One-Pot Biosynthesis of l-Aspartate from Maleate via an Engineered Strain Containing a Dual-Enzyme System

l-Aspartate is currently produced from fumarate by biological methods, and fumarate is synthesized from maleate by chemical methods in industry. We established a biosynthesis method to produce l-aspartate from maleate that is environmentally responsible, convenient, and efficient. Compared to conventional l-aspartate production, no separation and purification of intermediate products is required, which could greatly improve production efficiency and reduce costs. As environmental issues are attracting increasing attention, conventional chemical methods gradually will be replaced by biological methods. Our results lay an important foundation for the industrialization of l-aspartate biosynthesis from maleate.

ABSTRACT

l-Aspartate has been widely used in medicine and the food and chemical industries. In this study, Serratia marcescens maleate cis-trans isomerase (MaiA) and Escherichia coli aspartase (AspA) were coupled and coexpressed in an engineered E. coli strain in which the byproduct metabolic pathway was inactivated. The engineered E. coli strain containing the dual-enzyme system (pMA) was employed to bioproduce l-aspartate from maleate with a conversion of 98%. We optimized the activity ratio of double enzymes through ribosome binding site (RBS) regulation and molecular modification of MaiA, resulting in an engineered strain: pMA-RBS4-G27A/G171A. The conversion of l-aspartate biotransformed from maleate using the pMA-RBS4-G27A/G171A strain was almost 100%. It required 40 min to complete the whole-cell catalysis, without the intermediate product and byproduct, compared to 120 min before optimization. The induction timing and the amount of inducer in a 5-liter fermentor were optimized for scale-up of the production of l-aspartate. The amount of produced l-aspartate using the cells obtained by fermentation reached 419.8 g/liter (3.15 M), and the conversion was 98.4%. Our study demonstrated an environmentally responsible and efficient method to bioproduce l-aspartate from maleate and provided an available pathway for the industrial production of l-aspartate. This work should greatly improve the economic benefits of l-aspartate, which can now be simply produced from maleate by the engineered strain constructed based on dual-enzyme coupling.

IMPORTANCE l-Aspartate is currently produced from fumarate by biological methods, and fumarate is synthesized from maleate by chemical methods in industry. We established a biosynthesis method to produce l-aspartate from maleate that is environmentally responsible, convenient, and efficient. Compared to conventional l-aspartate production, no separation and purification of intermediate products is required, which could greatly improve production efficiency and reduce costs. As environmental issues are attracting increasing attention, conventional chemical methods gradually will be replaced by biological methods. Our results lay an important foundation for the industrialization of l-aspartate biosynthesis from maleate.

INTRODUCTION

As environmental issues are attracting increasing attention, conventional chemical methods gradually will be replaced by biological methods, including enzymatic conversion, whole-cell catalysis, and fermentation methods. The biosynthesis of l-aspartate has been developed since the 1970s. Sato et al. carried out the biosynthesis of l-aspartate from fumarate and ammonia through the use of cellular aspartase in immobilized Escherichia coli cells (1). A high-production E. coli strain was obtained by irradiation with UV light, and the production of l-aspartate achieved within 1 h was 0.19 to 0.35 g/min/g (dry mass of cells) (2). The cells were then immobilized and used for continuous production of l-aspartate, and the production of l-aspartate reached 6 g/h/g (dry mass of cells) under the optimal conditions (3). Aspartase catalyzes a reversible conversion from l-aspartate to fumarate and ammonia. Under the alkaline condition, the reaction proceeds toward the synthesis of l-aspartate from fumarate and ammonia (4). For decades, the characteristics of aspartase have been investigated, including its origin, crystal structure, biochemical properties, and catalytic mechanism (5). Aspartase has been found in various bacteria, plants, and animals, and aspartase from E. coli (AspA) is widely employed in industrial applications due to its superiority and cost.

l-Aspartate is currently produced through recombinant cells containing AspA by use of fumarate as the substrate, and fumarate is synthesized from maleate by chemical methods in industry. An alternative biosynthesis of l-aspartate is proposed through enzyme cascades with the use of maleate as the substrate (Fig. 1). Maleate cis-trans isomerase is the enzyme that catalyzes the transformation of maleate into fumarate and was first characterized from Pseudomonas fluorescens in 1969 (6). Subsequently, maleate cis-trans isomerases from Alcaligenes faecalis, Bacteria stearothermophilus, Serratia marcescens, etc., were identified (7–9), and the active sites and catalytic mechanisms were analyzed (10–12). Maleate cis-trans isomerase from Serratia marcescens (MaiA) was found to be suitable for industrial application due to its enzymatic properties and high expression level in recombinant E. coli cells (8), but no subsequent studies were conducted.

