Creation of an α-mannosynthase from a broad glycosidase scaffold.

Glycosynthases provide a rare and practical source of catalytic synthetic utility for oligosaccharide and other glycoside synthesis. These biocatalysts, nucleophile-less mutants of retaining glycosidases, are typically exquisitely stereoselective (inverting) enzymes and some can also attach monosaccharides with good regioselectivity. Many examples now exist, but in most cases b-glycosides are synthesized; a-glycosynthases are rare and have not been applied widely to practical synthetic use, and some important linkages in biology, such as a-mannosides, have been inaccessible. In synthesis glycosynthases advantageously obviate some of the typical need in oligosaccharide chemistry for hydroxy protection. However, when glycosynthases are derived from exo-glycosidases (Figure 1a) their mode of action often leads to repetitive condensation of these unprotected substrates, sometimes resulting uncontrollably in the creation of a mixture of varying oligomers as products 17–20] (Figure 1b). This limitation may, at first sight, seem an inherent consequence of reversing the natural exo-glycosidase mechanism (Figure 1a), since an enzyme that evolved to bind the repetitive sugar structure may then readily accommodate product containing this sugar type at its nonreducing terminus. However, we describe herein a strategy for discrete, controlled glycoside synthesis in which a glycosynthase is created that can bind one sugar type as a donor substrate in the 1 (or D) subsite but then bind it less favorably as an acceptor when displayed in any ensuing product (Figure 1d). This strategy has allowed the discovery of a novel a-mannosynthase that does not uncontrollably catalyze oligomerization. Our design necessitated the combination of a glycosidase and a monosaccharide with good mutual affinity at the donor binding site (D site) but poor affinity at the acceptor-binding site (A site). This goal could, in principle, be achieved by identifying a nonpreferred substrate for a glycosidase with plasticity in its D site (Figure 1c,d). Validation for this approach came from the contrasting behavior of the Agrobacterium sp. b synthase towards Gal and Glc substrates. Although a priori prediction of enzyme sugar specificity and plasticity is difficult, Nishio et al. have reported an interesting feature of the glycoside hydrolase family 31 (GH31) in the carbohydrate-active enzyme (CAZy) classification: these hydrolases can display their activity toward both a-glucosides and 2-deoxy derivatives. This activity suggested that the 2OH group of a-d-glucose might be less important for substrate recognition and implied that the epimer of a-dglucose at the 2-OH position, a-d-mannose, might also be accommodated at the D site of these enzymes. If this were true, and d-mannose is also not favorable as an acceptor, our design concept (Figure 1d) might be realized. To test this hypothesis, we selected and expressed GH31 open reading frames (orfs) and discovered previously unknown a-mannosidase activity for the product of GH31 gene malA 25] from Sulfolobus solfataricus. The enzyme s Figure 1. exo-Glycosidase trims oligosaccharides sequentially from nonreducing termini (a), and its glycosynthase can catalyze the reverse reaction, which may result in the formation of unwanted oligomers (b). When the enzyme has plasticity at its D site (c,d), the selection of an appropriate donor could, in principle, prevent oligomerization (d). A site and D site denote acceptor binding site(s) (+1 (,2,3...)) and donor binding site ( 1), respectively. Green circle=preferred glycan substrate residues; yellow motif=donor substrate that binds to the donor site but less favorably to the acceptor site; red triangle indicates the site of hydolysis between the sites 1 and +1.