Improving thermostability of (R)-selective amine transaminase from

Aspergillus terreus through introduction of disulfide bonds

1 1 2 1 1 1
Dong-Fang Xie , Hui Fang , Jia-Qi Mei , Jin-Yan Gong , Hong-Peng Wang , Xiu-Ying Shen , Jun

Huang1*, Le-He Mei3*

1
School of Biological and Chemical Engineering, Zhejiang University of Science and Technology,

Hangzhou 310023, PR China

2
Department of Chemical Engineering, University of Utah, Salt Lake City, Utah 84102, United States

3
Department of Biological and Pharmaceutical Engineering, Ningbo Institute of Technology, Zhejiang

University, Ningbo 315100, PR China

Running Title

Engineering disulfide bonds to AT-ATA

This article has been accepted for publication and undergone full peer review but has not been through
the copyediting, typesetting, pagination and proofreading process, which may lead to differences
between this version and the Version of Record. Please cite this article as doi: 10.1002/bab.1572.

This article is protected by copyright. All rights reserved.

*Corresponding authors:

Jun Huang

Tel.: 86-571-85070370; Fax: 86-571-85070370

E-mail address: [email protected]

Le-He Mei

Tel: +86-571-87953161; Fax: +86-571-7951982

E-mail address: [email protected]

This article is protected by copyright. All rights reserved.

2
 Abstract

To improve the thermostability of (R)-selective amine transaminase from Aspergillus terreus

(AT-ATA), we used computer software Disulfide by Design (DbD) and Modelling of Disulfide Bonds in

Proteins (MODIP) to identify mutation sites where the disulfide bonds were most likely to form. We

obtained three stabilised mutants (N25C-A28C, R131C-D134C, M150C-M280C) from seven

candidates by site-directed mutagenesis. Compared to the wild-type, the best two mutants

N25C-A28C and M150C-M280C showed improved thermal stability with a 3.1-fold, and 3.6-fold

increase in half-life (t1/2) at 40 °C, and a 4.6 °C, and 5.1 °C increase in T5010. In addition, the

combination of mutant R131C-D134C and M150C-M280C displayed the largest shift in

thermostability with a 4.6-fold increase in t1/2 at 40 °C and a 5.5 °C increase in T5010. Molecular

dynamics (MD) simulation indicated that mutations of N25C-A28C and M150C-M280C lowered the

overall root mean square deviation (RMSD) for the overall residues at elevated temperature, and

consequently increased the protein rigidity. The stabilised mutation of R131C-D134C was in the

regions of high mobility and on the protein surface, and the disulfide bond constraints the flexibility

of loop 121-136.

Keywords: amine transaminase; disulfide bond; root mean square deviation (RMSD); molecular

dynamics (MD); thermostability

This article is protected by copyright. All rights reserved.

3
List of abbreviations: amine transaminase from Aspergillus terreus (AT-ATA); dimethylsulphoxide

(DMSO); molecular dynamics (MD); root mean square deviation (RMSD); root mean square

fluctuations (RMSF); methylbenzylamine (MBA); pyridoxal-5’-phosphate (PLP); Disulfide by

Design (DbD); Modelling of Disulfide Bonds in Proteins (MODIP).

1. Introduction

Chiral amine compounds are important intermediates for the pharmaceutical, chemical,

and biochemical industries [1, 2, 3, 4]. Transaminases are highly stereoselective and

regioselective when catalysing the transfer of an amino group from a primary amino donor

onto a carbonyl moiety, and utilise pyridoxal-5’-phosphate (PLP) as a cofactor [5, 6, 7, 8, 9].

The reaction mixture consists of two amines (an amino donor and a product) and two

carbonyl compounds (a ketone substrate and a by-product) (Scheme 1). Transaminases are

used for the asymmetric synthesis of enantiomerically pure amines and unnatural amino

acids in the pharmaceutical and fine chemical industries [10, 11, 12, 13]. Stabilization of

transaminase is highly desirable because it allows tolerating harsh reaction conditions and

has beneficial effects on long-term storage and recycling usage of the enzyme.

This article is protected by copyright. All rights reserved.

4
 Scheme 1 Asymmetric synthesis of (S)- and (R)-amines using transaminase

Several rational design methods have been developed and applied to increase the

thermostability of transaminase. Huang et al. [14] used rational design to improve the

thermostability of amine transaminase from Aspergillus terreus (AT-ATA) based on the

flexibility of the amino acid residues (B-factor values) and folding free energy calculations.

