In silico design of ankyrin repeat proteins that bind to the insulin-like growth factor type 1 receptor

Abstract

Ankyrins are proteins widely distributed in nature that mediate protein‒protein interactions. Owing to their outstanding stability and ability to recognize targets, ankyrins have been used as therapeutic and diagnostic tools in several diseases, including cancer. Insulin-like growth factor type 1 receptor (IGF-1R) is overexpressed in a variety of cancers, making it an attractive molecular target. Advances in anticancer treatment have focused on inhibiting the binding between IGF-1R and its natural ligand, IGF1. In this work, three ankyrins were designed to interact with IGF-1R, and molecular models using AlphaFold were generated. The designed ankyrin sequences included amino acids of IGF1 that recognize IGF-1R: a two-module ankyrin (DAN2SON), a loop ankyrin (Loop-DAN2SON) and a bispecific ankyrin (BI-DAN2SON-D1). Models with the best results from the predicted local distance difference test and predicted assigned error values were used to perform rigid binding tests with the ClusPro server. The best complexes were selected based on the binding energies. Further analysis of the interactions was performed with the PDBsum server. The three IGF1-R complexes showed negative free binding energies, indicating that the binding of these proteins could be energetically favorable. Molecular binding assays revealed that DAN2SON and Loop-DAN2SON bind to IGF-1R at the natural ligand binding site via hydrogen bonds and salt bridge interactions. This work shows that using artificial intelligence to generate protein models allows prediction of interactions between ankyrins and the IGF-1R, to be confirmed in subsequent studies using both in vitro and in vivo models.