Benzoates are a class of natural products containing compounds of industrial and strategic importance. In plants, the compounds exist in free form and as conjugates to a wide range of other metabolites such as glucose, which can be attached to the carboxyl group or to specific hydroxyl groups on the benzene ring. These glucosylation reactions have been studied for many years, but to date only one gene encoding a benzoate glucosyltransferase has been cloned. A phylogenetic analysis of sequences in the Arabidopsis genome revealed a large multigene family of putative glycosyltransferases containing a consensus sequence typically found in enzymes transferring glucose to small molecular weight compounds such as secondary metabolites. Ninety of these sequences have now been expressed as recombinant proteins in Escherichia coli, and their in vitro catalytic activities toward benzoates have been analyzed. The data show that only 14 proteins display activity toward 2-hydroxybenzoic acid, 4-hydroxybenzoic acid, and 3,4-dihydroxybenzoic acid. Of these, only two enzymes are active toward 2-hydroxybenzoic acid, suggesting they are the Arabidopsis salicylic acid glucosyltransferases. All of the enzymes forming glucose esters with the metabolites were located in Group L of the phylogenetic tree, whereas those forming O-glucosides were dispersed among five different groups. Catalytic activities were observed toward glucosylation of the 2-, 3-, or 4-hydroxyl group on the ring. To further explore their regioselectivity, the 14 enzymes were analyzed against benzoic acid, 3-hydroxybenzoic acid, 2,3-, 2,4-, 2,5-, and 2,6-dihydroxybenzoic acid. The data showed that glycosylation of specific sites could be positively or negatively influenced by the presence of additional hydroxyl groups on the ring. This study provides new tools for biotransformation reactions in vitro and a basis for engineering benzoate metabolism in plants.

The complete sequence of the Arabidopsis genome enables definitive characterization of multigene families and analysis of their phylogenetic relationships. Using a consensus sequence previously defined for glycosyltransferases that use small-molecular-weight acceptors, 107 gene sequences were identified in the Arabidopsis genome and used to construct a phylogenetic tree. Screening recombinant proteins for their catalytic activities in vitro has revealed enzymes active toward physiologically important substrates, including hormones and secondary metabolites. The aim of this study has been to use the phylogenetic relationships across the entire family to explore the evolution of substrate recognition and regioselectivity of glucosylation. Hydroxycoumarins have been used as the model substrates for the analysis in which 90 sequences have been assayed and 48 sequences shown to recognize these compounds. The study has revealed activity in 6 of the 14 phylogenetic groups of the multigene family, suggesting that basic features of substrate recognition are retained across substantial evolutionary periods.

No abstract