AI protein structure prediction-based modeling and mutagenesis of a protostome receptor and peptide ligands reveal key residues for their interaction

Authors

Shi-Qi Guo, Nanjing University
Ya-Dong Li, Nanjing University
Ping Chen, Nanjing University
Guo Zhang, Nanjing University
Hui-Ying Wang, Nanjing University
Hui-Min Jiang, Nanjing University
Wei-Jia Liu, Nanjing University
Ju-Ping Xu, Nanjing University
Xue-Ying Ding, Nanjing University
Ping Fu, Nanjing University
Ke Yu, Nanjing University
Hai-Bo Zhou, Nanjing University, Peng Cheng Laboratory
James W. Checco, University of Nebraska-LincolnFollow
Jian Jing, Nanjing University, Peng Cheng Laboratory, Icahn School of Medicine at Mount Sinai

Date of this Version

8-30-2022

Citation

J. Biol. Chem. (2022) 298(10) 102440. https://doi.org/10.1016/j.jbc.2022.102440

Comments

Open access.

Abstract

The protostome leucokinin (LK) signaling system, including LK peptides and their G protein-coupled receptors, has been characterized in several species. Despite the progress, molecular mechanisms governing LK peptide–receptor interactions remain to be elucidated. Previously, we identified a precursor protein for Aplysia leucokinin-like peptides (ALKs) that contains the greatest number of amidated peptides among LK precursors in all species identified so far. Here, we identified the first ALK receptor from Aplysia, ALKR. We used cell-based IP1 activation assays to demonstrate that two ALK peptides with the most copies, ALK1 and ALK2, activated ALKR with high potencies. Other endogenous ALK-derived peptides bearing the FXXWX-amide motif also activated ALKR to various degrees. Our examination of cross-species activity of ALKs with the Anopheles LK receptor was consistent with a critical role for the FXXWX-amide motif in receptor activity. Furthermore, we showed, through alanine substitution of ALK1, the highly conserved phenylalanine (F), tryptophan (W), and C-terminal amidation were each essential for receptor activation. Finally, we used an artificial intelligence– based protein structure prediction server (Robetta) and Autodock Vina to predict the ligand-bound conformation of ALKR. Our model predicted several interactions (i.e., hydrophobic interactions, hydrogen bonds, and amide-pi stacking) between ALK peptides and ALKR, and several of our substitution and mutagenesis experiments were consistent with the predicted model. In conclusion, our results provide important information defining possible interactions between ALK peptides and their receptors. The workflow utilized here may be useful for studying other ligand–receptor interactions for a neuropeptide signaling system, particularly in protostomes.