Appl Microbiol Biotechnol DOI 10.1007/s00253-016-7613-6

BIOTECHNOLOGICALLY RELEVANT ENZYMES AND PROTEINS

Identification of catalysis, substrate, and coenzyme binding sites and improvement catalytic efficiency of formate dehydrogenase from Candida boidinii Wei Jiang 1,2 & Peng Lin 1,2 & Ruonan Yang 1,2 & Baishan Fang 1,2,3

Received: 31 March 2016 / Revised: 24 April 2016 / Accepted: 3 May 2016 # Springer-Verlag Berlin Heidelberg 2016

Abstract Formate dehydrogenases (FDHs) are continually used for the cofactor regeneration in biocatalysis and biotransformation with hiring NAD(P)H-dependent oxidoreductases. Major weaknesses of most native FDHs are their low activity and operational stability in the catalytic reaction. In this work, the FDH from Candida boidinii (CboFDH) was engineered in order to gain an enzyme with high activity and better operational stability. Through comparing and analyzing its spatial structure with other FDHs, the catalysis, substrate, and coenzyme binding sites of the CboFDH were identified. To improve its performance, amino acids, which concentrated on the enzyme active site or in the conserved NAD+ and substrate binding motif, were mutated. The mutant V120S had the highest catalytic efficiency (kcat/Km) with COONH4 as it enhanced the catalytic velocity (kcat) and kcat/Km 3.48-fold and 1.60-fold, respectively, than that of the wild type. And, the double-mutant V120S-N187D had the highest kcat/Km with NAD+ as it displayed an approximately 1.50-fold increase in kcat/Km. The mutants showed higher catalytic efficiency than other reported FDHs, suggesting that the mutation has achieved good results. The single and double mutants exhibited higher thermostability Electronic supplementary material The online version of this article (doi:10.1007/s00253-016-7613-6) contains supplementary material, which is available to authorized users. * Baishan Fang [email protected] 1

Department of Chemical and Biochemical Engineering, College of Chemistry and Chemical Engineering, Xiamen University, Xiamen 361005, China

2

The Key Lab for Synthetic Biotechnology of Xiamen City, Xiamen University, Xiamen 361005, China

3

The Key Laboratory for Chemical Biology of Fujian Province, Xiamen University, Xiamen, Fujian 361005, China

than the wild type. The structure-function relationship of single and double mutants was analyzed by homology models and site parsing. Asymmetric synthesis of L-tert-leucine was executed to evaluate the ability of cofactor regeneration of the mutants with about 100 % conversion rates. This work provides a helpful theoretical reference for the evolution of an enzyme in vitro and promotion of the industrial production of chiral compounds, e.g., amino acid and chiral amine. Keywords Biocatalyst . Molecular modeling . Biocatalysis . Chiral compounds . Cofactor regeneration

Introduction NAD(P)H-dependent oxidoreductases are meritorious biocatalysts for the industrial production of chiral compounds, such as chiral amine and amino acid (Fischer and Pietruszka 2010; Hoelsch et al. 2012; Wohlgemuth 2010). Efficient methods for cofactor regeneration are required as the high cost of coenzymes (Donk et al. 2003; Wichmann and Vasic-Racki 2005). The common solution is the addition of a second oxidoreductase in the reaction system to regenerate the consumption coenzyme at the cost of a cheap substrate. Formate dehydrogenase (FDH; EC 1.2.1.2), which catalyzes formate to CO2, is one of the most appropriate enzymes for this aim (Donk et al. 2003; Tishkov and Popov 2006). And, the bioconversion of formate to carbon dioxide is typically reaction in biological systems by FDH because the formate is cheap and the generating of byproduct (CO2) effectively volatilizes to promote the reaction to complete. Moreover, CO2 is inert compound and does not inhibit or inactivate the enzyme in the reaction system. The carbon dioxide is effortlessly removed from the system, so it neither interferes and advantageous to the purification of products. In addition, a representative biotechnological application of this

Appl Microbiol Biotechnol

enzyme is to measure the oxalic acid and formic acid in solution, while it is also a powerful tool for NAD+ regeneration system (Jiang et al. 2015; Koutinas et al. 2014; Shaked and Whitesides 1980; Yoshimoto et al. 2010), especially after immobilization (Akers et al. 2005; Bolivar et al. 2006; Demir et al. 2011; Lu et al. 2006). Using alginate gels as backing material, the best recyclability property has been achieved (Addo et al. 2010). Another important characteristic of FDH is its role as a critical catalyst in the enzymatic bioconversion of CO2 to methanol (Hong et al. 2004; Jeon et al. 2004; Munro et al. 1975), which plays a significant role in development of new methods to fabrication of chemicals and fuels. Based on these advantages, the FDH as a kind of excellent biological catalyst should be focused on and studied. Several methods and material supports have already been studied to enhance FDH’s activity (Ansorge-Schumacher et al. 2006; Carter et al. 2014; Hoelsch et al. 2012; Netto et al. 2012), but the activity and performance of FDH still cannot meet the needs of scientific research and industrial production as it also displays some disadvantages, e.g., low operational stabilities and specific activities. The directed evolution of FDH through error-prone PCR had been executed (AnsorgeSchumacher et al. 2006; Carter et al. 2014); however, the mutants’ activity was still not very high and it included a long time-consuming screening process. Improving the activity and performance of FDH is particularly urgent as the requirements of environmentally friendly industrial model, and it is one of the important materials for providing coenzyme regeneration. And, a systematic research focusing on the improvement of the stability and enzyme activity is yet missing. Rational design and site-directed mutagenesis, two powerful tool in anatomizing the structure and function of proteins, were used to enhance FDH’s activity and performance in this work. Herein, we identified the catalysis, substrate, and coenzyme binding sites of the CboFDH from Candida boidinii (C. boidinii) by comparing the sequence and spatial structure with other FDHs. Then, the important sites, 120 concentrated on the enzyme active site and 187 near to the coenzyme binding site (Fig. 4), for affect substrate and coenzyme binding were identified and mutated by rational design and sitedirected mutagenesis. The characterization of the CboFDH, single and double CboFDH mutants, was implemented with respect to their stability and kinetic properties.

