DNA-encoded libraries

Among the earliest examples of combinatorial libraries were peptide libraries prepared on solid phase via split-mix synthesis. The number of library compounds was high on the other hand, the actually prepared amounts were minute. Therefore, methods to tag each solid-phase bead were developed. 

The complete library is added to the target protein. Each member is present only at very low concentrations. Binders are affinity captured on the target and can then be separated from the mixture of unbound library members. The structure of the binding compounds is determined by readout of the tag via PCR and sequence analysis [158]. 

Utilizing the double-strand nature of DNA, it is also possible to direct a singlestrand tagged diversity reagent to an antisense single-strand tagged scaffold. In this case, the DNA tag of the final product will not be built sequentially but will be complete from the start. This can be understood as a yoctoliter-scale reactor [160]. 

Another concept utilizing antisense are dual pharmacophore libraries, in which two fragments, each on its own single-strand DNA tag, are spatially aligned via antisense regions close to the fragments [161]. 

---

The nonSELEX concept provides the opportunity for the selection of affinity ligands from DNA-encoded libraries of small molecules and peptides. DNA tags in such libraries encode the information, which allows one to identify structures of corresponding small molecules, when the DNA tags are PCR amplified and sequenced. Owing to the small size of the molecules with respect to that of the covalently attached DNA tag, such libraries are expected to have electrophoretic properties identical to those of DNA libraries. SELEX is not applicable to such libraries since small molecules and peptides cannot be amplified by PCR. 

However, one key issue has prevented Nature s approach being applied to small molecules the information coded in DNA or RNA base sequences cannot be directly translated into small molecule structures with Nature s machinery. New technologies and platforms need to be designed and implemented. The DNA-encoded library has emerged as one of the most intriguing approaches towards this goal. 

In this chapter, we will first discuss the basic principles of DNA-templated organic synthesis, as it is the mechanistic foundations for DNA-encoded libraries then we will discuss the progressive development of DNA-templated libraries and the evolution into a novel drug discovery tool. We will also discuss the DNA-recorded library, which also encodes library molecules with DNA but is conceptually different. These discussions will naturally involve specific drug discovery programs in which these libraries were applied and finally, we will discuss the outlook of DNA-eneoded libraries in the future of drug discovery. 

Our focus in this chapter is the DNA-encoded library as a tool in exploring chemical space for drug discovery. For the applications of DNA-templated synthesis in other fields, such as the origin of life, biodetection, nano-technol-ogy, etc., we refer readers to various excellent reviews. ... 

In a DNA-templated library, the DNA template serves two purposes. First, it is a nano-reactor virtually separated from each other spatially so that parallel reactions can proceed simultaneously. Second, it encodes the information of the specific chemical structure to be synthesized on the template in form of DNA sequences. In other words, the sequence of each template pre-determines the specific structures to be synthesized on it. This is analogous to the genetic information translation from mRNA to proteins the sequence of a particular mRNA determines the specific protein synthesized on it. This feature differentiates a DNA-templated library from other types of DNA-encoded libraries, which will be discussed in later sections. 

DNA-encoded libraries come in many different forms, but they share many common features, such as chemical reactions used, selection and amplification method, hit identification, etc. These approaches are different in their solutions to address these following questions ... 

Albeit on a miniaturized scale, the DNA-encoded libraries can reach a diversity level far beyond normal HTS libraries. For example, a library by Harbury s DNA routing approach has 100 million compounds Nuevolution announced on their website that their Chemetics library passed the 1 billion compound mark while GSK s DEL-B library has more than 800 million different structures. These numbers are on a par with biopolymer libraries prepared by genetics techniques such as phage-display and mRNA display, while the largest synthetic combinatorial library is around merely 2 million. Therefore, DNA-encoded library is truly an expansion of HTS into a much higher level of molecular diversity. 

Even with limited usable chemical reactions, in some cases, DNA-encoded libraries can be a unique tool to access chemical structures previously difficult... 

All the molecules in a DNA-encoded library are selected simultaneously. As shown in Ensemble Discovery and GSK s data analyses, a comprehensive view of SAR can be rapidly generated from the lines and planes formed in the SAR plot, high priority structural elements can be identified and applied to the next generation of library design, or to the synthesis of discrete compounds. In addition, a clear correlation between related structural elements with binding potency strongly validates the selection itself. 

One of the most critical limitations of a DNA-encoded library is DNA itself It requires an aqueous media and mild environmental conditions to ensure DNA s chemical stability and hybridization fidelity. Naturally, many useful chemical reactions cannot be directly applied. Liu and co-workers have adapted a variety of organic reactions into DNA-compatible format however, a survey of the structures synthesized by DNA encoding indicates the compounds made by DNA-templated chemistry are still being confined within the paradigm of pep-tidic structures, an undesired structural feature for drug molecules. 

Therefore, expanding DNA-compatible chemical reactions is one of the most critical tasks for further development of this technology for DNA-encoded library to generate hit molecules with drug-like or even lead-like structures. Liu and co-workers have attempted DNA-templated reaction in organic solvents assisted by detergents" and achieved some success, which may be a viable approach to address this issue. 

Naturally, not all the goals above can be feasibly achieved with the current state of the art, but we believe technological advancement and ingenious human input, as they have already done, will eventually accomplish these tasks and make DNA-encoded libraries indispensable tools in drug discovery. 

As sampling of chemical space will be even more sparse with increasing size of the molecules, it may be required to revert to other technologies like fragment-based screen, natural products, DNA-encoded libraries, or dynamic combinatorial chemistry to generate starting points for lead identification. 

Scheme 5.12 Inhibitors of carbonic anhydrase IX from a DNA-encoded library [159).

---