Formate Dehydrogenase from Rhodococcus jostii (RjFDH) – A High-Performance Tool for NADH Regeneration

Article abstract of DOI:10.1002/adsc.202000536

The use of formate dehydrogenases (FDHs) for regeneration of the important cofactor NADH in enzyme-catalysed synthetic reactions has several advantages over alternative systems. However, a major bottleneck for broad industrial applications is the low specific activity of the currently used FDHs. In this study, we introduce a novel NAD-dependent formate dehydrogenase from Rhodococcus jostii (RjFDH) with both high specific activity and stability. The enzyme was identified in a targeted database research and recombinantly obtained from Escherichia coli. RjFDH is a homodimer with a monomeric molecular mass of 44.7 kDa. The homology model shows that all amino acid residues of the NAD-dependent formate dehydrogenases are usually concerned with catalytic activity, substrate acceptance, and cofactor binding. The only substrate oxidised by these enzymes is formate. RjFDH had a specific activity of 19.9 U mg^?1 at 22 °C along with unimpaired activity and high stability over a broad pH range. The K_m values for formate and NAD⁺ were 7.3 and 0.098 mmol L^?1, respectively. The optimum temperature was found to be 50 °C, at which the enzyme activity increased to about 318%. Both activity and thermal stability were higher than those of the FDH from Candida boidinii (CbFDH), which is the standard enzyme currently in use for cofactor regeneration. Different solvents roughly had the same impact on the activity and stability of both RjFDH and CbFDH. The superior performance of RjFDH over CbFDH as a regeneration system for NADH was demonstrated for the synthesis of L-tert-leucine as well as (S)-1-phenylethanol. In both systems, the concentration of RjFDH used was only one-third of the concentration of CbFDH required to achieve comparable conversion rates. Rational designing provided a promising NADP-accepting variant. Thus, RjFDH has a great potential to serve as an alternative system for NADH regeneration in enzyme-catalysed synthetic reactions. (Figure presented.).

Full text of DOI:10.1002/adsc.202000536

Title: Formate dehydrogenase from Rhodococcus jostii (RjFDH) - A
high-performance tool for NADH regeneration
Authors: Alexander Boldt and Marion Ansorge-Schumacher
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Adv. Synth. Catal. 10.1002/adsc.202000536
Link to VoR: https://doi.org/10.1002/adsc.202000536

Advanced Synthesis & Catalysis
10.1002/adsc.202000536
Very Important Publication - VIP
FULL PAPER
DOI: 10.1002/adsc.202000536
Formate Dehydrogenase from Rhodococcus jostii (RjFDH) -
A High-Performance Tool for NADH Regeneration
a
a,
Alexander Boldt and Marion B. Ansorge-Schumacher *
a
Chair of Molecular Biotechnology, TU Dresden, Zellescher Weg 20b, 01217 Dresden
Fax: +49 351 463-39520; phone: +49 351 463-39518; e-mail: marion.ansorge@tu-dresden.de
Received: ((will be filled in by the editorial staff))
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201######.((Please
delete if not appropriate))
Abstract. The use of formate dehydrogenases (FDHs) for
regeneration of the important cofactor NADH in enzyme-
the enzyme activity increased to about 318%. Both activity
catalysed synthetic reactions has several advantages over
alternative systems. However, a major bottleneck for broad
industrial applications is the low specific activity of the
currently in use for cofactor regeneration. Different solvents
currently used FDHs. In this study, we introduce a novel
NAD-dependent formate dehydrogenase from Rhodococcus
jostii (RjFDH) with both high specific activity and stability.
The enzyme was identified in a targeted database research and
was demonstrated for the synthesis of L-tert-leucine as well
recombinantly obtained from Escherichia coli. RjFDH is a
homodimer with a monomeric molecular mass of 44.7 kDa.