Material and methods Chemicals, plasmids, and strains Enzymes, e.g., DNA polymerase, Ex Taq™ DNA, DpnI, and restriction endonucleases, were obtained from TaKaRa Co., Ltd. (Dalian, China). Ligation solution I, Agarose Gel DNA, Mutan BEST Kit, Purification Kit version 2.0, and Agarose

Gel DNA Fragment Recovery Kit version 2.0 were purchased from Omiga Co., Ltd. (Xiamen, China). Reagents and buffers, chromatographically pure or analytically graded, were from either Omiga China or Sigma China. Strains were cultivated in a Luria-Bertani (LB) medium at 37 °C. Escherichia coli DH5α (E. coli DH5α; TaKaRa) was used for general cloning. E. coli BL21 (DE3; Novagen, USA) was used as the host strain for protein production. The pET-28aleudh containing the LeuDH gene (1101 bp, Gene ID: NC_004722.1) from Bacillus cereus were constructed in our group (data unpublished). Plasmid pET-28a was used for expression of CboFDH and its mutants. The sequence information of the FDH gene from bacterium C. boidinii (ATCC 32195) was obtained by previously reported (Kula et al. 2003; Slusarczyk et al. 2000). And, the CboFDH gene was synthesized by Sangon Biotech (Shanghai, China) Co., Ltd. Site-directed mutagenesis The CboFDH catalysis, substrate, and coenzyme binding sites should be confirmed for improving the efficiency of mutation to obtain the mutant with high activity and superior performance. The FDH from Pseudomonas sp. 101 (Psp101-FDH) was selected as the template for its crystal structure and active sites have been reported (Filippova et al. 2005; Nilov et al. 2012). The three-dimensional structure alignment of the CboFDH and Psp101-FDH was implemented, and the result was shown in Fig. 1. The corresponding sites in CboFDH to the Psp101-FDH catalysis and substrate binding sites were highlighted (Fig. 1). PDB files of the CboFDH (2fss) and Psp101-FDH (2GO1) were obtained from the PDB database. The amino acid sequence alignments of CboFDH and other FDHs were executed by ClustalW in MegAlign program. According to the above results, two distinct sites (Fig. S1) with Val at position 120 and Asn at 187 were observed for the next experiment. The plasmid pET-28a-fdh as the template and primers were designed from pairs of complementary oligonucleotides including the purpose mutants (Table S1). The site-directed mutagenesis was implemented using a one-step overlap extension PCR (overlap PCR) with the following program: (i) 97 °C for 2 min; (ii) 20 cycles of 95 °C for 20 s, 55 °C for 20 s, and 72 °C for 5 min 30 s; and (iii) 72 °C for 7 min (Jiang et al. 2014; Xu et al. 2014). PCR mixture (40 μL) was composed of 5-ng template pET-28a-fdh, 0.3 μM each oligonucleotide, 0.8 μM dNTP, and 1 U of Pfu DNA polymerase (TransGen, China). The mutant V120S was obtained using the wild pET-28a-fdh as template with the forward primer V120S-F and the reverse primer V120S-R, so did the mutant N187D with the forward primer N187D-F and the reverse primer N187D-R. The double mutation, V120S/N187D, was generated with the circled pET-28aV120S as template and the primers N187D-F/R. Then, the PCR product was purified and treated with DpnI (10 U/L) to remove the circled template through incubation at 37 °C overnight and

Appl Microbiol Biotechnol Fig. 1 The three-dimensional structure alignment of the CboFDH and Psp101-FDH. PDB files of the CboFDH (2fss) and Psp101-FDH (2GO1) were obtained from PDB database. a The three-dimensional structure alignment of the CboFDH (red) and Psp101-FDH (green). b The residues essential for FDH catalysis and substrate binding are highlighted in the threedimensional structure alignment of the CboFDH (red) and Psp101FDH (blue). c Locations for the residues, which are essential for FDH catalysis and substrate binding, are marked in the threedimensional structure alignment of the CboFDH (red) and Psp101FDH (blue). d The essential residue alignment of the CboFDH and Psp101-FDH

Appl Microbiol Biotechnol

then transformed into DMT competent cells. The raring mutants were selected, confirmed by DNA sequencing, and the successfully desired mutation was designated as pET-28a-V120S, pET28a-N187D, and pET-28a-V120S-N187D. After that, the plasmids were transformed into BL21 (DE3) for protein expression and enzymatic property study. Heterologous expression and purification of wild-type CboFDH and mutants The desired plasmids were transformed into BL21 (DE3) for protein expression, and the strains were cultivated in LB medium (containing 100 mg/mL kanamycin). After culture at 37 °C for 14–18 h, the mixture was transferred to a new LB medium (1.5:100 dilution, containing 100 μg/mL kanamycin) and cultivated at 37 °C for 2–3 h until the optical density reached 0.5– 0.6 at 600 nm. A final concentration of 0.15 mM isopropylbeta-D-thiogalactopyranoside (IPTG) was added. After 10-h growths at 21 °C, the cells were collected centrifugation at 8000 rpm for 3 min, washed triple, and resuspended in PBS buffer (Shanghai Biological Engineering Co., Ltd., China). The treated cells were suspended in 2-[4-(2-hydroxyethyl)-1piperazinyl-ethanesulfonicacid (HEPES; pH 7.2), disrupted by an Ultrasonic Cell Disruptor, and the crude enzyme was collected, followed by centrifugation 30 min at 4 °C, 8000g. Finally, these enzymes were purified by AKTA Purification System. The results of the purification were detected by SDSPAGE using a 12 % (w/v) polyacrylamide separating gel, and the concentrations of the pure proteins were detected using the Bradford reagent. Enzyme activity assays and kinetic parameter studies Enzyme activity was detected from the NADH absorbance at 340 nm (molar extinction coefficient of 6.22 mM−1 cm−1, ε = 6220 M −1 cm −1 ) with a Tecan Infinite M200PRO (Salzburg, Austria) at 30 °C (Hoelsch et al. 2012). The results are average values, detection at least three times. One unit (U) of the enzyme activity was defined as the number of micromoles of NADH consumed in 1 min by 1-mg enzyme. Kinetic parameters (Km, Vmax, kcat, and kcat/Km) were confirmed using different concentration of formate or NAD+ in accordance with Lineweaver–Burk plot. All the tests were performed at optimum conditions of the different enzymes. Measurement of temperature and pH influences The optimal temperature of the different enzymes was investigated by detecting the enzyme activity at a different temperatures range of 20–70 °C after pre-incubating the reaction system (without enzyme NADH) at corresponding temperatures for 5 min. And, the thermal stability of the different

