Production of (R)-3-Quinuclidinol by E. coli Biocatalysts Possessing NADH-Dependent 3-Quinuclidinone Reductase (QNR or bacC) from Microbacterium luteolum and Leifsonia Alcohol Dehydrogenase (LSADH)

We found two NADH-dependent reductases (QNR and bacC) in Microbacterium luteolum JCM 9174 (M. luteolum JCM 9174) that can reduce 3-quinuclidinone to optically pure (R)-(−)-3-quinuclidinol. Alcohol dehydrogenase from Leifsonia sp. (LSADH) was combined with these reductases to regenerate NAD+ to NADH in situ in the presence of 2-propanol as a hydrogen donor. The reductase and LSADH genes were efficiently expressed in E. coli cells. A number of constructed E. coli biocatalysts (intact or immobilized) were applied to the resting cell reaction and optimized. Under the optimized conditions, (R)-(−)-3-quinuclidinol was synthesized from 3-quinuclidinone (15% w/v, 939 mM) giving a conversion yield of 100% for immobilized QNR. The optical purity of the (R)-(−)-3-quinuclidinol produced by the enzymatic reactions was >99.9%. Thus, E. coli biocatalysis should be useful for the practical production of the pharmaceutically important intermediate, (R)-(−)-3-quinuclidinol.


Introduction
The chemical processes for obtaining optically pure compounds include enantiomer separation from a racemic mixture, derivation from natural substances and asymmetric synthesis. Of these, asymmetric organic synthesis is the most efficient and useful method of producing chiral synthons. Chiral metal catalysts such as BINAP-Ru [1] and chiral Co (II) salen complex [2,3] have been successfully used as chemical catalysts for synthesizing chiral alcohols or chiral diols from various ketones or epoxides in a number of cases. However, trace metal contamination left in the products and the high cost of catalysts are unresolved difficulties affecting many reactions. To overcome the issues of conventional processes, biocatalytic transformation using enzymes or microorganisms has been applied to the asymmetric reduction of ketones. These biocatalytic processes are more environmentally sustainable and thus more attractive for pharmaceutical manufacturing [4]. In recent years, the asymmetric reduction of ketones with biocatalysts has been reported for the production of chiral alcohols and applied to industry [5][6][7][8][9].
(R)-(−)-3-Quinuclidinol, which has a bicyclic structure with a bridgehead nitrogen, is a valuable intermediate for pharmaceuticals. It has been used as the chiral synthon for a cognition enhancer, a bronchodilator and a urinary incontinence agent [10]. The optical resolution of (±)-3-quinuclidinol esters by the hydrolysis reaction of protease has been reported [11]. Moreover, a number of enzymes have been reported as catalysts for the reduction of 3-quinuclidinone to (R)-(−)-3-quinuclidinol [12][13][14]. To accelerate the bioreduction process, it is necessary to regenerate NAD(P) + to NAD(P)H in situ.
There have been many reports on the reproduction of NAD(P)H via coupling reactions using formate/formate dehydrogenase (FDH) [15] and glucose/glucose dehydrogenase (GDH) [16]. In the regeneration of NAD(P)H, 2-propanol is another suitable hydrogen donor because of its chemical properties and low cost [17,18]. Recently, our research group reported an excellent alcohol dehydrogenase from Leifsonia sp. strain S749 (LSADH) for the synthesis of chiral alcohols and concomitant regeneration of NADH with 2-propanol [19][20][21]. Biocatalytic reaction using whole cells is more stable than isolated enzyme reaction, and can reduce the cost of catalysts and cofactors [21,22]. In this study, we report a bioreduction system for synthesizing (R)-(−)-3-quinuclidinol using a recombinant E. coli cell biocatalyst possessing 3-quinuclidinone reductase (QNR or bacC) from M. luteolum JCM 9174 and LSADH ( Figure 1). The bacC gene, which consists of 768 nucleotides corresponding to 255 amino acid residues and is a constituent of the bacilysin synthetic gene cluster, was obtained by PCR based on homology with known genes [23]. The qnr gene consisted of 759 nucleotides corresponding to 252 amino acid residues [24]. Both enzymes belong to the short-chain alcohol dehydrogenase/reductase (SDR) family. QNR showed 37% homology with the bacC isolated from M. luteolum. The resulting E. coli biocatalyst exhibited a high level of production of (R)-(−)-3-quinuclidinol and enantioselectivity (150 mg/mL, >99.9% e.e.).

