Publication
Bicyclic peptides binding to targets of interest can be isolated from combinatorial libraries using a phage display-based approach. In a first project of this thesis, we aimed at generating bicyclic peptide inhibitors of matrix metalloproteinases (MMPs), a family of proteases that share high structural similarities and have been difficult to target in a specific manner. From phage-encoded combinatorial libraries, we isolated a bicyclic peptide that inhibits specifically the gelatinase MMP-2 with a Ki of 2 μM. This inhibitor was less potent than the best peptidic MMP-2 inhibitor, the linear APP-derived inhibitory peptide (Ki = 13 nM). However, due to its conformational constraint, the bicyclic peptide is expected to be significantly more stable and hence more suitable for applications in biological systems. It may be used as a research tool alternatively to the broadly applied monocyclic peptide MMP-2 inhibitor CTT which is less potent (Ki of 142 μM). If affinity matured, the bicyclic peptide MMP-2 inhibitor might even be developed into an anti-cancer therapeutic. In a second project, bicyclic peptide binders to the centriolar protein SAS-6 were generated. SAS-6 plays an important structural role in centrioles and bicyclic peptide binders are currently applied in cellular assays by the laboratory of Professor Pierre Gönczy (EPFL) to study the process of centriole formation. In a third project of this thesis, we established an enzymatic method to selectively cyclise peptides with unprotected side chains using a transglutaminase (TGase). TGases catalyse the formation of stable amide bonds between the side chains of glutamine and primary amines. They have been used extensively in biotechnological applications to cross-link peptides and proteins or to attach labels to proteins. We found that the microbial transglutaminase (MTGase) of Streptomyces mobaraensis can cyclise quantitatively peptides with an N- terminal glutamine-donor sequence and a C-terminal lysine residue in an intramolecular reaction. Experiments with minimised glutamine-donor substrates revealed that peptides with only an Ala-Leu dipeptide N-terminal to the glutamine and a randomly chosen peptide between the glutamine and lysine residues are efficiently cyclised. By combining the MTGase-based cyclisation strategy with a chemical thiol-based cyclisation reaction, we were able to generate tricyclic peptide structures. It remains to be tested if this method can be applied to cyclise combinatorial peptide libraries on the surface of phage. In a fourth and last project, we developed a method to follow qualitatively and quantitatively chemical or enzymatic modifications applied to peptides displayed on phage. In this method, peptides on phage were chemically modified, cleaved off by a protease, purified, concentrated and analysed by mass spectrometry. The method will help to assess new reactions applied to phage peptides such as the MTGase-based enzymatic cyclisation reaction.