Comprehensive Transformation of Escherichia coli for Nicotinamide Mononucleotide Production

: Nicotinamide mononucleotide (NMN) is a key precursor of nicotinamide adenine dinucleotide and an important source of cellular energy. It can prevent neuronal mitochondrial defects and alleviate heart ﬁbrosis. Strategies to improve NMN production have important implications for human health. Through plasmid expression technology and CRISPR/Cas9 technology, we engineered Escherichia coli for efﬁcient NMN production. First, we upregulated the expression of genes encoding key enzymes in the NMN synthesis pathway, enabling E. coli to directly produce NMN, and established the important role of the nicotinamide mononucleotide transporter in the transport of NMN from cells. The content of NMN was 0.24 g · L − 1 at 24 h. Second, we strengthened the adenosine triphosphate (ATP) cycle, and the concentration of NMN was 0.49 g · L − 1 at 24 h. Third, we increased the synthesis of the NMN precursor 5-phosphate ribose-1-phosphate and obtained an NMN content of 0.49 g · L − 1 at 12 h and 1.11 g · L − 1 at 24 h. Fourth, we introduced nicotinamide riboside kinase (NRK) and found that it was effective only for a period of time. The content of NMN was 0.54 g · L − 1 at 12 h but only 1.05 g · L − 1 at 24 h. Finally, we combined these strategies to regulate the whole metabolic ﬂow, revealing that integrating multiple pathways promoted NMN production. During fermentation, we added 1 g · L − 1 nicotinamide and 10 g · L − 1 glucose, yielding an extracellular NMN concentration of 1.11 g · L − 1 .


Introduction
Nicotinamide mononucleotide (NMN) is the precursor of nicotinamide adenine dinucleotide (NAD + ). β-NMN and NAD + metabolisms are linked [1]. In addition to being directly converted to NAD + , NMN has many physiological functions; for example, it delays cell aging, maintains redox homeostasis, and has therapeutic effects on degenerative diseases and cardio-cerebral diseases [2][3][4][5]. NAD + is a coenzyme of dehydrogenase and participates in hundreds of physiological and biochemical reactions in metabolic pathways. The NAD + content decreases gradually with age and cannot be directly replaced via supplementation. However, the absorption of NMN in the human body is extremely fast. It can enter the blood in approximately 2 min, increase the tissue NMN concentration in 15 min, and rapidly generate NAD + in the blood, liver, and other organs, subsequently exerting physiological functions. Increasing evidence of the beneficial effects of NMN on human health has led to its increased commercial availability as a dietary supplement. Hence, the development of environmentally friendly methods to improve the production of NMN and reduce production costs has gradually become a major focus of research.
Many microorganisms have NMN-related genes and are able to produce NMN [6]; however, the genome of most microorganisms is still largely unknown. Although Escherichia coli does not directly synthesize NMN, it presents a synthetic pathway for 5-phosphate overexpressing three genes, the intracellular content of the substrate PRPP increased, thereby increasing NMN production. To sum up, by increasing the contents of two substrates, ATP supply, and key enzyme expression levels, metabolism in E. coli is expected to flow toward NMN production; this comprehensive strategy involves the regulation of the overall pathway for NMN production rather than the regulation of individual genes [20]. Finally, by adding nicotinamide and glucose during fermentation, we obtained a newly constructed strain that produced 1.11 g·L −1 NMN.

