3 Results and discussion
3.1 Design of NMN biosynthetic pathways
In cells, NMN is an intermediate in NAD+ biosynthesis produced from nicotinamide (NAM) and phosphoribosyl pyrophosphate (PRPP) by nicotinamide phosphoribosyltransferase (Nampt, EC 2.4.2.12) (Lin et al., 2016; Poddar et al., 2019). However, the direct use of PRPP as a substrate for NMN preparation is not acceptable because it is not only expensive but also unstable (Hove-Jensen et al., 2017). To avoid this problem, three possible NMN biosynthesis pathways were designed, in which adenosine, adenosine monophosphate (AMP), or ribose was used as the starting material to generate PRPP, respectively (Figure 2a). In pathway I, adenosine was first converted to AMP by the catalysis of adenosine kinase (Adk, EC 2.7.2.10), and then PRPP and adenine were synthesized through AMP and pyrophosphate (PPi), which was catalyzed by adenine phosphoribosyltransferase (Apt, EC 2.4.2.7). In pathway II, the conversion of adenosine into AMP was omitted, and PRPP was directly generated from AMP and PPi. In pathway III, ribose was converted into PRPP through two catalytic steps. First, ribokinase (Rbk, EC 2.7.1.15) catalyzed the phosphorylation of ribose using ATP as a donor of the phosphate group to generate ribose-5-phosphate (R5P). Second, R5P was pyrophosphorylated by ATP to form PRPP catalyzed by phosphoribosyl pyrophosphate synthetase (Prs, EC 2.7.6.1). The last step for all three pathways was the same. In this step, the synthesized PRPP reacted with NAM to form NMN and by-product PPi, which was catalyzed by Nampt.
After designing three possible biosynthetic pathways for NMN production, we next aimed to verify the feasibility of these pathways. As the thermodynamic analysis was a straightforward and reliable way to determine whether a metabolic reaction or pathway was feasible or not (Flamholz et al., 2012), the standard Gibbs free energy changes (ΔrG’°) for the three pathways were calculated by using the eQuilibrator website (equilibrator.weizmann.ac.il/) at pH 8.0 and 0.05 M of ionic strength. The ΔrG’° of pathways I, II, and III was -4.2 kJ/mol, +11.8 kJ/mol, and -45.7 kJ/mol, respectively (Figure 2b). Although Apt could convert AMP to generate PRPP in one step, the ΔrG’° of this step is +26.3 kJ/mol, which means that this reaction is difficult to occur. On the other hand, the ΔrG’° of all the three steps in pathway III was negative, indicating that the overall reaction of this pathway was thermodynamically favorable. Therefore, the pathway I and II containing Apt were discarded, and the most thermodynamically favorable pathway III with ribose as the co-substrate together with NAM was selected as a promising route for NMN synthesis.
We next investigated the capacity of this pathway for producing NMN experimentally. Two homologs for each reaction step were selected as initial enzyme sets. They were Rbk fromEscherichia coli (Maj & Gupta, 2001) andHomo sapiens (Park et al., 2007), Prs from Pyrobaculum calidifontis (Bibi et al., 2016) andHomo sapiens (Nosal et al., 1993), and Nampt from Meiothermus ruberand Homo sapiens (Hara et al., 2011). All enzymes were expressed in Escherichia coli BL21(DE3) and purified using a Ni-NTA affinity column. By combining different enzyme homologs, each possible combination was constructed using 1 µM of each enzyme (8 unique pathway combinations), and the performance in the synthesis of NMN of each pathway combination was tested. NMN was successfully detected in all 8 pathway combinations (Figure 2c), which strongly suggested that pathway III is feasible for producing NMN. The NMN titers produced by these initial enzyme sets ranged from 12 mg/L to 101 mg/L, and the two highest NMN titers were obtained from the pathway combinations in which both Prs and Nampt are from prokaryotic organisms (that is, Prs from Pyrobaculum calidifontis and Nampt fromMeiothermus ruber ). This result not only implied that prokaryotic enzymes are more likely to be productive for NMN synthesis but also highlighted the importance of testing enzyme homologs to enhance pathway performance.