Characteristics of L-threonine transaldolase for asymmetric synthesis of β-hydroxy-α-amino acids

Lian Xuab, Li-Chao Wangab, Xin-Qi Xub and Juan Lin*ab
aCollege of Chemical Engineering, Fuzhou University, Fuzhou 350116, China. E-mail: ljuan@fzu.edu.cn
bCollege of Biological Science and Engineering, Fuzhou University, Fuzhou 350116, China

Received 11th August 2019 , Accepted 16th September 2019

First published on 17th September 2019


L-Threonine transaldolase (LTTA) is a putative serine hydroxymethyltransferase (SHMT) that can catalyze the trans-aldehyde reaction of L-threonine and aldehyde to produce L-threo-β-hydroxy-α-amino acids with excellent stereoselectivity. In the present study, an L-threonine transaldolase from Pseudomonas sp. (PsLTTA) was mined and expressed in Escherichia coli BL21 (DE3). A substrate spectrum assay indicated that PsLTTA only consumed L-threonine as the donor substrate and could accept a wide range of aromatic aldehydes as acceptor substrates. Among these substrates, PsLTTA could catalyze p-methylsulfonyl benzaldehyde and L-threonine to produce L-threo-p-methylsulfonylphenylserine with a high conversion rate (74.4%) and a high de value (79.9%). The conversion and stereoselectivity of PsLTTA were found to be dramatically influenced by the concentration of the whole cell, the co-solvent and the reaction temperature. Through conditional optimization, L-threo-p-methylsulfonylphenylserine was obtained with 67.1% conversion and a near-perfect de value (94.5%), the highest stereoselectivity for an L-threo-β-hydroxy-α-amino acid so far reported by enzymatic synthesis. Finally, synthesis of L-threo-p-methylsulfonylphenylserine at a 100 mL scale by whole-cell biocatalysis was conducted. This is the first systematic report of L-threonine transaldolase as a robust biocatalyst for preparation of β-hydroxy-α-amino acids, which can provide new insights for β-hydroxy-α-amino acids synthesis.


1. Introduction

β-Hydroxy-α-amino acids are comprised of numerous important building blocks, which are essential for the synthesis of agriculturally or medically bioactive products.1–3 For example, some β-hydroxy-α-amino acids, such as L-threo-p-methylsulfonylphenylserine and L-threo-p-nitrophenylserine, are known to be key intermediates for the synthesis of thiamphenicol and chloramphenicol, respectively.4,5 L-Threo-phenylserine can be found in cyclomarin A, a marine cyclopeptide with antitubercular and antimalarial activities.6 In addition, L-threo-3,4-dihydroxyphenylserine (droxidopa) serves as a crucial drug for the treatment of Parkinson's disease.7 As a result, β-hydroxy-α-amino acids have attracted much attention because of their important role in chemical synthesis and pharmaceutical manufacturing.

β-Hydroxy-α-amino acids containing two chiral centers and four isomers (L-threo, D-threo, L-erythro and D-erythro) are produced simultaneously during the chemical process.8,9 This makes it difficult to synthesize optically pure of β-hydroxy-α-amino acids. Many approaches, including Sharpless asymmetric dihydroxylation, epoxidation or enzymatic aldol addition, have been developed for the synthesis of β-hydroxy-α-amino acids.10–12 Among these, the enzymatic approach is the most promising as it can establish two new chiral centers in a single reaction, with inexpensive substrates, as well as under mild and unprotected conditions.12 Threonine aldolases (TAs) are examples of such enzymes that can catalyze the aldol addition of glycine and various aldehydes to produce β-hydroxy-α-amino acids. TA is a pyridoxal 5′-phosphate (PLP)-dependent enzyme and can be classified into L-threonine aldolase (LTA) and D-threonine aldolase (DTA) based on the stereospecificity at the Cα atom of its products.13–15 In most cases, TA is highly stereoselective (>99% enantiomeric excess, ee) at the Cα, while showing moderate stereoselectivity at Cβ (about 10–45% diastereomeric excess, de).16 As a result, both threo and erythro isomers are produced in the TA-catalyzed reaction. This ‘Cβ problem’ has greatly hampered and limited the synthetic application of TA. One possible reason for the poor stereoselectivity at Cβ is that two active histidine residues in TA can interact with the hydroxyl group on Cβ from opposite directions.16 It is noteworthy that, up to now, attempts to improve the Cβ stereoselectivity of TA by protein engineering have ended in failure, mainly due to complicated interactions in the enzyme-active center participating in the aldol addition reactions.14,17,18 DTA was reported to be able to obtain high stereoselectivity by keeping the reaction under kinetic control (usually under conditions of low temperature and low amounts of enzyme).19,20 Unfortunately, almost all of the active intermediates were L-threo isomers and were only produced by LTA.4–7 Recently, Chen et al. have developed a high-throughput screening method and improved the stereoselectivity of LTA. However, the de value of the mutant enzyme was up to about 74%, which is still too low for industrial application.18 As a result, mining and characterization of novel L-threonine aldolase-like enzymes with high stereospecificity is required to meet the needs of industrial production, as well as to expand the toolkit for asymmetric synthesis of β-hydroxy-α-amino acids.