enzymes was determined by pre-incubating the enzyme at temperatures from 30 to 70 °C for 1 h, and then, the residual enzyme activity was detected under the described above method. The optimal pH of the four enzymes was studied in different buffers at pH 6.0–11.0, namely, 50 mM glycine-NaOH buffer (pH 9.0–11.0), 50 mM Tris–HCl buffer (pH 8.0–9.0), and 200 mM phosphate buffer (pH 6.0–8.0). The pH stability was measured of the four enzymes by pre-incubating them in different pH value of pH 6.5–11.0 at 4 °C for 30 h. Then, the residual activity of the incubated samples was investigated according to the standard method. Molecular modeling To date, a significant section of the information securable on the structure-function relationships of FDH has been confirmed from crystallographic structures. The crystal structure and active sites of the FDH from Pseudomonas sp. 101 (Psp101-FDH) have been reported (Filippova et al. 2005; Nilov et al. 2012; Popov and Lamzin 1994). And, the important sites for coenzyme binding of the FDH from Mycobacterium vaccae N10 (MycFDH) had been labeled (Hoelsch et al. 2012). Even though the CboFDH’s crystal structure had been reported (Schirwitz et al. 2007), important sites of the CboFDH are rarely parsed. Based on the high identity of CboFDH, MycFDH, and Psp101-FDH, the information of the MycFDH and Psp101-FDH was used to identify the catalysis, substrate, and coenzyme binding sites of the CboFDH. Swiss-model (http://swissmodel.expasy.org/) was selected to predict protein structures through identifying FDH senior structural homologue (Mourad et al. 2011). Using the crystal structure of FDHs from C. boidinii and Pseudomonas sp. 101 (PDB ID 2fss and 2GO1) as templates, homology structure modeling of the CboFDH mutants was generated. Detection of the ability of coenzyme regeneration of the CboFDH and mutants The purpose plasmids, pET-28a-fdh, pET-28a-V120S, pET28a-N187D, and pET-28a-V120S-N187D, were transformed into BL21(DE3); then, the engineering bacteria were inoculated, induced, cultivated, collected, centrifuged (8000 rpm, 3 min), and washed by phosphate buffer (0.2 M, pH 7.0), tripled, collected (each was 6, 20 mg, wet weigh), and stored at −80 °C overnight (Jiang et al. 2015). After that, cells were thawed at 20 °C and re-suspended with NH3 · H2O-NH4Cl buffer (1 M, pH 9.0) and introduced into a 10-mL reaction system which composed of ammonium formate (200 mM), TMA (100 mM), NH3 · H2O-NH4Cl buffer (1 M, pH 9.0), NAD+ (50 mM), and moderate enzymes. The enzymatic reactions were performed at 30 °C, 200 rpm. Two-hundred microliter was sampled during 0–24 h from the reaction system. Meanwhile, pH was adjusted to 8.0–9.5. Finally, HPLC was

Appl Microbiol Biotechnol

utilized to analyze the L-tert-leucine concentration in the enzymatic reaction systems (Jiang et al. 2015).

Results Identification of catalysis, substrate, and NAD+ binding sites The success or failure of rational design depends upon the determination of the enzyme active site and the selection of mutant site. Although the CboFDH’ crystal structure had been reported (Schirwitz et al. 2007), the analysis of the important sites of this enzyme is insufficient, suggesting that it would be benefited to enhance its activity and improve its performance by confirming the catalytic site of the CboFDH. Base on the information of the active sites and crystal structure of the Psp101-FDH and MycFDH (Filippova et al. 2005; Hoelsch et al. 2012; Nilov et al. 2012; Popov and Lamzin 1994), the catalysis, substrate, and NAD+ binding sites of the CboFDH were identified. The three-dimensional structure of Psp101FDH and CboFDH was compared (Fig. 1), and it was shown that the two enzymes have high similarity on the structure, especially in some key areas, suggesting that their conservative regions are similar. The catalysis and substrate binding sites, 67P, 98P, 99F, 93V, 94G, 119N, 123V, 256D, 257A, 260A, 283V, 287Q, 311H, 313S, 314G, and 321T of the CboFDH, were confirmed by the space and the multiple-sequence alignment (Figs. 1 and 2) and some pre-reports (Filippova et al. 2005; Hoelsch et al. 2012; Nilov et al. 2012; Popov and Lamzin 1994). And, these sites, 93V, 119N, 257A, and 311H, which correspond to the important site of the Psp101-FDH and MycFDH (Filippova et al. 2005; Hoelsch et al. 2012; Nilov et al. 2012; Popov and Lamzin 1994), were demonstrated as the critical sites for enzyme catalysis and substrate binding. Kathrin Hoelsch et al. demonstrated the cofactor binding sites of the MycFDH to alter its substrate specificity (Hoelsch et al. 2012), and the molecular modeling was used to analyze the substrate channel and coenzyme binding sites of the MycFDH (Nilov et al. 2012). According to the information, the cofactor binding sites of the CboFDH were highlighted in Fig. 2. Site-directed mutagenesis, expression, purification of CboFDH, and mutants According to the results in the BSite-directed mutagenesis^ and BIdentification of catalysis, substrate, and NAD+ binding sites^ sections, two distinct sites, 120 Val and 187 Asn, were observed to mutate by site-directed mutagenesis for improvement in enzyme activity and property. Through overlap PCR, two single-site mutants (V120S, N187D) and one chimeric mutant (V120S-N187D) were obtained. Sequence