Construction of Expression Vector of QNR and LSADH
pET28a (Merck KGaA, Darmstadt, Germany) and pRSFDuet-1 (Merck) vectors have the kanamycin resistance gene. Therefore, it was not suitable to select the colony having both plasmids. Thus, the qnr gene was subcloned into NdeI-HindIII sites of pETDuet-1, which has an ampicillin resistance gene, from pET28-QNR [24] to give pETDuet-QNR. The lsadh gene was amplified by PCR with the following primers and using pKELA [20] as a template: LSADHforNde (5'-GAGATCATATGGCTCAGTACGACGTC-3') (the NdeI site is underlined) and LSADHrevSal (5'-TTTGTCGACTCACTGGGCGGTGTAG-3') (the SalI site is underlined) under the following conditions: 94 °C for 2 min, followed by 98 °C for 10 s, 60 °C for 30 s and 68 °C for 1 min for a total of 30 cycles in accordance with the manufacturer's protocol for KOD FX DNA polymerase (Toyobo, Osaka, Japan). The PCR fragments were digested with NdeI and SalI and inserted into NdeI and XhoI sites of pRSFDuet-1 (Merck) to obtain pRSFDuet-LSADH. For the co-expression of the qnr and lsadh, E. coli BL21 (DE3) was transformed with pETDuet-QNR and pRSFDuet-LSADH. Unfortunately, we could not construct pETDuet-QNR-LSADH.

Construction of Expression Vector of bacC and LSADH
The PCR fragment of lsadh mentioned above was digested with NdeI and SalI, and then cloned into the NdeI and XhoI sites of pETDuet-1 to obtain pETDuet-LSADH. Then, to add the T7 promoter region into the upstream of the lsadh gene, the fragment was amplified again by PCR with the pETDuet-LSADH as a template and the following primers: pETUpstream-69214-3 (5'-ATGCGTCCGGCGTAGA-3') and LSADHrevSal. The amplicon was digested with HindIII and SalI, and cloned into HindIII and XhoI sites of the pET28-bacC [23] to generate pET28-bacC-LSADH, in which the bacC and lsadh genes were connected in this order and each gene was under the control of the T7 promoter ( Figure 2).  Table 1 shows the results of the enzymatic activity of each E. coli biocatalyst. The activities of E. coli BL21(DE3)/pET28-QNR, E. coli BL21(DE3)/pET28-bacC and E. coli BL21/pKELA, in which the gene was independently expressed, indicated higher 3-quinuclidinone-reducing/2-propanol oxidizing activity than E. coli BL21 (DE3)/pETDuet-QNR/pRSFDuet-LSADH and E. coli BL21 (DE3)/pET28-bacC-LSADH, in which the two genes were co-expressed. The results suggested that the combination of two E. coli biocatalysts, for example, E. coli BL21(DE3)/pET28-QNR and E. coli BL21/pKELA, should be suitable for the conversion of 3-quinuclidinone. Moreover, in order to increase the durability of the biocatalyst, we immobilized the recombinant E. coli cells including E. coli BL21(DE3)/pET28-QNR, E. coli BL21(DE3)/pET28-bacC and E. coli BL21/pKELA by coating the cell surface with polyethyleneimine (PI) and glutaraldehyde (GA) [25]. Unfortunately, the activity of immobilized biocatalyst was not determined because the enzyme could not be extracted from the immobilized cells.

Evaluation of E. coli Biocatalyst and Optimization of the Reaction
Based on the enzyme activity shown in Table 1, we applied various combinations of biocatalysts to (R)-(−)-3-quinuclidinol synthesis. Table 2 summarizes the production level of (R)-(−)-3-quinuclidinol and the molar conversion yield. In all cases, the absolute configuration of (R)-(−)-3-quinuclidinol produced was >99.9%. Conversions of almost 100% were attained except for the combination of E. coli/pET28-bacC and E. coli/pKELA (ratio: 1:4), when the final 10% (w/v, 626 mM) of 3-quinuclidinone was employed in the reaction mixture. Low production was due to the insufficient 3-quinuclidinone-reducing activity of this combination of biocatalysts. Neither QNR nor bacC catalyzes the reverse reaction, (R)-(−)-3-quinuclidinol oxidation to 3-quinuclidinon [23,24]; therefore, the unfavorable equilibrium between alcohol and ketone was negligible. However, it was noted that the conversion yield decreased when 15% (w/v, 939 mM) 3-quinuclidinone was employed in the reaction mixture, even if twice the amount of biocatalyst was added. During the course of the study, we observed that more than 5% of 3-quinuclidinone in the reaction mixture seemed to inhibit the reaction rate, suggesting that a high concentration of 3-quinuclidinone has an inhibitory effect on the enzyme catalysts. Thus, we adopted a method of consecutive additions of 3-quinuclidinone (5% w/v) and 2-propanol (5% v/v) to the reaction mixture at 8 h intervals. Under the optimized conditions, combinations of E. coli/pET28-QNR and E. coli/pKELA (4:1), E. coli/pET28-QNR and E. coli/pKELA (1:1) and E. coli/pETDuet-QNR/pRSFDuet-LSADH biocatalysts gave complete conversions of 10% 3-quinuclidinone to (R)-(−)-3-quinuclidinol. We measured the activity and stability of purified QNR and bacC [23,24] at various concentrations to check the effects on the enzymes of 2-propanol and the acetone produced. Polar organic solvents such as 2-propanol and acetone are known to inhibit enzyme activity [25]. QNR showed greater stability than bacC in 2-propanol-KPB (pH 7.0) or acetone-KPB (pH 7.0) medium ( Figure 3). BacC was especially unstable in both 10% mediums. Moreover, the specific activity of bacC in 2-propanol-KPB (pH 7.0) or acetone-KPB (pH 7.0) medium decreased as the concentrations of 2-propanol and acetone increased, while QNR barely decreased (Figure 4). The decline in specific activity of bacC in these media was much greater than that of QNR. Thus, the data confirmed that QNR was superior to bacC for the production of (R)-(−)-3-quinuclidinol in 2-propanol-KPB (pH 7.0) medium. The results also suggested that it is important to keep the concentrations of 2-propanol and acetone lower than 10% during the reaction. Therefore, consecutive additions of 3-quinuclidinone (5% w/v) and 2-propanol (5% v/v) to the reaction mixture at 8 h intervals suppressed the inhibitory effect of the polar organic solvent. In addition, an open reaction system without sealing was adopted to promote the vaporization of acetone from the reaction mixture. Itoh et al. reported that aeration is another effective method in asymmetric bioreduction processes using LSADH with 2-propanol to reduce the concentration of acetone [21].  However, it was impossible to accumulate more than 10% of (R)-(−)-3-quinuclidinol product in the reaction mixture using intact E. coli cells biocatalysts even after consecutive additions of the substrate and 2-propanol ( Table 2).

