There are ongoing
interests in improving the galactooligosaccharide (GOS) synthesis
efficiency of β-galactosidase by protein engineering. In this
study, an intelligent double-hydrophobic amino acid scanning strategy
was proposed and employed to target nine residues forming the glycon-binding
site (−1 subsite) of β-galactosidase Bgal1-3. Two mutants
C510V and H512I with significantly improved GOS synthesis efficiency
were obtained. When 40% (w/v) lactose was used as a substrate, Bgal1-3
reached a maximum GOS yield of 45.3% at 16 h, while the mutants reached
higher yields in a much shorter time (59.1% at 10 h for C510V, 51.5%
at 2 h for H512I). When skim milk was treated with these enzymes,
more GOS was produced (19.9 g/L for C510V, 12.7 g/L for H512I) than that for Bgal1-3
(10.3 g/L) at a lactose conversion of 90%. These results validated
hydrophobicity scanning as an efficient method to engineer β-galactosidases
into promising catalysts for the preparation of GOS and GOS-enriched
milk.