The rational engineering of disulfide bonds into proteins is another promising strategy. The

disulfide bond, or S-S key, is formed by the oxidation of two thiol groups in cysteine, forming

a cross-linked covalent bond between the sulphur atoms. Disulfide bonds–which introduce

conformational constraints to a polypeptide backbone–play an important role in stabilisation

of proteins, catalysis, and regulation ofprotein activity [15, 16, 17].

To minimise screening time and effort, some computational softwares (SSBOND,

MODIP and DbD2) have been used to select mutation sites where disulfide bonds are most

likely to form [17-20]. Firstly, Hazes and Dykstra devised the SSBOND algorithm, which

determined if the two Sγ from a residue pair were properly oriented to form a disulfide bond

[17]. Based on native disulfide geometry, MODIP model includes not only SSBOND algorithm

function but also consideration of steric interactions between the modelled cysteines and

This article is protected by copyright. All rights reserved.

5
surrounding residues [18]. The DbD2 algorithm can accurately and rapidly estimate the Cβ–

Sγ–Sγ–Cβ (χ3) torsion angle which represents rotation of the Cβ atoms, and analyse B-factors

of candidate disulfide regions to help identify those that may confer increased stability to

the protein [19, 20]. DbD and MODIP have been previously applied to successfully improve

the stability of lipase B in Candida Antarctica [21], acetylcholine esterase in Drosophila

melanogaster [22], and Type A feruloylesterase in A.usamii [23].

Molecular dynamics (MD) is a computer simulation method used to study the physical

movements of atoms and molecules. It offers a detailed description of thermal motion at

different temperatures and follows conformational changes as a function of small time

intervals [24]. By calculating the root mean square deviation (RMSD) and root mean square

fluctuation (RMSF) values for backbone molecules, thermally sensitive or conformationally

flexible regions of wild-type and mutant proteins can be calculated and compared.

In this study, the computational software DbDV2.12 and MODIP were applied to predict

where disulfide bonds were most likely to form in residue pairs within AT-ATA. Subsequently,

the predicted stabilising effects of these mutations were experimentally verified by thermal

inactivation, and the possibility of additive or synergistic effects from combined single

disulfide bond mutations was further investigated. In addition, using homology modelling

and MD simulations, conformational changes of the stabilised mutants were investigated by

calculating RMSF values for backbone atoms.

This article is protected by copyright. All rights reserved.

6
2. Materials and methods

2.1 Materials

The AT-ATA cDNA from A. terreus sequence, including the NcoI and XhoI restriction

sites, was synthesised at General Biosystems (AnHui) Co., Ltd (Chuzhou, China), and plasmid

pET28a(+) was used for gene cloning and DNA sequencing. All PCR primers were synthesised

at GenScript Corp. (NanJing, China). PrimeSTAR® Max DNA polymerase was obtained from

Takara Biotechnology (Dalian, China) for PCR. Yeast extract, tryptone, kanamycin sulphate

and isopropyl-β-D-thiogalactoside (IPTG) were obtained from Sangon Inc. (Shanghai, China).

DNA ladder, protein marker, and the Ni-IDA-Sefinose™ resin kit were purchased from Bio

Basic Inc. (Toronto, Canada). SDS-PAGE Gel Kit, DNA Clean-up Kit and Pure Plasmid Mini Kit

were purchased from CWBIO, Co. Ltd. (Beijing, China).

2.2 Location of the Disulfide bond Sites

We used the crystal structure of AT-ATA (PDB ID: 4CE5) from the Protein Data Bank

(http://www.rcsb.org) to design thermostable AT-ATA [25]. The 3D structure of AT-ATA was

visualised using PyMOL software (http://pymol.org). DbD V2.12

(http://cptweb.cpt.wayne.edu/DbD2/) and MODIP

(http://caps.ncbs.res.in/dsdbase/modip.html) were used to predict which amino acid pairs

were most likely to form disulfide bonds. Based on the physicochemical properties of the

This article is protected by copyright. All rights reserved.

7
amino acids and their positions within the 3D structure of AT-ATA, seven pairs of amino

acids were selected for substitution with cysteine residues to create a series of candidate

variants of AT-ATA. YASARA (Version 16.4.6) software (http://www.yasara.org) was used to

model substitutions of cysteine in the crystal structures of AT-ATA.

