1. Introduction
l-pipecolic acid (l-PA), a non-proteinogenic compound, is a precursor for ropivacaine, bupivacaine, and chloroprocaine, which are widely used in local anesthetics. l-PA is also a precursor for many macrolide-pipecolate natural products including rapamycin, tacrolimus (FK-506), and ascomycin (FK-520), which are indispensable immunosuppressors in the clinic. Furthermore, l-PA is a precursor for the anticancer agent VX-710, piperidine alkaloids, and peptide antibiotics (Cheng et al., 2018; Gatto, Boyne, Kelleher, & Walsh, 2006; Y. H. Kim et al., 2014; Muramatsu et al., 2006; Pérez-García, Max Risse, Friehs, & Wendisch, 2017; Tani et al., 2015; Ying et al., 2017). l-PA is a key chiral intermediate in the synthesis of many important drugs as mentioned above.
Chirality, which is naturally determined in organisms, plays a key role in the pharmaceutical, agricultural, and chemical industries. Especially in the pharmaceutical industry, production and preparation of single enantiomers from chiral intermediates are important, since chiral drugs account for more than half of the drugs produced (Ying et al., 2017). Similar to the preparation of many other enantiomeric organic compounds, a pure enantiomer of l-PA has been prepared by stereoselective transformation, biosynthesis, and chemical procedures, including resolution of racemates and diastereoselective and enantioselective syntheses (Fujii, Aritoku, Agematu, & TSUNEKAWA, 2002; Tani et al., 2015; Ying et al., 2015). However, the chemical procedures for commercial-scale preparation of l-PA are complex, tedious, expensive, low-yielding, and environmentally unfriendly (Cheng et al., 2018; Ying et al., 2017; Ying et al., 2015). Compared to chemical methods, biosynthesis of drug-related chiral intermediates is a more environmentally friendly approach for the preparation of affordable pharmaceuticals with higher efficiency (Ying et al., 2017). Since the biological activity of these compounds is dependent on the stereochemistry, there is an increasing demand for an efficient and applicable method to synthesize enantiomerically pure l-PA.
Previous studies have shown two basic routes from l-lysine to l-PA: the d-1-piperidine-2-carboxylic acid (P2C) pathway and the d-1-piperidine-6-carboxylic acid (P6C) pathway (Fujii, Mukaihara, Agematu, & Tsunekawa, 2002; He, 2006; Ying et al., 2017). In the P2C pathway, the α-amino group of l-lysine is lost, and the ε-nitrogen is incorporated into l-PA. In the P6C pathway, the ε-nitrogen is lost, and the α-nitrogen is incorporated into l-PA (Ying et al., 2017). Typically, five major steps are involved in these two pathways (Fig. 1). The first requires apoptosis-inducing protein (rAIP) and P2C reductase (DpkA) (Cheng et al., 2018; Muramatsu et al., 2006; Tani et al., 2015). The second belonged to P2C pathway involves the enzymes aminotransferase AGD2-like defense response protein 1 (ALD1) and systemic acquired resistance- deficient 4 (SARD4) (Ding et al., 2016; Hartmann et al., 2017; Xu et al., 2018). The third one requires lysine cyclodeaminase (LCD), and this enzyme is the focus of this study. LCD alone can directly catalyze the conversion of l-lysine to l-PA through the P6C pathway and has advantages, including being more suitable for complicated upstream processes because of its simplicity (Cheng et al., 2018; Gatto et al., 2006; Tsotsou & Barbirato, 2007; Ying et al., 2017; Ying et al., 2018; Ying et al., 2015). The fourth one involves l-lysine 6-aminotransferase (LAT) fromFlavobacterium lutescens (IFO 3084 strain) and pyrroline-5-carboxylate (P5C) reductase encoded by proC gene ofE. coli through the P6C pathway (Fujii, Aritoku, et al., 2002; Fujii, Mukaihara, et al., 2002). Finally, another route in the P6C pathway using P5C reductase requires lysine dehydrogenase (LysDH) from Silicibacter pomeroyi . In this route, LysDH converts l-lysine to α-aminoadipic semialdehyde. Then, α-aminoadipic semialdehyde is spontaneously transformed to P6C by cyclization with dehydration (Pérez-García et al., 2017; Pérez-García, Peters-Wendisch, & Wendisch, 2016).
Up to now, two of the five enzymes, rAIP/DpkA and LCD, have been used in high production systems (Table 1). Tani et al. (2015) used purified rAIP/DpkA enzymes to produce 45.1 g/L of pipecolic acid with racemic catalysis. With the same enzymes, Cheng et al. (2018) applied the fed-batch fermentation system using recombinant cadaverine decarboxylase (cadA)-knockout strains containing lysine permease (LysP) and glucose dehydrogenase (GDH) and reported a titer of 46.7 g/L pipecolic acid from l-lysine HCl solution. Ying et al. (2015) produced 17.25 g/L of pipecolic acid using E. coli with LCD as the whole-cell biocatalyst and demonstrated that it has sustained enzymatic activity and self-regenerates NAD+ by cyclization, offering a better alternative to the two-step enzymatic process with rAIP and DpkA. To overcome substrate inhibition, they manipulated the enzyme structure related with the substrate and product delivery tunnels via saturation mutagenesis, resulting in 73.4 g/L of pipecolic acid using whole-cell reaction (Ying et al., 2019).
In this study, we selected LCD because of its advantages for l-PA production and used a reinforced whole-cell system by introducing multiple copies of pipA to improve overall conversion without any complex manipulation of LCD itself. After optimization of the reinforced whole-cell reaction, we report the highest titer (93.5 g/L) to date, increased 5.4-fold over using the same LCD reported previously. A major factor improving l-PA production in the whole-cell system was fine-tuning of gene copy numbers.