In-silico modeling was conducted using Schrödinger Suite (BioLuminate, Maestro). Protein-protein interactions between AChR and IgG1 (PDB: 5HBT), were analyzed to identify key residue interactions. Residue scanning was performed to predict beneficial mutations enhancing stability and Prime-MMGBSA calculations were used to evaluate the binding free energies of the engineered decoys.

Protein interaction analysis identified key antibody-binding regions at residues 30–39 and 8–10 of the AChR receptor. This guided residue scanning and engineering of α1-subunit fragments into minimized Pepenzymes. A redesigned decoy was created through targeted truncation and incorporation of nonnatural amino acids to improve stability and binding. Residues 62 and 63 were deleted, residue 65 was replaced with AIB, residue 70 mutated to proline, residue 71 to arginine, and residue 75 to AIB. Nonnatural residues such as AIB were substituted for similar non-polar residues like leucine or glycine. Prime-MMGBSA calculations showed the engineered Pepenzyme achieving a binding free energy of −67.38 kcal/mol compared to −29.36 kcal/mol for wild-type AChR, indicating stronger antibody interactions. Structural visualization confirmed the designed decoys preserved critical epitope structural elements.

This study demonstrates that computationally designed nanobody decoys can mimic the AChR α1-subunit and preferentially bind to pathogenic IgG1 autoantibodies, as the decoy had 2.25x the binding free energy compared to the wild type. By diverting antibody activity away from native receptors, this approach offers a targeted, non-immunosuppressive therapeutic strategy for Myasthenia Gravis.