PCR-based AT-ATA site-directed mutagenesis was conducted as previously described by

Georgescuet al. [26]. All primers for site-directed mutagenesis are listed in the

supplementary material (Table S1).The reaction mixture consisted of DNA template

(pET28a(+)-AT-ATA, 100 ng), forward and reverse primers (10 μM, 1 μL each), 2x PrimeSTAR

Max Premix (25 μL), with water added to obtain a final volume of 50 μL. PCR was performed

as follows: 98 °C for 1 min; 30 cycles of 98 °C for 15 s, 55 °C for 15 s and 72 °C for 3 min; a

final extension at 72 °C for 10 min. The PCR products were digested with DpnI and incubated

for 3 h at 37 °C before purification; purified products were used to transform competent E.

coli DH5α bacteria by heat shock. The transformed bacteria were spread on a Luria–Bertani

(LB) plate containing 50 µg/mL kanamycin and incubated at 37 °C overnight. Plasmid DNA

was isolated from several colonies using the Pure Plasmid Mini Kit, and presence of the

desired mutations was verified by sequencing.

This article is protected by copyright. All rights reserved.

8
2.3 Protein Expression and Purification

E. coli BL21 (DE3) with wild-type or mutant pET28a(+)-AT-ATA plasmid was grown in LB

media (10 g/L tryptone, 5 g/L yeast extract and 10 g/L NaCl, pH 7.0) with kanamycin (50

µg/mL) at 37 °C until the OD600 reached 0.6–0.8; the media was then cooled to 25 °C. Protein

expression was induced by adding IPTG at a final concentration of 1 mM. After culturing for

18 h at 25 °C with shaking, the cells were harvested by centrifugation at 6000 × g for 8 min

at 4 °C and resuspended in buffer A (50 mM sodium phosphate buffer, 300 mM NaCl, 20

mM imidazole, 0.1 mM PLP, pH 8.0). The cells were then lysed in an iced bath by ultrasound

(output magnitude 400 w, duty time 3 s, interval time 6 s, total time 15 min). After

centrifugation at 12,000 × g at 4 °C for 30 min, the supernatant was used for protein

purification. Recombinant proteins were purified by Ni-affinity chromatography using the

Ni-IDA-Sefinose™ resin kit. Purified proteins were eluted in buffer B (50 mM sodium

phosphate, 300 mM NaCl, 250 mM imidazole, 0.1 mM PLP, pH 8.0), and concentrated using

an Amicon Ultra-15 centrifugal filter unit with a 10 kDa membrane molecular weight cut-off

(Millipore, Billerica, MA, USA). The purity of purified proteins was analysed by sodium

dodecyl sulphate polyacryl-amide gel electrophoresis (SDS-PAGE). The protein

concentrations were determined using the Bradford method with bovine serum albumin

(BSA) as the standard [27].

This article is protected by copyright. All rights reserved.

9
2.4 Thermostability of the AT-ATA and Its Variants

The wild-type and mutant proteins were incubated at 40 °C for time intervals ranging

from 0 to 50 min, and then cooled on ice for 10 min. The enzyme activity was assayed at 25

°C. The data were fitted to a four-parameter Boltzmann sigmoidal function by nonlinear

regression using Origin 8.0. The thermal inactivation half-life (t1/2) was defined as the time

when the residual activity of AT-ATA was reduced to 50% of its original activity under

standard conditions. The data were fitted to the equation y=exp(-kd·t) and the first-order

rate constants (kd) were determined by nonlinear regression using Origin 8.0.

Alternatively, the enzyme solution was incubated for 10 min at temperatures ranging

from 20 °C to 60 °C, and then cooled on ice for 10 min. The T5010 value was defined as the

temperature at which enzyme activity was reduced to 50% of its original activity. The data

were fitted to a four-parameter Boltzmann sigmoidal function. Temperature profiles of

wild-type and mutant AT-ATA were evaluated by measuring their activity, as described

above, at temperatures ranging from 20-60 °C.

2.5 Enzyme Activity Assay

The activities of the wild-type and mutant AT-ATA were measured as previously

described by Schätzle et al. [28]. An assay solution was prepared containing 0.25% dimethyl

sulphoxide (DMSO), 2.5 mM (R)-α-methylbenzylamine (MBA) and 2.5 mM pyruvatein

This article is protected by copyright. All rights reserved.

