Driving Transamination Irreversible by Decomposing Byproduct Α-Ketoglutarate into Ethylene Using Ethylene-Forming Enzyme

Article abstract of DOI:10.1007/s10562-018-2552-8

The transformations of transaminases have been extensively studied as an approach to the production of chiral amino moieties. However, the low equilibrium conversion of the reaction is a critical disadvantage to transaminase application, and a strategy for shifting the reaction equilibrium is essential. Herein, we have developed a novel method to effectively prevent the reversibility of transamination by fully decomposing byproduct α-ketoglutarate into ethylene and carbon dioxide in situ using ethylene-forming enzyme (EFE). Two transaminases and one EFE were expressed in E. coli and purified to be used in the cascade reaction. After optimal reaction conditions were determined based on the enzymatic properties, a cascade reaction coupling transaminase with EFE was conducted and showed high efficiency in the synthesis of l-phosphinothricin. Finally, using this approach with only an equivalent amount of amino donor l-glutamate increased the conversions of various keto acids from < 60% to > 99%. This strategy shows great potential for transamination using glutamate as the amino donor.

Full text of DOI:10.1007/s10562-018-2552-8

Catalysis Letters
https://doi.org/10.1007/s10562-018-2552-8
Driving Transamination Irreversible by Decomposing Byproduct
α-Ketoglutarate into Ethylene Using Ethylene-Forming Enzyme
Li‑Jun Meng1 · Ya‑Yun Liu1 · Hai‑Sheng Zhou1 · Xin‑Jian Yin1 · Jian‑Ping Wu1 · Mian‑Bin Wu1 · Gang Xu1 ·
Li‑Rong Yang1
Received: 30 June 2018 / Accepted: 11 September 2018
© Springer Science+Business Media, LLC, part of Springer Nature 2018
Abstract
The transformations of transaminases have been extensively studied as an approach to the production of chiral amino moie-
ties. However, the low equilibrium conversion of the reaction is a critical disadvantage to transaminase application, and a
strategy for shifting the reaction equilibrium is essential. Herein, we have developed a novel method to effectively prevent
the reversibility of transamination by fully decomposing byproduct α-ketoglutarate into ethylene and carbon dioxide in situ
using ethylene-forming enzyme (EFE). Two transaminases and one EFE were expressed in E. coli and purified to be used
in the cascade reaction. After optimal reaction conditions were determined based on the enzymatic properties, a cascade
reaction coupling transaminase with EFE was conducted and showed high efficiency in the synthesis of l-phosphinothricin.
Finally, using this approach with only an equivalent amount of amino donor l-glutamate increased the conversions of various
keto acids from <60% to >99%. This strategy shows great potential for transamination using glutamate as the amino donor.
Graphical Abstract
Keywords Transaminase · Ethylene-forming enzyme · Reaction equilibrium · Asymmetric synthesis · Amino acids
1 Introduction
Electronic supplementary material The online version of this
article (https://doi.org/10.1007/s10562-018-2552-8) contains
Transaminases (TAs, EC 2.6.1.X) catalyze the transfer of
supplementary material, which is available to authorized users.
an amino group from a donor substrate to a ketone acceptor
using pyridoxal 5′-phosphate (PLP) as cofactor (Scheme 1).
TAs have been widely used to produce chiral amino moie-
ties, including amino acids, chiral amines, amino alcohols,
* Li-Rong Yang
lryang@zju.edu.cn
Extended author information available on the last page of the article
Vol.:(0123456789)
1 3

L.-J. Meng et al.
Scheme 1 General reaction
scheme for transaminase
and amino sugars, which are valuable key intermediates or
starting materials in the synthesis of chiral drugs and agri-
cultural products [1–4].
produced in 94% yield owing to its poor solubility, which
resulted in in-situ product precipitation [12, 13]. Further-
more, A 92% yield of l-2-aminobutyrate was achieved by
recycling the α-ketoglutarate (α-KG) byproduct back to l-
Glu in situ [14–18]. These strategies all effectively shifted
the equilibrium to some extent.
