Genome-wide identification and biochemical characterization of the UGT88F subfamily in Malus x domestica Borkh

Abstract

The UDP-glycosyltransferase UGT88F subfamily has been described first in Malus x domestica with the characterization of UGT88F1. Up to now UGT88F1 was one of the most active UGT glycosylating dihydrochalcones in vitro. The involvement of UGT88F1 in phloridzin (phloretin 2′-O-glucoside) synthesis, the main apple tree dihydrochalcone, was further confirmed in planta. Since the characterization of UGT88F1, this new UGT subfamily has been poorly studied probably because it seemed restricted to Maloideae. In the present study, we investigate the apple tree genome to identify and biochemically characterize the whole UGT88F subfamily. The apple tree genome contains five full-length UGT88F genes out of which three newly identified members (UGT88F6, UGT88F7 and UGT88F8) and a pseudogene. These genes are organized into two genomic clusters resulting from the recent global genomic duplication event in the apple tree. We show that recombinant UGT88F8 protein specifically glycosylates phloretin in the 2′OH position to synthetize phloridzin in vitro and was therefore named UDP-glucose: phloretin 2′-O-glycosyltransferase. The Km values of UGT88F8 are 7.72 μM and 10.84 μM for phloretin and UDP-glucose respectively and are in the same range as UGT88F1 catalytic parameters thus constituting two isoforms. Co-expression patterns of both UGT88F1 and UGT88F8 argue for a redundant function in phloridzin biosynthesis in planta. Contrastingly, recombinant UGT88F6 protein is able to glycosylate in vitro a wide range of flavonoids including flavonols, flavones, flavanones, chalcones and dihydrochalcones, although flavonols are the preferred substrates, e.g. Km value for kaempferol is 2.1 μM. Depending on the flavonoid, glycosylation occurs at least on the 3-OH and 7-OH positions. Therefore UGT88F6 corresponds to an UDP-glucose: flavonoid 3/7-O-glycosyltransferase. Finally, a molecular modeling study highlights a very high substitution rate of residues in the acceptor binding pocket between UGT88F8 and UGT88F6 which is responsible for the enzymes divergence in substrate and regiospecificity, despite an overall high protein homology.