In recent years, L-threonine transaldolase (LTTA) has been reported as a member of the antibiotic gene cluster and participates in the biosynthesis of antibiotics.21–23 The characteristic of LTTA is that it can catalyze the formation of β-hydroxy-α-amino acids with high stereoselectivity at Cβ using L-threonine and aldehyde as substrates.22 Phylogenetic analysis indicated that LTTA, LTA and serine hydroxymethyl transferase (SHMT) shared a common evolutionary origin and thus might have a similar catalytic mechanism.22 In the present study, we characterized an LTTA from Pseudomonas sp. (PsLTTA) with excellent stereoselectivity for β-hydroxy-α-amino acid synthesis. We showed that PsLTTA only consumed L-threonine as the donor substrate and could accept a wide range of aromatic aldehydes as acceptor substrates. The conversion and de values of L-threo-p-methylsulfonylphenylserine by PsLTTA were up to 67.1% and 94.5%, respectively, following optimization of the reaction conditions. Finally, synthesis of L-threo-p-methylsulfonylphenylserine at a 100 mL scale by whole-cell biocatalysis was conducted. Taken together, our results systematically evaluate the catalytic properties of PsLTTA and suggest that PsLTTA could serve as a promising biocatalyst for preparation of L-threo-β-hydroxy-α-amino acids.

2. Results and discussion

2.1 Sequence analysis and expression of PsLTTA

PsLTTA was identified as an SHMT-like enzyme, since it showed about 26–33% amino acid sequence identity with SHMT (Table S1). Multiple sequence alignments demonstrated that the PLP-binding residues were conserved among PsLTTA and SHMTs, suggesting that they might have a similar catalytic mechanism (Fig. S1). This is consistent with the previous report that LTTA and SHMT share a common evolutionary origin.22

PsLTTA expression was induced under optimum conditions (0.2 mM isopropyl β-D-1-thiogalactopyranoside (IPTG), 28 °C, 16 h) and further purified with His-tag affinity chromatography. The purified PsLTTA revealed a pink color in solution (Fig. S2A). Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) indicated that most of the PsLTTA was soluble protein and its monomer molecular weight was about 48 kDa, which is in accordance with the predicted molecular size (Fig. S2B). PsLTTA could catalyze the trans-aldehyde reaction of L-threonine and aldehyde to produce β-hydroxy-α-amino acids and acetaldehyde (Fig. S3). A transaldolase activity assay using p-methylsulfonyl benzaldehyde and L-threonine as substrates demonstrated that the cell-free extracts and purified enzyme had specific activities of 0.87 U mg−1 and 1.32 U mg−1, respectively.

2.2 Enzymatic properties

The thermostability profile indicated that PsLTTA was quite stable in the temperature range 4–30 °C. The enzymatic activity of PsLTTA rapidly decreased when the temperature was above 37 °C, with complete inactivation at 60 °C (Fig. 1A). The optimum reaction temperature for PsLTTA was determined to be 30 °C (Fig. 1B). PsLTTA was active at pH 6.0–8.0, when more than 80% activity could be detected (Fig. 1C). The maximum activity of PsLTTA was observed in 100 mM potassium phosphate buffer with a pH of 7.0 (Fig. 1D). These results suggest that PsLTTA is a pH- and temperature-sensitive enzyme. The effects of metal ions on the activity of PsLTTA were further investigated at concentrations of 1 and 10 mM. As shown in Fig. 1E, the transaldolase activity of PsLTTA was promoted slightly by Mg2+ and Li+ when compared with the control group. The presence of Ni2+, Cu2+, Fe2+ and Zn2+ could inhibit PsLTTA at both 1 mM and 10 mM concentrations. Other metal ions, such as Cs2+, Ba2+, Ca2+ and Mn2+, showed little effect on PsLTTA in 1 mM solution, but moderate inhibition in 10 mM solution. The PLP concentration assay suggested that PLP was required for the catalytic activity of PsLTTA, since a PLP-free group showed 65.8% activity compared with the control group (250 μM PLP) (Fig. 1F). This result is similar to the previous report that LTTA is a PLP-dependent enzyme.24 The partial activity in the PLP-free group was probably due to the enzyme itself combining with a portion of PLP during the process of intracellular folding.
image file: c9cy01608b-f1.tif
Fig. 1 Enzymatic properties of PsLTTA. (A) Temperature stability profile of PsLTTA after incubation of enzyme solution at various temperatures (4–80 °C) for 60 minutes. (B) Detection of the optimal reaction temperature of PsLTTA. (C) pH stability profile of PsLTTA after incubation of enzyme solution at various pH (3–11) for 12 hours at 4 °C. (D) Detection of the optimal reaction pH of PsLTTA. (E) Effects of metal ions (1 mM and 10 mM) on the trans-aldehyde activity of PsLTTA. (F) Effects of PLP concentration on the trans-aldehyde activity of PsLTTA. All assays were performed three times.

2.3 Substrate specificity of PsLTTA

The donor substrate specificity of PsLTTA was evaluated with p-methylsulfonyl benzaldehyde as the acceptor. The Km and kcat of PsLTTA using L-threonine as donor substrate were 9.85 mM and 63.4 min−1, respectively, while PsLTTA was not active towards L-allothreonine, D-threonine, D-allothreonine, L-serine and glycine (Table S2 and Fig. S4). Similarly to the enzyme assay results, high pressure liquid chromatography (HPLC) analysis indicated that PsLTTA could catalyze p-methylsulfonyl benzaldehyde and L-threonine to form L-threo-p-methylsulfonylphenylserine with high stereoselectivity (80.1% de) (Fig. 2A). However, no products were detected when using L-allothreonine, D-threonine, D-allothreonine or glycine as the donor substrate (Fig. 2C–F). PsLTTA could also catalyze the trans-aldehyde action of p-methylsulfonyl benzaldehyde and L-serine; however, the transaldolase activity declined sharply and the major product was L-erythro-p-methylsulfonylphenylserine (−43.8% de) (Fig. 2B). These data suggest that PsLTTA is a highly selective donor substrate. Coincidentally, the configuration of L-threonine was the L-threo form, which was the same as the major product, and this finding allowed us to speculate that strict donor specificity might contribute to the high stereoselectivity of PsLTTA.
image file: c9cy01608b-f2.tif
Fig. 2 Donor substrate specificity of PsLTTA detected by HPLC after OPA/NAC derivatization: (A) L-threonine, (B) L-serine, (C) D-threonine, (D) L-allothreonine, (E) D-allothreonine, (F) glycine. Product L-p-methylsulfonylphenylserine (tL-threo = 7.1 min, tL-erythro = 8.4 min).

