Selection of DNA-encoded dynamic chemical libraries for direct ligand discovery

Abstract

In the event of genomic and proteomic prosperity in the 21st century, the discovery of potential small molecules to cure diseases has been more expedited. Dynamic combinatorial chemistry (DCC) has evolved to bridge the biological tests and organic synthesis into one pot. Dynamic combinatorial libraries (DCLs) are generated via reversible reactions among the components under thermodynamic control. Target protein served as a template is then added to move the equilibrium to generate more compounds binding to the protein, resulting in the enrichment of high-affinity binders. The entire process can be performed without any separation of components and directly recognize the capable ligands; nevertheless, the diversity of DCLs is generally limited to hundreds of compounds, which impedes the utilization of DCLs. Typically, the analysis tools for DCLs are HPLC, NMR, LC-MS etc., which are not enabled to identify each homogeneous component. Brenner and Lerner prominently proposed the concept of DNA encoded chemical libraries (DELs) for the recognition of compounds in an ultra-large library. Utilized the distinct feature of unique sequence, the compounds covalently linked with DNA tags could be paralleled recognized by next generation sequencing (NGS). As DNA sequence could be amplified with PCR technology, the amount of DELs are only needed in pmol range and the consuming protein targets are less than that in DCLs. The fast and efficient procedure of DEL selection promotes it becoming a prevalent way for ligand discovery. To combine the merits of DCLs and DELs, an advanced concept named DNA encoded dynamic chemical libraries (DEDLs) was put forward. In present DEDLs, the dynamic formation is by using short DNA hybridization. The addition of protein target can change the dynamic equilibrium to the formation with high affinity-binding compounds. After selection, the efficient compound pairs are identified and further linker optimization between pairs are required. As known to all, the exploration of suitable linkers off DNA is unpredictable whether the linkers selected are benefited for compound pairs. We proposed a novel DEDL method for direct ligand identification without any further optimization. Different to present DEDL methods, our proposed DEDL method utilized reversible imine reaction to constitute the dynamic libraries. A known compound to the protein target as “anchor” was applied to direct the formation of highly efficient ligands under the target template. Subsequently, the non-binders were captured and removed by streptavidin beads. The flow-through with selected mixture were then amplified and detected with NGS. In chapter 2, we validated the feasibility of this method by applying it to target CA-Ⅱ. Later, we investigated the ligands in BRD4 BD1/BD2 with 67,600-member libraries. To expand the potential with different anchors, 17.576 million-member libraries were constituted, and specific ligands towards BD1, BD2 separately were discovered. Distinct from general DELs are applied with different targets without any explicitness, our method can notably select the ligands in a specific target by altering the anchor to form a specific DEDL towards the targets. Consequently, another proteins AChE and XIAP were used for novel ligands discovery.