Metalloenzymes are rich targets for protein engineering. The scope of directed evolution with the 20 canonical amino acids is, however, limited for direct mutation of the primary coordination sphere and is often not suitable for making small controlled changes. Through genetic code expansion, the incorporation of non-canonical amino acids into proteins can help address this limitation. Herein, we describe the design of new genetic code expansion tools to test structure-function relationships in metalloenzymes and semi-rationally engineer these enzymes for improved activity. From these studies, we find that changes in enzymatic activity can often be predicted from fundamental inorganic coordination chemistry principles, providing support for the use of semi-rational mutagenesis with non-canonical amino acids.