The acceptor substrate specificity was further assessed using L-threonine as the donor. We evaluated the catalytic activity of PsLTTA on aliphatic aldehydes, such as formaldehyde or acetaldehyde. The results showed that PsLTTA could not consume these two aldehydes, mainly due to their oxidizability. PsLTTA could utilize a wide range of aromatic aldehydes as acceptor substrate to produce β-hydroxy-α-amino acids. As expected, all the major products were L-threo forms (Table 1). This is important since the active intermediates of thiamphenicol, chloramphenicol and droxidopa are L-threo isomers.4–7 We found that the substituent type on the phenyl ring of the aromatic aldehydes could affect the activity of PsLTTA. For example, PsLTTA could catalyze aromatic aldehydes with electron-withdrawing substituents, such as –CH3SO2, –NO2, –F, –Cl and –Br, while electron-donating substituents (–CH3, –OCH3 and –OH) could not serve as substrates for PsLTTA. For Cl- and Br-substituted benzaldehydes, it was shown that PsLTTA only catalyzed o-substituted benzaldehydes, suggesting that the position of the substituent also had an effect on PsLTTA. Similar results were obtained when using DTA to catalyze aromatic aldehydes and glycine. Chen et al. reported that the type of substitution, as well as the position on the phenyl ring, played an essential role in the catalytic performance of DTA by affecting the interaction between the β-OH group of the substrate and the manganese ion during the process of substrate recognition.20 This conclusion seems to be relevant for PsLTTA and could be the basis for protein engineering of PsLTTA.

Table 1 Acceptor substrate specificity of PsLTTA

image file: c9cy01608b-u1.tif

R Conversion (%) de (%) R Conversion (%) de (%)
Reaction conditions: 1 mL reaction mixture containing 40 mM aromatic aldehydes, 100 mM L-threonine, 0.2 mM PLP, 1 mM MgCl2 and 25 mg mL−1 wet cells in Tris-HCl buffer (100 mM Tris-HCl, 10% CH3CN, pH 7.0) at 30 °C for 4 h. ND, not detected.
H 51.0 77.2 o-Br 43.3 80.8
o-NO2 40.5 50.2 m-Br ND ND
m-NO2 25.2 29.9 p-Br ND ND
p-NO2 59.4 76.7 p-MeSO2 74.4 79.9
o-F 43.5 69.2 p-I ND ND
m-F 44.6 66.4 o-OH ND ND
p-F ND ND m-OH ND ND
o-Cl 37.9 80.3 p-OH ND ND
m-Cl ND ND o-CH3 ND ND
p-Cl ND ND p-O(CH3) ND ND


It is worth mentioning that PsLTTA can catalyze p-methylsulfonyl benzaldehyde with a high conversion rate (74.4%) and high stereoselectivity (79.9% de). The product, L-threo-p-methylsulfonylphenylserine, is the main intermediate for thiamphenicol and florfenicol. L-Threonine aldolase has also been reported to be able to catalyze the formation of p-methylsulfonylphenylserine using p-methylsulfonyl benzaldehyde and glycine as substrates.25 We assessed the catalytic activity and stereoselectivity of PsLTTA and an LTA from Clavibacter michiganensis (ClLTA). The conversion of p-methylsulfonyl benzaldehyde by PsLTTA and ClLTA was 75.2% and 85.6%, respectively. However, the de value for L-threo-p-methylsulfonylphenylserine using PsLTTA was 79.6%, which is much higher than that of ClLTA, with a de value of 28.6% (Fig. S5). This result suggested that PsLTTA was more stereoselective than ClLTA. Furthermore, the reverse activities of PsLTTA and ClLTA were evaluated using L-threo-p-methylsulfonylphenylserine as substrate. As shown in Fig. S6, ClLTA could change L-threo-p-methylsulfonylphenylserine into p-methylsulfonyl benzaldehyde and glycine completely within 3 hours. This result is consistent with the conclusion that the LTA-catalyzed reaction is reversible.26 To our surprise, the reverse transaldol activity of PsLTTA was almost undetectable when using L-threo-p-methylsulfonylphenylserine and acetaldehyde as substrates, suggesting that the LTA-catalyzed reaction might not be a reversible reaction. In the LTA-catalyzed reaction, excess substrate (usually, glycine was 10-fold more than the aldehyde) was required to make sure that the reaction equilibrium shifted in the direction of producing β-hydroxy-α-amino acids, which resulted in waste of the substrate.18 In PsLTTA, use of superfluous substrate could be avoided because of its poor reverse transaldol activity. Considering stereoselectivity and substrate availability, this shows that PsLTTA is more suitable for industrialized applications for producing β-hydroxy-α-amino acids.

2.4 Effect of reaction conditions on conversion and stereoselectivity of PsLTTA

Although the enzyme activity assay indicated that PsLTTA achieved maximum activity with phosphate buffer (Fig. 1D), it was revealed that the transformation of β-hydroxy-α-amino acid by whole-cell biocatalysis could be improved more than 30% when using Tris-HCl (100 mM Tris-HCl, pH 7.0) as the reaction buffer. This result suggested that Tris-HCl buffer was more suitable for whole-cell biocatalysis and thus it was used in the following reactions. Acetonitrile (10%, v/v) was selected as the co-solvent, since p-methylsulfonyl benzaldehyde dissolved well in acetonitrile, and the optimal reaction temperature (30 °C) was used as the initial reaction temperature.