FIG 1.

l-Aspartate biosynthesis pathway via a dual-enzyme cascade composed of S. marcescens maleate cis-trans isomerase (MaiA) and E. coli aspartase (AspA).

Artificially designed cascades have attracted scientists’ interest because of their many advantages. Compared to conventional enzymatic biosynthesis, no separation and purification of intermediate products is required, which could greatly improve production efficiency and reduce costs. Additionally, it is usually simpler to design a synthetic cascade than to perform metabolic engineering. Rieckenberg et al. established an in vitro system containing three enzymes, glycerol dehydratase, propanediol oxidoreductase isoenzyme, and hydrogenase I, to transform glycerol into 1,3-propanediol with a yield of 1 mol/mol (13). A dual-enzyme cascade combining an alcohol dehydrogenase with an amine dehydrogenase was applied to aminate aromatic and aliphatic alcohols, yielding up to 96% conversion and 99% enantiomeric excess (14). AspA coupling with l-aspartate-α-decarboxylase from Corynebacterium glutamicum has been constructed into a two-enzyme cascade to efficiently biosynthesize β-alanine from fumaric acid (15). Furthermore, a synthetic pathway coupling five enzymes has been designed to synthesize terephthalic acid using ρ-xylene as a substrate in E. coli (16).

Here, we designed an enzymatic cascade that catalyzes maleate transformation to l-aspartate in an engineered E. coli strain. Two enzymes were employed in our system: MaiA, catalyzing maleate to fumarate, and AspA, catalyzing fumarate to l-aspartate. In cells, the intermediate fumaric acid is converted to malic acid in the tricarboxylic acid (TCA) cycle by fumarase. There are three fumarases (encoded by three genes: fumA, fumB, and fumC) in E. coli. FumA and FumC function in the TCA cycle under aerobic cell growth conditions, while FumB functions during anaerobic growth (17). Our previous study demonstrated that almost no by-product was detected when we used the E. coli BL21 ΔfumA ΔfumC strain overexpressing MaiA; however, malic acid remained at the same level in the fumB gene-deleted strain as that in the original strain (18). In the present study, we overexpressed MaiA and AspA in the E. coli BL21 ΔfumA ΔfumC strain and completed the bioproduction of l-aspartate from maleate. The activity ratio of double enzymes was optimized by ribosome binding site (RBS) regulation. The productivity of biosynthesis was improved through molecular modification of the rate-limiting enzyme.

RESULTS

Construction of an engineered strain containing a dual-enzyme system.

Aiming to biosynthesize l-aspartate from maleate, a dual-enzyme system containing MaiA and AspA was constructed (Fig. 2a). MaiA from S. marcescens (encoded by the gene maiA) was chosen, as it exhibits better catalytic activity and expression level than the other MaiA proteins that have been characterized (6–9). AspA from E. coli (encoded by the gene aspA) has been proven in industrial applications (2). Each gene was located downstream of a P_T7 promoter and an RBS. E. coli BL21 ΔfumA ΔfumC was used as the host cell, in which two genes (fumA and fumC) responsible for the conversion of fumarate to malate were deleted (18). As a result, a recombinant strain (pMA, E. coli BL21 ΔfumA ΔfumC overexpressing maiA and aspA) was constructed by transforming the dual-enzyme plasmid into the host cell.

FIG 2.

