Genetic code expansion to enable the development of short, diverse peptide libraries

Project Summary Many intracellular cancer targets involve protein-protein interactions that are ?undruggable? because the binding surfaces are too large and featureless to be blocked by a standard rule-of-5 compliant small molecule.

Recently, there have been attempts to catalog molecules that are orally bioavailable but lie beyond the rule of five (bRo5).

Macrocyclic peptides can inhabit this bRo5 space, and a key advantage to using peptides as bRo5 molecules is that there are many mature techniques for finding peptide binders from vast libraries.

Arguably, the most powerful of these techniques is mRNA display, which allows creation of peptide libraries containing over 10 trillion variants, 6-7 orders of magnitude larger than a standard peptide library prepared on beads. The extreme diversity of these libraries has enabled many successes in inhibitor development.

Yet these successes are disconnected from real drug discovery, because the peptides uncovered are much too large to be bRo5 compliant.

Libraries that are short in sequence and bRo5 compliant can be created by mRNA display, but these libraries lack the diversity needed to uncover potent inhibitors because standard mRNA display is limited by the genetic code to ~20 variants at each position. Addition of unnatural base pairs (UBP)s offers great potential to address this problem.

In fact, the addition of a single UBP into the genetic code at a single codon position opens 32 new empty codons.

In principle, these codons can be exploited to encode novel non-canonical amino acids which in turn will dramatically enhance the potential diversity of short macrocyclic peptide libraries.

Still, genetic code expansion on this scale has not before been attempted, and, therefore, the key goals of this proposal are to prepare, validate, and optimize the mRNAs, tRNAs, and non- canonical amino acids required to build this system.

The validated system will be able to create macrocyclic peptide libraries that are short, yet contain billions of variants for the discovery of bRo5 compliant inhibitors to undruggable cancer targets.