The GT1 family of glycosyltransferases controls the synthesis and bioavailability of small molecule signaling hormones, but we lack a detailed understanding of their molecular precision. GT1s are crucial for bioactive small molecule synthesis in many organisms as well as the majority of phase 2 drug metabolism in humans, and the ability to predict substrate utilization patterns would benefit the fields of systems biology, biocatalysis, and medicinal chemistry. However, functional analysis based on sequence alignment alone has failed to provide an overall model robust enough for the prediction of substrate patterns. To overcome the problem of limited functional data for GT1 enzymes, we profiled the full, 107-member GT1 family of the model plant Arabidopsis thaliana with broad substrate libraries using a high-throughput mass spectrometry (HT-MS) platform. Over 13,000 screening events led to an unprecedented look at overall GT1 activity patterns. Though the sugar donor and acceptor scopes far surpass prior screens, analysis of our dataset by phylogenetic clustering once again supported few robust conclusions based solely on sequence alignment. To circumvent this bottleneck, we developed quantitative structure-activity relationship (QSAR)-like models for functional analysis using the physicochemical properties inherent in our libraries. These tools, which we are packaging into a free distribution (“GT Predict”), include modules for the correlations of physicochemical properties with protein activities, the prediction of new substrates for Arabidopsis GT1s, and the projection of potential activities onto novel input GT1 sequences. The activity patterns we observed reinforce several evolutionary clades of GT1s in plants, and further highlight new clusters of activities. Several active site structural features including key polar sites or hydrophobic pockets were identified from this data and confirmed using homology models generated from the few known GT1 structures. Finally, new GT1 substrates were identified beyond the original scope of this dataset. This combined functional screening and bio/cheminformatic approach therefore spotlights highly active GT1 enzymes for further biocatalyst design and utilization, and enables informed studies of metabolic and sugar signaling pathways in plants.

Structural insight from activity: Functional screening of the entire Arabidopsis GT1 family enables cheminformatic-bioinformatic predictions of glycosyltransferase reactions and protein features

253rd National Meeting (2017)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.