However, the unfavorable thermodynamic equilibrium of
transamination is a significant obstacle to the application of
TAs [5–7]. For α-TAs, l-glutamate (l-Glu) can be used as a
universal amino donor, with transamination typically end-
ing at a thermodynamic equilibrium with an equilibrium
constant of K≈1 [8]. For example, the equilibrium constant
(K) for transamination between trimethylpyruvic acid and
l-Glu is 0.67, meaning that only 45% of trimethylpyruvic
acid can be converted into l-tert-leucine using an equivalent
amount of l-Glu [9].
While investigating α-KG catabolism, we found that
an ethylene-forming enzyme (EFE) can decompose α-KG
into ethylene and carbon dioxide with high specificity for
α-KG [19–21]. This indicated that α-KG generated in the
transamination reaction could be decomposed simultane-
ously by adding an EFE to TA in a single reaction vessel.
Accordingly, we have developed a novel approach to pre-
vent reverse transamination through in-situ decomposition
of byproduct α-KG using EFE. This cascade can reach a
theoretical 100% keto acid conversion. Several typical keto
acids were selected as amino acceptors to verify this novel
concept in transaminase-catalyzed asymmetric synthesis and
demonstrate its versatility (Scheme 2).
Several attempts have been made to shift this neutrally
positioned equilibrium towards product formation, includ-
ing using a stoichiometric excess of amino donor and in-situ
product/byproduct removal [10]. A four-fold molar excess of
l-Glu with a keto acid was used to produce l-phosphinothri-
cin in more than 90% yield [11]. l-Homophenylalanine was
Scheme 2 α-TA reactions for the asymmetric synthesis of chiral
α-amino acids driven to completion by coupling with an ethylene-
forming reaction. 1a 2-Oxo-4-[(hydroxy)(methyl)phosphinoyl]butyric
acid; 1b 2-oxobutyric acid; 1c 2-oxovaleric acid; 1d trimethylpyruvic
acid; 1e 4-methyl-2-oxopentanoic acid; 1f 3-methyl-2-oxopentanoic
acid; 1g 3-methyl-2-oxobutanoic acid; 2a l-phosphinothricin; 2b
ll--i2s-oalmeuincionbeu;t2ygralte-v; a2lcinle-norvaline; 2d l-tert-leucine; 2e l-leucine; 2f
1 3

Driving Transamination Irreversible by Decomposing Byproduct α-Ketoglutarate into Ethylene…
analyzed using 1% (w/v) agarose gel electrophoresis. Next,
2 Materials and Methods
gabT and bcta were digested using their respective restric-
tion endonucleases. The digested DNA fragments were then
inserted into the same restriction sites of the pET-28 a(+)
expression vector under the control of the T7 promoter. The
constructed recombinant plasmids were pET-28-GabT and
pET-28-BCTA.
2.1 Reagents, Strains, and Plasmids
PrimeSTAR Max DNA polymerase and restriction endo-
nucleases were purchased from Takala (Dalian, China).
Yeast extract, tryptone, antibiotics, and isopropyl β-d-1-
thiogalactopyranoside (IPTG) were purchased from San-
gon Biotech Co., Ltd. (Shanghai, China). Compound 1a
(Scheme 2) was synthesized and purified in our laboratory
according to previous reports [22, 23]. Other chemical rea-
gents were of analytical grade and obtained from standard
commercial sources.
2.3 Expression of Recombinant Enzymes
Recombinant E. coli was cultivated at 37 °C in Luria–Ber-
tani medium with kanamycin (50 µg/mL). The cells were
induced for 16 h using IPTG (0.2 mM) at 25 °C (for GabT)
or 18 °C (for BCTA and EFE), at which point the OD600 had
reached approximately 0.6. The cells were then harvested by
centrifugation (8000×g, 5 min, 4 °C), washed with phos-
phate buffer (100 mM, pH 7.4), and resuspended in half
the volume of buffer. The cell suspension was subjected to
ultrasonic cell disruption to obtain cell lysate.
Escherichia coli W3110 was maintained in our labora-
tory and used to amplify TA genes, including branched-
chain amino acid TA (BCTA, GenBank: CAQ34114.1)
and 4-aminobutyrate-α-KG TA (GabT, GenBank:
WP_001087611.1). E. coli BL21 (DE3) was used for gene
expression (Novagen, USA).
