“Inverse Drug Discovery” strategy to identify proteins that are targeted by latent electrophiles – exemplified by arylfluorosulfates

• Published in: Journal of the American Chemical Society 2017-12, Vol.140 (1), p.200-210
• Main Authors: Mortenson, David E., Brighty, Gabriel J., Plate, Lars, Bare, Grant, Chen, Wentao, Li, Suhua, Wang, Hua, Cravatt, Benjamin F., Forli, Stefano, Powers, Evan T., Sharpless, K. Barry, Wilson, Ian A., Kelly, Jeffery W.
• Format: Article
• Language: English
• ISSN: 0002-7863; 1520-5126
• DOI: 10.1021/jacs.7b08366

Drug candidates are generally discovered using biochemical screens employing an isolated target protein or by utilizing cell-based phenotypic assays. Both non-covalent and covalent hits emerge from such endeavors. Herein, we exemplify an “Inverse Drug Discovery” strategy in which organic compounds of intermediate complexity harboring weak, but activatable, electrophiles are matched with the protein(s) they react with in cells or cell lysate. An alkyne substructure in each candidate small molecule enables affinity chromatography–mass spectrometry, which produces a list of proteins that each distinct compound reacts with. A notable feature of this approach is that it is agnostic with respect to the cellular proteins targeted. To illustrate this strategy we employed arylfluorosulfates, an underexplored class of sulfur(VI) halides, that are generally unreactive unless activated by protein binding. Reversible arylfluorosulfate binding, correct juxtaposition of protein side chain functional groups, and transition state stabilization of the S(VI) exchange reaction all seem to be critical for conjugate formation. The arylfluorosulfates studied thus far exhibit chemoselective reactivity toward Lys and, particularly, Tyr side chains, and can be used to target non-enzymes (e.g., a hormone carrier or a small-molecule carrier protein) as well as enzymes. The “Inverse Drug Discovery” strategy should be particularly attractive as a means to explore latent electrophiles not typically used in medicinal chemistry efforts, until one reacts with a protein target of exceptional interest. Structure-activity data can then be used to enhance the selectivity of conjugate formation or the covalent probe can be used as a competitor to develop non-covalent drug candidates. Here we use the “Inverse Drug Discovery” platform to identify and validate covalent ligands for 11 different human proteins. In the case of 2 of these proteins, we have identified and validated ligands for the first time.