The widespread adoption of peptide therapeutics has been restricted by their poor stability and therefore their limited in vivo half-life and oral availability. Members of the knottin family of peptides possess inherent stability, which makes them attractive scaffolds for the development of therapeutic agents. Given its remarkable stability against digestive proteases, the cyclic knottin peptide kalata B1 was employed as a scaffold to create a large peptide library displayed on the surface of E. coli. Fluorescence-activated cell sorting was used to isolate peptide ligands for two distinct protein targets: human thrombin and human neuropilin-1.

Thrombin is a well-characterized enzyme involved in the blood coagulation cascade and was selected as a proof-of-concept target with which to validate our methodology for altering kalata B1 to bind a desired molecule. A series of thrombin ligands was successfully isolated from the naive kalata B1 bacterial display library. The thrombin ligands exhibited high nanomolar affinities in solution with slow dissociation rates and also inhibited thrombin's enzymatic activity. Importantly, 80% of one of the knottin-based thrombin inhibitors remained intact after a two hour incubation both with trypsin and with chymotrypsin, demonstrating that modifying the kalata B1 sequence did not compromise its stability properties. In addition, the kalata B1 scaffold mediated a 20-fold enhanced affinity to the thrombin ligand when compared to the same binding epitope constrained by a single disulfide bond, which highlighted the importance of screening for ligands within the context of the kalata B1 scaffold. These results indicated that peptide libraries derived from the kalata B1 scaffold can yield high-affinity ligands that retain the protease resistance associated with the parent peptide.

Neuropilin-1 is an interesting therapeutic target for cancer and age-related macular degeneration since the receptor helps regulate angiogenesis. A neuropilin-binding ligand isolated from the original kalata B1-based library was used to construct a second-generation library with the goals of improving both the affinity and solubility of the originally identified peptide. The second-generation neuropilin-1 ligands exhibited solution affinities in the low nanomolar range, which was at least a ten-fold improvement over previously reported peptide ligands of neuropilin-1. Furthermore, the kalata B1-based neuropilin ligands showed low cytotoxicity towards mammalian cells. Overall, the strategies discussed in this dissertation may prove useful in the development of stable peptide ligands suitable for a variety of in vivo applications.