The expression of recombinant protein was examined
using sodium dodecyl sulfate polyacrylamide gel electropho-
resis (SDS-PAGE). Samples for SDS-PAGE were prepared
in two steps. First, cell lysate was centrifuged (12,000 rpm,
5 min, 4 °C) and the supernatant was collected and stored
below 4 °C, while the sediment was resuspended in phos-
phate buffer. Both the supernatant and resuspended sediment
were then added into loading buffer, followed by denatura-
tion at 99 °C for 10 min.
Vector pET-28a (+) (Novagen, USA) was used to con-
struct recombinant vector for gene expression. Recombinant
vector pET-28-EFE was obtained from Sangon Biotech
Co., Ltd. (Shanghai, China), where the gene of EFE from
Pseudomonas syringae pv. phaseolicola PK2 (Genbank:
D13182.1) was synthesized and inserted between EcoRI and
HindIII restriction sites of vector pET-28a(+). Recombinant
vectors pET-28-GabT and pET-28-BCTA were constructed
in our laboratory (Fig. 1).
2.2 Construction of Recombinant Vectors
2.4 Purification of Recombinant Enzymes
Genes gabT and bcta were amplified by polymerase chain
reaction (PCR) using genomic DNA of E. coli W3110 as
the template. The primers and restriction endonucleases
were 5′-CCGGAATTCATGAACAGCAATAAAGAGTTA
ATG-3′ (EcoRI) and 5′-CCCAAGCTTCTACTGCTTCGC
CTCATCAAAAC-3′ (HindIII) for gabT, and 5′-CCGGAA
TTCATGACCACGAAGAAAGCTGATTAC-3′ (EcoRI)
and 5′-CCGCTCGAGTTATTGATTAACTTGATCTAA
CCAGC-3′ (XholI) for bcta. The PCR products were
The harvested cells were resuspended in equilibration buffer
(20 mM phosphate buffer, 500 mM NaCl, 50 mM imidazole,
pH 7.5) and lysed using ultrasonication. The supernatant
(crude enzyme) was then loaded onto a Ni–NTA (nitrilo-
triacetic acid) chelating affinity column equilibrated with
equilibration buffer. The column was eluted with a stepwise
gradient of imidazole (50–500 mM) in 15–20 column vol-
umes of equilibration buffer. Fractions containing the eluted
product were identified by SDS-PAGE. The protein was then
Fig. 1 Vector map for recombinant vectors expressing GabT, BCTA and EFE
1 3

L.-J. Meng et al.
desalted and replaced with phosphate buffer (100 mM, pH
7.4) by ultrafiltration.
The concentration and enantiomeric excess (ee) of the
amino acids (l-Glu and products 2a–g in Scheme 2) were
determined by analytical HPLC using a C18 column (Pntu-
lips QS-C18, 5 µm, 4.6×250 mm) after derivatization with
o-phthalaldehyde and N-acetyl-l-cysteine [25]. The flow rate
was maintained at 0.85 mL/min, with detection at 338 nm,
and the column temperature was constant at 30 °C. l-Glu
and 2a were eluted with 50 mM sodium acetate buffer/ace-
tonitrile (95:5, v/v), 2b and 2 h were eluted with 50 mM
sodium acetate buffer/acetonitrile (85:15, v/v), and 2c–2e
and 2 g were eluted with 50 mM sodium acetate buffer/ace-
tonitrile (80:20, v/v).
2.5 Activity Assays and Biocatalytic Reaction
The assay solution for TA activity measurements contained
phosphate buffer (50 mM, pH 8.0), PLP (1 mM), l-Glu
(100 mM), and 1a (50 mM, adjusted to pH 7 with NaOH;
Scheme 2). The reaction was initiated by adding TA and
allowed to react for 15 min at 45 °C. The enzyme reaction
was then quenched by adding 1 M HCl and the amount of
α-KG was detected by HPLC. The TA activity was defined
as the amount of enzyme that catalyzed the formation of
1 µmol of α-KG from l-Glu per minute.
The assay solution for EFE activity measurements con-
tained phosphate buffer (50 mM, pH 7.4), FeSO4 (1 mM),
l-arginine (l-Arg, 5 mM), and α-KG (5 mM). The reaction
was initiated by adding EFE and allowed to react for 10 min
at 25 °C. The enzyme reaction was then quenched by adding
16% (w/v) trichloroacetic acid. The decrease in the amount
of α-KG was detected through derivatization with 2,4-dini-
trophenylhydrazine using a colorimetric method [24]. The
EFE activity was defined as the amount of enzyme that cata-
lyzed the consumption of 1 µmol of α-KG per minute.