Introduction of Nampt and PnuC Confers NMN Accumulation Ability
E. coli is deficient in Nampt. We optimized the selected gene and cloned it into the pACYCDuet plasmid, which had the Multiple-cloning-site (MCS). We inserted the Nampt gene into MCS after the T7 promoter (a strong promoter) in the pACYCDuet plasmid by single-fragment homologous recombination. The constructed plasmid was transformed into receptive BL21(DE3) cells to obtain strain B1 (Table 1). However, extracellular NMN was undetectable (Figure 2), suggesting that a transporter was needed to transfer NMN out of the cell. After the cellular NMN concentration reaches a certain level, NMN is degraded and transformed via metabolic processes that are not conducive to its accumulation and acquisition. Therefore, the transfer of NMN out of the cell is critical. To address this issue, we added the nicotinamide mononucleotide transporter encoded by the PnuC gene. We selected highly expressed PnuC from Bacillus mycoides [10]. In particular, we connected the PnuC gene to Nampt through the connecting sequences (SD-AS), connected the whole unit to the pACYCDuet plasmid, transformed receptive cells, and obtained strain B2 (Table 1). As determined using high-performance liquid chromatography Metabolic strategies for enhancing nicotinamide mononucleotide (NMN) production in Escherichia coli BL21(DE3). Purple font indicates overexpression of the corresponding gene. "X" indicates the deletion of the relevant gene. Enzymes involved in the reactions are nicotinamide phosphoribosyltransferase (Nampt; derived from Chitinophaga pinensis); nicotinamide mononucleotide transporter (PnuC; derived from Bacillus mycoides); nicotinamide riboside kinase (NRK; derived from Kluyveromyces marxianus); nicotinamide transporter (Niap; derived from Burkholderia cenocepacia); glucose 6-phosphate dehydrogenase (Zwf); 6-phosphogluconate dehydrogenase (Gnd); ribose-phosphate diphosphokinase (Prs); nucleosidase (amn); adenosine deaminase (add); adenosine kinase (Ado1; derived from Saccharomyces cerevisiae). Unlabeled genes are from E. coli. Abbreviations: Nam, nicotinamide; NR, nicotinamide riboside (NR); PRPP, 5-phosphoribosyl 1-pyrophosphate; ATP, adenosine triphosphate; ADP, adenosine diphosphate; AMP, adenosine monophosphate; ade, adenine; ado, adenosine; Hx, hypoxanthine; ino, inosine; PPi, pyrophosphoric acid.

Introduction of Nampt and PnuC Confers NMN Accumulation Ability
E. coli is deficient in Nampt. We optimized the selected gene and cloned it into the pACYCDuet plasmid, which had the Multiple-cloning-site (MCS). We inserted the Nampt gene into MCS after the T 7 promoter (a strong promoter) in the pACYCDuet plasmid by single-fragment homologous recombination. The constructed plasmid was transformed into receptive BL21(DE3) cells to obtain strain B1 (Table 1). However, extracellular NMN was undetectable (Figure 2), suggesting that a transporter was needed to transfer NMN out of the cell. After the cellular NMN concentration reaches a certain level, NMN is degraded and transformed via metabolic processes that are not conducive to its accumulation and acquisition. Therefore, the transfer of NMN out of the cell is critical. To address this issue, we added the nicotinamide mononucleotide transporter encoded by the PnuC gene. We selected highly expressed PnuC from Bacillus mycoides [10]. In particular, we connected the PnuC gene to Nampt through the connecting sequences (SD-AS), connected the whole unit to the pACYCDuet plasmid, transformed receptive cells, and obtained strain B2 (Table 1). As determined using high-performance liquid chromatography (HPLC), the NMN content in the fermentation broth was 0.24 g·L −1 , which proved that the addition of PnuC promoted NMN transfer out of cells ( Figure 2). In the process of NMN production from PRPP and Nam, we added Nam (which can enter the cell spontaneously) via artificial feeding. Based on the results of fermentation experiments, the use of the nicotinamide transporter (Niap; derived from Burkholderia cenocepacia) reduced NMN production by 12.1%; thus, the Niap gene was not introduced.

