Bicyclic peptide ligands are promising molecules for the development of new therapeutics. They combine advantages from large protein therapeutics and small molecule drugs. Large combinatorial libraries of bicyclic peptides can be generated and screened by phage display using a recently developed strategy. Potent and selective bicyclic peptide inhibitors against several therapeutic targets have already been developed and their therapeutic potential is currently being evaluated in animal models. One aim of my thesis was the exploration of ring size diversity in bicyclic peptides. I generated a set of libraries of the format Cys-(Xaa)m-Cys-(Xaa)n-Cys, where 'm' and 'n' = 3, 4, 5 or 6, and performed affinity selections against the serine protease urokinase-type plasminogen activator. Bicyclic peptide inhibitors from virtually all ring size combinations were isolated, suggesting that many peptide formats can be accommodated in the active site of this enzyme. Moreover, they showed a large variety of consensus sequences and several of the identified consensus sequences were exclusively found in bicyclic peptides having specific ring size combinations. Having available multiple leads isolated from such bicyclic peptide libraries with variable ring sizes could be a great asset for the generation of high affinity binders. Additionally, other targets may have preferences for specific peptide constraints and the availability of these libraries increases the chances to isolate high affinity binders to any desired target. A second goal of my thesis was to apply high throughput sequencing technologies to phage display selections of bicyclic peptides, in order to identify a larger number of specific target-binding sequences and motifs. I developed a procedure to efficiently compare the sequences of large numbers of phage-selected peptides to identify target-binding peptide motifs based on abundance and sequence similarity. Applying this approach to phage isolated in selections against five different protein targets, I was able to identify rare target-binding peptide motifs and could more precisely define groups and sub-groups of consensus sequences. This information is valuable to choose peptide leads for drug development and facilitates the identification of epitopes. Moreover, binding motifs could be identified after a single round of phage panning. The final aim of my thesis was to discover bicyclic peptides that could be used as new antibiotics. Towards this end, I combined the newly generated variable ring size libraries and high-throughput sequencing procedures. I focused on the development of inhibitors of Staphylococcus aureus sortase A, an antivirulence target for which no potent and specific inhibitors have been reported. For the isolation of bicyclic peptide inhibitors to this target, the ring size diversity of the libraries turned out to be key. Inhibitors all shared the same motif in a loop of 5 residues. Further characterization of their effects on S. aureus showed that they could inhibit sortase-mediated incorporation of external substrates on the staphylococcal cell wall. However, they were not sufficiently potent to compete with the native substrates of the enzyme. More potent inhibitors are needed to effectively inhibit sortase A on S. aureus cells, and the bicyclic peptide inhibitors isolated constitute promising leads for the development of future antisortase therapeutics.
EPFL2014