Biocatalytic reactions were performed in 15 × 150 mm
test tubes at 25 °C and 220 rpm using a constant temperature
shaker. Other reaction conditions are shown in the respec-
tive tables.
3 Results and Discussion
3.1 Establishment of Cascade Reaction Conditions
Initially, the feasibility of the cascade reaction in one-pot
was assessed based on the compatibility of the involved
enzymes. GabT has been reported to exhibit a maximum
reaction velocity at 55 °C and pH 8.0–9.5 [11], while insuf-
ficient information exists on the optimal reaction conditions
for BCTA, although BCTA-catalyzed reactions have been
indicated to take place at 37 °C and pH 7.5–8.0 [26]. The
highest activity of EFE from Pseudomonas syringae pv.
phaseolicola PK2 has been reported at 20–25 °C and pH
7.0–7.5 [19, 27].
These three enzymes were cloned and expressed in E. coli
BL21 (DE3) and then purified to electrophoretic purity (see
Figs. S1 and S2). The specific activities of GabT and BCTA
were 65.5 1.1 and 41.3 2.0 U/mg, respectively, when
using 2a as a model product. However, the specific activity
of EFE was 0.14 0.04 U/mg, which was significantly lower
than that of the TAs.
2.6 Analysis
The keto acids (α-KG and compounds 1a–g in Scheme 2)
were detected by analytical HPLC using a reverse-phase
C18 column (Ultimate AQ-C18, 5 µm, 4.6 × 250 mm).
The flow rate was maintained at 1 mL/min with detection
at 205 nm, and the column temperature was maintained at
40 °C. The mobile phase was composed of buffer A (50 mM
(NH4)2HPO4 and 0.1% (w/v) tetrabutylammonium hydrox-
ide, adjusted to pH 3.6 with 50% (w/v) phosphoric acid)
and acetonitrile. α-KG, 1a, 1b, and 1 g were eluted with
buffer A/acetonitrile (90:10, v/v), while 1c–f were eluted
with buffer A/acetonitrile (80:20, v/v).
To select the optimal reaction conditions for the cascade
reaction, the enzymatic properties of the three recombinant
enzymes were characterized. The results indicated that both
GabT and BCTA were more active and stable than EFE at
high temperatures and pH levels, while EFE retained its high
activity and stability at 25–30 °C and pH 7.0–7.5 (Table 1).
Due to EFE having a much lower specific activity than TAs,
Table 1 Enzymatic properties
Enzymes
GabT
BCTA
EFE
of three enzymes involved in
this study
Optimal reaction temperature
Optimal reaction pH
Thermal stability
75 °C
75 °C
30 °C
pH 7.5
Stable below 25 °C
Stable at pH 7.0–7.5
pH 9.0–11.0
pH 6.0–12.0
72.7 °Ca
67.9 °Ca
pH stability
Stable at pH 6.0–12.0
Stable at pH<10.0
See Figs. S3 and S7
aThermal stability is indicated by the kinetic thermal temperature, which means that the residual activity
was 50% when incubated at this temperature for 15 min
1 3

Driving Transamination Irreversible by Decomposing Byproduct α-Ketoglutarate into Ethylene…
Table 3 Versatility of equilibrium shifting effect of TA and EFE cas-
the reaction conditions of the cascade were set to pH 7.5 and
25 °C, because EFE performed better under these conditions
while the TAs still showed high activities (Figs. S3 and S7).
cade
Entry Substrate Conversion of Conversion of
1 (%) by TAa 1 (%) by TA &
EFEb
ee of 2 (%)
by TA &
EFE
3.2 Effect of TA and EFE Cascade on the Conversion
1
2
3
4
5
6
7
1a
1b
1c
1d
1e
1f
55.8 0.8
58.3 0.8
55.2 0.3
44.6 1.2
42.3 1.7
50.0 1.7
45.4 1.4
>99
>99
>99
>99
>99
>99
>99
>99
>99
>99
>99
>99
>99
>99
of 1a
The effect of coupling EFE with TA on the reaction equi-
librium position was first evaluated by converting 1a. l-Arg
was added to the reaction system because it was essential
for α-KG decomposition, but did not serve as an amino
donor for TAs [27–30]. Due to its rather low specific activ-
ity, large amounts of EFE were added to reduce the reaction
time. However, the decomposition of α-KG catalyzed by
EFE remained the rate-limiting step in the cascade reaction.
