(119b) Epitope Expansion Evolution for the Affinity Maturation of Ligand-Receptor Interactions

Molecular recognition underlies nearly every event in living systems. Engineering enhanced molecular recognition has been widely used in the biotechnology arena to develop protein-based biosensors, imaging agents, and therapeutics. Current approaches for affinity maturation of protein-protein interactions often focus on optimization of that particular interaction. Such approaches include optimizing loops of antibody CDR regions, or applying directed evolution to enhance molecular complementarity of an existing ligand-receptor interaction. In nature, however, molecular recognition often occurs at the interface of multiple domains. For example, vascular endothelial growth factor functions as a homodimer with both monomers coming together to form the receptor binding interface and multiple domains of hepatocyte growth factor engage in weak interactions with the Met receptor to result in a high affinity HGF-Met interactions. This suggests nature may have evolved for enhanced recognition by expanding the epitope of an initial interaction through inclusion and evolution of an additional domain. We have developed epitope expansion evolution as a new technique for enhancing ligand-receptor interactions, and selected the Gas6-Axl system for testing this new approach. Axl is a transmembrane tyrosine kinase receptor that is emerging as an important therapeutic target for cancer metastasis, and agents that disrupt binding of the Gas6 ligand to the Axl receptor hold promise as therapeutic candidates.

To enhance the Gas6-Axl interaction, a synthetic binding domain was added to the Axl receptor and evolved for binding to Gas6. We selected Ecballium elaterium trypsin inhibitor 2 (EETI-II) as the synthetic binding peptide as this has recently been shown by our lab to be an effective scaffold for evolving molecular recognition. A library of EETI-II mutants containing randomized amino acids in loops 1 and 3 was fused to the N-terminus of wild-type Axl receptor and expressed as a covalent fusion to the yeast cell surface. Fusion of this EETI library to the Axl N-terminus did not alter the strength of Axl binding to soluble Gas6, suggesting the structure of Axl was not perturbed. The yeast displayed EETI-Axl fusion library was screened for enhanced binding to Gas6 using fluorescence activated cell sorting (FACS). To increase the stringency of screening, we employed ?off-rate' sorts, in which equilibrium binding was followed by an unbinding step in the presence of excess competitor. Clones that retained binding following the unbinding step were collected using FACS.

Pooled library products retained binding to soluble Gas6 after a 48h unbinding step, in contrast to wild-type Axl which completely dissociates by 4h. We identified three individual EETI-Axl mutants with up to 4-fold improvement in Gas6 binding affinity compared to wild-type Axl. This enhanced affinity was a direct result of improved off-rate. Each of the library products contained wild-type Axl sequence and EETI with mutated loops 1 and 3, indicating improved affinity was a direct result of optimized EETI synthetic binding peptide and not mutations that may have accumulated within the Axl receptor. Given that the native Gas6-Axl interaction lies in the single-digit nanomolar range, this system provides a rigorous test for whether epitope expansion evolution would be useful for enhancing the affinity of ligand receptor interactions. We envision epitope expansion evolution can be a useful complement to the traditional affinity maturation approaches and can also be a useful tool for probing potential evolutionary pathways for molecular recognition.