comparison of wild type and the mutants has shown a Ser substitution of Val at 120 position and a Asn substitution of Asp at 187 position, followed by measuring their effects on enzymatic property. After expressions of the CboFDH, V120S, N187D, and V120S-N187D, these enzymes were harvested, purified, and detected. And, they resolved to a single band (Fig. S2). It can be seen that the change of site residues did not affect the expression of the protein (Fig. S2). These sites, 120 and 187, were selected as the 120 site concentrated on the enzyme active site and 187 near to the coenzyme binding site (Fig. 4). The change of these sites would be affect substrate and coenzyme binding. Biochemical properties of CboFDH and mutants The optimal temperature for mutant V120S, mutant N187D, and mutant V120S-N187D was 60, 65, and 60, respectively (Fig. 3a). The three mutants exhibited higher activity than the wild type (Fig. S3a), especially during 20–30 and 65–70 °C. After pre-treated at a different temperature for 1 h, the three mutants exhibited high thermostability during 30–45 °C. V120S, N187D, and V120S-N187D exhibited higher thermostability at 40–45 °C than the CboFDH, while V120S displayed the highest thermostability at 45 °C. V120S and V120S-N187D exhibited the highest activity at pH 9.5 as same as the wild-type CboFDH, while N187D exhibited its optimal activity at pH 10.0 (Fig. S3a). However, V120S, N187D, and V120S-N187D displayed higher activity than the CboFDH (Fig. S3a). The three mutants showed the similar pH stability profiles to the CboFDH, but the V120S-N187D has shown better pH stability than the CboFDH and two other mutants over a pH value ranging from 9.5 to 11.0 (Fig. S3b). The three mutants exhibited high stability under the weak acid and alkaline condition. Steady-state Michaelis–Menten kinetic parameters The steady-state kinetic constants were measured to contrast the catalytic reactions between the CboFDH and three mutants (Table 1). Mutant V120S showed an approximately 3.48-fold increase in kcat and 1.60-fold increase in catalytic efficiency (kcat/Km) for COONH4 than that of the wild type (WT; CboFDH). The results indicated that the catalytic velocity and catalytic efficiency were significantly enhanced by replacing the valine with serine at the 120 position. It was confirmed that the catalytic velocity and catalytic efficiency were significantly improved by replacing the asparagine with aspartate in the 187 site as the kcat and kcat/Km for COONH4 of the N187D was 2.14-fold and 1.16-fold than that of the WT. Mutant V120S-N187D showed a 31 % decrease in Km and 0.92-fold decrease in the kcat, resulting that its kcat/Km was 0.65-fold of the WT, suggesting that the affinity of the substrate and

Appl Microbiol Biotechnol

Fig. 2 Multiple FDH sequence comparison by ClustalW. The residues essential for FDH catalysis and substrate binding are highlighted in blue box, and particularly important sites are marked by pentagon (red). The residues essential for coenzyme NAD+ binding are highlighted in red box, and particularly important sites are marked by quadrangle (yellow).

Sequences are FDH from Mycobacterium vaccae N10 (Myc10-FDH) [Hoelsch, K. et al. 2012], FDH from Pseudomonas sp. 101 (Psp101FDH) [Serov AE et al. 2002], FDH from M. vaccae (MycFDH) [NCBI database], and FDH from Candida boidinii (CboFDH) [Andreadeli A et al. 2008]

enzyme was significantly improved by combining the two single-point mutations. Mutant V120S showed a 182 % increase in the catalytic turnover number (kcat) and a 238 % increase in Km, leading to about a 24 % decrease in (kcat/Km) for NAD+, suggesting that this mutant can enhance the reaction velocity of CboFDH. N187D displayed about no change in Km and a 130 % increase in k cat for NAD +, resulting in about a 148 % increase in catalytic efficiency. The results showed

that the catalytic velocity and catalytic efficiency were significantly improved by replacing the asparagine with aspartate at the 187 position. Mutant V120S-N187D displayed a 40 % decrease in Km and 0.16-fold decrease in kcat for NAD+, leading to approximately 1.50-fold increase in catalytic efficiency. Even though the kcat of the mutant was enhanced by a factor of 5.8, the kcat (75 min−1) value (Carter et al. 2014) was still less than the three mutants. The kinetic parameters of wild-type CboFDH were

Fig. 3 Activity and thermostability of the CboFDH and mutants at different temperatures. a Effect of temperature on the activity of the CboFDH and mutants. The optimal temperature was determined by measuring the activity at temperatures from 20 to 70 °C. The maximal activity for each one was taken as 100 %. b For the thermal stability of the CboFDH and mutants, the purified enzyme was pre-treated at a different

temperature for 1 h. The activity of the enzyme without pre-incubation was defined as 100 %. Black del operator represents the CboFDH, black circles represent the V120S, black squares represent the N187D, and black triangle represents the V120S-N187D. All experiments were performed three times, and each time had three to five parallel experiments. Error bars represent the standard deviation

Appl Microbiol Biotechnol Table 1 The kinetic parameters for the CboFDH and mutants

Enzyme

Substrate

Km (mM)

kcat (min−1)

CboFDH V120S N187D

COONH4

3.63 ± 0.10

92.05 ± 2.1

25.38

7.90 ± 0.32 6.70 ± 0.21

320.13 ± 12.7 197.13 ± 5.21

40.55 29.46

2.50 ± 0.15 3.40 ± 0.15

41.80 ± 0.62 1338.08 ± 32.8

16.51 397.10

V120S

8.81 ± 0.18

2437.97 ± 0.001

300.49

N187D V120S/N187D

3.80 ± 0.22 2.02 ± 0.11

1739.15 ± 3.36 1121.05 ± 23.49

588.48 596.05

V120S/N187D CboFDH

NAD+

kcat/Km (mM−1 min−1)

The pure CboFDH and mutants were obtained with His-tag. All experiments were performed three times, and each time had three to five parallel experiments. Data represent the mean ± standard deviation of triplicate samples

different, maybe because the measurement condition for the optimum pH of the CboFDH was about 9.5 and it exhibited low activity during pH 7.0 and 7.5 (Fig. S3).