We first evaluated the concentration of wet cells (3 to 50 mg ml−1) and substrate (10 to 40 mM) on the conversion and stereoselectivity of PsLTTA using p-methylsulfonyl benzaldehyde and L-threonine as substrates. As shown in Fig. 3, similar trends were observed in that the increase in wet-cell concentration led to a decrease in stereoselectivity in PsLTTA under different substrate concentrations. The de values of the products were about 86.1% when the wet-cell concentrations were 3.0 and 6.25 mg ml−1. Then, the de values declined from 86.1% to 72.0% with an increase in wet-cell concentration from 6.25 up to 50 mg ml−1. These results strongly indicate that a low dose of wet cells contributed to formation of the L-threo isomer. The wet-cell concentrations of 3.0 and 6.25 mg ml−1 showed little difference in the stereoselectivity of PsLTTA, but greatly affected its conversion (about 5–22% improvement in conversion), suggesting that 6.25 mg ml−1 of wet cells was suitable for achieving high stereoselectivity performance. Thus, the concentration of substrates affected the conversion of PsLTTA but not its stereoselectivity. At a wet-cell concentration of 6.25 mg ml−1, the conversion rate fell from 90.1% to 50.5% when the substrate concentrations ranged from 10 mM to 40 mM, and we found that the products reached maximum yield at 30 mM substrate concentration, with a conversion of 69.5%. Based on these results, 6.25 mg ml−1 and 30 mM were selected as the final concentrations for wet cells and substrate, respectively.


image file: c9cy01608b-f3.tif
Fig. 3 Concentration effect of wet cells and substrates on the conversion and stereoselectivity of PsLTTA. The concentrations of p-methylsulfonylbenzaldehyde were 10 mM (A), 20 mM (B), 30 mM (C) and 40 mM (D), respectively. The reaction mixtures contained 10% (v/v) CH3CN as co-solvent and were incubated at 30 °C for 4 hours.

The effect of co-solvents was also investigated. As shown in Table 1, the group with no organic solvents gave a comparatively low conversion (41.6%) and de value (71.3%), suggesting that organic solvents were required for the catalysis of PsLTTA. The results demonstrated that a low percentage of organic solvent (10%) promoted the conversion of PsLTTA, while high percentages of organic solvents (30%) led to a decline in conversion. Among the solvents tested, high conversion (80.4%) and de values (90.2%) were obtained using 10% ethyl acetate as co-solvent. MeOH (10%) and DMSO (10%) exerted promotion effects on the conversion, but showed little enhancement to the stereoselectivity of PsLTTA when compared with 10% CH3CN. The addition of 10% acetone and 10% EtOH lowered both the conversion and the stereoselectivity of PsLTTA. Thus, 10% ethyl acetate was selected as the optimal co-solvent for the enzymatic reaction.

The trans-aldehyde reaction of PsLTTA was evaluated under different temperature conditions, for 6 hours. As shown in Fig. 4, the highest conversion of products (79.8%) was observed at 30 °C, which was consistent with the conclusion that the optimum reaction temperature for PsLTTA was 30 °C. For stereoselectivity, we surprisingly found that increasing the temperature from 10 °C to 35 °C led to a reduction in stereoselectivity for PsLTTA. The de values showed little change (about 94.5%) but the conversion rates increased rapidly from 44.7% to 67.8% at temperatures of 10, 15 and 20 °C. The de values declined sharply (from 94.5% to 87.2%) when the temperature was 25, 30 and 37 °C. These results indicate that a lower temperature contributed to formation of the L-threo isomer and suggested that 20 °C was the best temperature for the catalysis reaction of PsLTTA.


image file: c9cy01608b-f4.tif
Fig. 4 Effect of temperature on the conversion and stereoselectivity of PsLTTA. (A) The conversion and de values of L-threo-p-methylsulfonylphenylserine at various temperatures. (B) Determination of stereospecificity of PsLTTA by HPLC analysis after OPA/NAC derivatization. Product L-p-methylsulfonylphenylserine (tL-threo = 7.1 min, tL-erythro = 8.4 min).

Taken together, our results strongly suggest that a low concentration of wet cells (6.25 mg ml−1), 10% ethyl acetate as co-solvent and low temperature (20 °C) could dramatically improve the stereoselectivity of PsLTTA, giving a de value from 71.3% up to 94.5%. Our conclusion that a low concentration of wet cells and a low temperature contributed to the stereoselectivity of PsLTTA and was exactly similar to that found for DTA.20 Under such conditions, the catalysis reaction is under kinetic but not thermodynamic control, which is beneficial for producing the L-threo isomer.27

In recent years, lots of attention has been paid to TAs, as they have been the only reported aldolases for production of β-hydroxy-α-amino acids using cheap substrates (glycine and aldehyde) and under mild conditions. However, the poor stereoselectivity at Cβ (about 10–45% de) has severely hindered their industrial application.16 In the present study, the PsLTTA characterized not only displayed TA-like merits, such as inexpensive substrates and mild conditions, but also showed excellent stereoselectivity at Cβ (94.5% de) and poor reverse transaldol activity. Under the optimized conditions, we further assessed the stereoselectivity of PsLTTA and ClLTA on other aromatic aldehydes (Table 3 and Fig. S7–S14). Similar to p-methylsulfonyl benzaldehyde, PsLTTA could consume multiple aromatic aldehydes to produce β-hydroxy-α-amino acids with excellent stereoselectivity. Among those products, L-threo-p-nitrophenylserine (92.8% de), L-threo-phenylserine (89.1% de), L-threo-o-chlorophenylserine (87.1% de) and L-threo-o-bromophenylserine (86.9% de) were obtained with excellent stereoselectivity. The de values of these products were much higher than for catalysis by PsLTTA before reaction optimization (Table 2), or by ClLTA (Fig. S7–S14). These results demonstrate that the optimized reaction conditions determined were universally applicable for aromatic aldehydes with different substituents. Considering that L-threo-p-nitrophenylserine is widely reported as an intermediate of chloramphenicol and L-threo-phenylserine can be found in serval bioactive substances,4–6 this strongly suggests that PsLTTA might take the place of TAs as a promising and robust biocatalyst for production of useful β-hydroxy-α-amino acids.