Improvement in ATP Supply Using CRISPR/Cas9
Increasing the energy supply in cells is a common strategy for increasing the biosynthesis of target products [21,22]. The ATP content is dynamically stable in cells. Our strategy aimed to improve the synthesis level of AMP by regulating AMP-related genes. Because the AEC value is relatively stable, some of the AMP was converted to ATP and ADP. The content of ATP and ADP was ultimately increased [23].
In order to achieve the above purpose, we deleted the adenosine deaminase gene (add; EC 3.5.4.4), which deaminates adenosine to inosine. We identified AMP nuclease in E. coli, which can break down AMP into adenosine. AMP nuclease is encoded by the amn gene; therefore, we knocked out amn (EC 3.2.2.4) based on the knockout of the add gene. Furthermore, adenosine kinase (Ado1; EC 2.7.1.20) in the purine recovery pathway of Saccharomyces cerevisiae can directly convert adenosine into a molecule of AMP and a molecule of ADP by consuming a molecule of ATP; this pathway is shorter than the synthesis pathway in E. coli. Therefore, we used the add gene as the target site to insert the Ado1 gene from S. cerevisiae [24]. To express Ado1 efficiently, we first connected it to the pet28a plasmid, inserted the T7 promoter, and then inserted it into the add site together with the T7 promoter. Finally, we obtained strain A0, in which add and amn were knocked out, and Ado1 was inserted with the T7 promoter. We copied the plasmids from B1 and B2 and transformed them into A0 receptive cells, respectively, and obtained the A1 and A2 strains. During experimental verification, the A2 strain produced 0.49 g·L −1 NMN, which was 99.9% higher than that for B2 ( Figure 2). These results show that promoting the ATP cycle can effectively increase NMN production.

Strengthening the PRPP Pathway to Increase NMN Production
Ribose-phosphate diphosphokinase is encoded by the Prs gene. To increase the expression of the Prs gene, we connected it to the pCDFDuet plasmid through homologous recombination and then transformed the plasmid into A2 to obtain A4. However, by observing the state of cell growth, we found that the OD value of cells with elevated Prs gene expression was 12.5% lower than that of cells without enhanced Prs gene expression. This may be due to the strict regulation of PRPP encoded by Prs. The increase in Prs affected the overall metabolic flow, resulting in a slight decrease in the cell growth rate. However,

Improvement in ATP Supply Using CRISPR/Cas9
Increasing the energy supply in cells is a common strategy for increasing the biosynthesis of target products [21,22]. The ATP content is dynamically stable in cells. Our strategy aimed to improve the synthesis level of AMP by regulating AMP-related genes. Because the AEC value is relatively stable, some of the AMP was converted to ATP and ADP. The content of ATP and ADP was ultimately increased [23].
In order to achieve the above purpose, we deleted the adenosine deaminase gene (add; EC 3.5.4.4), which deaminates adenosine to inosine. We identified AMP nuclease in E. coli, which can break down AMP into adenosine. AMP nuclease is encoded by the amn gene; therefore, we knocked out amn (EC 3.2.2.4) based on the knockout of the add gene. Furthermore, adenosine kinase (Ado1; EC 2.7.1.20) in the purine recovery pathway of Saccharomyces cerevisiae can directly convert adenosine into a molecule of AMP and a molecule of ADP by consuming a molecule of ATP; this pathway is shorter than the synthesis pathway in E. coli. Therefore, we used the add gene as the target site to insert the Ado1 gene from S. cerevisiae [24]. To express Ado1 efficiently, we first connected it to the pet28a plasmid, inserted the T 7 promoter, and then inserted it into the add site together with the T 7 promoter. Finally, we obtained strain A0, in which add and amn were knocked out, and Ado1 was inserted with the T 7 promoter. We copied the plasmids from B1 and B2 and transformed them into A0 receptive cells, respectively, and obtained the A1 and A2 strains. During experimental verification, the A2 strain produced 0.49 g·L −1 NMN, which was 99.9% higher than that for B2 ( Figure 2). These results show that promoting the ATP cycle can effectively increase NMN production.

