Engineering acid-sensing ion channel 1a with disrupted desensitization: A computer-guided approach

Human Acid-Sensing Ion Channel 1a (hASIC1a) is a complex, multi-state membrane macromolecule whose rapid desensitization limits its application as a natural biosensor. To manipulate these structural dynamics, we designed and applied an integrative computational approach, combining protein residue network analysis and a machine learning-based functional site model, to identify functionally important positions governing hASIC1a desensitization. Focusing on structural communication regions, we pinpointed novel regulatory positions modulating the channel's conformational dynamics involving the desensitized state. Through the experimental characterization of a targeted library of 26 variants across nine predicted positions, we discovered three critical regulatory functional sites (Y426, L453, and Y456). Mutations at these sites yielded eight variants that successfully impaired or disrupted desensitization over a wide pH range, resulting in sustained current properties. This study demonstrates how our integrative computational strategy can be successfully applied to dynamically intricate membrane proteins, efficiently identifying mutational hotspots to engineer and fine-tune complex macromolecular functional properties.