Evaluation of Immobilized E. coli Biocatalyst and the Conversion Time Course
Immobilization of recombinant E. coli cells often increases the operational stability of biocatalysts in the synthetic reaction. Itoh et al. reported a simple immobilization method for recombinant E. coli biocatalyst (pKELA, LSADH) using PI and GA [26], and revealed its superior operational stability for the bioreduction of 4-hydroxy-2-butanone to (R)-1,3-butanediol in 10% 2-propanol-KPB (pH 7.0) medium [25]. We applied this method for the immobilization of E. coli/pET28-QNR, E. coli/pET28-bacC and E. coli/pKELA. Complete conversion of 15% (w/v) (150 mg, 939 mM) 3-quinuclidinone to (R)-(−)-3-quinuclidinol was attained by a combination of immobilized E. coli/pET28-QNR and E. coli/pKELA, although the reaction time was extended ( Figure 5 and Table 2). Low production of immobilized E. coli/pET28-bacC and E. coli/pKELA was probably due to the loss of bacC activity during the immobilization procedure. We speculated that the PI and GA polymer matrix constructed on the E. coli cell surface has cationic properties due to unreacted amino and imino groups of PI and hinders the access of cationic 3-quinuclidinone to the enzymes in the cells, making it possible to overcome the inhibitory effect of a high concentration of substrate. However, the lower accessibility of substrate to immobilized cells compared with intact cells prolonged the reaction time.
The production level of (R)-(−)-3-quinuclidinol in this study was the highest ever reported, indicating that E. coli biocatalysis would provide a practical method of producing important chiral compounds. Figure 5. Time course of the production of (R)-(−)-3-quinuclidinol by immobilized E. coli biocatalysts. The concentration of (R)-(−)-3-quinuclidinol is represented by a solid line and that of 3-quinuclidinone by a dashed line. Data are the mean value of three independent measurements with standard deviation as shown.

Enzyme Assay
A spectrophotometric assay of 3-quinuclidinone reductase activity was performed by measuring the decrease in absorbance of NADH at 340 nm (ε = 6.22 mM −1 cm −1 ). The assay was performed in a reaction mixture consisting of 10 μmol of substrate, 0.3 μmol of NADH, 50 μmol of KPB (pH 7.0) and 10 μL of enzyme solution in a total volume of 1.0 mL. LSADH activity with 2-propanol (50 mM) was measured in the oxidative reaction. The reaction mixture consisted of 1 μmol NAD + , 50 μmol 2-propanol, 100 μmol KPB (pH 7.0) and 10 μL enzyme solution in a total volume of 1.0 mL. One enzyme unit was defined as the amount of enzyme that converted 1 μmol of NADH or NAD + per min at 25 °C.

Effect of Organic Solvent for Enzymes
The stability of the purified enzymes (500 µg) [23,24] was measured after incubation in 2-propanol/acetone and 100 mM KPB (pH 7.0) medium at various concentrations in a total volume of 1 mL for 2, 4 and 6 h at 25 °C.

Product Analysis by GC
The enantiomer of 3-quinuclidinol was analyzed by a GC system (HP 6890, Hewlett Packard, CA, USA) equipped with a chiral capillary column (CP-cyclodextrin-β-236-N19, 0.25 mm × 25 m, Varian, CA, USA) with a flame ionization detector. The GC conditions were as follows: the column temperature program ramped from 70 °C to 180 °C at 10 °C min −1 , the injection and detection temperatures were 250 °C, and the He flow rate was 3.3 mL min −1 with a linear velocity of 50 cm s −1 and a split ratio of 50. The retention times were 8.17 min for the 1-octanol internal standard, 10.94 min for 3-quinuclidinon, 12.69 min for (S )-(+)-3-quinuclidinol and 12.77 min for (R)-(−)-3-quinuclidinol.