Strengthening the PRPP Pathway to Increase NMN Production
Ribose-phosphate diphosphokinase is encoded by the Prs gene. To increase the expression of the Prs gene, we connected it to the pCDFDuet plasmid through homologous recombination and then transformed the plasmid into A2 to obtain A4. However, by observing the state of cell growth, we found that the OD value of cells with elevated Prs gene expression was 12.5% lower than that of cells without enhanced Prs gene expression. This may be due to the strict regulation of PRPP encoded by Prs. The increase in Prs affected the overall metabolic flow, resulting in a slight decrease in the cell growth rate. However, the NMN content was 0.54 g·L −1 (representing an 11.3% increase over that for A2), indicating that increasing Prs was beneficial to NMN production ( Figure 3).
Catalysts 2023, 13, x FOR PEER REVIEW 6 of 13 the NMN content was 0.54 g·L −1 (representing an 11.3% increase over that for A2), indicating that increasing Prs was beneficial to NMN production ( Figure 3). In addition, glucose conversion to PRPP involves the dehydrogenation and oxidation of glucose 6-phosphate and gluconate 6-phosphate. Glucose-6-phosphate dehydrogenase (EC1.1.1.49) is encoded by Zwf, while 6-phosphogluconate dehydrogenase (EC1.1.1.44) is encoded by Gnd. We used the connecting sequences (SD-AS) to connect Prs, Gnd, and Zwf, and used the same method to connect this to the pCDFDuet plasmid. Finally, the plasmid was transformed into A2 to obtain the A6 strain. When the Prs-Zwf gene was inserted, the NMN content of the constructed strain increased by 79.4% relative to that of the strain without Prs-Zwf. After the Gnd gene was further inserted, the NMN content increased by 126.5% relative to that of the strain without an altered PRPP pathway, which proved the necessity of promoting the PRPP pathway.
After introducing the above genes, we obtained a strain with increased PRPP pathway activity. We cultured this strain and found that when 1 g·L −1 Nam and 10 g·L −1 glucose were added during induction, the NMN yield could reach 1.1 g·L −1 in 24 h, supporting the effectiveness of our approach.

Effect of the Introduction of NRK from Kluyveromyces marxianus on NMN Yield
NR is present in E. coli and can directly synthesize NMN through phosphorylation. Because NR is transformed quickly in the cell, the introduction of NRK alone would not have a significant effect. As NRK consumes ATP, we introduced NRK into strains with increased ATP and NMN production capacities. NRK was connected to the pET-28a plasmid through single-fragment homologous recombination, and the pACYCDuet plasmid was connected with Nampt-SD-AS-PnuC. These two plasmids were transformed into the A0 strain to obtain A9. After shake-flask fermentation, the NMN content of strain A9 at 12 h was 111.2% of that of the control strain, while the NMN content of strain A9 at 24 h was only 94.8% of that of the control strain ( Figure 4). These results indicate that the addition of NRK increases the synthesis of NMN for a certain period of time; however, over time, the presence of NRK may affect metabolic flow, resulting in a decrease in the total yield In addition, glucose conversion to PRPP involves the dehydrogenation and oxidation of glucose 6-phosphate and gluconate 6-phosphate. Glucose-6-phosphate dehydrogenase (EC1.1.1.49) is encoded by Zwf, while 6-phosphogluconate dehydrogenase (EC1.1.1.44) is encoded by Gnd. We used the connecting sequences (SD-AS) to connect Prs, Gnd, and Zwf, and used the same method to connect this to the pCDFDuet plasmid. Finally, the plasmid was transformed into A2 to obtain the A6 strain. When the Prs-Zwf gene was inserted, the NMN content of the constructed strain increased by 79.4% relative to that of the strain without Prs-Zwf. After the Gnd gene was further inserted, the NMN content increased by 126.5% relative to that of the strain without an altered PRPP pathway, which proved the necessity of promoting the PRPP pathway.
After introducing the above genes, we obtained a strain with increased PRPP pathway activity. We cultured this strain and found that when 1 g·L −1 Nam and 10 g·L −1 glucose were added during induction, the NMN yield could reach 1.1 g·L −1 in 24 h, supporting the effectiveness of our approach.

Effect of the Introduction of NRK from Kluyveromyces marxianus on NMN Yield
NR is present in E. coli and can directly synthesize NMN through phosphorylation. Because NR is transformed quickly in the cell, the introduction of NRK alone would not have a significant effect. As NRK consumes ATP, we introduced NRK into strains with increased ATP and NMN production capacities. NRK was connected to the pET-28a plasmid through single-fragment homologous recombination, and the pACYCDuet plasmid was connected with Nampt-SD-AS-PnuC. These two plasmids were transformed into the A0 strain to obtain A9. After shake-flask fermentation, the NMN content of strain A9 at 12 h was 111.2% of that of the control strain, while the NMN content of strain A9 at 24 h was only 94.8% of that of the control strain ( Figure 4). These results indicate that the addition of NRK increases the synthesis of NMN for a certain period of time; however, over time, the presence of NRK may affect metabolic flow, resulting in a decrease in the total yield of NMN. Therefore, the fermentation time should be taken into account when considering the utility of NRK. of NMN. Therefore, the fermentation time should be taken into account when considering the utility of NRK.