Effect of RBS on the expression of MaiA. (a) Schematic diagram of reconstructed expression cassette. Two genes, the promoter (P_T7), and the RBS are represented by gray and black. (b) Relative activities of the cells regulated by four kinds of RBS compared to that regulated by the original RBS. The activity of the cells harboring the original RBS was defined as 100%. (c) SDS-PAGE analysis of cell lysate. Lane M is the standard marker protein. Lane 1, cells harboring the original RBS; lane 2, cells harboring RBS1; lane 3, cells harboring RBS2; lane 4, cells harboring RBS3; lane 5, cells harboring RBS4. The bands corresponding to MaiA and AspA were indicated by arrows separately. (d) Production of l-aspartate using a recombinant strain containing RBS4. The concentrations of l-aspartate (solid circle), fumarate (solid square), and maleate (open circle) were analyzed by HPLC.

The pMA strain was used to bioproduce l-aspartate from maleate (Table 1). The pMA strain successfully catalyzed maleate transformation into l-aspartate, while the conversion of maleate to l-aspartate was maintained above 98%. As maleate increased, the conversion time increased from 120 min to 390 min. The intermediate product fumarate barely remained even when the substrate concentration was increased to 3.2 M, suggesting that AspA had much better catalytic ability than MaiA. Thus, the rate-limiting enzyme was MaiA in this dual-enzyme cascade.

TABLE 1.

Production of l-aspartate from maleate using the engineered strain pMA

Substrate concn (M)	Reaction time (min)	Product concn (M)			Conversion (%)
		Maleate	Fumarate	l-Aspartate
1.6	120	0	0.03	1.57 ± 0.06	98.1
2.4	270	0	0.02	2.36 ± 0.09	98.3
3.2	390	0	0.02	3.15 ± 0.16	98.4

Optimization of RBS sequences for MaiA.

The efficiency of whole-cell catalysis will be enhanced if the expression level of MaiA is higher. We optimized the sequence of RBS to improve the expression level of MaiA using RBS Calculator 2.0 website software. Four RBS sequences with different translation initiation rates were chosen to enhance the MaiA expression, and they are listed in Table 2. The expression levels of recombinant MaiA initiated by different RBS sequences were evaluated using whole-cell catalysis. The cell activity was measured in phosphate buffer (50 mM, pH 8.0) at 40°C, and the activity of cells harboring the original RBS was defined as 100%. As shown in Fig. 2b, the highest activity was exhibited by the cells containing RBS4 (pMA-RBS4). The substrate of 1.6 M maleate was completely converted to l-aspartate within 80 min using the pMA-RBS4 strain (Fig. 2d). The intermediate product of fumarate was still not left, and the conversion reached 99.4%. The productivity was increased 50% through RBS optimization. The expression levels of recombinant enzymes were observed by SDS-PAGE (Fig. 2c).

TABLE 2.

Description of RBS selected in this study^a

Primer	Sequence (5′–3′)	Translation initiation rate
RBS1-F	CGAAAATCCCTAAGGAGCTTAAGCATGGGCAGCAGCCATCACCATCATCACC	90,504.51
RBS1-R	GCTTAAGCTCCTTAGGGATTTTCGATTAAAGTTAAACAAAATTATTTCTACAGGGGAATTGTTATCCGCTC
RBS2-F	CATCACCGTTAGACGAGGAGGTATCCTATGGGCAGCAGCCATCACCATCATCACC	212,828.18
RBS2-R	AGGATACCTCCTCGTCTAACGGTGATGATTAAAGTTAAACAAAATTATTTCTACAGGGGAATTGTTATCCGCTC
RBS3-F	AATACCCTACTAAGGAGGTAAGCATGGGCAGCAGCCATCACCATCATCACC	318,342.84
RBS3-R	GCTTACCTCCTTAGTAGGGTATTATTAAAGTTAAACAAAATTATTTCTACAGGGGAATTGTTATCCGCTC
RBS4-F	GAACTCGAACATAGTCTTAAGGAGGTTCAAATGGGCAGCAGCCATCACCATCATCACC	370,978.54
RBS4-R	TTGAACCTCCTTAAGACTATGTTCGAGTTCATTAAAGTTAAACAAAATTATTTCTACAGGGGAATTGTTATCCGCTC

^aThe sequences of RBS are underlined.

Molecular modification and characterization of MaiA variants.

