Structural and Biochemical studies of Arabidopsis thaliana Glycosyltransferases

Abstract Background: Glycosylation of secondary metabolites involves plant UDP-dependent glycosyltransferases (UGTs). UGTs have shown promising potential as drug targets as well as catalysts in the synthesis of glycosides of medicinal importance. However, limited understanding at the molecular level due to insufficient biochemical and structural information has hindered potential applications of most of these UGTs. For example, only two crystal structures of Arabidopsis thaliana UGTs are currently solved of the 122 genes available. In addition, more than half of these UGTs are yet to be biochemically characterised. Aims: This research aims to i) investigate qualitatively substrate specificities of Arabidopsis thaliana UGTs from selected families using mass spectrometry (MS) based methods and to study the kinetic parameters via bioluminescence (Chapter 3); ii) produce and study homology models of the UGTs (novel) to further understand their substrate preferences and key catalytic amino acid residues involved (Chapter 4); and lastly iii) manipulate rationally the active sites of UGTs to engineer mutant UGTs of improved donor substrate activity (Chapter 5). Methodology: Direct monitoring of products of glycosylation was done using triple quadrupole mass spectrometry (QQQ-MS) as it involves limited substrate modification. Full scan and product ion screening modes identifies the potential glycosylated product and confirms the product formation respectively. The kinetic data of the UGTs was determined via UDP-Glo glycosyltransferase assay which measured the amount of UDP released as a function of time (Chapter 3). Homology modeling was employed in the absence of experimental crystal structures to identify structural differences in these UGTs which drive substrate preferences. Docking of ligand substrates into the model UGTs was done to understand interactions at the molecular level (Chapter 4). Site directed mutagenesis was used to produce mutant UGTs to substantiate the functional roles of potential key amino acids. These mutations were rationally (sequence-based and structure-based) designed (Chapter 5). Results and conclusions: 22 recombinant UGTs from groups L, H and D were selected for substrate screening. 15 of these were successfully expressed while 8 UGTs show glycosylation activity. 76E1 displayed the highest acceptor substrate recognition while both 76E5 and 76E1 showed highest donor recognition. Very low Km at μM scale suggests enzymes good affinity for the donor substrates with 76E5 showing stronger preference for UDP-Gal and UDP-GlcNAc (Chapter 3). Homology models of five group H UGTs were constructed, validated and substrate ligands docked into them. With a focus on donor sugar interactions, key amino acid residues interacting at specific positions of each model UGT were shown. In addition, a major structural difference in N3/Nα3 region of 76E1 was found which may be responsible for its higher acceptor substrate recognition (Chapter 4). 4 The usefulness and predictive power of these models helped design mutant UGTs. Rationally designed mutant UGTs such as 76E2 N320S, 76E4 K275L, 76D1 P129T and 76E2 D374E displayed improved substrate recognition which also highlights the functional roles of those amino acid residues (Chapter 5)