Table 2 Effects of co-solvents on the conversion and stereoselectivity of PsLTTA
Co-solvents Conversion (%) de (%) Co-solvents Conversion (%) de (%)
Reaction conditions: 1 mL reaction mixture containing 30 mM p-methylsulfonyl benzaldehyde, 100 mM L-threonine, 0.2 mM PLP, 1 mM MgCl2 and 6.25 mg mL−1 wet cells in Tris-HCl buffer (100 mM Tris-HCl, pH 7.0) at 30 °C for 4 h.
None 41.6 71.3 10% CH3CN 70.4 84.9
10% MeOH 81.1 85.3 20% CH3CN 45.1 82.1
20% MeOH 76.9 86.9 30% CH3CN 26.2 79.0
30% MeOH 44.6 82.6 10% DMSO 78.2 84.2
10% EtOH 45.1 80.8 20% DMSO 69.9 85.7
20% EtOH 55.9 81.0 30% DMSO 56.0 84.6
30% EtOH 17.7 80.2 10% ethyl acetate 80.4 90.2
10% acetone 64.5 80.0 20% ethyl acetate 74.8 89.4
20% acetone 46.2 85.1 30% ethyl acetate 69.2 89.4
30% acetone 39.7 84.5 40% ethyl acetate 61.1 89.9


Table 3 Stereoselectivity of PsLTTA and ClLTA
R de (%) R de (%)
PsLTTA ClLTA PsLTTA ClLTA
Reaction conditions: for PsLTTA, 1 mL reaction mixture containing 30 mM aromatic aldehydes, 100 mM L-threonine, 0.2 mM PLP, 1 mM MgCl2 and 6.25 mg mL−1 wet cells in Tris-HCl buffer (100 mM Tris-HCl, 10% ethyl acetate, pH 7.0) at 20 °C for 6 h; for ClLTA, 1 mL reaction mixture containing 30 mM aromatic aldehydes, 300 mM glycine, 0.2 mM PLP and 6.25 mg mL−1 wet cells in Tris-HCl buffer (100 mM Tris-HCl, 10% ethyl acetate, pH 7.0) at 20 °C for 6 h.
H 89.1 17.0 o-F 78.1 38.5
o-NO2 77.7 35.0 m-F 80.4 20.4
m-NO2 67.9 25.6 o-Cl 87.1 53.9
p-NO2 92.8 28.3 o-Br 86.9 53.8


2.5 Synthesis of L-threo-p-methylsulfonylphenylserine by whole-cell biocatalysis

A 100 mL scale reaction was then carried out under the optimal reaction conditions, and time courses for transformation of L-threo-p-methylsulfonylphenylserine were monitored, from 0 to 24 h (Fig. 5). As expected, L-threo-p-methylsulfonylphenylserine was produced with excellent stereoselectivity (90.4% de). The conversion rate was 67.1% and the concentration of p-methylsulfonyl benzaldehyde decreased from 31.0 mM to 10.2 mM within 24 hours. The synthesis of L-threo-p-methylsulfonylphenylserine reached 13 mM (95.1% de) during the first 2 hours, indicating that the beginning reaction was efficient. The products added up to 21 mM at 12 h, and did not increase in yield from 12 to 24 h. However, the de value at 12 h was 93.1% and subsequently decreased to 90.4% at 24 h. In our opinion, L-threo-p-methylsulfonylphenylserine was synthesized preferentially in the initial reaction stage. As time went by, the reaction tended to produce more L-erythro-p-methylsulfonylphenylserine, which led to the decline in de value. These results suggest that a short reaction time contributed to the formation of L-threo-p-methylsulfonylphenylserine. Considering that L-threo-p-methylsulfonylphenylserine is unstable, we converted it into the L-threo-p-methylsulfonylphenylserine ethyl ester by esterification. The final purified product was 189 mg of white solid, with a yield of 33.0% and was further confirmed by high-resolution mass spectroscopy (HRMS) (Fig. S15), 1H NMR (Fig. S16) and 13C NMR (Fig. S17) analysis. As far as we knew, these results represent the highest de value reported so far for asymmetric catalysis of L-threo-p-methylsulfonylphenylserine.
image file: c9cy01608b-f5.tif
Fig. 5 Time course for synthesis of L-threo-p-methylsulfonylphenylserine at 100 mL scale.

3. Conclusions

An L-threonine transaldolase from Pseudomonas sp. (PsLTTA) was mined and expressed in Escherichia coli BL21 (DE3). A substrate spectrum assay indicated that PsLTTA consumed only L-threonine as the donor substrate and could accept a wide range of aromatic aldehydes as the acceptor substrate. Through reaction condition optimization, the reaction occurred under kinetic control and L-threo-p-methylsulfonylphenylserine was obtained at more than 60% conversion and near-perfect stereoselectivity (94.5% de), the highest reported stereoselectivity so far obtained for L-threo-β-hydroxy-α-amino acids by enzymatic synthesis. Finally, synthesis of L-threo-p-methylsulfonylphenylserine at a 100 mL scale by whole-cell biocatalysis was conducted. Thus, for the first time, it has been systematically demonstrated that L-threonine transaldolase is a powerful biocatalyst for preparation of L-threo-β-hydroxy-α-amino acids. However, it should be noted that the production of L-threo-p-methylsulfonylphenylserine by PsLTTA was comparatively low, since the maximum concentration was only about 20 mM under the optimized conditions. In our view, this moderate production by PsLTTA was mainly due to its less efficient catalytic ability. Furthermore, the acetaldehyde produced during the reaction could also inhibit the synthesis of L-threo-p-methylsulfonylphenylserine (data not shown). Therefore, further studies, such as engineering PsLTTA or establishing an acetaldehyde removal system, are under way to improve L-threo-β-hydroxy-α-amino acid production.

