Accurate computational design of artificial metalloproteins using Metal-Installer

Metalloproteins have evolved through the selective, strong, and precise interactions between metal ions and polypeptides. The coordination of metal ions to amino acid side chains is therefore a fundamental prerequisite for designing metalloproteins. Here, we present Metal-Installer, a user-friendly in silico tool that proposes probable mutation sites for metal ligation. This tool employs four geometric restraints with defined numerical ranges, derived from the chemically guided curation and in-depth analysis of comprehensive metalloprotein structures, and further refines the candidates using probability maps. To validate its utility and accuracy, we applied Metal-Installer to two unrelated protein scaffolds. We successfully created 13 artificial mononuclear and dinuclear metalloproteins incorporating zinc, manganese, iron, or copper. The X-ray crystal structures, biochemical properties, spectroscopic features, and catalytic activities of these metalloproteins closely match our predictions, demonstrating that Metal-Installer enables the accurate and broadly applicable design of artificial metalloproteins across diverse protein scaffolds, metal elements, ligand combinations, and nuclearities.