New Strategies to Discover Small-Molecule Probes and Clinical Candidates

small molecules;chemical library;diversity-oriented synthesis;DOS;DNA-encoded library;DEL;cell painting;morphological profiling;mechanism of action;target identification

Hua, Bruce ; Schreiber, Stuart L.,Myers, Andrew G.,Balskus, Emily P.

University:Havard University

Department:Chemistry and Chemical Biology

关键词: small molecules;    chemical library;    diversity-oriented synthesis;    DOS;    DNA-encoded library;    DEL;    cell painting;    morphological profiling;    mechanism of action;    target identification;   

Others  :  https://dash.harvard.edu/bitstream/handle/1/37365952/HUA-DISSERTATION-2020.pdf?sequence=1&isAllowed=n

美国|英语

来源: Digital Access to Scholarship at Harvard

【 摘 要 】

Small molecules—through their engagement with molecular targets—can be used to elucidate biological pathways, test therapeutic hypotheses, and treat human diseases. These compounds are often discovered via high-throughput screening, where large chemical libraries are tested for an activity of interest. However, the synthetic design of these libraries has tended to generate molecules that suffer from high structural similarity and lack three-dimensionality, resulting in redundancy in biological performance. Instead, diversity-oriented synthesis (DOS) can be used to the access compound collections comprising a wide range of complex structures with an emphasis on scaffold diversity. At the same time, the discovery of selective bioactive compounds can benefit greatly from the rapid annotation of their biological activity. Here, I describe different strategies whereby DOS libraries can be synthesized and tested to identify useful chemical matter.In Chapter 1 of this dissertation, I discuss the design and synthesis of a DOS library that relies on a key photocyclization whereby N-substituted pyridinium salts can be converted to bicyclic aziridines or a variety of 5,5-fused bicycles. The library design allowed the synthesis of a collection of DOS compounds containing rigid, bicyclic ring systems, in three or fewer synthetic steps from simple starting materials. One of these library members, BRD3795, was found to exhibit potent and selective killing of patient-derived organoid models of pancreatic ductal adenocarcinoma (PDAC). Subsequent target elucidation studies indicated that BRD3795 likely binds to and inhibits sulfide:quinone oxidoreductase (SQOR), a mitochondrial enzyme that has no known inhibitors and has also attracted interest as a potential drug target for heart failure. These experiments serve as an example where the principles of DOS helped enable the discovery of a class of compounds with a novel mechanism of action (nMoA).In Chapter 2 of this dissertation, I characterize three separate compound collections via “real-time biological annotation,” a concept that allows the annotation of compound activities in the same time frame in which they are synthesized. First, I describe the real-time biological annotation of a set of constitutional isomers, where we first present the feasibility of this concept and show that cell painting is a viable method to generate multidimensional profiles of compound biological activities in an unbiased manner. Next, I describe the annotation of a set of stereoisomers and show that stereochemical and appendage-diversity both contribute to the biological-performance diversity of a compound collection, and that the greatest benefit occurs when the two are used in conjunction. Lastly, I describe the biological annotation of a collection of “nuisance” compounds and show how the resulting cell-painting signatures can be used to flag compounds that exhibit nonspecific activities in cell-based assays.DNA-encoded library (DEL) technology allows the efficient identification of high-affinity binders to targets of interest. In Chapter 3 of this dissertation, I discuss the design and synthesis of a 35,712-membered DEL that incorporates the principles of DOS and benefits from on-DNA scaffold generation via split-and-pool synthesis. The library is also designed such that off-DNA synthesis can be completed in just three synthetic steps from commercial building blocks. I screened this library against a collection of 23 targets, including hydrolases, kinases, lipases, sirtuins, chaperones, transcription factors, and protein complexes. The DEL screen against undecaprenyl pyrophosphate synthase (UppS) revealed a collection of enriched library members which exhibited strong structure-activity relationship (SAR) trends. These results validated the successful construction of the DEL and highlighted the importance of isomers for the interpretation of DEL screening data.