Multiple Strategies for Improving NMN Production in E. coli
Nampt expression and the timely transfer of NMN out of cells promoted the synthesis of NMN. Increasing the expression of genes involved in the PRPP synthetic pathway enhanced the content of the substrate PRPP and had a beneficial impact on the NMN output. By regulating genes involved in the adenine metabolic pathway, the ATP cycle was strengthened, which also increased the synthesis of NMN. Finally, we found that the introduction of NRK was conducive to the generation of NMN for a certain period of time. All of the above strategies were beneficial when carried out alone. However, metabolic processes are highly complex, and the regulation of a single factor or pathway may not be optimal. Thus, we further designed a strategy to regulate the whole metabolic flow.
First, we expressed Nampt in wild-type E. coli and introduced PnuC. The strain could produce NMN and transfer it out of the cell, demonstrating the roles of Nampt and PnuC. In the second step, we knocked out the add and amn genes from the strain obtained in the first step and transferred the Ado1 gene from Saccharomyces cerevisiae to increase the ATP levels, effectively increasing NMN production. Third, we introduced the pCDFDuet plasmid with Prs-SD-AS-Gnd-SD-AS-Zwf into the strain obtained in the second step. This increased the synthesis of PRPP, the precursor of NMN, and finally increased the NMN yield. In the fourth step, based on the third step, the NRK gene was introduced. The addition of 2 g·L −1 and 1 g·L −1 Nam in the culture of strain A6 yielded 1.01 g·L −1 and 1.11 g·L −1 of NMN, respectively. This implied that 2 g·L −1 NAM has an inhibitory effect on NMN production. Fermentation culture of each strain was performed by adding 1 g·L −1 Nam and 10 g·L −1 glucose. The production of strain A6 was highest at 24 h ( Figure 5), proving the validity of our overall metabolic pathway modification.

Multiple Strategies for Improving NMN Production in E. coli
Nampt expression and the timely transfer of NMN out of cells promoted the synthesis of NMN. Increasing the expression of genes involved in the PRPP synthetic pathway enhanced the content of the substrate PRPP and had a beneficial impact on the NMN output. By regulating genes involved in the adenine metabolic pathway, the ATP cycle was strengthened, which also increased the synthesis of NMN. Finally, we found that the introduction of NRK was conducive to the generation of NMN for a certain period of time. All of the above strategies were beneficial when carried out alone. However, metabolic processes are highly complex, and the regulation of a single factor or pathway may not be optimal. Thus, we further designed a strategy to regulate the whole metabolic flow.
First, we expressed Nampt in wild-type E. coli and introduced PnuC. The strain could produce NMN and transfer it out of the cell, demonstrating the roles of Nampt and PnuC. In the second step, we knocked out the add and amn genes from the strain obtained in the first step and transferred the Ado1 gene from Saccharomyces cerevisiae to increase the ATP levels, effectively increasing NMN production. Third, we introduced the pCDFDuet plasmid with Prs-SD-AS-Gnd-SD-AS-Zwf into the strain obtained in the second step. This increased the synthesis of PRPP, the precursor of NMN, and finally increased the NMN yield. In the fourth step, based on the third step, the NRK gene was introduced. The addition of 2 g·L −1 and 1 g·L −1 Nam in the culture of strain A6 yielded 1.01 g·L −1 and 1.11 g·L −1 of NMN, respectively. This implied that 2 g·L −1 NAM has an inhibitory effect on NMN production. Fermentation culture of each strain was performed by adding 1 g·L −1 Nam and 10 g·L −1 glucose. The production of strain A6 was highest at 24 h ( Figure 5), proving the validity of our overall metabolic pathway modification.

