High-Level Expression of Nitrile Hydratase in Escherichia coli for 2-Amino-2,3-Dimethylbutyramide Synthesis

Abstract

In the synthesis of imidazolinone herbicides, 2-Amino-2,3-dimethylbutyramide (ADBA) is an important intermedium. In this study, the recombinant production of nitrile hydratase (NHase) in Escherichia coli for ADBA synthesis was explored. A local library containing recombinant NHases from various sources was screened using a colorimetric method. NHase from Pseudonocardia thermophila JCM3095 was selected, fused with a His-tag and one-step purified. The enzymatic properties of recombinant NHase were studied and indicated robust thermal stability and inhibition of cyanide ions due to substrate degradation. After systematic optimization of fermentation conditions, the OD₆₀₀ (optical density at 600 nm), enzyme activity and specific activity of recombinant strain E. coli BL21(DE3)/pET-28a+NHase reached 19.4, 3.72 U/mL and 1.04 U/mg protein at 42 h, representing 5.86-, 26.6- and 4-fold increases, respectively. These results offered an efficient recombinant whole-cell biocatalyst for ADBA synthesis.

1. Introduction

Imidazolinone herbicides are an important family of herbicidal agents with effective selectivity, broad-spectrum activity and desired environmental safety [1,2]. The first commercial herbicide was developed in the 1980s. They are widely used to manage weeds throughout the world at reasonable prices and their market position was consolidated by the development of herbicide-resistant crops [1,2]. In the synthesis of imidazolinone herbicides, 2-Amino-2,3-dimethylbutyramide (ADBA) is an important intermedium which can be produced by the hydration of 2-amino-2,3-dimethylbutyronitrile (ADBN) via conventional chemical processes or newly developed biological ones [3,4,5].

Chemical hydrations of ADBN are often conducted under harsh conditions, such as high temperature (100 °C) or strong acid or alkali conditions, inevitably rendering them energy-intensive and thermally hazardous [4]. Hydrogen peroxide, sodium hydroxide, concentrated ammonia and sulfuric acid are required in the time-consuming processes [4,5,6]. In addition, side-product formation makes the downstream procedures, including isolation and purification, difficult. Large amounts of organic solvents, such as petroleum ether and ethyl acetate, are used in downstream procedures and cause serious environmental pollution.

Biological methods, using highly efficient enzymes or whole cells with regio- and stereo-selectivity, can be conducted under mild conditions (e.g., ambient temperatures and neutral pHs), [7]. They are eco-friendly and energy-saving and have been considered as promising alternatives to chemical processes [8,9]. Biological hydration of ADBN can be catalyzed by nitrile hydratase (NHase, EC 4.2.1.84) (Figure 1). NHase is one type of metalloenzyme that acts on the triple bond of nitriles and transforms the substrates into valuable amides. Most characterized NHases coordinate either iron or cobalt ions at their catalytic centers. Gene structures of most Co-type NHases contain a β-subunit, an α-subunit and an activator [7]. The catalytic mechanism of NHase has not been fully clarified. It is believed that metal ions at the active site improve the hydration process and aid in the enzyme folding [7].

Whole cells of Rhodococcus boritolerans CCTCCM 208108 harboring NHase were exploited for biosynthesis of ADBA [10,11,12]. In a biphasic system (n-hexane and water), a yield of 91% was reached after 19 h of reaction at 10 °C in fed-batch mode [10]. However, in the original strains, NHases are often accompanied by amidases and nitrilases which could transform the amide products into carboxylic acids [13]. With the rapid development of cloning techniques, recombinant expressions of NHases in host cells, such as Escherichia coli and Pichia pastoris, are expected to reduce amidase and nitrilase contamination. However, a complex posttranslational procedure, including ion insertion and cysteine oxidation, is needed for NHase maturation and its mechanism has not been fully understood yet [7]. Recombinant expression of NHase in E. coli was usually limited by the low expression level as well as the formation of inactive and insoluble proteins [14,15,16]. To overcome these obstacles, ribosome-binding site engineering, molecular chaperone co-expression and tag fusion strategy are employed [7]. Additionally, fermentation conditions are optimized using strategies such as auto-induction and fed-batch culture [7,14].