The homology model shows that all amino acid residues of
CbFDH required to achieve comparable conversion rates.
the NAD-dependent formate dehydrogenases are usually
The optimum temperature was found to be 50 °C, at which
and thermal stability were higher than those of the FDH from
Candida boidinii (CbFDH), which is the standard enzyme
roughly had the same impact on the activity and stability of
both RjFDH and CbFDH. The superior performance of
RjFDH over CbFDH as a regeneration system for NADH
as (S)-1-phenylethanol. In both systems, the concentration of
RjFDH used was only one-third of the concentration of
Rational designing provided a promising NADP-accepting
concerned with catalytic activity, substrate acceptance, and
variant. Thus, RjFDH has a great potential to serve as an
cofactor binding. The only substrate oxidised by these
alternative system for NADH regeneration in enzyme-
catalysed synthetic reactions.
enzymes is formate. RjFDH had a specific activity of 19.9 U
-
1
mg at 22 °C along with unimpaired activity and high stability
over a broad pH range. The K
values for formate and NAD+
m
Keywords: Biocatalysis, Cofactors, Green chemistry,
Oxidoreduction, Protein expression
-
1
were 7.3 and 0.098 mmol L , respectively.
[1].
In fact, for chemical reactions with distinct
Introduction
+
substrates, NAD(P) acts as a stoichiometric agent.
However, due to its high cost, it is usually not
economically feasible to use stoichiometric amounts
Cofactors including non-protein chemicals and metal
ions are mandatory for the catalytic activity of many
enzymes. Hence, their appropriate provision plays an
important role in synthetic biocatalysis. Most
+
[1,2]
of NAD(P) in industrial processes . Thus, efficient
strategies for the regeneration of NAD(P)+ or
NAD(P)H are crucial for large-scale synthetic
applications [3]; therefore, various chemical,
electrochemical, microbial, and enzymatic methods
prominently, nicotinamide adenine dinucleotide
+
(
phosphate), NAD(P) , or its reduced form, NAD(P)H,
is involved in the production of important building
blocks and fine chemicals employing oxidoreductases
[3,4]
have been developed for this purpose
. For
industrial purposes, the enzymatic approach is often
1
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.202000536
preferred owing to its high selectivity and efficiency
include enzymatic synthesis of chiral aromatic α-keto
[1,2].
However, till date, only a few enzymes suitable
esters or L-lactic acid, with xylose reductase and lactic
[18,19]
for the enzymatic regeneration of NAD(P)H are
known; among them are NAD-dependent formate
dehydrogenases.
acid dehydrogenase as catalysts, respectively
.
First described in 1950, NAD-dependent formate
dehydrogenases (FDH, EC 1.2.1.2) catalyse the
[5,6,7]
oxidation of formate to carbon dioxide
. This
+
oxidation is coupled to the reduction of NAD to
NADH. During the catalytic step, direct transfer of
hydride ion from formate to NAD takes place [8].
Several advantages of FDHs over alternative enzymes
make them particularly desirable for cofactor
regeneration: 1. During formate oxidation and
+
Figure 1. L-tert-leucine formation from trimethylpyruvic
acid catalysed by leucine dehydrogenase (LeuDH) and
NADH regeneration with formate dehydrogenase (FDH).
associated NAD+ reduction, only inert CO
produced as a by-product. The CO produced can
easily be removed from the medium pushing the
2
is
2
[1]
reaction equilibrium towards product formation . 2.
The thermodynamics of formate oxidation are quite
favourable because the back reaction i.e. fixation of
To date, the main bottleneck of cofactor regeneration
employing CbFDH is its low specific activity, of about
-1 [20]
6 U mg
. Alternative NAD-dependent FDHs
CO
2
to formate, is prevented by thermodynamic
capable of replacing CbFDH in cofactor regeneration
have not been described in the literature, to the best of
our knowledge. Although enzymes with higher
formate-oxidising activity were discovered, they
considerably lacked stability. A recent example is the
[1,9]
barriers under standard conditions . 3. Formate as a
substrate is innocuous for most enzymes as well as
inexpensive [1,10,11]. 4. NAD-dependent formate
dehydrogenases do not incorporate sensitive metal
ions or prosthetic groups and are known to have a wide
pH optimum which enables coupling with a broad
range of NADH-dependent oxidoreductases.
