Fragment screening of infectious disease targets in a structural genomics environment

Abstract

Structural genomics efforts have traditionally focused on generating single protein structures of unique and diverse targets. However, a lone structure for a given target is often insufficient to firmly assign function or to drive drug discovery. As part of the Seattle Structural Genomics Center for Infectious Disease (SSGCID), we seek to expand the focus of structural genomics by elucidating ensembles of structures that examine small molecule-protein interactions for selected infectious disease targets. In this chapter, we discuss two applications for small molecule libraries in structural genomics: unbiased fragment screening, to provide inspiration for lead development, and targeted, knowledge-based screening, to confirm or correct the functional annotation of a given gene product. This shift in emphasis results in a structural genomics effort that is more engaged with the infectious disease research community, and one that produces structures of greater utility to researchers interested in both protein function and inhibitor development. We also describe specific methods for conducting high-throughput fragment screening in a structural genomics context by X-ray crystallography.

Figure 1

Small molecules from ligand-bound structures super-imposed onto the apo structure of IspF from Burkholderia pseudomallei, with key side chains (lines) and the catalytic zinc ion (sphere) depicted. Fragment screening by NMR spectroscopy and X-ray crystallography led to three distinct subsets of ligands which bind in the cytosine sub-pocket (above left), coordinate to catalytic zinc (above right), and bind a hydrophobic cleft external to the active site (bottom left). Two lead compounds were designed by linking cytidine to zinc site-binding fragments (bottom right). While their cytosine moieties bind in the same manner, only one lead compound recapitulates the binding mode of its fragment precursor by coordinating to zinc through a heteroaromatic nitrogen. Figure generated using PyMol (DeLano, 2008) using PDB IDs 3F0E, 3IKE, 3IEQ, 3F0F, 3IEW, 3IKF, 3JVH, 3K14, 3KE1, #### and #### (Begley, manuscript in preparation).

Figure 2

GDP-bound structure of a protein monomer from Mycobacterium tuberculosis annotated as an arginine/ornithine transporter with ATPase activity. Diffraction-quality crystals were not obtainable in the presence of ADP or ATP, while hydrolysis of GTP to GDP was observed crystallographically. The N1/N3 interaction between GDP and the Asp207 side chain confirm a highly specific guanine binding pocket. Figure generated using PyMol (DeLano, 2008) from PDB ID 3MD0 (Edwards, manuscript in preparation).