In this study, a recombinant NHase library constructed in the lab was screened for ADBN hydration. The enzymatic properties of selected recombinant NHase were studied using ADBN as substrate. The culture condition of the selected recombinant strain was optimized to obtain a robust biocatalyst for ADBA synthesis.

2. Materials and Methods

2.1. Materials

Racemic ADBN and ADBA were purchased from Adamas (Shanghai, China) and Topbiochem Technology Co., Ltd. (Shanghai, China), respectively. The components of the fermentation medium were all of industrial grade and purchased from a local Chinese market. The Plasmid Mini kit, gel extraction kit and synthesized oligonucleotides were obtained from Sangon Company (Shanghai, China). Taq DNA polymerase, T4 DNA ligase and restriction enzymes came from Takara Biotechnology Co., Ltd. (Dalian, China). An Unstained Protein Molecular Weight Marker (#26610) came from Thermo Fisher Scientific Inc. A Ni-NTA chelating metal affinity column was purchased from GE healthcare Co., Ltd. (Piscataway, NJ, USA).

2.2. Recombinant NHase Library

The NHase library, containing 24 recombinant NHases from different sources, was reserved in this lab (Table S1). The original nitrile converting strains were obtained from the China General Microbiological Culture Collection Center (CGMCC), the German Collection of Microorganisms and Cell Cultures (DSMZ) or the American Type Culture Collection (ATCC). NHase genes were amplified by PCR using the genomes of corresponding strains as templates and inserted into pET or pQE plasmids, as indicated in Table S1. Recombinant pET and pQE plasmids were transformed into E. coli BL21(DE3) and M15 for protein production, respectively.

2.3. Cultivation Conditions of Recombinant NHase

Glycerol stocks of the recombinant NHase strains were spread on Luria–Bertani (LB) agar plates. Single clones were inoculated into LB medium containing 0.025 g/L kanamycin or ampicillin depending on the plasmid type. After shaking overnight at 37 °C and 200 rpm, 500 μL of seed culture was inoculated to 50 mL LB medium containing 0.025 g/L kanamycin or ampicillin. The cells were then cultivated at 37 °C and 200 rpm until the optical density at 600 nm (OD₆₀₀) reached 0.8–1.0. Then, 0.1 mM isopropyl-β-D-thiogalactopyranoside (IPTG) was added to induce protein production along with 0.5 mM Co²⁺ or Fe³⁺ depending on the type of NHase. After shaking at 20 °C, 200 rpm for 24 h, cells were harvested by centrifugation at 11,644× g for 10 min, then washed twice with phosphate-buffered saline (PBS).

2.4. Screening of the Recombinant NHase Library

The preliminary screening was conducted based on ferrous and ferric ion chromogenic reactions, as described previously [11]. The fresh cells were resuspended in 5 mL PBS buffer (100 mM, pH 7.0) to a final concentration of 20 g/L (wet weight). After preheating at 30 °C for 5 min, 20 μL ADBN was added to start the reaction. After incubation at 30 °C for 1 h, the reaction was stopped by centrifugation at 7378× g for 5 min. Then, 5 μL FeSO₄ (1 M) and FeCl₃ (1 M) solutions were added to 0.3 mL supernatant successively. ADBA production was primitively identified by color change and precipitation formation and further confirmed by gas chromatography (GC) [11].