Fig. 4 The locations of the FDH catalysis, substrate, and coenzyme NAD+ binding sites. a The locations of the FDH catalysis and substrate binding sites are marked as mazarine, and particularly important sites are marked as brownish red. b The locations of the coenzyme NAD+ binding sites are marked as yellow, and particularly important sites are marked as red. c The locations of the FDH catalysis, substrate, and coenzyme NAD+ binding sites. Particularly important sites are marked as brownish red (catalysis and substrate binding sites) and red (coenzyme NAD+ binding sites)

These results indicated that the N187D and double mutant were helpful to improve the ability of CboFDH of catalysis NAD+ to NADH.

Appl Microbiol Biotechnol

Molecular modeling

Asymmetric synthesis with coenzyme regeneration

Homology structure modeling of the three mutants was generated by using the crystal structure of CboFDH and Psp101FDH (PDB ID 2fss and 2GO1, respectively) as templates. It was shown in Fig. 1 that the two enzymes have a high overlap of three-dimensional structures, especially in the key section. Then, the locations and residues of the FDH catalysis, substrate, and coenzyme NAD+ binding sites were identified and displayed (Figs. 4 and 5). The distance and spatial distribution of the mutations and active sites were highlighted to explain why the mutation can influence the enzyme activity and stability to further resolve the relationship of structure and function (Fig. 6). It could be seen from the chart that the 120 site was located on the catalysis and substrate-binding site, close to the active sites (119N and 93V) and the coenzyme binding site 118S. To 187 site of the CboFDH, the improvement of the catalytic efficiency for cofactors that may be attributed to this site was close to the coenzyme binding site 161D.

Asymmetric synthesis with cofactor regeneration was implemented to detect the ability of the three mutants of providing coenzyme regeneration (Scheme 1). The catalytic reactions were executed and L-tert-leucine was measured by HPLC. The V120S, N187D, and V120S-N187D that exhibited improved performance than that of the WT for the percent conversions were 53.71 % (CboFDH), 99.24 % (V120S), 100.00 % (N187D), and 81.21 % (V120S-N187D; Fig. 7).

Fig. 5 The residues of the FDH catalysis, substrate, and coenzyme NAD+ binding sites. a, b The locations of the 120 (cyan) and 187 sites (blue). c The residues of the FDH catalysis, substrate, and coenzyme NAD+ binding sites. d The residues of the coenzyme NAD+ binding sites. e The important residues of catalysis, substrate, and coenzyme NAD+ binding sites

Discussion The FDH from C. boidinii was chosen as its prominent value as a coenzyme regeneration donor in biosyntheses (AnsorgeSchumacher et al. 2006; Jiang et al. 2015), and only a very few researches deal with enhancing its enzymatic activities. As an alternative, improvement of the biocatalyst might be achievable through mutagenesis of the protein sequence.

Appl Microbiol Biotechnol

Fig. 6 The distribution of the FDH catalysis, substrate, and coenzyme NAD+ binding sites, 120 (cyan) and 187 residues (blue). a–c The distribution of the FDH catalysis and substrate binding residues is

displayed. d–f The distribution of the coenzyme NAD + binding residues is displayed. g–j The distribution of the important residues for FDH catalysis, substrate, and coenzyme NAD+ binding is displayed

Thermal stability of the V120S was enhanced mainly due to the improvement of the enzymatic activity. Thermal stabilities of the N187D and V120S-N187D that were improved mainly attribute to the N187D located far from the active site. Previous researches have reported similar results that the thermal stability

or solvent tolerance was improved mainly because the beneficial mutations often located far from the active site (Carter et al. 2014). It had been demonstrated that the replacement of one or more amino acids on the surface of an enzyme had different influences on its stability according to the surroundings of the mutation site(s) (Bhardwaj et al. 2010; Funahashi et al. 2000; Perl et al. 2000). Since aspartate is an acidic amino acid and asparagine is a neutral amino acid, the substitution of aspartate for asparagine might lead to the variation in the total number of acidic amino acids on the surface of the enzyme. Therefore, N187D and V120S-N187D are promising to help the enzyme maintain its activity at a higher-alkaline environment and protect the enzyme core from the OH attack. Except for V120S, N187D was situated on the surface of CboFDH (Fig. 5b).

Scheme 1 Asymmetric synthesis with cofactor regeneration

Appl Microbiol Biotechnol

Fig. 7 Detection of the ability of the CboFDH and mutants to provide coenzyme regeneration. HPLC was used to analyze the L-tert-leucine concentration in the enzymatic reaction systems under 30 °C during 0– 4 h. The 10-mL reaction system composed of ammonium formate (200 mM), TMA (100 mM), NH3 · H2O-NH4Cl buffer (1 M, pH 9.0), NAD+ (50 mM), and moderate enzymes. a The L-tert-leucine concentration in the enzymatic reaction systems under 30 °C during 0–4 h. b The L-

tert-leucine concentration in the enzymatic reaction systems under 30 °C at 0.5 h. Black del operator represents the CboFDH, black circles represent the V120S, black squares represent the N187D, and black triangle represents the V120S-N187D. All experiments were performed three times, and each time had two to three parallel experiments. Note that error bars represent the standard deviation