To further improve the catalytic ability of MaiA, fifty single-point variants of MaiA were constructed based on the evolutionary information from thermophilic bacteria, on which surface lysine and glycine were replaced by arginine and alanine, respectively (19–21). Six glycines and nine lysines on the molecular surface of MaiA (see Fig. S1a in the supplemental material) were selected for mutation according to the prediction by the GETAREA website. The expression levels of MaiA and its variants were estimated by SDS-PAGE, and crude enzymes that were adjusted to the same concentrations were used for screening. The results of screening showed that five single-point variants (G27A, K47R, K51R, K104R, and G171A) might have improved catalytic ability, and K104R was the variant with the best activity and longest half-life (Fig. S1b).

Combinatorial mutagenesis was carried out to further improve the catalytic ability of MaiA. Ten double-point variants were constructed, and the relative enzyme activities and the half-lives at 55°C were measured for screening (Fig. S1c). Two of these dual-point variants, G27A/K104R and G27A/G171A, which had much better thermostability and activity, were selected for purification and characterization. The optimum temperature of the wild type was 40°C, whereas G27A/K104R and G27A/G171A showed their highest activity at 45°C (Fig. 3a). The G27A/K104R and G27A/G171A variants were much more thermostable than the wild type, retaining 81% and 70% activity, respectively, after treatment for 180 min at 55°C. The wild type only retained 46% activity after treatment for 180 min at 55°C (Fig. 3b). The kinetic constants of the double-point variants were analyzed and are summarized in Table 3. The K_m values of the two variants were higher than that of the wild type, suggesting a lower affinity to the substrate. As we expected, the catalytic efficiencies of the two double-point variants were increased by the increasing values of k_cat.

FIG 3.

Effect of temperature on the enzymatic activity of MaiA and its variants. (a) The optimum temperature was determined by the enzyme activity at the corresponding temperature, and the highest activity was plotted as 100%. (b) Thermostability was determined after treatment for the corresponding time at 55°C, and the residual activity was plotted as a percentage of the initial activity.

TABLE 3.

Kinetic parameters of MaiA and its variants

Enzyme	Km (mM)	kcat (min−1)	kcat/Km (min−1 · mM−1)
Wild type	0.6 ± 0.02	19.1 ± 0.10	31.8
G27A/K104R	1.1 ± 0.04	35.2 ± 1.56	32.0
G27A/G171A	1.2 ± 0.04	51.6 ± 2.01	43.0

Structure assay of MaiA and its variants.

Structure analysis of thermoresistant variants (G27A, K104R, and G171A) were carried out to explore the inherent mechanism (Fig. 4). Homology modeling of S. marcescens MaiA was implemented using SWISS-MODEL. The MaiA from Pseudomonas putida S16 (PDB entry 4FQ7) was used as a template, and it shares 73% identity with S. marcescens MaiA. Residue Gly27 was involved in forming two pairs of hydrogen bonds in the wild type (Fig. 4a). The replacement of Gly27 with alanine did not change the hydrogen bond interactions in the G27A variant (Fig. 4b). The additional methyl group of Ala27 introduced side chain hydrophobicity (Fig. 4b), which might improve the stability of MaiA.

FIG 4.

Structure analysis of the region around mutated resides in the MaiA variants. (a and b) Interactions between the 27th residue with Pro23 and Gln30 in the wild type (a) and the G27A variant (b). (c and d) Interactions between the 104th residue with the surrounding residues in the wild type (c) and the K104R variant (d). (e and f) Interactions between the 171st residue with Asn167 and Leu168 in the wild type (e) and the G171A variant (f). The red dotted lines represent hydrogen bonds. Side chain sticks for hydrophobic residues were prepared using the PyMOL program.

An interesting phenomenon has been found in the analysis of hydrogen bond networks near the 140th residue of K104R and the wild type. Residue Lys104 is located at the top of an α-helix in the wild type, and two hydrogen bonds were formed between Lys104 and Gln107, along with Lys104 and Ala100 (Fig. 4c). Conversely, the replacement of Lys104 with arginine in the K104R variant resulted in three novel hydrogen bonds among Lys104, Gln107, Gln101, and Ala100 (Fig. 4d). The novel hydrogen bonds formed a hydrogen bond network, which might stabilize the α-helix and the loop. The stability of K104R might be enhanced due to the increased rigidity of the entire enzyme structure.

