Macrocycles by the trillions By Joanne Kotz, Senior Editor Japanese researchers have developed a method for creating large libraries of N-methylated peptide macrocycles,1 a class of molecules that are sized between small molecules and biologics and have the potential to combine the pharmacological versatility of the former and the therapeutic specificity of the latter. PeptiDream Inc. has exclusively licensed the technology and entered collaborations with six pharmas. Small molecules are typically less than 500 Da in size, whereas biologics start at approximately 5,000 Da and cover a range of one to two orders of magnitude. Macrocycles, a structure frequently found in natural products, range in size from 500–2,000 Da. Due to their structural rigidity and high target affinity, the hope has been for these compounds to be able to modulate challenging targets typically reserved for biologics, such as protein-protein interactions, while retaining the advantages of small molecules—cell permeability and oral bioavailability.2 One challenge has been generating sufficiently large and diverse libraries of synthetic macrocycles to identify drug leads. To overcome this hurdle, a team led by Hiroaki Suga has now merged two technologies— one developed by Suga in 2008 to incorporate non-natural amino acids into large libraries of peptide macrocycles3 and the other developed independently in 1997 by Jack Szostak and Hiroshi Yanagawa for displaying natural peptide and protein libraries on mRNA.4,5 Suga is a professor in the Department of Chemistry at The University of Tokyo. Yanagawa is a professor at Keio University. Szostak is a professor at Harvard Medical School and Massachusetts General Hospital. The integrated method, called RaPID, starts from a cDNA library encoding non-natural peptides of 8–15 residues. The peptides contain a random mixture of 12 natural amino acids and 4 N-methylated (nonnatural) amino acids. Non-natural amino acids can increase the types of chemical groups in peptides beyond what is offered by natural amino acids. In particular, N-methylated amino acids increase a peptide’s cell permeability. The cDNA library is then transcribed to an mRNA library and linked to a second mRNA oligonucleotide that ends in a puromycin residue. The residue causes the growing peptide chain to be covalently connected to and displayed by its own mRNA template. In vitro translation of these mRNAs resulted in a library of about 1,012 N-methylated peptide macrocycles displayed on mRNA (see Figure 1, “Making macrocycles RaPID-ly”). As proof of principle, the Japanese team used the library to select for macrocycles that bound the ligase domain of ubiquitin protein ligase E3A (UBE3A; E6AP). Identifying potent and selective inhibitors that interfere with the protein-protein interactions made by ubiquitin ligases has proven challenging. The team identified three distinct macrocycles that each bound to E6AP with low or subnanomolar affinity. Results were published in Chemistry & Biology. “This is a very clever technology. Synthetic tRNAs to introduce new amino acids in combination with mRNA allows you to create vast libraries of novel peptides,” said Nick Terrett, CSO of macrocycle company Ensemble Therapeutics Corp. To isolate drugs with more complex function, such as disrupting protein-protein interactions, researchers are turning to larger molecules like peptides. “The real value of [this] work is that extremely large and diverse libraries can be produced and used in affinity-based selections, which greatly increases the likelihood that highly potent proteinprotein interaction inhibitors can be isolated. These arguments also apply to cases where highly specific drugs are needed, for example, when attempting to inhibit one member in a family of closely related proteins,” said Douglas Treco, president, CEO and cofounder of Ra Pharmaceuticals Inc. Ra Pharma is using in vitro display technologies to produce libraries of cyclic peptidomimetics to identify therapeutics in a variety of diseases. The company was cofounded by Szostak. “The affinities described for the ubiquitin ligase are impressive. It will be key to demonstrate that high-affinity binders can be isolated for a broad range of targets,” said Christian Heinis, assistant professor in the Institute of Chemical Sciences and Engineering at the Swiss Federal Institute of Technology Lausanne. “I like the RaPID approach very much since it allows the facile incorporation of non-natural amino acids.” Heinis is a cofounder of Bicycle Therapeutics Ltd., a company developing cyclic peptide therapeutics using phage display. However, Terrett said the molecules generated by RaPID could be too big to become drug leads. “The molecular weight is too high,” as the macrocycles described in the paper were around 2,000 Da, he said. “My guess would be that cell membrane permeability and the in vivo bioavailability would be low.” Ensemble’s macrocycles, which Terrett said are typically cell permeable, are in the range of 600–1,000 Da. For example, Ensemble’s macrocyclic IL-17 antagonists are 700 Da and have low nanomolar potency. The compounds are in preclinical development for autoimmune and inflammatory diseases. Ensemble synthesizes macrocycles using a process called DNAprogrammed chemistry that relies on DNA as a template but generates