The results showed that, when TA was coupled with EFE,
byproduct α-KG was fully decomposed and the conversion
of 1a (Scheme 2) was improved from 55.6% to over 99%
using only one equivalent of amino donor (Table 2).
1 g
aReaction mixtures contained l-Glu (10 mM), keto acids (10 mM,
1a–1 g), and PLP (0.2 mM) in sodium phosphate buffer (50 mM, pH
7.4). Purified BCTA (3.6 g/L) was added into the reaction. Conver-
sion and ee were analyzed by HPLC after reacting for 6 h
bReaction mixtures contained L-Glu (10 mM), keto acids (10 mM,
1a–1 g), PLP (0.2 mM), l-Arg (3.5 mM), and Fe2+ (0.5 mM) in
sodium phosphate buffer (50 mM, pH 7.4). Purified BCTA (3.6 g/L)
and purified EFE (10.6 g/L) were added into the reaction. Conversion
and ee were analyzed by HPLC after reacting for 6 h
3.3 Versatility of TA and EFE Cascade
different conversions when the reaction was catalyzed by
BCTA alone (Table 3). The conversion of TA and the EFE
cascade proved that decomposing byproduct α-KG using
EFE was an excellent method for making transamination
irreversible. Furthermore, the decomposition of α-KG
might also reduce byproduct inhibition of TAs (Fig. S5).
This TA and EFE cascade shows versatility and applica-
tions of TA-catalyzed asymmetric synthesis using only
equivalent amount of l-Glu as amino donor are currently
being investigated.
Owing to the success of combining GabT with EFE when
converting 1a into 2a, the substrate scope of the cascade
reaction was further investigated using BCTA [26]. The
conversion of keto acids 1a–1 g was conducted using an
equivalent amount of l-Glu. Among the products of this
study, four (2a–2d) were unnatural amino acids and the other
three (2e–2 g) were natural hydrophobic amino acids.
As shown in Table 3, the results of the BCTA and EFE
cascade showed that keto acid conversion was substantially
enhanced to over 99%, with perfect enantioselectivity
(> 99%), in each case. However, BCTA and EFE showed
4 Conclusions
Table 2 Equilibrium shifting effect of coupling EFE with TA
In summary, a novel strategy to make the transamination
reaction in the asymmetric synthesis of optically pure
amino acids irreversible was developed. Compared with
reported procedures, this approach using EFE has the
major advantage that the keto acid amino acceptor can be
nearly completely converted using only one equivalent of
amino donor, l-Glu.
Enzymes
TAa
TAb
TA & EFEb
Conversion of 1a (%)
Yield of 2a (%)
α-KG concentration (mM)
55.6 0.6
54.5 0.8
5.5 0.2
56.0 0.4 >99c
54.9 1.2 98.2 0.9
5.3 0.3 N.D.d
aReaction mixtures contained l-Glu (10 mM), 1a (10 mM), and
PLP (0.2 mM) in sodium phosphate buffer (50 mM, pH 7.4). Puri-
fied GabT (1.28 g/L) was added to the reaction. The reaction mixtures
were analyzed by HPLC after reacting for 6 h
Furthermore, this strategy might also be used in the
synthesis of d-amino acids catalyzed by d-amino acid TA
and using d-Glu from racemic Glu as the amino donor.
This strategy shows great potential for the production of
chiral amino acids using Glu as an amino donor.
bReaction mixtures contained l-Glu (10 mM), 1a (10 mM,
Scheme 2), PLP (0.2 mM), l-Arg (3.5 mM), and Fe2+ (0.5 mM) in
sodium phosphate buffer (50 mM, pH 7.4). Purified GabT (1.28 g/L)
with or without purified EFE (10.6 g/L) were added into the reaction.
The reaction mixtures were analyzed by HPLC after reacting for 6 h
Acknowledgements This work was financially supported by the
National Natural Science Foundation of China (No. 21476199).
cNo obvious 1a was detected
dNo obvious α-KG was detected
1 3

L.-J. Meng et al.
