2-O-α-D-glucopyranosyl-L-ascorbic acid (AA-2G) is a stable derivative of L-ascorbic acid (L-AA), which has been widely used in food and cosmetics industries. Sugar molecules, such as glucose and maltose produced by cyclodextrin glycosyltransferase (CGTase) during AA-2G synthesis may compete with L-AA as the acceptors, resulting in low AA-2G yield. Multiple sequence alignment combined with structural simulation analysis indicated that residues at positions 191 and 255 of CGTase may be responsible for the difference in substrate specificity. To investigate the effect of these two residues on the acceptor preference and the AA-2G yield, five single mutants Bs F191Y, Bs F255Y, Bc Y195F, Pm Y195F and Pm Y260F of three CGTases from Bacillus stearothermophilus NO2 (Bs), Bacillus circulans 251 (Bc) and Paenibacillus macerans (Pm) were designed for AA-2G synthesis. Under optimal conditions, the AA-2G yields of the mutants Bs F191Y and Bs F255Y AA-2G were 34.3% and 7.9% lower than that of Bs CGTase, respectively. The AA-2G yields of mutant Bc Y195F, Pm Y195F and Pm Y260F were 45.8%, 36.9% and 12.6% higher than those of wild-type CGTases, respectively. Kinetic studies revealed that the residues at positions 191 and 255 of the three CGTases were F, which decreased glucose and maltose specificity and increased L-AA specificity. This study not only proposes for the first time that the AA-2G yield can be improved by weakening the acceptor specificity of CGTase toward sugar byproducts, but also provides new insight on the modification of CGTase that catalyze the double-substrate transglycosylation reaction.

Understanding the mechanisms of enzyme specificity is increasingly important from a fundamental viewpoint and for practical applications. Transglycosylation has attracted many attentions due to its importance in improving the functional properties of acceptor substrates both in vivo and in vitro. Cyclodextrin glucanotransferase (CGTase) is one of the key enzymes in transglycosylation, it has a broad substrate spectrum and utilizes sugar as the donor. However, little is known about the acceptor selectivity of CGTase, which greatly hampers efforts toward the rational design of desirable transglycosylated derivatives. In this study, we found that the CGTase from Bacillus circulans, BcCGTase, was able to form glycosylated products with diverse ginsenosides. In particular, it not only carries out diverse mono-, di-, and even higher-order glycosylations via the transfer of glucose moieties to the COGlc positions, but also can glycosylate the C3-OH position of ginsenosides. In contrast, another CGTase from Bacillus licheniformis (BlCGTase) showed relatively specific acceptor preference with only several ginsenosides. Structural comparison between BcCGTase and BlCGTase revealed that the Arg74/K81 position within the acceptor-binding sites of BcCGTase/BlCGTase was responsible for the differences in catalytic specificity for ginsenoside F1. Further mutagenesis confirmed their roles in the acceptor selection. In conclusion, our study not only demonstrates the acceptor selectivity of CGTases, but also provides insight into the catalytic mechanism of CGTases, which will potentially increase the utility of CGTase for biosynthesis of new, rationally designed transglycosylated derivatives.