Residue Gly171 was located in a loop and was involved in forming two hydrogen bonds in the wild type (Fig. 4e). Although only one hydrogen bond interaction in the G171A variant was observed, an additional methyl group brings Ala171 much closer to Leu168 and Asn167 (Fig. 4f). Hydrophobic interactions that were caused by the overlapping regions between these hydrophobic side chains might contribute to the stability. Residue Asn167 is strictly conserved in maleate cis-trans isomerase and plays an important role in recognizing the substrate maleate (11). The enzyme kinetic stability of G171A might be improved by the rigidity of the flexible loop within the active-site residues (22).

Production of l-aspartate using an engineered strain expressing a MaiA variant.

Two engineered strains, pMA-RBS4-G27A/K104R and pMA-RBS4-G27A/G171A, were constructed for the bioproduction of l-aspartate, since the two variants exhibited better catalytic activity than the wild type (Fig. 5). The production of l-aspartate through biotransformation from maleate using whole-cell catalysis was examined in a beaker containing recombinant E. coli cells (optical density at 600 nm [OD₆₀₀] of 40). Upon adding maleate (1.6 M) to the reaction system once, the yield of l-aspartate bioproduced using recombinant cells expressing G27A/K104R reached 1.59 M (211.7 g/liter) after 60 min, and the substrate (maleate) and the intermediate product (fumarate) were almost depleted (Fig. 5a). Moreover, recombinant cells expressing G27A/G171A exhibited the highest production efficiency: 1.5 times that of pMA-RBS4-G27A/K104R, 2 times that of pMA-RBS4 (expressing the wild-type MaiA) (Fig. 2c), and 3 times that of pMA (harboring the original RBS and the wild-type MaiA) (Table 1). The yield of l-aspartate reached 1.60 M (212.4 g/liter) after 40 min, and the substrate (maleate) and the intermediate product (fumarate) were almost exhausted (Fig. 5b). The conversions of both catalytic systems were close to 100%.

FIG 5.

Production of l-aspartate by whole-cell catalysis. (a) The recombinant strain expressing the G27A/K104R variant. (b) The recombinant strain expressing the G27A/G171A variant. The concentrations of l-aspartate (solid circle), fumaric acid (solid square), and maleate (open circle) were analyzed by HPLC.

Our results indicated that the engineered strain pMA-RBS4-G27A/G171A showed the highest productivity, so the scale-up production of l-aspartate using the engineered strain pMA-RBS4-G27A/G171A was completed in a 5-liter fermentor according to the fermentation results of our previous report (23). As shown in Fig. 6, the cell density (OD₆₀₀) and the cell activity were determined for fermentation evaluation. When the OD₆₀₀ reached 80 at 19 h, we applied 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG) to induce the expression of MaiA and AspA enzymes. The activity of the recombinant cells significantly increased and reached a peak value of 240.6 U/ml at 47 h (Fig. 6a). The cell density was maintained at 80 after IPTG was added.

FIG 6.

Scale-up of production of l-aspartate using the pMA-RBS4-G27A/G171A strain. (a) Fermentation of the pMA-RBS4-G27A/G171A strain in a 5-liter fermentor. The inducer IPTG at 0.5 mM was added when the cell density (OD₆₀₀) reached 80. Cell growth (open square) and cell activity (solid circle) were measured. (b) Optimized fermentation. The inducer IPTG at 0.8 mM was added when the cell density (OD₆₀₀) reached 120. Cell growth (open square) and cell activity (solid circle) were determined. (c) SDS-PAGE analysis of the expression of MaiA and AspA. The induction times are indicated as 3 to 21 h. The bands corresponding to MaiA and AspA are indicated by arrows separately. Lane M is the standard marker protein. (d) Whole-cell catalysis. The reaction mixture contained 3.2 M substrate and recombinant cells (OD₆₀₀ of 40). The concentrations of l-aspartate (solid circle), fumaric acid (solid square), and maleate (open circle) were analyzed by HPLC.