Discussion
In the current society, NMN has raised interest because of its important role in inhibiting aging. However, chemical synthesis and other methods are overly complex, so we combined two different methods to modify Escherichia coli to produce a high NMN yield. Based on previous studies on the effectiveness of various plasmids [15], we selected several plasmids to increase the expression of genes involved in the NMN synthesis process. We also used CRISPR/Cas technology, initially established by Jiang et al. [25], to modify the strain. Compared with other metabolic modifications, this study achieved a more comprehensive regulation of the whole metabolic process.
To improve NMN synthesis, we introduced Nampt and PnuC. Enzymes from different organisms are expressed to different degrees in E. coli. Shoji et al. [10] compared 10 Nampt enzymes and found that the activity of Nampt from C. pinensis was highest. They also screened out a highly active PnuC. In the present study, Nampt was constructed into the pACYCDuet plasmid. Moreover, the PnuC gene from B. mycoides, which was the best, was connected to Nampt in the pACYCDuet plasmid. The introduction of these two genes ensured the synthesis and transport of NMN. ATP provides energy or phosphate groups for enzymatic reactions via its hydrolysis; accordingly, the regulation of ATP content is expected to affect NMN production. ATP releases energy for chemical reactions through the hydrolysis of its three phosphate bonds. ATP can break a high-energy phosphate bond to become adenosine diphosphate (ADP), which hydrolyzes another high-energy phosphate bond to become adenosine monophosphate (AMP); AMP breaks the last phosphate bond to become phosphate [26]. According to a large number of previous studies, ATP, AMP, and ADP have a dynamic

Discussion
In the current society, NMN has raised interest because of its important role in inhibiting aging. However, chemical synthesis and other methods are overly complex, so we combined two different methods to modify Escherichia coli to produce a high NMN yield. Based on previous studies on the effectiveness of various plasmids [15], we selected several plasmids to increase the expression of genes involved in the NMN synthesis process. We also used CRISPR/Cas technology, initially established by Jiang et al. [25], to modify the strain. Compared with other metabolic modifications, this study achieved a more comprehensive regulation of the whole metabolic process.
To improve NMN synthesis, we introduced Nampt and PnuC. Enzymes from different organisms are expressed to different degrees in E. coli. Shoji et al. [10] compared 10 Nampt enzymes and found that the activity of Nampt from C. pinensis was highest. They also screened out a highly active PnuC. In the present study, Nampt was constructed into the pACYCDuet plasmid. Moreover, the PnuC gene from B. mycoides, which was the best, was connected to Nampt in the pACYCDuet plasmid. The introduction of these two genes ensured the synthesis and transport of NMN. ATP provides energy or phosphate groups for enzymatic reactions via its hydrolysis; accordingly, the regulation of ATP content is expected to affect NMN production. ATP releases energy for chemical reactions through the hydrolysis of its three phosphate bonds. ATP can break a high-energy phosphate bond to become adenosine diphosphate (ADP), which hydrolyzes another high-energy phosphate bond to become adenosine monophosphate (AMP); AMP breaks the last phosphate bond to become phosphate [26]. According to a large number of previous studies, ATP, AMP, and ADP have a dynamic quantitative relationship, which can be described as (ATP+1/2ADP)/(ATP+ADP+AMP), referred to as the adenosine energy charge (AEC). This also means that simply increasing the ATP content may trigger a violent reaction in the cell and initiate rapid ATP consumption and a lack of ATP accumulation [27]. The content of AMP could be increased by regulating AMP-related genes. Due to the fact that the AEC value was dynamically stable, some of the AMP could be converted to ATP and ADP [28][29][30]. We deleted the add and amn genes using the CRISPR/Cas9 gene editing method. Moreover, the Ado1 gene with a T 7 promoter was integrated into the add locus to regulate adenylate metabolism.
We focused on expanding the pool of precursor materials. The addition of 2 g·L −1 Nam caused a decrease in NMN yield; accordingly, we finally adopted the addition of 1 g·L −1 Nam. PRPP is an important metabolite of purine nucleoside and pyrimidine nucleoside synthesis and is also a substrate involved in NMN synthesis; the concentration of PRPP is closely related to the synthesis of NMN. For PRPP, we transformed the pCDFDuet plasmid with Prs-SD-AS-Gnd-SD-AS-Zwf into cells using plasmid expression. The utilization of PRPP does not result in a cellular metabolic burden [31]. The supplementary concentration of glucose was 10 g·L −1 . After fermentation, the remaining glucose was around 6.76 g·L −1 , and the Nam was used up.
NRK from Kluyveromyces marxianus has high activity and was therefore used in this study [32]. Over a certain period, NRK and Nampt of strain A11 havd a cumulative effect, resulting in a 11.2% higher NMN production than strain A6 at 12 h. However, with time, the accumulation of NMN reached levels lower than those in the control strains. We hypothesized that the introduction of NRK changed the carbon flux distribution in the metabolic pathways of E. coli.
Previous research has focused on enzymatic synthesis and screening of highly active key enzymes to enhance the NMN content [8,[32][33][34]. This study applied some of the effective enzymes screened by previous researchers and utilized two metabolic modification techniques, comprehensively regulating the metabolism pathway by multiple strategies. With fewer steps and lower costs, this study constructed a high-yield strain of NMN.