10
phosphate buffer (50 mM, pH 8.0).The enzyme reaction mixture, containing 2-4 µg purified

enzyme (20 μL) and 180 μL assay solution, was measured in UV 96-well microtiter plates at

25 °C at 245 nm (ε245nm=12 mM-1cm-1) for 3 min using an MD 190 photometer (Molecular

Devices, Sunnyvale, CA). One unit of activity was defined as the amount of enzyme that

produced 1 μmol acetophenone from (R)-α-MBA per minute. Km for (R)-α-MBA (Kmα-MBA) and

pyruvate (Kmpyruvate) were determined by measuring the activity at varied substrate

concentrations (0–2.5 mM). The enzyme activity assay was repeated three times for each

protein and reported as the mean and standard deviation. The data were fitted to the

Michaelis–Menten equation by nonlinear regression using Origin 8.0 (Northampton, MA,

USA).

2.6 Substrate specificity of pyruvate derivatives

To determine the substrate specificity of different pyruvate derivatives, 2-4 µg purified

enzyme (20 µL) was added to 180 µL reaction phosphate buffer (50 mM, pH 8) including

0.25% DMSO, 2.5 mM (R)-α-MBA and 2.5 mM pyruvate derivatives (methyl pyruvate, ethyl

pyruvate, 3-bromo pyruvate, 3-bromo ethyl pyruvate, phenylpyruvic acid, hydropyruvic acid,

and methyl trifluoropyruvate, respectively). The transamination reaction was measured in

UV 96-well microtiter plates at 25 °C at 245 nm (ε245nm=12 mM-1 cm-1) for 3 min using an MD

190 photometer (Molecular Devices, Sunnyvale, CA), and the consumed (R)-α-MBA was

analysed to measure the initial reaction rate.

This article is protected by copyright. All rights reserved.

11
2.7 MD Simulation for AT-ATA and the Stabilised Variants

MD simulation was performed to analyse the thermal fluctuation of AT-ATA wild-type,

N25C-A28C, R131C-D134C, and M150C-M280C at 313 K and 400 K for 10 ns using YASARA

with Amber14 force field. The structures were initially cleaned for added hydrogens or

incomplete side-chain atoms using YASARA. In the MD simulation, the structures were put

in a cubic box with a volume of 10×10×10 Å3, and the box was filled with explicit water

molecules to a density of 0.98 g/L. The system was neutralised by adding counter ions, and

all ionisable protein groups were protonated according to their tabulated pKa values at pH

8.0 of the medium. The cut-off value for van der Waals interactions was 7.86 Å, and the

long-range electrostatics of particle mesh Ewald (PME) were employed. The time step was

2.5 fs, and trajectories were saved every 25 ps. Trajectory analysis of data was performed

with YASARA, and the RMSD and RMSF values for backbone atoms were calculated.

3. Results and Discussion

3.1 Selection of Disulfide Bond Sites

To predict candidate residue pairs with extra disulfide bonds and analyse amino acid

residue flexibility, computational tools MODIP and DbD V2.12 were used. MODIP predicted

153 possible residue pairs ranked from grades A to D (data not shown), and DbD V2.12

predicted 40 potential residue pairs for disulfide bonds in wild-type AT-ATA (Fig. 1). To

This article is protected by copyright. All rights reserved.

12
obtain more reliable residue pairs for the possible formation of disulfide bonds, the

potential residue pairs predicted by DbD V2.12 were compared with the grade A predictions

of MODIP. Seven potential residue pairs for disulfide bonds were predicted by both

computational tools.

3.2 Thermostability of the AT-ATA and Its Variants

Of the seven mutants screened, the mutants D113C-Y146C and A161C-Y201C did not

express as soluble protein due to misfolding or the formation of incorrect disulfide

cross-links. The remaining five purified enzymes exhibited a single band near 36 kDa in

agreement with the calculated molecular weight of the wild-type enzyme based on amino

acid composition (Fig. 2). However, A32C-F142C and E213C-V234C showed very low catalytic

activity. As shown in Fig. 3, E213C-V234C pairs were located near the active site (K180)

within10 Å, which could influence the enzymatic activity. A32C-F142C is situated in a

β-sheet, which could affect the protein conformation and decrease the catalytic activity of

AT-ATA.

To evaluate inactivation of the enzymes, the T5010 values and half-life (t1/2) were

determined for the purified wild-type enzyme and mutant enzymes (N25C-A28C,

R131C-D134C and M150C-M280C). As shown in Table 1 and Fig.4, the T5010value and half-life

(t1/2) of the wild-type AT-ATA were determined to be 38.5 °C and 6.9 min, respectively. The

This article is protected by copyright. All rights reserved.

