An Artificial Pathway for N-Hydroxy-Pipecolic Acid Production From L-Lysine in Escherichia coli

N-hydroxy-pipecolic acid (NHP) is a hydroxylated product of pipecolic acid and an important systemic acquired resistance signal molecule. However, the biosynthesis of NHP does not have a natural metabolic pathway in microorganisms. Here, we designed and constructed a promising artificial pathway in Escherichia coli for the first time to produce NHP from biomass-derived lysine. This biosynthesis route expands the lysine catabolism pathway and employs six enzymes to sequentially convert lysine into NHP. This artificial route involves six functional enzyme coexpression: lysine α-oxidase from Scomber japonicus (RaiP), glucose dehydrogenase from Bacillus subtilis (GDH), Δ1-piperideine-2-carboxylase reductase from Pseudomonas putida (DpkA), lysine permease from E. coli (LysP), flavin-dependent monooxygenase (FMO1), and catalase from E. coli (KatE). Moreover, different FMO1s are used to evaluate the performance of the produce NHP. A titer of 111.06 mg/L of NHP was yielded in shake flasks with minimal medium containing 4 g/L of lysine. By this approach, NHP has so far been produced at final titers reaching 326.42 mg/L by 48 h in a 5-L bioreactor. To the best of our knowledge, this is the first NHP process using E. coli and the first process to directly synthesize NHP by microorganisms. This study lays the foundation for the development and utilization of renewable resources to produce NHP in microorganisms.


INTRODUCTION
Plant metabolites play an important role in plant defense, since they can directly harm attacking pathogens, preventing pathogens from entering plant tissues (Chezem et al., 2017). Systemic acquired resistance (SAR) is one of the inducible forms of defense in plants, which produces long-lasting and broad-spectrum immunity against secondary infections at remote locations other than the initial infection site (Fu and Dong, 2013). In dicotyledonous and monocotyledonous plants, enhanced disease resistance is very effective in suppressing infections caused by biological and semibiological vegetative pathogens . Many compounds have been identified as mobile signals of SAR, including salicylic acid (Metraux et al., 1990), methyl salicylate, pipecolic acid (Pip) (Ding et al., 2016), azelaic acid (Vicente et al., 2012), glycerol-3-phosphate (Chanda et al., 2011), nicotinamide adenine dinucleotide, N-hydroxy-pipecolic acid (NHP) (Guerra and Romeis, 2020), nicotinamide adenine dinucleotide phosphate, and dehydroabietinal (Chaturvedi et al., 2012).
Recent research have shown that the lysine-derived metabolites Pip and SAR signal molecule NHP are key activators of the Arabidopsis system immunity, and their endogenous biosynthesis are essential for pathogen-induced SAR (Hartmann and Zeier, 2018). Pip is a cyclic amino acid, a common lysine metabolite in plants and animals (Pérez-García et al., 2019). Pip is an important precursor of some drugs, such as bupivacaine, rapamycin (Eggeling and Bott, 2015), and sandramycin (He, 2006). Importantly, Pip was a compatible solute for Escherichia coli, Silicibacter pomeroyi, and Sinorhizobium meliloti (Gouesbet et al., 1994;Gouffi et al., 2000).
The natural occurrence of SAR signal molecule NHP in organisms was first discovered in Arabidopsis (Hartmann and Zeier, 2018). NHP accumulated significantly after Pseudomonas syringae infection in locally inoculated Arabidopsis (Hartmann and Zeier, 2018). The biosynthesis of NHP in Arabidopsis involved AGD2-like defense response protein 1 (ALD1), SARdeficient 4 (SARD4), and flavin-dependent monooxygenase 1 (FMO1) (Hartmann and Zeier, 2019). External application of NHP can improve the disease resistance of crops and achieve the purpose of disease prevention and control (Holmes et al., 2019). The phylogenetic analysis of the open genome based on the sequence information of Arabidopsis shows that homologs of the core genes of the NHP biosynthesis pathway are widespread throughout the plant kingdom (Holmes et al., 2019). Holmes et al. (2019) found that NHP accumulated in tobacco and tomato when infected with DC3000 bacteria. Therefore, monocots and dicots have the ability to trigger the biosynthesis of NHP in response to the infection of exogenous microorganism.
H 2 O 2 could be exported from the cell, but its rapid production could lead to excessive accumulation of H 2 O 2 in host cells, which would bring harmful effects (Korshunov and Imlay, 2010). Lv et al. (2018) reported that the activity of vanillyl alcohol oxidase PsVAO was significantly decreased over 20 µM of H 2 O 2 by incubating PsVAO in H 2 O 2 . Hossain et al. (2016) established a whole-cell biotransformation system to form pyruvic acid in E. coli expressing L-amino acid deaminase. Finally, 14.57 g/L of pyruvic acid was achieved in the biocatalytic system (Hossain et al., 2016). Zhou et al. (2021) introduced a catalase ScCTA1 to scavenge H 2 O 2 to further improve the titer of coniferyl alcohol to 53.90 g/L. Moreover, some catalases from Saccharomyces cerevisiae are introduced into Hansenula polymorpha and Pichia pastoris to scavenge harmful H 2 O 2 (Gellissen et al., 1996;Payne et al., 1997).
In this work, an artificial route for the production of SAR signal molecule NHP with RaiP, DpkA, GDH, LysP, FMO1, and catalase (KatE) overexpression was first successfully established in E. coli, as seen in Figure 1. First, lysine was oxidatively deammoniated, cyclized spontaneously, and reduced to generate Pip by RaiP and DpkA . Subsequently, Pip was hydroxylated by FMO1 to form NHP. Engineering NHP biosynthesis into microorganism would be an attractive approach to enhance a plant's exogenous ability to respond to pathogens.