4. Experimental

4.1 Materials

Escherichia coli BL21 (DE3) and plasmid pET28a were kept in our laboratory and served as host strain and expression vector, respectively. Yeast alcohol dehydrogenase (ADH), pyridoxal 5′-phosphate (PLP) and NADH were purchased from Aladdin (China). o-Phthaldialdehyde (OPA) and N-acetyl-cysteine (NAC) were obtained from Sigma-Aldrich (USA). L-Threo-p-methylsulfonylphenylserine standard was purchased from Shanghai Yuanye Bio-Technology Co., Ltd (China). All other chemicals and reagents used in this work were obtained commercially with the highest purity, unless otherwise stated.

4.2 Gene cloning, expression and purification

The DNA sequences of Pseudomonas sp. L-threonine transaldolase (PsLTTA) and Clavibacter michiganensis low specificity L-threonine aldolase (ClLTA), based on their protein sequences (GenBank No. WP_065936857 and WP_011931605), were optimized and synthesized by Wuhan GeneCreate Biological Engineering Co., Ltd. (China) and cloned into plasmid pET28a. The recombinant plasmids were further transformed into E. coli BL21 (DE3).

E. coli BL21 (DE3) cells (500 mL) harbouring pET28a-PsLTTA or pET28a-ClLTA were grown in LB medium containing 50 μg mL−1 kanamycin. Expression of PsLTTA and ClLTA was induced at 28 °C with 0.2 mM IPTG for 16 hours after the optical density at 600 nm (OD600) reached 0.5. Cells were pelleted at 6000g at 4 °C and were resuspended with Tris-HCl buffer (100 mM Tris-HCl, 200 mM NaCl, 25 mM imidazole, pH 7.0). Cells were lysed by ultrasonication and the soluble PsLTTA protein was purified by a Ni-IDA Prepacked Column (Sangon, China) according to the manufacturer's instructions. The purified enzyme was exchanged into Tris-HCl buffer (100 mM Tris-HCl, 20% glycerol, pH 7.0) by a PD-10 desalination column (GE Healthcare, USA) and further assessed by SDS-PAGE. Protein concentration was determined using a BCA Protein Assay Kit (Beyotime, China). For whole-cell biocatalysis, wet cells were collected, weighed and resuspended in Tris-HCl buffer (100 mM Tris-HCl, pH 7.0) with a concentration of 250 mg ml−1. The purified enzymes and wet cells were stored at −80 °C until use.

4.3 Enzyme activity assay

Enzyme activities of PsLTTA were determined by an assay with NADH coupled with ADH, as previously described.28 The reduction of NADH was detected by monitoring the decrease in absorbance at 340 nm (ε = 6220 M−1 cm−1) using a Multiskan Go microplate reader (Thermo Scientific, USA) at 25 °C. Reaction mixtures (180 μL) consisting of 10 mM p-methylsulfonyl benzaldehyde, 30 mM L-threonine, 0.2 mM PLP, 0.2 mM NADH and 10 U ADH in Tris-HCl buffer (100 mM Tris-HCl, pH 7.0, 10% acetonitrile) were incubated at 25 °C for 1 min. Reactions were initiated by introduction of 10 μg PsLTTA enzyme (in 20 μL Tris-HCl buffer) and monitored at 340 nm for 2 min. A boiled enzyme sample was used as a negative control. One unit (U) of transaldolase activity was defined as the amount of enzyme catalyzing the conversion of 1 μmol of p-methylsulfonylphenylserine per minute.29 All experiments were conducted in triplicate.

4.4 Determination of kinetic parameters

The kinetic parameters were determined by measuring the initial rate of enzymatic reaction at 10 mM p-methylsulfonyl benzaldehyde and varying the concentration of L-threonine (0.1–200 mM).29 Five independent replicates were performed for each concentration of L-threonine assayed and the data were fitted to the Michaelis–Menten equation using Origin 8.0 software.

4.5 Effect of temperature and pH on the stability and activity of PsLTTA

The pH stability of PsLTTA was evaluated by investigating the transaldolase ability after incubation at 4 °C for 12 h at various pH values using 100 mM citric acid–citrate (pH 3.0–6.0), phosphate (pH 6.0–8.0), Tris-HCl (pH 8.0–9.0) and carbonate–bicarbonate (pH 9.0–11.0) as buffer. The thermostability of the enzyme was assessed after 1 h incubation at temperatures ranging from 4 to 80 °C.

The pH dependence of the enzyme activity was studied at various pH values (pH 6.0–8.0) in 100 mM phosphate buffer. The effect of temperature on enzyme activity was assessed at temperatures ranging from 25 to 45 °C. The enzyme activity was measured as described above. All experiments were conducted in triplicate.

4.6 Effect of metal ions on transaldolase activity of PsLTTA

The influence of different metal ions on the transaldolase activity of PsLTTA was studied in the presence of the specific metal ions at final concentrations of 1 mM and 10 mM. The activity measured in the absence of metal ion was taken as a negative control and all experiments were performed in triplicate.