2.5. Construction of His-Tag Fused NHase

The recombinant strain E. coli BL21(DE3)/pET-21a+NHase containing the NHase gene from Pseudonocardia thermophila JCM3095 was selected for further study. In plasmid pET-21a+NHase, the NHase gene was inserted between the restriction enzyme sites Nde I and Hind III; thus, no His-tag was fused with this NHase (Table S1). To facilitate the protein purification process, the gene was inserted into plasmid pET-28a between the same restriction enzyme sites, so that an N-terminal His-tag could be fused. Plasmid pET-21a+NHase was digested by restriction enzymes Nde I and Hind III to acquire the NHase gene. Empty plasmid pET-28a was digested by the same enzymes to acquire the backbone of plasmid pET-28a. The NHase gene and backbone of plasmid pET-28a were purified using a gel extraction kit and ligated by T4 DNA ligase. The products were transformed into E. coli DH5α competent cells, and a monoclonal colony was selected for sequencing. Plasmid pET-28a+NHase was extracted and transformed into E. coli BL21(DE3) for protein production.

2.6. Recombinant NHase Preparation and Purification

The collected recombinant cells were resuspended in PBS buffer (100 mM, pH 7.0) and lysed by ultrasonic disruption (400 W, 99 cycles, working 5 s and interval 5 s as one, ice-water bath). The insoluble debris was removed by centrifugation at 11,644× g for 20 min and the supernatants were collected for protein purification and activity assay.

The supernatants were applied to the Ni-NTA column equilibrated by 5 column volumes of start buffer (50 mM potassium phosphate, pH 8.0, 0.3 M NaCl). The column was then washed with 5 column volumes of start buffer. The target protein was purified by gradient elution (50 mM potassium phosphate, pH 8.0, 0.3 M NaCl, 10–100 mM imidazole).

The eluted fractions were collected and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Coomassie brilliant blue staining. The gel for SDS-PAGE analysis was prepared using a preparation kit (C631100) from the Sangon Company (Shanghai, China). The concentrations of separating and stacking gel were 10% and 5%, respectively, and 2× Protein Loading Buffer from the Sangon Company (Shanghai, China) was used. Protein content was determined by the Bradford method [17].

2.7. GC Analysis and NHase Activity Assay

GC was performed on Agilent 7820A (Agilent, Palo Alto, CA, USA) equipped with an AT·FFAP column (30 m × 0.25 mm × 0.25 μm; Lanzhou Institute of Chemical Physics, Lanzhou, China). The temperature of the injector and detector was 250 °C. Nitrogen was used as the carrier gas at a flow rate of 1.2 mL min⁻¹. After a 3 min solvent delay time at 150 °C, the oven temperature was increased to 200 °C and then maintained for 5 min. The injection volume was 0.2 μL under a splitless mode.

The activity assay solution contained 80 mM ADBN, 0.1 M PBS (pH 7.0) and a certain amount of enzymes or whole cells in a total volume of 400 μL. After pretreating for 5 min, the reaction was started by the addition of ADBN. After incubation at 35 °C for 6 min, the reaction was terminated by the addition of 10 μL hydrochloric acid (6.0 M). After centrifugation at 10,625× g for 10 min, the supernatant was analyzed by GC to determine the ADBA formation [10,11]. One unit of NHase activity (U) was defined as the amount of whole cells or free enzymes needed to produce 1 μmol of ADBA per minute at 35 °C and pH 7.0. Protein content was determined by the Bradford method [17]. Specific activity was denoted as U/mg protein.