The widely employed enzyme for the
regeneration of NADH is the formate dehydrogenase
of the yeast Candida boidinii (CbFDH)
prominent industrial application of this enzyme is the
production of L-tert-leucine [12,13,14]. Because of its
bulky and inflexible tert-butyl side chain, L-tert-
leucine is widely used in the pharmaceutical and
cosmetic industry as a precursor molecule for various
drugs such as HIV protease inhibitors or tumour-
fighting agents [14,15,16]. Since chemical methods of L-
tert-leucine production have major drawbacks, a
biocatalytic approach has been established, yielding
excellent enantioselectivity under mild conditions
[9]
formate dehydrogenase of Ancylobacter aquaticus .
In the present study, we report a novel NAD-
dependent formate dehydrogenase with both high
catalytic activity and stability. This enzyme was
derived from a rational database research based on the
amino acid sequences of three NAD-dependent FDHs
with the highest reported activities so far. The coding
gene was cloned and expressed in Escherichia coli,
and the resulting enzyme was described. The cofactor
regeneration performance of this enzyme was
investigated for the synthesis of L-tert-leucine as well
as (S)-1-phenylethanol, and benchmarked against the
established formate dehydrogenase from C. boidinii
(CbFDH). Rational designing provided a promising
NADP-accepting variant with high cofactor
regeneration ability.
[1]
. A
[
16,17]. In this process, enantiopure L-tert-leucine is
synthesised on a ton-scale through reductive amination
of trimethylpyruvic acid catalysed by leucine
dehydrogenase (LeuDH) from Lysinibacillus
sphaericus, previously Bacillus sphaericus (Figure 1)
The process, however, became industrially feasible
only through the establishment of CbFDH-catalysed
cofactor regeneration. Other applications of CbFDH
Results and Discussion
Based on the assumption that FDHs with increased
formate-oxidising activity will be structurally related
to already known FDHs having high activity, a
database research was performed using the amino acid
sequences of the highly active NAD-dependent FDHs
[1].
2
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.202000536
from A. aquaticus, Thiobacillus KNK65M, and
Pseudomonas sp. 101 [8,9,21,22] as templates. The three
sequences were aligned against various NAD-
dependent FDHs using the BLASTP database (Figure
S1, supplementary material). The search was restricted
to bacterial FDHs because bacterial FDHs have been
reported to have overall higher specific activities and
and Pseudomonas sp. 101, respectively. However,
sequence identity to the CbFDH was only 41%. One
subunit of this recombinant enzyme corresponds to a
molecular weight of 44.7 kDa (including the adjusted
His-Tag; Figure S2, supplementary material). A
homology model of RjFDH (Figure S3, supplementary
[25]
material) built with the SWISS-MODEL algorithm
[8]
superior stabilities as compared to eukaryotic FDHs .
Bacterial FDHs are often identified by a prolonged N-
terminal loop, which is considered to contribute to its
stability [23]. In addition, only candidates with more
than 385 amino acid residues were taken into
consideration assuming that the additional C-terminal
extension that was found in many highly active FDHs
might also play a role in its activity [24]. Among the
resulting hits, seven FDH candidates with 90–60%
amino acid sequence identity to the primary sequences
were chosen. Their genes were expressed in E. coli and
the activity of the recombinant enzymes was
determined in the crude cell extracts (Table 1). FDH
from
on the crystal structure of Pseudomonas sp. 101 FDH
(RCSB ID: 2NAD) shows a particularly high identity
between the active centres of the two enzymes in terms
of spatial structure as well as amino acid composition.
Figure 2-A illustrates the highly conserved amino
acids Arg285 and Asn147, which are essential for
formate binding, and Gln314 and His333, which are
crucial for catalysis [26]. Thr283, Asp309, Ser335, and
Gly336 stabilise the nicotinamide ring in the active
centre and are conserved in the FDHs from C. boidinii,
Moraxella sp., A. aquaticus, and Thiobacillus sp.
KNK65M (Figure S1, supplementary material).
A
R. jostii showed the highest enzyme activity. Notably,
it also exceeded the documented maximum activity of
FDH from C. boidinii [20], even in its crude form. To
the best of our knowledge, this FDH has not been
described in the literature so far.
Table 1. Amino acid identities with templates from
A. aquaticus, Thiobacillus sp. KNK65M, and Pseudomonas
sp. 101, and chain length and specific catalytic activities of
FDH candidates.