Further exploration should be implemented to clarify the mechanism, how the mutated residue influences the enzyme activity, while it still unclear. Based on comparison of three-dimensional structure and amino acids, it can be deduced that the serine is located on the catalysis and substrate-binding site and in the β-sheet strands and aspartate is located on the cofactor-binding site and in the β-sheet strands (Hoelsch et al. 2012; Popov and Lamzin 1994) (Fig. 5). Replacement of valine with a polar amino acid generated alters in the catalytic velocity (kcat), maybe because polar amino acids brought some different charges. The acidic amino acids and polar amino acid with a negative charge might increase the catalytic velocity. When valine was displaced in polar amino acids, the size and space location of the amino acid maybe had a major effect on catalytic velocity. This probably could be used to explain why V120S can enhance the catalytic velocity of substrates and coenzyme. The 2.60-fold increase in catalytic efficiency for COONH4 of the V120S that might be contributed to the 120 site was located on the catalysis and substrate-binding site and close to the active site (119N). As the V120S was close to the 118S, a coenzyme binding site, it could improve the catalytic velocity of the enzyme for 1.82-fold. When enzymes’ activity or enantioselectivity is enhanced, it is mainly because mutations are mostly concentrated on or near to the active site (Popov and Lamzin 1994). In a similar way, aspartate is located on the cofactorbinding site and in the β-sheet strands (Hoelsch et al. 2012; Popov and Lamzin 1994) (Fig. 5). It was conformed that the introduction of an aspartate, an acidic amino acid with a negative charge, at 187 positions could increase the catalytic velocity. And, an amino acid with a negative charge would be more easily combined with the coenzyme which brings

electropositive charge. For the interaction between enzyme and coenzyme, the N187D owns higher influence than the V120S as the catalytic efficiencies of the V120S, N187D, and V120S-N187D were 0.76-, 1.48-, and 1.50-fold than that of the WT. Even though the CboFDH activity was significantly increased, its thermal stability increase was not significant. It appears that the researches about mutants generating the improvement of multiple performance were rare (Baik et al. 2003), so improving the stability of the CboFDH can be served as the next research direction on the basis of this study. Kinetic parameters of COONH 4 and cofactor of the CboFDH and three mutants were compared with other FDHs (Table 2). The Km for COONH4 of the GraFDH from Granulicella mallensis MP5ACTX8 (Fogal et al. 2015) was 32-fold than that of the V120S-N187D, suggesting that the V120S-N187D had higher affinity with substrate. The kcat/ Km for COONH4 of the V120S and N187D were 1.99- and 1.44-fold than that of the BaFDH from Bacillus sp. and 11.26and 8.18-fold than that of the CmeFDH from Candida methylica. These results indicated that the V120S and N187D have high enzyme activity. The kcat for NAD+ of the V120S, N187D, and V120S-N187D were 4.94, 3.53, and 2.27 times to the MycFDH C145S/D221Q/C255V from M. vaccae N10 with mutant (Hoelsch et al. 2012), approximately 7.84-, 5.60-, and 3.60-fold to the MycFDH C145S/A198G/D221Q/ C255V from M. vaccae N10 with mutant and PseFDH mutant from Pseudomonas sp. 101, respectively (Hoelsch et al. 2012; Serov et al. 2002), and about 25.00, 18.00, and 11.00 times than that of the CmeFDH D195S from C. methylica with mutant and BstFDH from Burkholderia stabilis 15516 (GulKaraguler et al. 2001; Hatrongjit and Packdibamrung 2010). These results indicated that the three mutants have high catalytic velocity for NAD+ than other reported FDHs. The kcat/Km

Appl Microbiol Biotechnol Table 2

Kinetic parameters of COONH4 and cofactor-accepting FDHs

Enzyme

Km.COONH4 (mM)

kcat.COONH4 (min−1)

kcat/Km.COONH4 (mM−1 min−1)

Km,NAD+ (mM)

kcat.NAD+ (min−1)

kcat/Km.NAD+ (mM−1 min−1)

Reaction conditions

CboFDH V120S N187D V120S/N187D

3.63 ± 0.10 7.90 ± 0.32 6.70 ± 0.21 2.50 ± 0.15

92.05 ± 2.1 320.13 ± 12.7 197.13 ± 5.21 41.80 ± 0.62

25.38 40.55 29.46 16.51

3.40 ± 0.15 8.81 ± 0.18 3.80 ± 0.22 2.02 ± 0.11

1338.08 ± 32.8 2437.97 ± 12.1 1739.15 ± 3.36 1121.05 ± 23.49

397.10 300.49 588.48 596.05

RTa, pH 9.5 RTa, pH 9.5 RTa, pH 10.0 RTa, pH 9.5

CboFDH D195Q/ Y196Hb CboFDH 195Q/Y196R/ Q197Nd CmeFDH D195Se MycFDH C145S/ D221Q/C255Vf MycFDH C145S/A198G/ D221Q/C255Vf PseFDH, mutantg

80 ± 10

n.d.c

n.d.

1.8 ± 0.09

29.4 ± 1.8

16.2

30 °C, pH 7.5

n.d.

n.d.

n.d.

0.36 ± 0.03 37.2 ± 1.8

102

RTa, pH 7.5

n.d. 113 ± 11

n.d. n.d.

n.d. n.d.

4.70 ± 0.30 96 ± 6.0 1.09 ± 0.04 493.2 ± 60

20.4 452.4

20 °C, pH 8.0 30 °C, pH 7.0

98 ± 13

n.d.

n.d.

4.10 ± 0.17 310.8 ± 5.4

75.6

30 °C, pH 7.0

9±3 1000 ± 200

n.d. n.d.

n.d. n.d.

1.00 ± 0.15 300 ± 24 8.40 ± 0.90 7.2 ± 1.2

300 0.6

30 °C, pH 7.0 30 °C, pH 7.0

55.5 4.7 19.6 5.1 80.00

n.d. n.d. 400.2 18 n.d.

n.d. n.d. 20.4 3.6 n.d.