Materials
E. coli DH5α was used as the cloning host for plasmid construction, and E. coli BL21(DE3) served as the host strain for DNA cloning and protein expression. Primer STAR MAX DNA polymerase and T4 DNA ligase were obtained from Takara Bio. Inc. (Dalian, China). ClonExpress II was obtained from Vazyme Bio. Inc. (Nanjing, China). A FastPure Gel DNA Extraction Mini Kit was used for DNA extraction and purification. A FastPure Plasmid Mini Kit was used to extract the plasmid. All genes were codonoptimized for expression in E. coli and synthesized by Talen-Bio Scientific. Sequencing and primer synthesis were performed by Sangon Biotech. The strains and plasmids used in this study are shown in Table 1. The primers are listed in Table 2.

Strains and Medium
The plasmids of the CRISPR/Cas9 gene editing system were provided by the strain storage in our laboratory. The pTargetF plasmid included an sgRNA sequence, an N20 sequence, and multiple restriction sites. The repair template DNA was provided as a recombinant fragment. Plasmid pCas comprised the Cas9 gene, a temperature-sensitive replicon, and an arabinose inducible promoter λ-Red recombinant gene composition [25]. E. coli strains with the pCas plasmid were cultured in Luria-Bertani (LB) medium (5 g·L −1 yeast extract, 10 g·L −1 trypsin, and 10 g·L −1 NaCl) at 30 • C. Other strains used the same medium but were cultured at 37 • C. If the pCas plasmid needs to be discarded, it can be shaken overnight at 42 • C, and 50 mg·L −1 kanamycin, 50 mg·L −1 spectinomycin, and 1 mM Isopropyl β-D-thiogalactopyranoside (IPTG) can be added if necessary. IPTG is used to induce protein expression and eliminate the ptargetF plasmid. The pCDFDuet plasmid and the pACYCDuet plasmid used for traditional plasmid expression contain two T 7 promoters. The strains with these plasmids were cultured in LB medium at 37 • C. Antibiotics and inducers were added as required.

Construction of Plasmids
The genes needed for the experiment were obtained via Polymerase Chain Reaction (PCR) [35]. Mutation of some gene sites was accomplished by designing different primers. The partial fragment connection and target plasmid construction were spliced by overlapping amplification PCR (SOEPCR) and homologous recombination. The Ado1 gene was selected from Saccharomyces cerevisiae, and the plasmid pET-28a-Ado1 was constructed and used as a template for cloning T 7 -Ado1. T 7 -Ado1 was constructed as the homologous repair template of the CRISPR/Cas system. Together with the pTargetF plasmid with a specific sgRNA, they were shocked into the electroreceptor cells. Prs and other genes were constructed into the plasmid by homologous recombination. Except for the plasmids of the CRISPR/Cas system, the others were introduced into the cell by CaCl 2 chemical transformation.