Sequence
identity
Sequence
identity with
Thiobacillus
sp.
Sequence
Chain
length
[aa]
Specific
activity
[U mg-1
]
Donor
organism
identity with
Pseudomona
s sp.101 (%)
with A.
aquaticus
KNK65M
(
%)
(
%)
Cupriavidus
necator
61
64
76
77
72
74
86
67
65
78
77
76
75
84
68
386
386
399
402
402
400
396
0.03
1.6
2.5
1.3
0.4
6.3
0.8
Granulicella
mallensis
74
75
72
74
Hyphomicrobium
denitrificans
Methylohalobius
crimeensis
Methylotenera
versatilis
Rhodococcus jostii
8
4
Sciscionella
marina
64
RjFDH, having a chain length of 400 amino acids,
belongs to the longest NAD-dependent FDHs
described so far. The sequence of this enzyme is 84%,
86%, and 84% identical to the very active template
FDHs from A. aquaticus, Thiobacillus sp. KNK65M
3
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.202000536
+
B
other hand, K
m
values for both NAD and formate
[8]
were in the same range as those of other FDHs . This
makes RjFDH the most efficient enzyme among the
non-metal–containing
NAD-dependent
FDHs
described till date.
Table 2. Kinetic parameters for formate oxidation and
+
NAD reduction by RjFDH and CbFDH. Activity was
measured at 22 °C and pH 7.0.
NAD+
K
HCOO
-
k
cat / K
HCOO-
V
[
max
U mg-1
K
m
m
m
FDH
-1
-1
[mmol L min−1
-1
]
[mmol L ]
[mmol L ]
]
RjFDH
19.9 ± 2.1
6.0 [20]
0.098 ± 0.05
7.2 ± 1.8
117.4
CbFDH
0.05 [20]
5.6 [20]
25.4 [28]
An explanation for the differences in the activity of
NAD-dependent FDHs has not yet been established.
However, considering the high degree of conservation,
the differences in activity certainly cannot be related
to the amino acid composition which is directly
involved in the catalytic reaction. For the reverse
reaction, i.e. reduction of CO , it was assumed that
2
high activity was related to the presence of an
elongated C-terminus. In FDH from Thiobacillus sp.,
which is highly active, this elongation comprises 26
amino acids. The additional amino acids seem to mask
the active site of the enzyme, thus excluding solvent
Figure 2. Active centre (A) and cofactor-binding pocket (B)
of RjFDH. A: Superposition of the homology model of
RjFDH (white) with the crystal structure of FDH from
Pseudomonas sp. 101 (blue). B: Spatial enlargement of the
cofactor-binding pocket due to the exchange of three amino
acids. I: Asp222Ser, II: Glu261Gly, III: His380Arg. NAD+
[24]
molecules . Computer simulations show that this
constellation lowered the activation energy of CO2
is illustrated in turquoise, and the position of the bulky
phosphate residue is shown in black. The green mesh
illustrates the wild-type cofactor-binding pocket
-1
reduction by 10 kcal mol as compared to reactions
catalysed by enzymes without an elongated C-
terminus, similar to CbFDH.
As typical for most FDHs, this enzyme also accepted
only formate as a substrate; glycerol, 2-propanol,
acetaldehyde, ethanol, methanol, and formaldehyde at
Preliminary experiments in our lab indicate that
the C-terminal elongation might also benefit formate
oxidation. We created variants of RjFDH, in which the
C-terminus was truncated by 5, 15, 25, and 35 amino
acids, and determined their catalytic activity. Deletion
of 5 to 15 amino acids decreased the activity to about
-1
a concentration of 25 mmol L (pH 7,0; 22 °C) were
+
not converted by RjFDH. When NADP was used as
cofactor, RjFDH activity was approximately 4% of the
+
activity when NAD was used. The enzyme activity
9
0%, while truncation of 25 amino acids decreased the
was lost after 96 h of incubation in the presence of 1
mmol L Cu , but remained unaffected by other metal
-
1
2+
activity to 10%, and truncation of 35 amino acids
resulted in complete loss of enzyme activity. From the
homology model, it can be deduced that this effect is
not due to the loss of amino acids from the active
centre. The C-terminal elongation can be found in
many NAD-dependent FDHs having better formate-
oxidising activity than CbFDH, including FDHs from
Paracoccus sp-12-A, Pseudomonas sp. 101, and A.
aquaticus. On the other hand, these enzymes show
only half of the specific activity of RjFDH, which
suggests the contribution of factors other than the C-
2
+
2+
3+
+
2+
2+
ions (Ni , Fe , Fe , Li , Mn , and Mg ) under the
same conditions.