4.7 Chiral derivatization analysis of the stereospecificity of PsLTTA

The standard catalysis reaction was performed in 1 ml volume comprising 100 mM L-threonine, 40 mM p-methylsulfonyl benzaldehyde, 0.2 mM PLP, 1 mM MgCl2 and 2 μM PsLTTA (or 25 mg ml−1 wet cells) in Tris-HCl buffer (100 mM Tris-HCl, pH 7.0). The reaction was incubated at 30 °C for 4 h with constant shaking and terminated by addition of 2 ml MeOH. The reaction mixtures were incubated at 4 °C for 12 h and the supernatant was used for further study.

Determination of the conversion and stereospecificity of PsLTTA was performed by HPLC after derivatization with o-phthaldialdehyde/N-acetyl-cysteine (OPA/NAC).30 Briefly, the OPA/NAC reagent was obtained by dissolving 100 mg of NAC in 20 mL derivatization buffer (0.2 M boric acid, 0.2 M KCl) and then 25.6 mg of OPA in 5 mL MeOH was added. The OPA/NAC reagent was mixed with diluted reaction solution at a ratio of 4[thin space (1/6-em)]:[thin space (1/6-em)]1 and kept for 10 min at room temperature. Chromatographic analysis was carried out using a Shimadzu LC-20AT HPLC system (Japan) with a UV detector, at 236 and 340 nm. Column: Agilent ZORBAX reversed-phase columns (250 × 4.6 mm2, 5 μm), mobile phase: 50 mM KH2PO4, pH 8.0/acetonitrile (81/19), flow rate: 1 ml min−1, temperature: 30 °C.

4.8 Donor and acceptor substrate specificity

Donor substrate specificity of PsLTTA was evaluated using L-threonine, L-allothreonine, D-threonine, D-allothreonine, L-serine and glycine as donors and with p-methylsulfonyl benzaldehyde as the acceptor. For acceptor substrate specificity, a series of aromatic aldehydes and L-threonine were selected as acceptor and donor, respectively. Reaction mixtures were detected by enzyme activity assay or by analytical HPLC after OPA/NAC derivatization.

4.9 Effect of reaction conditions on the activity and stereospecificity of PsLTTA

The effect of reaction conditions on whole-cell biocatalysis was further determined. The initial catalysis was performed at 30 °C for 3 h, in a 1 ml volume comprising 100 mM L-threonine, 40 mM p-methylsulfonyl benzaldehyde, 0.2 mM PLP, 1 mM MgCl2 and 25 mg wet cells in Tris-HCl buffer (100 mM Tris-HCl, 10% CH3CN, pH 7.0). The reaction conditions, including amount of enzyme or substrate, co-solvents and reaction temperature, were evaluated further. The conversion and stereospecificity catalyzed by whole-cell biocatalysis were detected by analytical HPLC after OPA/NAC derivatization.

4.10 Reverse activity of PsLTTA

The reverse activity of PsLTTA was assessed using L-threo-p-methylsulfonylphenylserine and acetaldehyde as substrates. The reaction was performed in a 1 ml volume comprising 100 mM L-threo-p-methylsulfonylphenylserine, 50 mM acetaldehyde, 0.2 mM PLP, 1 mM MgCl2 and 25 mg ml−1 wet cells in Tris-HCl buffer (100 mM Tris-HCl, pH 7.0). The reaction was incubated at 30 °C for differing times with constant shaking, and was terminated by addition of 2 ml MeOH. The conversion was detected by analytical HPLC at 236 nm. The reaction catalyzed by ClLTA was performed as a positive control.

4.11 Synthesis of L-threo-p-methylsulfonylphenylserine in a large-scale 100 mL system by whole-cell biocatalysis

Transformation of L-threo-p-methylsulfonylphenylserine in a 100 mL system by whole-cell biocatalysis was performed under the optimized reaction conditions. Briefly, L-threonine (1.2 g, 10 mmol), p-methylsulfonyl benzaldehyde (0.55 g, 3 mmol), PLP (5 mg, 0.02 mmol), MgCl2 (10 mg, 0.1 mmol) and wet cells (0.625 g) were added to 100 ml Tris-HCl buffer (100 mM Tris-HCl, 10% ethyl acetate, pH 7.0). The reaction mixture was shaken constantly at 20 °C for 24 h. The time course of catalysis processing was measured in triplicate. The conversion and stereoselectivity were detected by analytical HPLC after OPA/NAC derivatization.

After reaction, the crude reaction supernatant was added to 20 mL ethanol and adjusted to pH 1.0 with concentrated sulfuric acid. For esterification, the mixture was incubated at 100 °C for 2 hours. The supernatant was extracted with chloroform, twice. The chloroform layer was combined and evaporated to dryness. The residue was purified using a C18 SPE column (CHCl3[thin space (1/6-em)]:[thin space (1/6-em)]MeOH = 2[thin space (1/6-em)]:[thin space (1/6-em)]1) to give 189 mg white solid, yield 33.0%. 1H NMR (400 MHz, CDCl3) δ 7.96 (d, J = 8.1 Hz, 2H), 7.62 (d, J = 8.1 Hz, 2H), 4.98 (d, J = 4.5 Hz, 1H), 4.19 (q, J = 7.0 Hz, 2H), 3.64 (d, J = 4.7 Hz, 1H), 3.08 (s, 3H), 1.23 (t, J = 7.1 Hz, 3H). 13C NMR (101 MHz, CDCl3) δ 172.79, 147.58, 139.91, 127.48, 127.26, 73.41, 61.64, 60.22, 44.53, 14.08. HRMS (m/z) (M+): calcd. for C12H18NO5S, 288.0897; found 288.0900.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

We thank Dr. Bing-De Zheng (College of Chemistry, Fuzhou University) for technical assistance with HRMS and NMR analysis. This study is financially supported by the Industry University Research project of Fuzhou University (No. 2019090501).