Subsequently, we optimized the induction timing and the amount of IPTG (Fig. S2). When the cell density reached 120 and 140, 0.5 mM IPTG was applied to induce the expression of enzymes. The activity of the recombinant cells increased to a peak value of 293.4 U/ml after the inducer was added and when the OD₆₀₀ was 120 (Fig. S2a). When the induction was started at an OD₆₀₀ of 140, the activity of the recombinant cells only reached 265.6 U/ml (Fig. S2b). These results showed that the best time for the application of IPTG was at an OD₆₀₀ of 120.

When the cell density was close to 120, the inducer IPTG at 0.8 mM was applied to induce enzyme expression, and the activity of the recombinant cells reached a peak value of 331.3 U/ml at 39 h (Fig. 6b). The expression level of the enzymes was analyzed by SDS-PAGE, and the highest level was reached after the recombinant cells were induced with IPTG for 18 h (Fig. 6c). When the amount of IPTG increased to 1 mM, the activity of the recombinant cells modestly increased to 349.2 U/ml (Fig. S2c), which was close to the activity induced by 0.8 mM IPTG. These results suggested that the optimal condition for the cell activity was to apply 0.8 mM IPTG when the OD₆₀₀ was 120.

The cells obtained by fermentation were used for the scale-up production of l-aspartate. As shown in Fig. 6d, when maleate was included at 3.2 M, the yield of l-aspartate reached 3.15 M (419.8 g/liter) after 240 min. The substrate (maleate) remained at 0.29 mM (0.03 g/liter) and the intermediate product (fumarate) remained at 7.70 mM (0.89 g/liter), and the conversion was 98.4%. The cells were recycled by centrifugation and were used to produce l-aspartate a second time. The recovery rate of the cell activity was 65.4%, and more than 1 M maleate remained after 240 min of catalysis (Table S1).

DISCUSSION

MaiA from S. marcescens has barely been applied to bioproduce fumarate, because fumarate was partly biotransformed into malic acid using E. coli cells, leading to a decreased conversion and an increased difficulty of product separation, and chemical synthesis was also economical. This study presented an engineered strain to eliminate by-product synthesis pathways. We uncovered an economical and convenient method to employ MaiA to produce l-aspartate from maleate by coupling with AspA. The amount of l-aspartate bioproduced by cells fermented in a 5-liter fermentor reached 419.8 g/liter, with a high conversion efficiency. Development of engineered strains that can produce other high-value-added products from maleate by one-pot biosynthesis will be an interesting topic for future study.

We found that cells reached growth arrest after adding inducer when the engineered strain pMA-RBS4-G27A/G171A was cultivated in a 5-liter fermentor (Fig. 6a and b; also see Fig. S2 in the supplemental material). Two of three genes encoding fumarase isozyme involved in the TCA cycle were inactivated in the engineered strain, and only one remained. We speculate that the remaining fumarase metabolizes a small amount of fumarate and maintains the cell growth. After the inducer was added, fumaric acid was converted to l-aspartate by the dual-enzyme cascade, which affected cell growth. It also suggested that the remaining fumarase failed to compete with the dual-enzyme cascade.

The multienzyme cascade requires precise regulation and compatibility to satisfy each reaction condition, including the expression level, stability, and catalytic ability of each enzyme. Expression levels of proteins could be changed by more than 100-fold by changing gene locations or RBS sites (24). We constructed several expression cassettes containing MaiA and AspA, including different tandem arrangements and addition of a promoter, RBS, or terminator (data not shown). The strain harboring the pRSF_Duet-1-maiA-aspA plasmid, in which each gene has a promoter and an RBS, exhibited the best conversion from maleate to l-aspartate. The expression level of MaiA was much higher than that of AspA in this dual-enzyme cascade, offsetting the situation in which the catalytic ability of MaiA was inferior to that of AspA. The whole-cell catalysis was implemented at pH 8.0, under which conditions MaiA exhibited the highest activity. In addition to the reaction conditions, all optimizations were aimed at the rate-limiting enzyme MaiA.

