Structure-Guided Investigation of Fungal Adenylation Domain Substrate Selectivity

Adenylation (A) domains are the primary gatekeepers to substrate selectivity in nonribosomal peptide synthetases (NRPSs). A significant body of work has investigated key structural motifs contributing to substrate activation in bacterial A domains, but these findings have often failed to hold true in fungal systems. Here, we employed homology modeling to interrogate the structures of bacterial and fungal A domains, aiming to improve bioinformatic predictions of substrate selectivity “codes,” particularly for uncommon and nonproteinogenic building blocks. After comparing five different homology modeling tools to an existing crystal structure, we generated 46 structural models for a variety of bacterial and fungal A domains, both characterized and uncharacterized. We were able to identify an 18 amino acid code by distance-mapping residues within 5 Å of a docked substrate, which acts as a more reliable predictor of substrate binding than existing in silico methods. This new 18 residue code was then confirmed through mutagenesis experiments, allowing for the activation of non-native substrates by a previously characterized fungal A domain.