Covalent drugs represent a significant portion of current pharmaceuticals, with the top 26 having sales of over $30B in the US alone. While in the past target-directed drug discovery projects rarely pursued covalent binders due to safety concerns, the industry is increasingly considering the use of targeted covalent inhibitors (TCIs) that utilize soft reactive warheads, weak electrophiles that become locally reactive only in the context of the topology of a binding site. Developing a TCI requires a potent and specific reversible scaffold and the ability to covalently bind it to a selective group, most frequently a cysteine side chain. Tools exist for the design of covalent inhibitors starting from known reversible inhibitors, but for many targets there are no appropriate starting compounds. The general goal of this Phase I SBIR proposal is to develop algorithms and software for the efficient design of covalent inhibitors that will be based on well-established principles of fragment-based drug discovery. The work will include three main steps as follows. (1) Potential target sites will be identified based on the protein mapping software developed by Acpharis. The software will determine if there is a druggable binding hot spot adjacent to a cysteine residue, enabling the design of covalent inhibitors. For complete characterization of the site we will develop extended probe libraries and an iterative mapping algorithm, resulting in distributions of functional groups that will be used as generalized pharmacophores in the design. (2) We develop a library of small "core" compounds that have known handles and chemistry for placing a warhead for attachment to a Cys side chain. The library will be screened to identify "core" fragments binding to the target site by fitting each compound from this library to the generalized pharmacophores using a fast Fourier transform based docking algorithm. Compounds will be selected based on a scoring function that measures the quality of the fit, and will be further filtered by their abilit to form a covalent bond (possibly with an appropriate linker) to the nearby Cys residue. (3) To extend the core fragments into potential leads we select a large set of compounds that include one of the fragments as a substructure. Each of these compounds will be docked to the site while constraining the position of the bound "core" fragment, and scored on the basis of its fit to both the site and the generalized pharmacophores. The algorithm will be tested by applications to proteins with known covalent inhibitors, as well as by designing, synthesizing, and testing covalent core compounds to inhibit KEAP1, a key target for anti-inflammatory drug discovery. The new fragment based algorithm will expand the design of covalent inhibitors to classes of targets such as protein-protein interactions that generally do not have a starting non-covalent inhibitor appropriate for adding a reactive warhead.
The general goal of this Phase I SBIR proposal is the development of algorithms and software for the efficient design of covalent inhibitors, expanding the technology to new classes of important pharmaceutical targets.