Item

Abstract

Growth factors are signaling molecules coordinating the complex functionality of multicellular organisms during development and homeostasis. The activity of signaling networks needs to be tightly controlled; otherwise, aberrant expression of growth factors can cause diverse disorders such as cancer, autoimmune and cardiovascular diseases. Being highly specific and able to target various molecular surfaces, protein-based binders constitute a powerful means to manipulate signaling interactions. In contrast to the common antibody development technologies, which are empirically guided and often yield binders to irrelevant epitopes, de novo protein design methods allow targeting arbitrarily selected functional sites. Additionally, de novo design offers control over protein sequence and topology, facilitating improvements in folding kinetics, protein stability, and solubility. Here, we present our work on in silico design of the epitope-directed inhibitors against vascular endothelial growth factor (VEGF), a key modulator of tumor progression. Taking advantage of a new computational approach for massivescale docking of a target epitope against a protein structure database, we selected several scaffolds with high shape complementarity to the receptor binding site of VEGF. After further interface design, aiming to minimize the estimated binding free energy, a small set of best candidates (16 proteins) was experimentally evaluated. Biophysical measurements revealed that the binding affinities of the designs to VEGF ranged from nano- to micromolar levels. X-ray structure determination of one of the candidates showed atomic-level agreement with the design model. Moreover, in vitro assays showed the ability of the best designs to inhibit proliferation of VEGF-dependent acute myeloid leukemia cells. Thus, our results not only provide potential anti-cancer therapeutic candidates, but also demonstrate the feasibility of the rational approach to design high-affinity protein binders against predefined conformational motifs of the target molecules. This generalizable approach can be deployed to generate novel leads for future research and therapeutic purposes.