13
T5010 value of the three mutant enzymes N25C-A28C, R131C-D134C, and M150C-M280C

increased to 43.1 °C, 39.6 °C, and 43.6 °C, respectively. In addition, the half-lives (t1/2) of

N25C-A28C, R131C-D134C, and M150C-M280C at 40 °C increased to 21.6 min, 10.4 min, and

24.7 min, respectively, compared to 6.9 min for wild-type AT-ATA (Fig. 4B). We also

investigated the possibility of additive or synergistic effect by combining the stabilising

mutations. The best stabilised mutant, R131C-D134C/M150C-M280C, showed the largest

increase in thermal stability with a 4.6-fold increase in half-life and a 5.5 °C increase in T5010

compared to the wild-type enzyme. The optimum catalytic temperature was also estimated

for both wild-type AT-ATA and four stabilised mutants (Fig. 4C). All mutants had the same

optimum catalytic temperature as the wild-type enzyme. The same was observed for similar

disulfide-bridge mutants in other studies [30, 31]. However, the four stabilised mutants

showed better catalytic performance than the wild-type at higher temperatures. These

results suggest that a properly engineered disulfide bond can improve AT-ATA

thermostability.

3.3 Kinetic Parameters

Using (R)-α-MBA and pyruvate as substrates, the steady state kinetic parameters of the

wild-type and the four stabilised mutant AT-ATAs were determined (Table 1). Among the

four stabilised mutants, the largest change in catalytic efficiency (kcat/Kmpyruvate) was an

increase from 2.22 to 2.75 s-1·mM-1 for mutant N25C-A28C, and a decrease to 1.95 s-1·mM-1

This article is protected by copyright. All rights reserved.

14
for mutant R131C-D134C/M150C-M280C. The largest change for the kcat/Kmα-MBA was an

increase from 2.82 to 5.55 s-1·mM-1 for mutant R131C-D134C, and a decrease to 2.70

s-1·mM-1 for mutant R131C-D134C/M150C-M280C. All four stabilised mutants showed lower

substrate-binding affinity (Kmpyruvate) than the wild-type, and the largest increase in Kmpyruvate

was an increase from 0.23 to 0.43 mM for R131C-D134C/M150C-M280C. N25C-A28C and

R131C-D134C had higher substrate-binding affinity (Kmα-MBA) than the wild-type. The largest

change for the Km α-MBA was decrease from 0.43 to 0.10 mM for the mutant N25C-A28C.

3.4 Substrate specificity of pyruvate derivatives

When (R)-α-MBA was used as the amino donor, substrate specificity of different

pyruvate derivatives as the amino acceptor was investigated. As shown in Fig S1, wild-type

AT-ATA and the four stabilised mutants had high catalytic activity for pyruvate, methyl

pyruvate and ethyl pyruvate. The catalytic activity of the stabilised mutants for pyruvate and

3-bromo pyruvate were higher than that of the wild-type AT-ATA. The largest specific

activities for pyruvate and 3-bromo pyruvate were a 1.4-fold increase for

R131C-D134C/M150C-M280C and a 2.2-fold increase for N25C-A28C, respectively. The

catalytic activity of the stabilised mutants for 3-bromo ethyl pyruvate was lower than that of

wild-type AT-ATA except for R131C-D134C/M150C-M280C. The wild-type and mutant

AT-ATAs showed no activity for bulky pyruvate derivatives such as phenyl pyruvic acid,

methyl trifluoropyruvate, and hydroxyl phenyl pyruvic acid. The binding of transaminase to

This article is protected by copyright. All rights reserved.

15
the substrate in the active site is based on a large (L) pocket and a small (S) pocket; lysine

binds the PLP cofactor prior to the entry of the substrate [2]. These results indicate that the

introduction of disulfide bonds have little effect on the substrate specificity of pyruvate

derivatives for the wild-type and mutant AT-ATAs, and the substrate preference is pyruvate

derivatives with the small substituent groups.