2.8. Characterization of Purified Recombinant NHase

The assay solution contained 80 mM ADBN, 4.1 mg/mL purified enzyme and 0.1 M PBS (pH 7.0). The optimum temperature of NHase was determined by conducting an activity assay at different temperature levels (20–45 °C). The optimum pH of NHase was determined by conducting an activity assay in citric acid and sodium citrate buffer (pH 5.5–6.5), PBS buffer (pH 6.5–8.0) and Tris–HCl buffer (pH 8.0–9.0), respectively. Extra 1 or 5 mM metal ions and ethylene diamine tetraacetic acid (EDTA) were added to the reaction solution to investigate their effect on NHase activity. The purified enzyme was incubated at 10–50 °C for 6 h, and samples were taken at different time points (0.25 h, 0.5 h, 1 h, 2 h, 3 h, 4 h, 6 h) to determine the residual activity. To determine pH tolerance, purified enzymes were incubated at pH 4.0–8.0, 4 °C for 6 h. The buffers used here were citric acid and sodium citrate buffer (pH 4.0–6.0) and PBS buffer (pH 6.5–8.0). Samples were taken at different time points (0.5 h, 2 h, 3 h, 4 h, 6 h) to determine the residual activity. Purified enzyme was incubated in various organic solvents at different concentrations (5, 15 and 50%, v/v) for 2 h to determine organic solvent tolerance.

2.9. Selection of Recombinant Strain

Two recombinant strains (E. coli BL21(DE3)/pET-21a+NHase and E. coli BL21(DE3)/pET-28a+NHase) were cultivated in LB, Terrific Broth (TB), tryptone–yeast (TY) and Super Broth (SB) media to determine the optimum plasmid and medium. The four basic culture media used were as follows: LB, 10 g/L tryptone, 10 g/L sodium chloride, 5 g/L yeast extract; TB, 24 g/L yeast extract, 8 g/L glycerol, 12 g/L tryptone, 17 mM KH₂PO₄, 72 mM K₂HPO₄; TY, 20 g/L tryptone, 10 g/L yeast extract, 5 g/L Na₂HPO₄, 8 g/L glycerol; SB, 32 g/L tryptone, 20 g/L yeast extract, 5 g/L sodium chloride.

2.10. Optimization of Fermentation Conditions

To avoid the formation of inclusion body at high temperatures, the effect of induction temperature (15–30 °C) was studied at first, with 0.1–1.25 mM IPTG or 0.5–4.0% (w/v) lactose tested to determine the optimal inducer. To reduce the process cost, the expensive imported components in basic medium were replaced by domestic raw materials at low cost. The nitrogen and carbon sources (glycerol, tryptone and yeast extract) in the TY medium were substituted with optimized concentrations of selected components. Glycerol (0.8%, w/v) in TY medium was substituted with 0.8% (w/v) of sorbitol, dextrin, starch, sucrose, glucose or lactose. Tryptone (2%, w/v) was substituted with 2.0% (w/v) of Ox bone peptone, fish peptone, peanut powder or corn steep liquor. Imported Oxoid yeast extract (2%, w/v) was substituted with gradient concentrations of domestic Angel yeast extract. Since NHase is a metalloenzyme, extra additions of 0.2 mM metal ions (Li⁺, Ca²⁺, Mg²⁺, Zn²⁺, Mn²⁺, Ni²⁺, Cu²⁺, Fe³⁺ or Fe²⁺) were also studied. The effect of phosphate buffer saline concentration (0–1.25%, w/v) was studied. The best induction time (OD₆₀₀) was determined, too.

The time course of fermentation using the optimal medium was analyzed to determine the time point for feeding. Then, different feeding strategies were tested. At first, 0.8% (w/v) glycerol, 1% (w/v) tryptone and 0.8% (w/v) glycerol plus 1% (w/v) tryptone were added to the medium at 8 and 16 h after induction, respectively. Then, 1% (w/v) tryptone was selected and feeding at more time points (4, 8, 12 and 16 h). Samples were collected at indicated time points to determine OD₆₀₀, enzyme activity, specific activity and productivity. Productivity was denoted as U/mL/h.

3. Results and Discussion

3.1. Screening of the NHase Library

A colorimetric screening method based on ferrous and ferric ions was employed for the primary screening [11]. Since most commercial imidazolinone herbicides are produced with racemic ADBA [3,4,5], the ADBN used for biosynthesis was a racemic mixture [10,11,12]. Among the 24 recombinant NHases in the library (Table S1), only one showed probable activity towards ADBN. Then, the production of ADBA by this recombinant NHase was confirmed by GC [11]. Sequence alignment indicated that this Co-depended NHase came from a thermophile named Pseudonocardia thermophila JCM3095 [18,19].