+
For reduction of NAD by RjFDH, a Vmax of 19.9
2.1 U mg (22 °C, pH 7.0) and a catalytic efficiency
of 117.4 mmol L min was obtained. This exceeds
the reported values of CbFDH by a factor of 3 and 4.5,
respectively (Table 2). RjFDH activity obtained was
also higher than that reported for NAD-dependent
-
1
±
-1
-1
FDHs from Paracoccus sp. 12-A, Pseudomonas sp.
8,9,27]
1
01, Moraxella sp., and A. aquaticus [
. On the
4
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.202000536
terminal length to catalytic activity. One such factor
might be the composition of the N-terminal region of
the enzymes, as it has been suggested that a special
loop on the prolonged N-terminus of bacterial FDHs
activity at different pH values. B: Residual activity after 24
h of incubation.
Maximum activity of RjFDH was obtained at
5
3
0 °C. At this temperature, the activity was about
20% of the activity at 22°C (Figure 4-A). The
[8,23]
contributes to the enzymatic reaction
.
The pI of RjFDH was calculated to be 5.4, and the
highest formate-oxidising activity was obtained at pH
temperature maximum is in the same range as that of
the FDH from A. aquaticus and slightly lower than that
of the FDHs from Thiobacillus sp. KNK65M and C.
7.0 (Figure 3-A). The pH optimum for this enzyme
was wide and ranged from 6.0 to 9.0, which complies
with the features of other NAD-dependent FDHs.
According to the findings of Tishkov et al. (1996), this
correlates with the presence of glutamine 97 and 342
in close proximity to the active centre [ . Nearly no
loss of activity was observed when RjFDH was
incubated at pH 6.0–9.0 for 24 h (Figure 3-B).
However, the activity dropped to 75% after 24 h of
incubation at pH 10.0, and to 16.6% and 0.75% at pH
boidinii [9,20,22]
.
Thermal stability was investigated for molecular
unfolding as well as deactivation at increasing
temperatures. The unfolding experiments utilised the
changes in intrinsic fluorescence of tryptophan and
tyrosine residues within the enzyme at an absorption
wavelength of 350 nm and an excitation wavelength of
29]
3
30 nm. Transitions in the overall folding state of
RjFDH were observed at an inflation temperature of
3.0 ± 0.19 °C. For CbFDH, a slightly lower inflation
5
.0 and 4.0, respectively. The wide pH optimum and
high stability of RjFDH in that pH range ensure broad
applicability for cofactor regeneration [30]
7
temperature of 70.8 ± 0.52 °C was observed.
Deactivation experiments revealed that the thermal
stability of RjFDH and CbFDH diverged at 50 °C and
above (Figure 4-B). At 55 °C, the residual activity of
RjFDH was about 70% after 20 min of incubation, and
.
50% activity was observed after 20 min of incubation
at 60 °C. In contrast, CbFDH only showed 20%
residual activity after 20 min of incubation at 60°C.
This is in accordance with previous studies, which
reported 50% residual activity of CbFDH after 20 min
[20,31]
of incubation at 55 °C
. Alternatively, FDHs from
A. aquaticus, Moraxella sp., and Thiobacillus sp.
KNK65M lost 40, 50, and 55% activity, respectively,
[9,22,32]
after only 10 min of incubation at 55 °C
The half-life time of the RjFDH was determined
to be 20.4 hours at a temperature of 50 °C (k = 0.034).
Under comparable conditions, a half-life time of only
.7 hours was reported for CbFDH [33]. All results
.
d
4
indicate that RjFDH is a suitably stable NAD-
dependent FDH for synthetic application.
