Reshaping Phosphatase Substrate Preference for Controlled Biosynthesis Using a “Design–Build–Test–Learn” Framework

Abstract Biosynthesis is the application of enzymes in microbial cell factories and has emerged as a promising alternative to chemical synthesis. However, natural enzymes with limited catalytic performance often need to be engineered to meet specific needs through a time‐consuming trial‐and‐error process. This study presents a quantum mechanics (QM)‐incorporated design–build–test–learn (DBTL) framework to rationally design phosphatase BT4131, an enzyme with an ambiguous substrate spectrum involved in N ‐acetylglucosamine (GlcNAc) biosynthesis. First, mutant M1 (L129Q) is designed using force field‐based methods, resulting in a 1.4‐fold increase in substrate preference ( k cat / K m ) toward GlcNAc‐6‐phosphate (GlcNAc6P). QM calculations indicate that the shift in substrate preference is caused by a 13.59 kcal mol −1 reduction in activation energy. Furthermore, an iterative computer‐aided design is conducted to stabilize the transition state. As a result, mutant M4 (I49Q/L129Q/G172L) with a 9.5‐fold increase in k cat‐GlcNAc6P / K m‐GlcNAc6P and a 59% decrease in k cat‐Glc6P / K m‐Glc6P is highly desirable compared to the wild type in the GlcNAc‐producing chassis. The GlcNAc titer increases to 217.3 g L −1 with a yield of 0.597 g (g glucose) −1 in a 50‐L bioreactor, representing the highest reported level. Collectively, this DBTL framework provides an easy yet fascinating approach to the rational design of enzymes for industrially viable biocatalysts.