1.43 0.07 0.09 5.5 6.50

69.6 n.d. 74 13.1 53.26

30 °C, pH 7.0 30 °C, pH 7.0 30 °C, pH 7.0 25 °C, pH 8.0 pH 7.0

SceFDH D196A/ Y197Rg BstFDH, wild typeh MeFDHi BaFDHi CmeFDHi GraFDHj

99.6 n.d. 6.67 72 346.2

FDHs from Candida methylica (CmeFDH), Burkholderia stabilis 15516 (BstFDH), Saccharomyces cerevisiae (SceFDH), Mycobacterium vaccae N10 (MycFDH), Pseudomonas sp. 101 (PseFDH), Candida boidinii (CboFDH), Methylobacterium sp. (MeFDH), Bacillus sp. (BaFDH), and Granulicella mallensis MP5ACTX8 (GraFDH) a

Room temperature

b

Andreadeli A, Platis D, Tishkov VI, Popov VO, Labrou NE (2008). FEBS J 275:3859–3869

c

Not determined

d

Wu W, Zhu D, Hua L (2009). J Mol Catal B: Enzym 61:157–161

e

Gul-Karaguler N, Sessions RB, Clarke AR, Holbrook J (2001). Biotechnol Lett 23:283–287

f

Hoelsch, K., Sührer, I., Heusel, M., and Weuster-Botz, D. (2012). Appl Microbiol Biotechnol 97, 2473–2481

g

Serov AE, Popova AS, Fedorchuk VV, Tishkov VI (2002). Biochem J 367:841–847

h

Hatrongjit R, Packdibamrung K (2010). Enzyme Microb Tech 46:557–561

i

These data were obtained from http://www.brenda-enzymes.org/enzyme.php?ecno=1.2.1.2

j

Stefano Fogal et al. (2015). Appl Microbiol Biotechnol 99:9541–9554

values for NAD+ of the N187D and V120S-N187D were 1.30- and 1.31-fold than that of the MycFDH C145S/ D221Q/C255V, which had the highest kcat/Km than other FDHs (Hoelsch et al. 2012), suggesting that the N187D and V120S-N187D have better ability to provide coenzyme regeneration. The above results indicated that the mutation, which was implemented by structural analysis and rational design, has achieved good results. Even though the N187D had the highest conversion rate, the V120S had the fastest conversion as the conversion rates were 50.04, 86.19, 83.40, and 83.46 % for CboFDH, V120S, N187D, and V120S-N187D at 0.5 h, respectively (Fig. 7b). The results were consistent with the results of kinetic constants.

The catalytic efficiency of N187D for substrates and coenzyme was all improved, so it had the highest conversion rate. Meanwhile, mutant V120S had the fastest conversion for its contribution to enhance the catalytic turnover number (kcat) for substrates and coenzyme. The enantiomeric excess’ value (e.e.) was over 99 % in all catalytic reactions, indicating the FDHs have no effect on the enantiomer choice of other enzymes, indicating that they can be utilized to provide expensive coenzyme for asymmetric synthesis. In this study, we have identified the catalysis, substrate, and coenzyme binding sites; improved catalytic efficiency of CboFDH; and demonstrated that structure analysis and rational design could be used to explore and enhance FDH

Appl Microbiol Biotechnol

substrate affinity and catalytic activity. Moreover, we also confirmed that the site 120 may play a pivotal role in regulating catalysis and substrate affinity while the site 187 may play a crucial role in cofactor affinity. Asymmetric synthesis was implemented to detect the ability of cofactor regeneration of the mutants, which had nearly 100 % of the conversion rate. The structure-function relationship analysis and experimental results obtained in this work would be helpful and provide reference to enhance FDH’s property and promote the industrial production of chiral compounds, such as chiral amine and amino acid. Acknowledgments This work was supported by the State Key Program of National Natural Science Foundation of China (No. 21336009), the National Natural Science Foundation of China (Nos. 41176111 and 41306124), the Fundamental Research Funds for the Central Universities (No. 2013121029), the Foundation of South Oceanographic Research Center of China in Xiamen (No. 14GYY011NF11), and the Public Science and Technology Research Funds Projects of Ocean (No. 201505032–6). Authors’ contributions WJ directed the research, performed the experiment, and wrote the manuscript. PL and RNY assisted with the steady-state kinetic study. BSF conceived the project, directed the research, and supervised the work. Compliance with ethical standards Ethical approval This article does not contain any studies with human participants or animals performed by any of the authors. Conflict of interest The authors state that they have no competing interests.

References Addo PK, Arechederra RL, Minteer S (2010) Evaluating enzyme cascades for methanol/air biofuel cells based on NAD+-dependent enzymes. Electroanalysis 22(7–8):807–812 Akers NL, Moore CM, Minteer SD (2005) Development of alcohol/O2 biofuel cells using salt-extracted tetrabutylammonium bromide/ Nafion membranes to immobilize dehydrogenase enzymes. Electrochim Acta 50(12):2521–2525 Ansorge-Schumacher MB, Slusarczyk H, Schümers J, Hirtz D (2006) Directed evolution of formate dehydrogenase from Candida boidinii for improved stability during entrapment in polyacrylamide. FEBS J 273(17):3938–3945 Baik SH, Ide T, Yoshida H, Kagami O, Harayama S (2003) Significantly enhanced stability of glucose dehydrogenase by directed evolution. Appl Microbiol Biotechnol 61(4):329–335 Bhardwaj A, Leelavathi S, Mazumdar-Leighton S, Ghosh A, Ramakumar S, Reddy VS (2010) The critical role of N- and C-terminal contact in protein stability and folding of a family 10 xylanase under extreme conditions. PLoS One 5(6):3039–3056 Bolivar JM, Lorena W, Susana Alicia F, Roberto FL, Guisan JM, Cesar M (2006) Stabilization of a formate dehydrogenase by covalent immobilization on highly activated glyoxyl-agarose supports. Biomacromolecules 7(3):669–673