Figure 3. Effect of pH on formate oxidation (A) and
stability (B) of RjFDH. Enzyme activity was determined
-
1
-1
using 3.5 µg mL RjFDH at 25 °C in 100 mmol L buffer
at different pH. Enzyme activity of fresh enzyme measured
at pH 7.0 was used as reference (100% relative activity
-
1
refers to 20.4 U mg at 25 °C). Each result represents the
average of three individual measurements. A: Residual
5
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.202000536
demonstrated a possible increase in FDH activity by
organic solvents [37]. A complete loss of RjFDH
activity was observed in 50% (v/v) acetone as well as
ethanol. Altogether, the relative activity of RjFDH was
very similar to that of CbFDH in different solvents
(
Figure 5). However, the RjFDH appears to be a bit
more tolerant towards acetonitrile, but slightly more
sensitive towards ethanol and ethylene glycol.
Figure 5. Activity of RjFDH and CbFDH in the presence of
water-miscible organic solvents. The solvents were added to
the assay mixture and enzyme activity was immediately
measured at standard conditions. The activity was related to
the maximum enzyme activity in a solvent-free system.
Each result represents the average of three individual
measurements.
After incubation of RjFDH in biphasic systems
consisting of water and methyl isobutyl ketone
(
MIBK), cyclopentyl methyl ether (CPME), ethyl
acetate, 2-methyltetrahydrofuran (2-methyl-THF),
hexane, toluene, or n-butanol, at a fixed ratio of 1:1
(v/v) at 22 °C, a decrease in activity was observed in
all systems (Figure 6). High residual activities were
obtained in systems containing n-butanol and 2-
methyl-THF, reaching 72% and 58%, respectively,
after 60 min. With MIBK, CPME, and ethyl acetate
systems, a very high initial activity drop of about 60%
was observed; however, the activity continued to
decrease gradually over time. With hexane and toluene
systems, the activity not only dropped to 40% within
the first 10 min, but continued to decrease rapidly even
after that. Interestingly, the solvent-tolerant NAD-
dependent FDH from Bacillus sp. F1 behaved similar
in hexane, although the residual activity in n-butanol
Figure 4. Effect of temperature on activity (A) and stability
(
B) of RjFDH. Enzyme activity was determined using 7 µg
-
1
mL RjFDH at pH 7.0 after 20 mins of incubation at the
selected temperature. Each measurement was performed
with fresh enzyme. The activity was related to the maximum
activity at 22 °C (100% relative activity refers to a specific
activity of 19.9 U mg ). Each result represents the average
of three individual measurements.
-
1
Considering the synthetic application of enzyme, the
impact of organic solvents on RjFDH in mono- and
biphasic reaction systems was determined. Both these
systems are frequently applied in biocatalysis for the
conversion or production of compounds with low
water solubility [34,35,36].
It was found that addition of water-miscible
organic solvents, including acetone, dimethyl
sulfoxide, ethylene glycol, ethanol, and acetonitrile, to
the standard reaction set-up had little impact on
RjFDH activity up to a solvent concentration of 5%
system was slightly higher (94%) [38]
.
(
v/v) (Figure 5). At 25% (v/v) solvent concentration,
the enzyme activity dropped to a varying extent,
except in the presence of DMSO, which increased the
activity of RjFDH to about 134%. This is consistent
with the report by Demchenko et al. (1990), which
6
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.202000536
Figure 6. Residual activity of RjFDH after exposure to non–
water-miscible organic solvents in a biphasic system. The
ratio of water to organic solvent was 1:1. The enzyme was
incubated in the system at 22 °C. Samples were collected
from the aqueous phase at defined time intervals and
transferred to a standard assay for activity measurements.
The activity was related to the maximum activity in a
solvent-free system. Each result represents the average of
three individual measurements.
The performance of RjFDH as a regeneration system
for NADH was demonstrated for the synthesis of L-
tert-leucine and (S)-1-phenylethanol, and compared to
the regeneration capacity of the commercially
established CbFDH under the same conditions.