3.2. Purification of Recombinant NHase

Plasmid pET-28a+NHase was constructed for His-tag fusion so that recombinant NHase could be one-step purified by Ni-NTA column. The collected protein fractions were examined by means of an SDS-PAGE analysis (Figure 2). It was found that NHase was mainly washed down by 75 mM imidazole. The specific activity of purified enzyme was 3.18 U mg⁻¹ protein. In accordance with a previous study, protein bands corresponding to β- and α-subunits (32 and 29 kDa, respectively) were observed [18].

3.3. Characterization of Recombinant NHase

Enzymatic properties of recombinant NHase were studied using ADBN as substrate. The apparent optimum reaction temperature of this enzyme was 35 °C (Figure 3a), while, using acrylonitrile as substrate, the reported value was 60 °C [18]. This should be ascribed to the difference in thermal stability of the substrates. In fact, α-aminonitriles, such as ADBN, are thermolabile, as they decompose spontaneously in aqueous phase to form aldehyde together with hydrogen cyanide in a temperature-dependent manner [10,12]. Released cyanide ions could inhibit NHase through their coordination with metallic ions at the active center [19,20]. Moreover, enzyme activity could be reduced by a decrease in the substrate concentration. Therefore, ADBN hydration at a low temperature was more favorable [10,12]. The optimum reaction pH was 6.5 (Figure 3b), lower than the pH of 7.0 of most reported NHases [21,22]. The low pH could stabilize the α-aminonitrile through the protonation of the highly active amino group [10,12], thus lowering the apparent optimum reaction pH of NHase. Enzyme activity increased significantly after Ni²⁺ addition (strong chelator of cyanide ion) [23] and decreased slightly after EDTA addition (Figure 3c).

After incubation at 50 °C for 7 h, the purified recombinant enzyme retained nearly 70% of its original activity (Figure 4a). This is superior to many other thermostable NHases [24,25]. After incubation at pH 6.0 or 7.0 for 6 h (Figure 4b), the purified enzyme retained more than 70% of its original activity. Its tolerance for organic solvents was investigated (

Table 1. The organic solvent tolerance of recombinant NHase.

Organic Solvent	Relative Activity (%)
	5% (v/v)	15% (v/v)	50% (v/v)
Isooctane	102.59 ± 5.31	101.61 ± 5.26	100.22 ± 4.86
Heptane	95.24 ± 5.45	81.20 ± 3.66	80.88 ± 4.74
Hexane	77.92 ± 4.67	77.05 ± 5.23	75.05 ± 3.93
Pentane	97.00 ± 6.06	95.00 ± 4.87	94.13 ± 5.29
Butyl acetate	72.63 ± 3.86	58.84 ± 3.30	nd 1
Octanol	84.00 ± 4.20	59.84 ± 3.13	57.87 ± 3.16
Dimethyl Sulfoxide	71.13 ± 3.65	59.46 ± 3.10	nd
Cyclohexane	84.88 ± 4.51	83.09 ± 3.89	74.35 ± 4.45
Toluene	81.66 ± 3.70	79.23 ± 4.51	73.22 ± 3.83
Dichloromethane	87.55 ± 4.04	69.28 ± 4.16	54.51 ± 3.33
Isopropyl ether	94.00 ± 5.78	90.00 ± 5.06	74.30 ± 3.63
Ethyl acetate	74.60 ± 3.70	nd	nd
Benzene	87.70 ± 5.54	86.30 ± 4.85	81.00 ± 4.23

¹ nd: not detected.

). This enzyme was stable in solutions containing isooctane, while other organic solvents inhibited it. These results suggested a promising recombinant enzyme with superior thermal stability, while the instability of ADBN substrate should be taken seriously.