References

  1. M. Hibi, T. Kasahara, T. Kawashima, H. Yajima, S. Kozono, S. V. Smirnov and J. Ogawa, Adv. Synth. Catal., 2015, 357, 767–774 CrossRef CAS.
  2. J. Steinreiber, K. Fesko, C. Mayer, C. Reisinger, M. Schürmann and H. Griengl, Tetrahedron, 2007, 63, 8088–8093 CrossRef CAS.
  3. S. F. Beaudoin, M. P. Hanna, I. Ghiviriga and J. D. Stewart, Enzyme Microb. Technol., 2018, 119, 1–9 CrossRef CAS PubMed.
  4. G. H. Zhao, H. Li, W. Liu, W. G. Zhang, F. Zhang, Q. Liu and Q. C. Jiao, Amino Acids, 2011, 40, 215–220 CrossRef CAS PubMed.
  5. I. B. Seiple, J. A. Mercer, R. J. Sussman, Z. Zhang and A. G. Myers, Angew. Chem., Int. Ed., 2014, 53, 4642–4647 CrossRef CAS PubMed.
  6. P. Barbie and U. Kazmaier, Org. Lett., 2015, 18, 204–207 CrossRef PubMed.
  7. H. J. Gwon and S. H. Baik, Biotechnol. Lett., 2010, 32, 143 CrossRef CAS PubMed.
  8. A. Geranurimi and W. D. Lubell, Org. Lett., 2018, 20, 6126–6129 CrossRef CAS PubMed.
  9. O. A. Levitskiy, Y. K. Grishin and T. V. Magdesieva, Eur. J. Org. Chem., 2019, 20, 3174–3182 CrossRef.
  10. H. C. Kolb, M. S. VanNieuwenhze and K. B. Sharpless, Chem. Rev., 1994, 948, 2483–2547 CrossRef.
  11. T. Katsuki and K. B. Sharpless, J. Am. Chem. Soc., 1980, 102, 5974–5976 CrossRef CAS.
  12. J. P. Adams, M. J. Brown, A. Diaz-Rodriguez, R. C. Lloyd and G. D. Roiban, Adv. Synth. Catal., 2019, 361, 2421–2432 CAS.
  13. M. L. Di Salvo, S. G. Remesh, M. Vivoli, M. S. Ghatge, A. Paiardini, S. D'Aguanno and R. Contestabile, FEBS J., 2014, 281, 129–145 CrossRef CAS PubMed.
  14. K. Fesko, Appl. Microbiol. Biotechnol., 2016, 100, 2579–2590 CrossRef CAS PubMed.
  15. Y. L. Du and K. S. Ryan, Nat. Prod. Rep., 2019, 36, 430–457 RSC.
  16. J. Blesl, M. Trobe, F. Anderl, R. Breinbauer, G. A. Strohmeier and K. Fesko, ChemCatChem, 2018, 10, 3453–3458 CrossRef CAS PubMed.
  17. K. Fesko, D. Suplatov and V. Švedas, FEBS Open Bio, 2018, 8, 1013–1028 CrossRef CAS PubMed.
  18. Q. Chen, X. Chen, J. Feng, Q. Wu, D. Zhu and Y. Ma, ACS Catal., 2019, 9, 4462–4469 CrossRef CAS.
  19. K. Fesko, C. Reisinger, J. Steinreiber, H. Weber, M. Schürmann and H. Griengl, J. Mol. Catal. B: Enzym., 2008, 52, 19–26 CrossRef.
  20. Q. Chen, X. Chen, Y. Cui, J. Ren, W. Lu, J. Feng and D. Zhu, Catal. Sci. Technol., 2017, 7, 5964–5973 RSC.
  21. R. Ushimaru and H. W. Liu, J. Am. Chem. Soc., 2019, 141, 2211–2214 CrossRef CAS PubMed.
  22. T. A. Scott, D. Heine, Z. Qin and B. Wilkinson, Nat. Commun., 2017, 8, 15935 CrossRef CAS PubMed.
  23. J. E. Schaffer, M. R. Reck, N. K. Prasad and T. A. Wencewicz, Nat. Chem. Biol., 2017, 13, 737 CrossRef CAS PubMed.
  24. S. Barnard-Britson, X. Chi, K. Nonaka, A. P. Spork, N. Tibrewal, A. Goswami and S. G. Van Lanen, J. Am. Chem. Soc., 2012, 134, 18514–18517 CrossRef CAS PubMed.
  25. K. Baer, N. Dückers, T. Rosenbaum, C. Leggewie, S. Simon, M. Kraußer and H. Gröger, Tetrahedron: Asymmetry, 2011, 22, 925–928 CrossRef CAS.
  26. K. Fesko, Front. Bioeng. Biotechnol., 2019, 7, 119 CrossRef PubMed.
  27. K. Fesko, G. A. Strohmeier and R. Breinbauer, Appl. Microbiol. Biotechnol., 2015, 99, 9651–9661 CrossRef CAS PubMed.
  28. S. L. Goldberg, A. Goswami, Z. Guo, Y. Chan and M. A. Schmidt, Org. Process Res. Dev., 2015, 19, 1308–1316 CrossRef CAS.
  29. S. Barnard-Britson, X. Chi, K. Nonaka, A. P. Spork, N. Tibrewal, A. Goswami and S. G. Van Lanen, J. Am. Chem. Soc., 2012, 134, 18514–18517 CrossRef CAS PubMed.
  30. J. Steinreiber, K. Fesko, C. Reisinger, M. Schürmann, F. van Assema, M. Wolberg and H. Griengl, Tetrahedron, 2007, 63, 918–926 CrossRef CAS.

Footnotes

Electronic supplementary information (ESI) available. See DOI: 10.1039/c9cy01608b
These authors contributed equally to this study.

This journal is © The Royal Society of Chemistry 2019