Construction of SgRNAs and DNA Templates
The linearization and homologous recombination of plasmids and target genes were carried out according to the instructions provided by the reagents (see the official website of Vazyme Bio. Inc. (Nanjing, China) for details). Specific sgRNAs, composed of a 20 bp guide sequence and a PAM sequence, were designed according to the sequence of the target gene. Publicly available websites, such as http://chopchop.cbu.uib.no//, accessed on 12 March 2023, were used to design the guide sequences and PAM sequences suitable for gene knockouts in E. coli. The gene knockout plasmid was constructed according to a previously reported protocol [18]. The sgRNAs were inserted into the plasmid using full-plasmid PCR and transformed into E. coli DH5α. To ensure that the plasmid of sgRNA was successfully constructed, the transformants were sent for sequencing. For the repair template, the sequences 500 bp to 600 bp upstream and downstream of the knockout gene were connected to the knock-in gene using overlap extension PCR. The successful connection of the three fragments was verified by nucleic acid gel electrophoresis and sequencing.

Production and Transformation of Receptive Cells
Common receptive cells were generated using the CaCl 2 method. For the production of CRISPR/Cas9 electroreceptor cells, the pCas plasmid was first transformed into receptive E. coli BL21 (DE3) by chemical methods. When the optical density (OD) reached 0.1-0.2, 30 mM arabinose was added for induction, and the cells were cultured to the logarithmic phase. Samples were cleaned using sterilized ultra-pure water and 10% sterilized glycerin and stored in 10% glycerin. The electroreceptor cells were used as soon as possible after preparation. The DNA repair template and pTargetF plasmid with sgRNA were prepared at a ratio of approximately 4:1, and the pTargetF plasmid was about 600 ng. They were added to the 0.2 cm-gap MicroPulser Electroporation Cuvette with the electroreceptor cells for electroporation at a voltage of 2.5 kV for 5 ms. This process was done using an Electroporator (Bio-Rad, Hercules, CA, USA). After electroporation, 1 mL of LB medium was added, followed by recovery for 2 h. The samples were centrifuged at 4000 g for 5 min, and 800 µL was removed. The remaining liquid was added to an agar plate with kanamycin and spectinomycin. Resistance came from the pCas and PtargetF plasmids, respectively. Positive transformants were screened according to antibiotic resistance, and the transformants were identified using colony PCR. To eliminate the PtargetF plasmid, 1 mM IPTG was used. If spectinomycin resistance remained, it could be eliminated through multiple passages after line separation. The elimination of the pCas plasmid was carried out by culturing cells at 42 • C for 12 h, and the culture time could be extended.

Shake-Flask Culture
First, the single colony was separated through plate scribing. The single colony was placed into a test tube containing 5 mL of LB medium. A volume of 0.5 mL of bacterial solution was transferred into a 25 mL medium triangular flask and cultured at 37 • C for 6-12 h. It was cultured for about 1.5-2 h until the bacterial absorbance was 0.6-0.8. Next, 1 mM of IPTG, 10 g·L −1 glucose, and 1 g·L −1 nicotinamide were added. After the bacteria were cultured in a 30 • C shaker for 12-24 h, the supernatant was collected by centrifugation. The NMN concentration in the supernatant was the extracellular NMN concentration.

Analytical Methods
To determine the growth process of bacteria, the growth density of cells was measured using a UV1800 spectrophotometer (Mapada, Shanghai, China). A wavelength of 600 nm was used for the measurements. The content of NMN in the supernatant of the fermentation broth was the extracellular content. This was determined as follows. After the culture was centrifuged at 9000× g for 10 min, the supernatant was collected as the test sample and filtered by 0.22 µM or 0.45 µM filter membranes before HPLC detection. A C18 column (EC-C18 3× 100 mm) was used in the HPLC system; the mobile phase was methanol and KH 2 PO 4 (30 mM, pH 7.5) solution, the ratio was 95:5, the flow rate was 1 mL·min −1 , the injection volume was 5 µL, the column temperature was 30 • C, the ultraviolet detector was used, and the wavelength was 254 nm. The standard purchased drug of NMN was configured into different concentrations to prepare for the standard curve.

Conclusions
We used the CRISPR/Cas9 technology and traditional plasmid expression technology to construct an E. coli strain capable of efficient NMN production. We used common strategies, including an increase in substrate content, overexpression of key enzymes, timely product transfer, and an increase in energy supply. Specifically, these strategies included: (1) introduction of Nampt and PnuC to enable the cell to directly produce NMN; (2) regulation of the supply of ATP; (3) overexpression of the PRPP pathway-related genes; and (4) introduction of the heterologous NRK. These attempts provided some experience