3.4. Selection of Recombinant Strain

Two recombinant strains (E. coli BL21(DE3)/pET-28a+NHase and E. coli BL21(DE3)/pET-21a+NHase) were constructed in this lab. The effect of plasmids on enzyme production was studied in four basic media (Figure 5). Much higher total activity and specific activity were obtained with recombinant strain E. coli BL21(DE3)/pET-28a+NHase. This could be mainly ascribed to the higher expression level of target protein with E. coli BL21(DE3)/pET-28a+NHase (Figure S1). The best medium for E. coli strain BL21(DE3)/pET-28a+NHase is TY, in which OD₆₀₀, enzyme activity and specific activity were 12.12, 0.63 U/mL and 0.37 U/mg protein, respectively, much higher than the 3.31, 0.14 U/mL and 0.26 U/mg protein obtained with LB medium. High-level production of recombinant protein was deemed a metabolic stress for host cells. TB or TY media rich in essential nutrients that could support cell growth and protein production were more favorable for recombinant protein expression compared with LB [26,27,28].

3.5. Optimization of Induction Temperature and Inducers

When higher induction temperatures (>20 °C) were employed, cell growth was inhibited. A lower induction temperature (15 °C) was favorable for recombinant protein production in view of higher specific activity but detrimental to biomass accumulation. So, the best induction temperature was 20 °C (Figure 6a). IPTG concentration had no significant effect on cell growth and protein production (Figure 6b). The change of IPTG to lactose had negative effects on bacterial growth, enzyme activity and specific activity (Figure 6c). This was different from our previous reports, in which lactose worked well for recombinant enzyme production [26,27,28], which suggested that the culture conditions are strain- or gene-specific to some extent. Lactose could be utilized as a carbon source as well as an inducer; excessive carbon sources might reduce the culture pH and slow down cell growth [26,27,28]. So, a relatively low concentration of IPTG (0.1 mM) was employed in the following experiments in view of process cost.

3.6. Optimization of Medium Components

Optimization of medium is always required to maximize the product yield for an industrial fermentation process. On the other hand, medium cost significantly affects overall process economics. The replacement of expensive components with raw materials of low cost is often tested [26,27,28].

In most cases, the substitution of glycerol in TY medium with other carbon resources had negative effects on cell growth, total activity and specific activity (Table S2). The only exception was sorbitol, in which the total activity increased from 0.63 to 0.88 U/mL. As cell growth was not significantly affected, this should be mainly ascribed to the increase of specific activity. Sorbitol is reported as a stable factor for some NHases, but hard to utilize for some microorganisms [29]. Further study suggested that the best concentration of sorbitol was 1.6% (w/v) (Figure 7a).

Oxoid tryptone in the TY medium constituted the majority of the medium cost. It was substituted by various nitrogen resources at a concentration of 2% (w/v) (Table S3). In most cases, the substitutions had negative effects, especially on cell growth, indicating the lack of growth-limiting nutrient. Compared with its counterparts, such as fish peptone or peanut powder, the carefully selected raw materials and strict production process increase the price of Oxoid tryptone but also guarantee high quality. Therefore, Oxoid tryptone was conserved in the medium. The best concentration of tryptone was 2% (w/v) (Figure 7b). Oxoid yeast extract is another important component in TY medium. To reduce the process cost, the expensive imported Oxoid yeast extract was substituted by homemade Angel yeast extract. Gradient concentrations of Angel yeast extract were tested (Table S4). Though the cell growth was not affected, the total activity decreased by about 40%. Therefore, the Oxoid yeast extract was also retained, and its optimal concentration was 1% (w/v) (Figure 7c). These results were different from our previous reports, in which the domestic materials of low cost worked well [26,27,28]. The posttranslational maturation of NHases involved complex steps, including ion insertion and cysteine oxidation, which might have contributed to the preference for high-quality nitrogen sources.