Protein engineering has been widely applied for enzyme activity, thermal stability, and substrate specificity. We chose to use a strategy in which we replaced lysine and glycine on the surface of the enzyme based on the successful results of previous studies (25–27). MaiA was chosen for molecular modification due to its lower stability than that of AspA, with a half-life of 28 days at room temperature (5). The thermal stability might have a great effect on the process of whole-cell catalysis, as the catalytic process takes place over a few hours. Compared with the activity of the wild type, the variant with better thermostability could maintain the activity for a longer time in the catalytic process. As a result, the production efficiency using the variants with better thermostability and activity was improved by 2 to 3 times (Fig. 2d and 5). The variant with improved thermostability might be more suitable for production in industry. The efficiency of free cell recycling does not seem to be applicable to industrial production, while cell immobilization is a considerable issue. Furthermore, the catalysis time and substrate concentration could be optimized depending on the actual needs of the product.

MATERIALS AND METHODS

Plasmids and strains.

The plasmid pRSF_Duet-1 was employed as the expression vector, and E. coli JM109 was used for plasmid construction. The S. marcescens maiA gene was cloned into the restriction enzyme sites of BamHI and HindIII, constructing pRSF_Duet-1-maiA plasmid. The N terminus of MaiA was designed to contain a histidine tag. After that, the pRSF_Duet-1-maiA-aspA plasmid was constructed by inserting the E. coli aspA gene into the restriction enzyme sites of NdeI and XhoI. E. coli BL21 ΔfumA ΔfumC was used for enzyme expression and biocatalysis, in which two genes (fumA and fumC) responsible for the conversion of fumarate to malic acid were deleted (18). Luria-Bertani (LB) medium containing 50 μg/ml kanamycin was used for plasmid construction and propagation. The solid medium contained 1.5% agar.

Modification of RBS sequence.

RBS sequences with different initial translation rates were calculated and predicted using RBS Calculator 2.0 web software (Salis Lab, Penn State University). Four RBS sequences were selected to replace the original RBS sequence located upstream of MaiA. The pRSF_Duet-1-maiA-aspA plasmid was used as the template, and the substitution was carried out by Megaprimer PCR of whole plasmid (MEGAWHOP) with a forward primer (X-F) and a reverse primer (X-R). Primers (see Table S1 in the supplemental material) were synthesized by GENEWIZ Biotech Co., Ltd. (Suzhou, China). The sequences of the transformants were confirmed by GENEWIZ Biotech Co., Ltd.

Structural analysis and site-directed mutagenesis of MaiA.

The crystal structure of maleate isomerase from Pseudomonas putida S16 (PDB entry 4FQ7) was used as the template. Homology modeling of MaiA was carried out using the SWISS-MODEL online software and uploaded to the GETAREA website (http://curie.utmb.edu/getarea.html) for surface amino acid prediction. The structural differences between the wild-type MaiA and its variants were illustrated using PyMol software.

All lysine and glycine residues on the surface of the MaiA molecule were replaced by arginine and alanine, respectively. The pET24a(+)-maiA plasmid was used as the template to construct mutants using the site-directed mutagenesis method. Primers used for site-directed mutation (Table 4) were synthesized by GENEWIZ Biotech Co., Ltd. The sequences of the transformants were confirmed by GENEWIZ Biotech Co., Ltd.

TABLE 4.

Primers for the construction of MaiA variants^a

^aThe nucleotide bases for the mutated residues are underlined.

Expression and purification of MaiA and its variants.

A single colony was inoculated into 5 ml of LB medium (including 50 μg/ml kanamycin) at 37°C, with shaking at 200 rpm. Cells cultured overnight then were transferred into 50 ml of LB medium with 1% inoculum and incubated at 37°C for cell growth. When the OD₆₀₀ reached 0.6 to 0.8, 0.2 mM IPTG was added to induce the expression of MaiA and its variants for 20 h at 20°C.

The purification of MaiA and its variants was performed using an AKTA purifier (GE Healthcare UK Ltd.). Harvested cells were suspended in Na₂HPO₄-KH₂PO₄ buffer (50 mM, pH 8.0) and then were sonicated on ice. Cell lysates were loaded onto a HisTrap HP column (GE Healthcare UK Ltd.) to absorb MaiA and its variants, and then the target proteins were eluted using elution buffer (20 mM Na₂HPO₄/KH₂PO₄, 500 mM NaCl, 500 mM imidazole, pH 7.5). Samples were dialyzed against 50 mM Tris buffer (pH 7.0) to eliminate imidazole. The protein concentrations were determined by the Bradford method (28).