14. Klaus Bartsch RS, Schulz A (1996) Appl Environ Microbiol
62:3794–3799
Compliance with Ethical Standards
15. Li T, Kootstra AB, Fotheringham IG (2002) Org Process Res Dev
6:533–538
Conflict of interest The authors declare no commercial or financial
conflict of interest.
16. Hong E, Cha M, Yun H, Kim B-G (2010) J Mol Catal B Enzym
66:228–233
17. Xian M, Alaux S, Sagot E, Gefflaut T (2007) J Org Chem
72:7560–7566
References
18. Schrittwieser JH, Velikogne S, Hall M, Kroutil W (2018) Chem
Rev 118:270–348
1. Kelly SA, Pohle S, Wharry S, Mix S, Allen CCR, Moody TS,
Gilmore BF (2017) Chem Rev 118:349–367
19. Nagahama K, Ogawa T, Fujii T, Tazaki M, Tanase S, Morino Y,
Fukuda H (1991) J Gen Microbiol 137:2281–2286
20. Fukuda H, Ogawa T, Ishihara K, Fujii T, Nagahama K, Omata T,
Inoue Y, Tanase S, Morino Y (1992) Biochem Biophy Res Com-
mun 188:826–832
2. Koszelewski D, Tauber K, Faber K, Kroutil W (2010) Trends
Biotechnol 28:324–332
3. Taylor PP, Pantaleone DP, Senkpeil RF, Fotheringham IG (1998)
Trends Biotechnol 16:412–418
21. Eckert C, Xu W, Xiong W, Lynch S, Ungerer J, Tao L, Gill R,
Maness PC, Yu J (2014) Biotechnol Biofuels 7:33
22. Yin X, Wu J, Yang L (2018) Appl Microbiol Biotechnol
102:4425–4433
4. Mathew S, Yun H (2012) ACS Catal 2:993–1001
5. Guo F, Berglund P (2017) Green Chem 19:333–360
6. Slabu I, Galman JL, Lloyd RC, Turner NJ (2017) ACS Catal
7:8263–8284
23. Meiji Seika Kaisha LTD (2010) US Patent 0,063,313
24. Pesez M (1959) J Pharm Pharmacol 11:475–476
25. Zeiss HJ (1994) Pestic Sci 41:269–277
7. Meier RJ, Gundersen MT, Woodley JM, Schürmann M (2015)
ChemCatChem 7:2594–2597
8. Höhne M, Bornscheuer UT (2012) In: Drauz K, Gröger H, May
O (eds) Enzyme catalysis in organic synthesis. Wiley, Weinheim
9. Park E-S, Dong J-Y, Shin J-S (2013) ChemCatChem 5:3538–3542
10. Abu R, Woodley JM (2015) ChemCatChem 7:3094–3105
11. Schulz A, Tripier PT,D, Bartsch K (1990) Appl Environ Microbiol
56:1–6
26. Yu X, Wang X, Engel PC (2014) FEBS J 281:391–400
27. Nagahama K, Ogawa T, Fujii T, Tazaki M, Goto M, Fukuda H
(1991) J Gen Microbiol 137:1641–1646
28. Martinez S, Hausinger RP (2016) Biochemistry 55:5989–5999
29. Martinez S, Fellner M, Herr CQ, Ritchie A, Hu J, Hausinger RP
(2017) J Am Chem Soc 139:11980–11988
12. Cho B-K, Seo J-H, Kang T-W, Kim B-G (2003) Biotechnol Bio-
eng 83:226–234
30. Zhang Z, Smart TJ, Choi H, Hardy F, Lohans CT, Abboud MI,
Richardson MSW, Paton RS, McDonough MA, Schofield CJ
(2017) Proc Natl Acad Sci 114:4667–4672
13. Hanzawa S, Oe S, Tokuhisa K, Kawano K, Kobayashi T, Kudo T,
Kakidani H (2001) Biotechnol Lett 23:589–591
Affiliations
Li‑Jun Meng1 · Ya‑Yun Liu1 · Hai‑Sheng Zhou1 · Xin‑Jian Yin1 · Jian‑Ping Wu1 · Mian‑Bin Wu1 · Gang Xu1 ·
Li‑Rong Yang1
1
Institute of Biological Engineering, College of Chemical
and Biological Engineering, Zhejiang University,
Hangzhou 310027, China
1 3