Phosphate is one of the basic elements in E. coli. An appropriate concentration of phosphate could maintain osmotic pressure and contribute to plasmid stability and protein production. The best concentration of PBS in TY medium was 0.25% (w/v) (Figure 7d), while excess Na₂HPO₄ inhibited bacterial growth and enzyme production due to high osmotic pressure.

Since this NHase is a Co-dependent metalloenzyme, the addition of other metal ions, such as iron and copper, were tested, but no positive results were observed (Table S5). The effects of induction time on cell growth and NHase expression were also studied (Table S6). The best induction time was 0.8–1.0 (OD₆₀₀).

3.7. Optimization of Feeding Strategy

After the component optimization, the time course of the fermentation process with the optimized medium was explored (Figure 8a). We found that process productivity began to decrease at 8 h after induction and finally reduced to 0.07 U/mL/h at 32 h, indicating a shortage of essential nutrients. To resume the process productivity, three types of medium were used for feeding at 8 and 16 h after induction (Table S7). The best result was obtained when the nitrogen source tryptone was used alone; the process productivity was 0.093 U/mL/h at 36 h. Carbon-source feeding is favorable for enzyme production at the early stage of fermentation, but rapid recession of enzyme activity was observed in the late period; the process productivity was 0.07 U/mL/h at 36 h. Then, nitrogen-source feeding was examined at more time points (4, 8, 12 and 16 h). The process productivity increased from 0.05 to 0.07 U/mL/h at 48 h. OD₆₀₀, enzyme activity and specific activity reached 19.4, 3.72 U/mL and 1.04 U/mg at 42 h, respectively (Figure 8b), indicated that the strategy of resuming process productivity is efficient. Compared with the results before optimization, these values increased 5.86-, 26.6- and 4-fold, respectively.

Hydration of ADBN to ADBA using this whole-cell biocatalyst was performed in a HFE-7100/H₂O (v/v, 10%) biphasic system [30]. As a water immiscible solvent, HFE-7100 served as a reservoir of ADBN and actually reduced ADBN decomposition in the aqueous phase. Moreover, amidase contamination was reduced with the recombinant strain. The average ADBA yield of the entire batch reaction reached 97.3%, higher than the previously reported 91% using original strains [10]. All these results suggested a robust whole-cell catalyst for ADBA synthesis.

4. Conclusions

In this study, a local library containing recombinant NHases was screened for ADBA biosynthesis using a colorimetric method. Recombinant NHase from Pseudonocardia thermophila JCM3095 was selected and ADBA formation was further confirmed by GC. This recombinant NHase was fused with a His-tag and one-step purified. Enzymatic properties of purified NHase were studied. Robust thermal stability suggested a promising biocatalyst for industrial purposes. Inhibition of cyanide ions due to spontaneous substrate degradation was also observed. Multilevel optimization of the fermentation process was then performed. The combination of recombinant strain E. coli BL21(DE3)/pET-28a+NHase and TY medium was chosen. After systematic optimization, the best medium for NHase expression contained 1.6% (w/v) sorbital, % (w/v) tryptone, 1% (w/v) Oxoid yeast extract and 0.25% (w/v) Na₂HPO₄. Protein production was induced by 0.1 mM IPTG along with 0.5 mM Co²⁺ at 20 °C when the cell density reached 0.8–1.0 (OD₆₀₀). Then, the feeding strategy at shake flask-level was explored; 1% (w/v) tryptone and 0.2 mM Co²⁺ were used for feeding at 4, 8, 12 and 16 h after induction. Under the optimized conditions, the OD₆₀₀, enzyme activity and specific activity reached 19.4, 3.72 U/mL and 1.04 U/mg, which represented 5.86-, 26.6- and 4-fold increases, respectively. In conclusion, this article presents a robust recombinant whole-cell catalyst for biosynthesis of ADBA.