Synthesis of L-tert-leucine was accomplished
with LeuDH from L. sphaericus, which complies with
the industrially established system [ . LeuDH was
added in excess over FDH to ensure that NADH
regeneration by FDH was the rate-limiting step in the
reaction. The decrease in substrate concentration was
1,39]
monitored, which, due to the documented selectivity
of LeuDH [
14,16,17,40]
, directly reflects L-tert-leucine
production. Figure 6-A illustrates the rapid substrate
consumption (initial substrate concentration: 10 mmol
-1
-1
L ) by RjFDH at a concentration of 25 µg mL . After
0 min, 95% of trimethylpyruvic acid (TMP) was
6
converted to L-tert-leucine. However, with the same
concentration of CbFDH, the substrate consumption
rate was considerably lower, and only 47% was
converted after 60 min, demonstrating the superior
catalytic efficiency of RjFDH. In fact, a three-fold
higher concentration of CbFDH was required to obtain
Figure 7. Catalytic performance of NADH-dependent
dehydrogenases with FDHs for cofactor regeneration. All
calculations were based on the average of three individual
measurements. A: TMP concentration in a combined system
of LeuDH and FDHs at different concentrations.
cofactor regeneration equal to that obtained by RjFDH. B: (S)-1-phenylethanol concentration in a combined system
of ADH-A and FDHs at different concentrations.
Very similar results were obtained when RjFDH and
CbFDH were coupled with the NAD-dependent
secondary ADH-A from R. ruber. ADH-A is a
versatile catalyst for the asymmetric reduction of
ketones [41] and was used for the synthesis of (S)-1-
phenylethanol. Like LeuDH, it was added in excess
over FDH. Figure 6-B illustrates the increasing (S)-1-
phenylethanol concentration over time (initial
-1
substrate concentration: 25 mmol L ). The use of 25
-1
µg mL RjFDH yielded about 75% of the product
within 120 min. In contrast, only 40% (S)-1-
phenylethanol was obtained with the same amount of
CbFDH within the same time period. Moreover, even
-1
with 75 µg mL CbFDH, the product yield was only
7
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.202000536
6
7%, that is, lower than that obtained by 25 µg mL-1
RjFDH.
Both model systems demonstrated superior
Unless stated otherwise, all chemicals and
oligonucleotides were purchased from Sigma-Aldrich
(St. Louis, MO, USA), Carl Roth GmbH & Co. KG
(Karlsruhe, Germany) or abcr GmbH (Karlsruhe,
Germany). Media components were purchased from
Carl Roth GmbH & Co. KG (Karlsruhe, Germany) and
enzymes for molecular cloning were obtained from
New England Biolabs Inc. (Ipswich, USA).
performance of RjFDH as compared to the standard
enzyme. Thus, RjFDH shows great potential to replace
the current standard enzyme in cofactor regeneration.
Studies on the homology model suggested the
+
possibility of improving RjFDH activity with NADP
as the cofactor by decreasing the steric hindrance of
the binding pocket for the phosphate moiety. Suitable
amino acid residues for this purpose and appropriate
substitutions were predicted using the online tool
CSR-SALAD [ . An RjFDH variant with the amino
acid exchanges Asp222Ser, Glu261Gly, and
His380Arg was created, in which the cofactor-binding
pocket was spatially enlarged, while all electrostatics
were maintained (Figure 2-B). The specific activity of
Construction of expression vectors
The genes coding for formate dehydrogenase (FDH) in
Rhodococcus jostii (RjFDH), Cupriavidus necator,
42]
Granulicella
mallensis,
Hyphomicrobium
denitrificans,
Methylohalobius
crimeensis,
Methylotenera versatilis, Sciscionella marina, and
leucine dehydrogenase (LeuDH) in Lysinibacillus
sphaericus were obtained via polymerase chain
reaction (PCR) of genomic DNA purchased from
DSMZ (Deutsche Sammlung von Mikroorganismen
und Zellkulturen GmbH, Braunschweig, Germany).
The sequences of the primers used are mentioned in
supplementary material (Table S1). The PCR mixture
+
this variant with NADP as the cofactor was 3.73 U
-1
mg , which was over 400-fold higher than that of the
wild-type enzyme (Table S4, supplementary material).