Culture Medium and Conditions
Escherichia coli ML03 cells carrying the corresponding plasmids were cultured in 100-ml flasks containing 10 ml of Luria-Bertani (LB) medium (containing 10 g/L of tryptone, 5 g/L of yeast extract, and 10 g/L of NaCl) with appropriate antibiotics for 12 h at 37 • C and 220 rpm. Then 20 µl of the culture was transferred into a 100-ml flask containing 20 ml of medium (composed of 15 g/L of glucose, 10 g/L of tryptone, 5 g/L of yeast extract, 0.5 g/L of K 2 PO 4 ·3H 2 O, 3 g/L of KH 2 PO 4 , 2.5 g/L of (NH 4 ) 2 SO 4 , 0.5 g/L of MgSO 4 ·7H 2 O, 0.1 g/L of FeCl 3 , 2.1 g/L of citric acid·H 2 O, 40 mM α-ketoglutarate, and 100 µg/ml of Amp and/or 50 µg/ml of kanamycin) at 37 • C and 220 rpm. After the OD 600 reached 0.6, 0.5 mM IPTG was added, and culture was continued at 30 • C for 48 h. L-lysine, Pip, and NHP were measured by HPLC.

Biotransformation of N-Hydroxy-Pipecolic Acid in a 5-L Bioreactor
The biotransformation of engineering strains for producing NHP were conducted in a 5-L bioreactor. The formula of the medium was 40 g/L of glucose, 7.5 g/L of K 2 HPO 4 ·3H 2 O, 1.6 g/L of (NH 4 ) 2 SO 4 , 1.6 g/L of MgSO 4 ·7H 2 O, 0.00756 g/L of FeSO 4 ·7H 2 O, 2 g/L of citric acid, 0.02 g/L of Na 2 SO 4 , 0.0064 g/L of ZnSO 4 , 40 mM α-ketoglutarate, 0.0006 g/L of Cu 2 SO 4 ·5H 2 O, 0.004 g/L of CoCl 2 ·6H 2 O, 100 µg/ml of Amp, and 50 µg/ml of kanamycin. For the biotransformation in a 5-L bioreactor, an E. coli colony was seeded in 25 ml of LB medium in a 250-ml flask and incubated at 37 • C and 220 rpm for 12 h. A 1-L flask containing 280 ml of LB medium was inoculated with the preculture and incubated at 37 • C for 12 h with shaking at 220 rpm. The medium [2.8 L in a 5-L fermenter (Baoxing, Shanghai, China)] was then inoculated with an aliquot (10%) of the seed culture for biotransformation at 30 • C. The pH of the medium was adjusted at 6.8 by NH 3 ·H 2 O during the biotransformation process. The biotransformation was carried out at 30 • C. After the OD 600 reached 18, 0.5 mM IPTG and 40 g/L of lysine were added. The samples were taken at specific time intervals, and the residual lysine, Pip, and NHP were analyzed using HPLC.

Analytical Methods
Lysine, Pip, and NHP were analyzed and quantitated by HPLC (Agilent Technologies 1200 series, Hewlett-Packard). The sample was derived with phenyl isothiocyanate (PITC) for the detection of lysine (Cheng et al., 2020). Pip and NHP contents were analyzed by HPLC with a Chirex R 3126 (D)penicillamine LC column (4.6 × 250 mm) as described by  Cheng et al. (2018). Samples were centrifuged and filtered using a 0.22-µm membrane.

Construction of a Heterogeneous Pathway for N-Hydroxy-Pipecolic Acid Production in Escherichia coli
In this study, a functional heterogeneous route for NHP production from lysine was first established in E. coli by a multienzyme expression system. The biosynthesis route is employed in Figure 1. Strain ML07 for the production of NHP precursor Pip was constructed previously . The designed heterogeneous route of NHP consists of two steps. The first step was to convert lysine into Pip in strain ML07 with coexpression of RaiP, DpkA, GDH, and LysP. The second step was to convert Pip into NHP mediated by FMO1. Although NHP can be produced in Arabidopsis, tomato (Holmes et al., 2019), and cucumber (Schnake et al., 2020), it has not been reported in microorganisms. Hence, we tried to introduce a plant FMO1 from Arabidopsis (ArFMO1) in combination with Raip, DpkA, GDH, and LysP for NHP production from lysine. First, a plasmid pZA22-Arfmo1 was constructed. Then plasmids pTrc99a-raiP-dpkA-gdh-lysP and pZA22-Arfmo1 were introduced into E. coli ML03 to obtain the engineered strain ML071. As shown in Table 2, we investigated the production of NHP from lysine in strain ML071. When 2 g/L of L-lysine was fed to the recombinant E. coli ML071, 29.36 mg/L of NHP was produced after 12 h ( Table 2). A titer of 86.48 mg/L of NHP could be obtained with 4 g/L of lysine addition after 48 h. Our data clearly proved that the heterogeneous pathway of NHP in E. coli is feasible.

