ABSTRACT: Lactulose, a synthetic disaccharide, has received increasing interest due to its role as a prebiotic, specifically proliferating Bifidobacilli and Lactobacilli and enhancing absorption of calcium and magnesium. The use of cellobiose 2-epimerase (CE) is considered an interesting alternative for industrial production of lactulose. CE reversibly converts D-glucose residues into D-mannose residues at the reducing end of unmodified ?-1,4-linked oligosaccharides, including ?-1,4-mannobiose, cellobiose, and lactose. Recently, a few CE 3D structure were reported, revealing mechanistic details. Using this information, we redesigned the substrate binding site of CE to extend its activity from epimerization to isomerization.Using superimposition with 3 known CE structure models, we identified 2 residues (Tyr114, Asn184) that appeared to play an important role in binding epilactose. We modified these residues, which interact with C2 of the mannose moiety, to prevent epimerization to epilactose. We found a Y114E mutation led to increased release of a by-product, lactulose, at 65 °C, while its activity was low at 37 °C. Notably, this phenomenon was observed only at high temperature and more reliably when the substrate was increased. Using Y114E, isomerization of lactose to lactulose was investigated under optimized conditions, resulting in 86.9 g/l of lactulose and 4.6 g/l of epilactose for 2 h when 200 g/l of lactose was used.These results showed that the Y114E mutation increased isomerization of lactose, while decreasing the epimerization of lactose. Thus, a subtle modification of the active site pocket could extend its native activity from epimerization to isomerization without significantly impairing substrate binding. While additional studies are required to scale this to an industrial process, we demonstrated the potential of engineering this enzyme based on structural analysis.