New Approaches to Discover Biologically Active Small Molecules I. Diversity-Oriented Synthesis Encoded by DNA Oligonucleotides II. Synergistic Coupling of Organic Synthesis and Biological Annotation

Abstract

Organic small molecules can be used to study and treat disease, improve our understanding of human biology, and address some of the most urgent problems in biomedical research. Biologically active compounds are often identified via high-throughput screening of large small-molecule libraries, but much of our existing chemical matter does not fully leverage the capabilities of modern synthetic organic chemistry. As a result, many screening libraries are enriched in compounds that lack three-dimensional complexity and diversity. These structural redundancies often lead to redundancy in biological performance. Modern asymmetric synthesis can yield topographically complex compounds that fill some of the gaps in current screening collections, and these novel chemistries have already generated valuable chemical probes and drug leads. Here, I present the development of two new approaches to use modern synthesis to discover new small-molecule modulators of challenging therapeutic targets. In Part I of this dissertation, I discuss the incorporation of diversity-oriented synthesis principles into the design of DNA-encoded libraries (DELs). DELs are powerful tools to discover small-molecule binders of immobilized biomacromolecular targets. Most DELs, however, exhibit a narrow range of structural features, which likely limits their potential to identify binders of challenging proteins. I expanded the types of compounds that can be included in DELs in two ways. First, I used stereospecific C–H arylation chemistry to generate all possible stereoisomers of chiral 2,3-disubstituted azetidine and pyrrolidine scaffolds, which formed the basis of a 107,616-member DEL (Chapter 2). I then screened this library against challenging protein targets implicated in a wide array of diseases, and I observed several sets of promising results (Chapter 3). Second, I developed a version of the [3+2] nitrone–olefin cycloaddition that generates fused tricyclic isoxazolidines and is suitable for DEL syntheses (Chapter 4). Both approaches—performing complexity-generating transformations off DNA or on DNA—can increase the variety of molecular architectures found in DELs. In Part II of this dissertation, I describe my studies of methods for real-time biological annotation of synthetic compounds, in which newly synthesized small molecules are tested in high-dimensional phenotypic assays to assess their biological activities. Currently, small-molecule biological activity is identified on a largely ad hoc basis; compounds in screening libraries are tested against individual protein targets over many years. This sequential process—chemistry then biology—prevents the results from those biological experiments from influencing synthesis efforts until years later. Instead, I performed a set of experiments in which synthetic chemistry and biological testing are performed in parallel. I first studied a remarkable photochemical rearrangement of N-substituted pyrroles into tricyclic aziridines (Chapter 6). The results of these studies guided the design of a pilot library comprising ten isomeric triads, which I tested in an imaging-based high-dimensional assay called “cell painting” that measures compound-induced changes in cell morphology (Chapter 7). Finally, I performed two rounds of synthesis and biological annotation to identify a set of tricyclic aziridines that act like ATPase inhibitors according to both the L1000 assay, which measures changes in gene expression, and cell painting (Chapter 8). These experiments establish that high-dimensional cellular assays can identify biologically active compounds rapidly and generate robust hypotheses regarding their mechanisms of action.