3.5 Conformational Changes at Molecular Level of AT-ATA Variants

The main effect of the disulfide cross-link on the protein stability results from a

decrease in the conformational entropy of the unfolded molecule. Pace et al. [34] reported

that the entropic effect of a covalent crosslink to protein stability may be predicted using

the following equation: ΔS=-2.1-(3/2) R ln (n), where n is the number of residues in the loop

formed by the crosslink, and R is the universal gas constant. The mutation of stabilised

M150C-M280C had longer loop lengths (130 residues), but the loop length of N25C-A28C

and R131C-D134C was only 3 residues. The stabilised mutation of M150C-M280C were

associated with longer loop lengths. The disulfide bridges most effective in increasing

stability are those which are located in flexible regions and surface of the protein and create

a large loop (>25 residues) [29]. The stabilising mutations of N25C-A28C and R131C-D134C

are located on the protein surface of AT-ATA. The residues 129-134, which had higher

B-factor values, were located in the loop (T121-V136) and relatively flexible [14, 25].

Elevated temperatures make protein flexible regions or segments unfold early. The disulfide

This article is protected by copyright. All rights reserved.

16
bond between R131C and D134C restricts motion of the unfolded loop T121-V136 to a far

greater extent than restrictions imposed upon the folded protein.

In general, lowering the overall RMSD is one way to modify the enzyme structure to

make it more rigid and consequently enhance its thermostability [32, 33]. The RMSF

reflected the individual residue stability of a protein, and a high RMSF corresponded to high

flexibility of a given residue [15]. To investigate the flexibility changes at high temperatures,

the RMSD and corresponding RMSF per residue of the wild-type and three stabilised mutant

AT-ATAs (N25C-A28C, R131C-D134C, M150C-M280C) were predicted by MD simulation at

313 K and 400 K for 10 ns (Fig.5 and Fig. S2). At 313K, the average RMSD for the overall

residues of wild-type and the stabilised mutants had no obvious difference. At 400 K, the

average RMSD for the overall residues of wild-type was higher than that of the best two

mutants N25C-A28C and M150C-M280C, and lower than that of the stabilized mutant

R131C-D134C. RMSF analysis of wild-type also reveals the highly flexible regions consisting

of residues 129-134 (Fig. 5B). Taking the RMSF value as the index (Fig. 5B and Fig. S2), the

difference of RMSF between wild-type and the stabilised mutants was not significant at

313K. However, the analysis of fluctuation illustrated that at the residue positions (147-158,

270-281) within 5Å introducing the disulfide bond, the RMSF of M150C-M280C was

reduced. Moreover, the loop region of residues 129-134 with higher B-factor values is more

flexible in wild-type than in R131C-D134C.

This article is protected by copyright. All rights reserved.

17
4. Conclusions

Engineering an extra disulfide bond to enhance the thermostability of a protein is a

simple but effective rational approach. The T5010 and t1/2 values of the stabilised mutants

(N25C-A28C, R131C-D134C, M150C-M280C) were increased by 1.1-5.1 °C and 3.5-17.8 min

respectively, compared to those of the wild-type enzyme. Remarkably, the combination of

the mutants R131C-D134C and M150C-M280C displayed the largest shift in thermostability

with a 4.6-fold increase in the t1/2 at 40 °C and a 5.5 °C increase in T5010. The stabilised

mutants introducing disulfide bonds had good thermostability while maintaining enzyme

activity and substrate specificity. The stabilised mutant M150C-M280C and N25C-A28C had

the lower RMSD for the overall residues at elevated temperature, and the RMSF of

M150C-M280C was decreased within 5 Å regions near the introduced disulfide bond.

Moreover, the stabilised mutation of M150C-M280C were associated with longer loop

lengths (130 residues). The stabilised mutation of R131C-D134C was in the regions of high

mobility and on the protein surface, and the disulfide bond constraints the flexibility of loop

121-136. Engineering the extra disulfide bonds will be useful to improve the thermostability

for other proteins.

This article is protected by copyright. All rights reserved.

18
Acknowledgments

This work was supported by grants from the National Natural Science Foundation of China

(NOs. 31470793, 31670804 and 31240054).

References

1. Bornscheuer, U. T., Huisman, G. W., Kazlauskas, R. J., Lutz, S., Moore, J. C., and Robins, K.
(2012) Nature 485, 185-194.

2. Fuchs, M., Farnberger, J. E., and Kroutil, W. (2015) Eur. J. Org. Chem. 2015, 6965-6982.

3. Mathew, S., and Yun, H. (2012) Org. Biomol. Chem. 10, 2482-2485.

4. Park, E. S., Dong, J. Y., and Shin, J. S. (2014) Appl. Microbiol. Biotechnol. 98, 651-660.

5.Koszelewski, D., Tauber, K., Faber, K., and Kroutil, W. (2010) Trends Biotechnol. 28,
324-332.

