define the plasticity and range of possible extensions to the scaffolds chosen. The secondary library compounds are also screened in shape-similar groupings of five that are chosen without the knowledge of binding affinity. Our approach is a completely orthogonal one from traditional high-throughput screening in finding novel compounds.