Biophysical Fragment Screening and Structure Determination For Class A, B, and D Carbapenemases

<p>The emergence of multi-drug resistant Gram-negative pathogens (especially Pseudomonas aeruginosa, Klebsiella pneumoniae, Acinetobacter baumannii, and the Enterobacteriaceae) is among the most serious challenges faced by infectious disease clinicians today. Many of these resistant pathogens now possess extended spectrum β-lactamases (ESBL) and carbapenemases, seriously limiting treatment options and increasingly producing clinical failures. Our long-term goal is to develop carbapenemase inhibitors from entirely novel chemotypes and for which resistance mechanisms are absent from microbial genomes. This focus on reversible, non-natural product based chemotypes is a key feature of our discovery approach, and one we have validated previously for the cephalosporinase CTX-M. Here we propose the use of Surface Plasmon Resonance (SPR) methods to screen a total of ~6,000fragments from UCSF and University of Dundee across representative Class A, B, and D carbapenemases (KPC-2, NDM-1, OXA-48) and representative ESBL that we have studied previously, CTX-M. We will further use SPR to validate fragment hits by determining Kd, kon, koff, and binding stoichiometry. Active site binding will then be assessed by competition SPR experiments with known active-site ligands. Finally, binding of validated fragments in the presence of saturating concentrations of specific, well-characterized fragments will be used to identify fragment pairs that bind the active site simultaneously and non-competitively. Individual fragments and the fragment pairs identified in this screening campaign will then be structurally characterized (X-ray) bound to multiple carbapenemases, greatly enabling future medicinal chemistry efforts to covert these fragments to leads (outside scope of this application). Three important innovations will result from the proposed project. First, the rich set of SPR and X-ray binding will be compared to blinded computational docking predictions for the same fragments, leading to improved docking methods for fragment screening. Second, the SPR binding data across diverse beta-lactamases will be used to test our outstanding hypothesis that carbapenemases are more "druggable" on account of more exposed hydrophobic surfaces in the active site (as compared to ESBLs and narrow spectrum β-lactamases). Finally, the X-ray structures of fragments bound to Class A, B and Denzymes, including ternary complexes of bound fragment pairs, will reveal consensus binding hot spots shared by diverse carbapenemases, informing a cogent approach to produce novel β-lactamase inhibitors with a clinically desirable cross-class spectrum.</p>