Carter JL, Bekhouche M, Noiriel A, Blum LJ, Doumèche B (2014) Directed evolution of a formate dehydrogenase for increased tolerance to ionic liquids reveals a new site for increasing the stability. ChemBioChem Demir AS, Talpur FN, Betul Sopaci S, Kohring GW, Ayhan C (2011) Selective oxidation and reduction reactions with cofactor regeneration mediated by galactitol-, lactate-, and formate dehydrogenases immobilized on magnetic nanoparticles. J Biotechnol 152(4):176–183 Donk WA, Der V, Huimin Z (2003) Recent developments in pyridine nucleotide regeneration. Curr Opin Biotechnol 14(4):421–426 Filippova EV, Polyakov KM, Tikhonova TV, Stekhanova TN, Boiko KM, Popov VO (2005) Structure of a new crystal modification of the bacterial NAD-dependent formate dehydrogenase with a resolution of 2.1 Å. Crystallogr Rep 50(5):796–800 Fischer T, Pietruszka J (2010) Key building blocks via enzyme-mediated synthesis. Springer, Berlin Fogal S, Beneventi E, Cendron L, Bergantino E (2015) Structural basis for double cofactor specificity in a new formate dehydrogenase from the acidobacterium Granulicella mallensis MP5ACTX8. Appl Microbiol Biotechnol 1–14 Funahashi J, Takano K, Yamagata Y, Yutani K (2000) Role of surface hydrophobic residues in the conformational stability of human lysozyme at three different positions. Biochemistry 39(47):14448–14456 Gul-Karaguler N, Sessions RB, Clarke AR, Holbrook JJ (2001) A single mutation in the NAD-specific formate dehydrogenase from Candida methylica allows the enzyme to use NADP. Biotechnol Lett 23(4): 283–287 Hatrongjit R, Packdibamrung K (2010) A novel NADP+-dependent formate dehydrogenase from Burkholderia stabilis 15516: screening, purification and characterization. Enzym Microb Technol 46(7): 557–561 Hoelsch K, Sührer I, Heusel M, Weuster-Botz D (2012) Engineering of formate dehydrogenase: synergistic effect of mutations affecting cofactor specificity and chemical stability. Appl Microbiol Biotechnol 97(6):2473–2481 Hong W, Huang S, Jiang Z (2004) Effects of modification of silica gel and ADH on enzyme activity for enzymatic conversion of CO2 to methanol. Catal Today 98(4):545–552 Jeon JK, Ihm SK, Park YK, Kim JS, Dong JI, Kim S, Kim JM, Kim SS, Yoo KS (2004) Membrane/PSA hybrid process for carbon dioxide recovery at low concentration. Stud Surf Sci Catal 153(04):543–546 Jiang W, Chen L, Yuan S, Li B, Liu Z (2014) A novel serine hydroxymethyltransferase from Arthrobacter nicotianae: characterization and improving catalytic efficiency by rational design. BMC Biotechnol 14(1):93 Jiang W, Sun D, Lu J, Wang Y, Wang S, Zhang Y, Fang B (2015) A coldadapted leucine dehydrogenase from marine bacterium Alcanivorax dieselolei: characterization and L-tert -leucine production. Eng Life Sci 12(11) Koutinas AA, Vlysidis A, Pleissner D, Kopsahelis N, Garcia IL, Kookos IK, Papanikolaou S, Kwan TH, Lin CSK (2014) Valorization of industrial waste and by-product streams via fermentation for the production of chemicals and biopolymers. Chem Soc Rev 43(8):2587–2627 Kula M-R, Pohl M, Slusarczyk H (2003) Mutanten der Formiatdehydrogenase aus Candida biodinii, Gensequenzen diese codierend sowie Verwendung dieser Formiatdeydrogenasen. EP Lu Y, Jiang ZY, Xu SW, Wu H (2006) Efficient conversion of CO2 to formic acid by formate dehydrogenase immobilized in a novel alginate–silica hybrid gel. Catal Today 115(s 1–4):263–268 Mourad B, Blum LCJ, Bastien D (2011) Contribution of dynamic and static quenchers for the study of protein conformation in ionic liquids by steady-state fluorescence spectroscopy. J Phys Chem B 116(1):413–423

Appl Microbiol Biotechnol Munro PA, Dunnill P, Lilly MD (1975) Magnetic particles as supports in immobilised enzyme reactors. Magn IEEE Trans 11(5):1573–1575 Netto CGCM, Nakamura M, Andrade LH, Toma HE (2012) Improving the catalytic activity of formate dehydrogenase from Candida boidinii by using magnetic nanoparticles. J Mol Catal B Enzym 84(4):136–143 Nilov DK, Shabalin IG, Popov VO, Svedas VK (2012) Molecular modeling of formate dehydrogenase: the formation of the Michaelis complex. J Biomol Struct Dyn 30(2):170–179 Perl D, Mueller U, Heinemann U, Schmid FX (2000) Two exposed amino acid residues confer thermostability on a cold shock protein. Nat Struct Mol Biol 7(5):380–383 Popov VO, Lamzin VS (1994) NAD+-dependent formate dehydrogenase. Biochem J 301((Pt 3)(3)):625–643 Schirwitz K, Schmidt A, Lamzin VS (2007) High-resolution structures of formate dehydrogenase from Candida boidinii. Protein Sci 16(6): 1146–1156 Serov A, Popova A, Fedorchuk V, Tishkov V (2002) Engineering of coenzyme specificity of formate dehydrogenase from Saccharomyces cerevisiae. Biochem J 367:841–847

Shaked Z, Whitesides GM (1980) Enzyme-catalyzed organic synthesis: NADH regeneration by using formate dehydrogenase. J Am Chem Soc 102(23):7104–7105 Slusarczyk H, Felber S, Kula MR, Pohl M (2000) Stabilization of NADdependent formate dehydrogenase from Candida boidinii by sitedirected mutagenesis of cysteine residues. Eur J Biochem 267(5): 1280–1289 Tishkov VI, Popov VO (2006) Protein engineering of formate dehydrogenase. Biomol Eng 23(2):89–110 Wichmann R, Vasic-Racki D (2005) Cofactor regeneration at the lab scale. Adv Biochem Eng/Biotechnol 92:225–260 Wohlgemuth R (2010) Asymmetric biocatalysis with microbial enzymes and cells. Curr Opin Microbiol 13(3):283–292 Xu HQY, Huang Z, Liu Z (2014) Characterization and site-directed mutagenesis of α-galactosidase from the deep-sea bacterium Bacillus megaterium. Enzym Microb Technol 56:46–52 Yoshimoto M, Yamasaki R, Nakao M, Yamashita T (2010) Stabilization of formate dehydrogenase from Candida boidinii through liposomeassisted complexation with cofactors. Enzym Microb Technol 46(7): 588–593