Comparison of N-Hydroxy-Pipecolic Acid Synthesis by Different Flavin-Dependent Monooxygenases in Escherichia coli Strains
Although the function of AfFMO1 has been primarily verified in this study, the synthesis efficiency of NHP was far from satisfactory. It was presumed that other alternative enzymes from other sources might work better. FMO1s from different sources can catalyze the direct formation of NHP from Pip. To evaluate the performance of the four FMO1s in E. coli, four recombinant strains overexpressing different FMO1s, i.e., ArFMO1 from Arabidopsis, BrFMO1 from B. rapa, GmFMO1 from G. max, and ZmFMO1 from Z. mays, were constructed in this work. As shown in Figure 2, the strains ML073 (GmFMO1) and ML074 (ZmFMO1) resulted in even less NHP production than ML071 (ArFMO1) (as 55.78 and 39.88 mg/L, respectively). The results improved greatly in ML072 (BrFMO1), with an enhancement to 111.06 mg/L of NHP, which was nearly 28% higher than ML071 (ArFMO1). The results indicated that BrFMO1 is slightly more active than AfFMO1, GmFMO1, and ZmFMO1 toward Pip (Figure 2). As a control, E. coli ML03 carrying the pTrc99a-raiP-dpkA-gdh-lysP vector did not produce any NHP.
One-Pot Biotransformation of N-Hydroxy-Pipecolic Acid Production at the 5-L Scale However, strain ML072 also produces H 2 O 2 by RaiP, which impairs cells and can further oxidize 2-keto-6-aminocaproate (2K6AC) to 5-aminovalerate (5AVA) as a byproduct (Cheng et al., 2021b). Therefore, we introduced katE encoding a catalase to decompose harmful H 2 O 2 and eliminate its side effects in strain ML08. The biotransformation of NHP in E. coli ML08 was executed in this study, and the results are shown in Figure 3. The data showed that the biomass increased from OD 600 5.38 to 65 between 8 and 24 h. Pip (38.42 g/L) was produced, and 286.42 mg/L of NHP was accumulated after 36 h. Further increasing the fermentation time to 48 h could obtain a little higher NHP titer of 326.42 mg/L. At the same time, 36.98 g/L of Pip remained as a byproduct, indicating that the catalytic efficiency of BrFMO1 is not high and needs to be further improved. Under optimal biotransformation conditions, 326.42 mg/L of NHP was obtained in a 5-L bioreactor from 40 g/L of L-lysine in 48 h.
Although H 2 O 2 is distributed in almost all organisms, when its concentration exceeds a certain threshold, it could cause adverse effects, such as cytotoxicity and enzyme inhibition (Valderrama, 2010). A high concentration of H 2 O 2 would affect cell growth and limited the production of target products (Niu et al., 2014;Cheng et al., 2021b). Liu et al. (2017) found that catalase KatE was introduced to decompose H 2 O 2 , which increased the titer of α-ketoglutaric acid to 77.4 g/L. The concentration of 5-aminovalerate was further improved by expressing catalase KatE in a novel synthetic pathway in our previous study (Cheng et al., 2021b). Therefore, the elimination of H 2 O 2 in situ is conducive to the production of the target product.

CONCLUSION
In conclusion, we described a biotransformation system utilizing RaiP, DpkA, GDH, LysP, KatE, and BrFMO1 to convert L-lysine to SAR signal molecule NHP in E. coli. This artificial pathway in this work announces a greener and healthier production of SAR signal molecule NHP. An engineered E. coli strain with RaiP, DpkA, GDH, LysP, KatE, and BrFMO1 overexpression can produce 326.42 mg/L of NHP from L-lysine in a 5-L bioreactor. This microbial transformation system utilizes a renewable substrate and has simple culture conditions. As shown in Figure 3, Pip is produced as the main byproduct, which is not an efficient way to produce NHP. So the next goal is to improve the catalytic activity of BrFMO1, then to achieve a more efficient NHP production.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.

AUTHOR CONTRIBUTIONS
ZL and ZW performed the experiments, analyzed the data, and drafted the manuscript. BW, LY, TB, JZ, and HL analyzed the data. ZZ, WW, and JC coordinated the study and finalized the manuscript. All authors contributed to the article and approved the submitted version.