6. Malik, M. S., Park, E. S., and Shin, J. S. (2012) Appl. Microbiol. Biotechnol. 94, 1163-1171.

7. Thomsen, M., Skalden, L., Palm, G. J., Höhne, M., Bornscheuer, U. T., and Hinrichs, W.
(2013) Acta. Crystallogr. 69, 1415-1417.

8. Ward, J., and Wohlgemuth, R. (2010) Curr. Org. Chem. 14, 1914-1927.

9. Willies, S. C., Galman, J. L., Slabu, I., and Turner, N. J. (2016) Phosphorus Sulfur. Silicon
Relat. Elem. 374, 20150084.

This article is protected by copyright. All rights reserved.

19
10. Fesko, K., Steiner, K., Breinbauer, R., Schwab, H., Schürmann, M., and Strohmeier, G. A.
(2013) J. Catal. 96, 103-110.

11. Höhne, M., Schätzle, S., Jochens, H., Robins, K., and Bornscheuer, U. T. (2010) Nat.
Chem. Biol. 6, 807-814.

12. Savile, C. K., Janey, J. M., Mundorff, E. C., Moore, J. C., Tam, S., and Jarvis, W. R. (2010)
Science 329, 305-309.

13. Steffenmunsberg, F., Vickers, C., Kohls, H., Land, H., Mallin, H., Nobili, A., Skalden, L.,
Bergh, T., Joosten, H., Berglund, P., Höhne, M., and Bornscheuer, U. T. (2015) Biotechnol.
Adv. 33, 566-604.

14. Huang, J., Xie, D. F., and Feng, Y. (2017) Biochem. Biophys. Res. Comm. 483, 397-406.

15. Bulaj, G. (2005) Biotechnol. Adv. 23, 87-92.

16. Wedemeyer, W. J., Welker, E., Narayan, M., and Scheraga, H. A. (2000) Biochemistry 39,
4207-4216.

17. Hazes, B., and Dijkstra, B. W. (1988) Protein Eng. 2, 119-125.

18. Dani, V.S., Ramakrishnan, C., and Varadarajan, R. (2003) Protein Eng. 16 (3), 187-193.

19. Dombkowski, A.A. (2003) Bioinformatics 19(14), 1852-1853.

20. Craig, D.B., and Dombkowski, A.A. (2013) BMC Bioinformatics 14 (1), 346.

21. Le, Q. A., Joo, J. C., Yoo, Y. J., and Kim, Y. H. (2012) Biotechnol. Bioeng. 109, 867-876.

22. Siadat, O., Lougarre, A., Lamouroux, L., Ladurantie, C., and Fournier, D. (2006) BMC
Biochem. 7, 12-19.

This article is protected by copyright. All rights reserved.

20
23. Yin, X., Hu, D., Li, J. F., He, Y., Zhu, T. D., and Wu, M. C. (2015) Plos One 10, 0126864.

24. Purmonen, M., Valjakka, J., Takkinen, K., Laitinen, T., and Rouvinen, J. (2007) Protein
Eng. Des. Sel. 20, 551-559.

25. Łyskowski, A., Gruber, C., Steinkellner, G., Schürmann, M., Schwab, H., Gruber, K., and
Steiner, K. (2014) Plos One e87350.

26. Georgescu, R., Bandara, G., and Sun, L. (2003) Methods Mol. Biol. 231, 75-83.

27. Bradford, M. M. (1976) Anal. Biochem. 72, 248–254.

28. Schätzle, S., Höhne, M., Redestad, E., Robins, K., and Bornscheuer, U. T. (2009) Anal.
Chem. 81, 8244-8248.

29. Dombkowski, A. A., Sultana, K. Z.,and Craig, D. B.(2014) FEBS Lett. 588, 206-212.

30. Turunen, O., Etuaho, K., Fenel, F., Vehmaanperä, J., Wu, X., Rouvinen, J., and Leisola, M.
A. (2001) J. Biotechnol. 88, 37-46.

31. Wakarchuk, W. W., Sung, W. L., Campbell, R. L., Cunningham, A., Watson, D. C., and
Yaguchi, M. (1994) Protein Eng. 7, 1379-1386.

32. Chen J, Yu H, Liu C, Liu J, and Shen Z. (2013) J Biotechnol. 164, 354–362.

33. Tu, T., Luo, H., Meng, K., Cheng, Y., Ma, R., Shi, P., Huang, H., Bai, Y., Wang, Y., Zhang, L.,
and Yao, B. (2015) Appl. Environ. Microbiol. 81, 6938–6944.

