Metabolic reprogramming and machine learning-guided cofactor engineering to boost nicotinamide mononucleotide production in Escherichia coli

Abstract

Nicotinamide mononucleotide (NMN) is a bioactive compound in NAD(P)⁺ metabolism, which exhibits diverse pharmaceutical interests. However, enhancing NMN biosynthesis faces the challange of competing with cell growth and disturbing intracellular redox homeostasis. Herein, we boosted NMN production in Escherichia coli by reprogramming central carbon metabolism with a machine learning (ML)-guided cofactor engineering strategy. Engnieering NMN biosynthesis-related pathway directed carbon flux toward NMN with the NADPH level increased by 73 %, which, although enhanced NMN titer (2.45 g/L), impaired cell growth. A quorum sensing (QS)-controlled cofactor engineering system was thus contructed and optimized by ML models to address redox imbalance, which led to 3.04 g/L NMN with improved cell growth. The final strain S344 produced 20.13 g/L NMN in fed-batch fermentation. This study showed that perturbation on cofactor level is a crucial limiting factor for NMN biosynthesis, and proposed a novel ML-guided strategy to manipulate intracellular redox state for efficient NMN production.