Determination of the activity of MaiA.

The enzyme activity of MaiA was determined by the increase of fumarate, which was converted from maleate. The reaction was performed in 50 mM phosphate buffer (pH 8.0) at 40°C for 10 min and then was boiled for 10 min. Fumarate and maleate were detected by high-pressure liquid chromatography (HPLC) with a Prevail organic acid column (5 μm, 4.6 mm by 250 mm; Grace Davison Discovery Sciences). The mobile phase was KH₂PO₄ solution (25 mM, pH 2.5) at a flow rate of 1 ml/min, and the UV detector wavelength was set to 210 nm. One unit of MaiA activity was defined as the amount of enzyme that released 1 mmol fumarate per minute at 40°C.

Assay of biochemical properties and kinetic parameters of MaiA and its variants.

The optimal temperatures for the activity of MaiA and its variants was determined in phosphate buffer (50 mM, pH 8.0). The enzymatic activity was measured at various temperatures, and the highest enzyme activity was defined as 100%.

To detect thermal inactivation, the purified enzyme was analyzed by measuring the remaining activity after incubation for the corresponding time at 55°C. The half-life time was calculated according to the following equations, where v indicates the residual activity after incubation for t minutes (60 min) at 55°C, v₀ denotes the initial activity, K_D represents the inactivation constant at 55°C, and t_1/2 refers to the half-life at 55°C.

$$\ln \frac{v}{v_0}=-K_{D}t$$ (1)

$$t_{1/2}=\frac{0.693}{K_D}$$ (2)

Kinetic parameters of MaiA and its variants were determined by the substrate (maleate) over a concentration range from 10 to 80 mM in phosphate buffer (50 mM, pH 8.0) at 40°C. The values of K_m and k_cat were calculated through the Lineweaver Burk method using the program GraphPad Prism 5. The data are presented as the means ± standard deviations.

Production of l-aspartate by whole-cell catalysis.

Recombinant cells were harvested and suspended in phosphate buffer (50 mM, pH 8.0). The substrate maleate dissolved in ammonia was added to the reaction mixture containing recombinant cells (OD₆₀₀ of 10). Detection of cell activity was performed at 40°C for 10 min. One unit of cell activity was defined as the amount of cells that produce 1 mmol l-aspartate per minute at 40°C.

The process for producing l-aspartate was performed in a 100-ml beaker containing a 30-ml reaction mixture with recombinant cells (OD₆₀₀ of 40) at 37°C, with shaking at 200 rpm. The substrate maleate dissolved in ammonia was added, and sample aliquots were withdrawn at intervals to detect the concentrations of the product. l-Aspartate was derivatized with phenyl isothiocyanate (PITC) according to a previously described method (29). The concentrations of l-aspartate were detected by HPLC with a C₁₈ column (5 μm, 4.6 by 250 mm; La Chrom), and the UV detector wavelength was set to 254 nm. The mobile phases A (80% acetonitrile) and B (3% acetonitrile mixed with 97% 0.1 M sodium acetate) were subjected to gradient elution at a flow rate of 0.6 ml/min. The gradient elution conditions were 0 to 35 min B, decreasing from 95% to 65%; 35 to 40 min B, increasing from 65% to 95%; and 40 to 45 min B, maintained at 95%.

Fed-batch cultivation in 5-liter fermentor.

Fed-batch cultivations in a 5-liter bioreactor were carried out based on a previously published method (23). The inoculation volume was 6%, and the pH was maintained at 7.0 through the automatic addition of 25% ammonia. During the cultivation process, the temperature was maintained at 37°C, and dissolved oxygen (DO) was maintained above 30% under the control of the inlet air and the exponential feeding of glucose and tryptone. When the OD₆₀₀ reached 120, a final concentration of 0.8 mM IPTG was added at a rate of 0.22 g/liter/h, and the temperature was gradually reduced to 30°C.