34. Pace, C. N., Grimsley, G. R., Thomson, J. A., and Barnett, B.J. (1988) J. Biol. Chem.
263(24), 11820–11825.

This article is protected by copyright. All rights reserved.

21
 Table 1. Stability and steady-state kinetic constants of wild-type and stabilised mutant AT-ATAs

T5010 t1/2 kcatpyruvate Kmpyruvate kcat/Kmpyruvate kcatα-MBA Kmα-MBA kcat/Kmα-MBA

AT-ATA
（°C） (min) (s-1) (mM) (s-1·mM-1) (s-1) (mM) (s-1·mM-1)

WT
38.5 ± 0.5 6.9 ± 0.6 0.50 ± 0.01 0.23 ± 0.02 2.22 0.64 ± 0.01 0.23 ± 0.03 2.82

N25C-A28C 43.1 ± 0.7 21.6 ± 1.4 0.66 ± 0.04 0.24 ± 0.07 2.75 0.58 ± 0.03 0.10 ± 0.03 5.55

R131C-D134C 39.6 ± 0.3 10.4 ± 0.6 0.67 ± 0.04 0.32 ± 0.07 2.07 0.59 ± 0.04 0.16 ± 0.05 3.72

M150C-M280C 43.6 ± 0.2 24.7 ± 0.9 0.65 ± 0.03 0.32 ± 0.07 2.00 0.65 ± 0.03 0.24 ± 0.06 2.74

R131C-D134C/M150C-M280C 44.0 ± 0.4 32.0 ± 0.7 0.84± 0.02 0.43 ± 0.06 1.95 0.66 ± 0.02 0.24 ± 0.04 2.70

This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading
process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1002/bab.1572.

This article is protected by copyright. All rights reserved.

FIGURE LEGENDS

Fig. 1 Alignment of amino acid sites for disulfide bridges predicted by DbD V2.12 and

MODIP. The amino acid sites predicted by DbD V2.12 are shown in the left oval. The amino

acid sites predicted by MODIP are shown in the right oval. The middle oval shows seven

amino acid sites that were predicted by both DbD V2.12 and grade A of MODIP to have the

highest chance of stable disulfide bridge formation.

This article has been accepted for publication and undergone full peer review but has not been through
the copyediting, typesetting, pagination and proofreading process, which may lead to differences
between this version and the Version of Record. Please cite this article as doi: 10.1002/bab.1572.

This article is protected by copyright. All rights reserved.

Fig. 2 Sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis of

the expression and purification of AT-ATA wild-type and mutants in E. coli BL 21. The protein

was visualised with Coomassie blue staining. Lanes: 1 Marker; 2 Wild-type; 3 N25C-A28C; 4

A32C-F142C; 5 D113C-Y146C; 6 R131C-D134C; 7 M150C-M280C; 8 A161C-Y201C; 9

E213C-V234C; 10 Marker; 11 N25C-A28C/M150C-M280C; 12 R131C-D134C/M150C-M280C.

24
Fig. 3 Modelled 3D structures of AT-ATA with disulfide bonds introduced.

25
Fig. 4 Stability of AT-ATA wild-type and stabilised mutant AT-ATAs. (A) Thermal inactivation

of wild-type and stabilised mutants at different temperatures over 10 min (T5010); (B)

Thermal inactivation half-life (t1/2) of wild-type and stabilised mutants at 40 °C; (C)

Temperature-dependent activity profiles of wild-type and stabilised mutants determined at

pH 8.0 for 3 min. Each data point represents the mean ± SD from three independent

samples. Square, wild-type; circle, mutant N25C-A28C; upward triangle, R131C-D134C;

downward triangle, M150C-M280C; left triangle, R131C-D134C/M150C-M280C. Error bars

are the standard deviation (SD) of three independent determinations; the solids lines are

fitted to a four-parameter Boltzmann sigmoidal function

26
27
Fig. 5 MD simulation analysis of AT-ATA wild-type and mutant M150C-M280C for 10 ns at

313 K and 400 K using YASARA. (A) RMSD values at 313 K and 400 K during a 10-ns

simulation for AT-ATA wild-type and mutant N25C-A28C; (B) RMSF values at 313 K and 400 K

during a 10-ns simulation for AT-ATA wild-type and mutant R131C-D134C.

28
29