The activity increase corresponded to a 10-fold
+
-1
decrease in the K value for NADP to 0.15 mmol L .
m
+
-1
Thus, an NADP -accepting RjFDH variant with an
activity almost similar to that of CbFDH in NADH
regeneration was obtained, indicating a potential for
further improvement.
contained 1-× Phusion HF buffer, 0.5 µmol L of each
-1
primer, 0.2 mmol L of each dNTP, 5% (v/v) dimethyl
sulfoxide (DMSO), 44–100 ng genomic DNA, and
-1
®
0.05 U µL Phusion high-fidelity DNA polymerase
New England Biolabs Inc., USA). Gibson assembly
(
was used to insert the genes into a pET-22b(+)
expression vector (Merck, Darmstadt, Germany)
according to the standard protocol to create proteins
with a C-terminal polyhistidine tag. The insertion was
carried out under the following conditions: 60 min at
Conclusion
[43]
NAD-dependent FDH from Rhodococcus jostii
(
RjFDH) was found to have high formate-oxidising
activity over a broad pH and temperature range with
good thermal stability. Thus, RjFDH can serve as a
suitable alternative system for NADH regeneration.
As only low concentrations of this enzyme are
50 °C followed by 15 min at 72 °C. After Gibson
assembly, E. coli TOP10 cells were transformed with
the gene construct and positive clones were screened
by colony PCR. Plasmids were isolated from the
positive clones, and sequenced by GATC Biotech
+
required for efficient NAD reduction in a coupled
synthetically active reaction system, RjFDH can
reduce production costs and contribute to the
expansion of biocatalytic processes. This could be
further enhanced with improvement of the enzyme
through molecular engineering Moreover, we obtained
a first variant of RjFDH with a promising acceptance
(
Konstanz, Germany). The plasmids pET22b_ADH-A,
harbouring the alcohol dehydrogenase A (ADH-A)
gene from Rhodococcus ruber, and pET22b_CbFDH,
harbouring the FDH gene from C. boidinii, were
obtained from our laboratory stock.
+
of NADP as a cofactor.
Recombinant expression of the FDH and LeuDH genes
Vectors harbouring the FDH and LeuDH genes were
transformed into E. coli BL21 (DE3) for expression.
The bacteria were cultured overnight at 37 °C in 5 mL
lysogeny broth (LB) containing 100 mg mL-1
ampicillin. A total volume of 500 mL of the main
culture was inoculated with 50 µL of preculture and
Experimental Section
Materials
8
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.202000536
the cells were grown at 37 °C until an optical density
of 0.6 was reached at 600 nm (OD600). Gene expression
was induced by addition of 0.1 mmol L Isopropyl β-
D-1-thiogalactopyranoside (IPTG), and the
temperature was decreased to 20 °C. After 40 h of
incubation, the cells were harvested by centrifugation
SDS-PAGE was performed according to the method of
Laemmli [44] on a SERVA BlueVertical™ PRiME™
(SERVA Electrophoresis GmbH, Heidelberg,
Germany) using 12% (v/v) acrylamide [44]. Samples
were incubated in Laemmli buffer (60 mM Tris-HCl
pH 6.8, 2% (m/v) sodium dodecyl sulfate, 10% (v/v)
-1
(
−
4000 rpm for 15 min at 4°C) and stored in a freezer at
20 °C.
glycerol,
β-mercaptoethanol, 0.025% bromophenol blue) at
5°C for 10 min (16 µL total volume) and loaded onto
5%
(v/v)
9
Enzyme purification
the gel after cooling on ice. The protein standard used
was 5 µL of protein colour standard (New England
Biolabs Inc., Ipswich, MA, USA). Electrophoresis was
performed for 80 min at 180 V. Proteins were
visualised by Coomassie brilliant blue R-250 staining.
Cell pellets obtained by centrifugation were
-1
resuspended in 50 mmol L potassium phosphate
buffer (pH 7.0 for FDH-containing cells and pH 8.0 for
LeuDH-containing cells). Phenylmethylsulfonyl
fluoride (PMSF) was added to the suspension at a final
-1
Determination of enzyme activity
concentration of 1 mmol L . A French press was
applied for cell lysis, utilizing three cycles at a
pressure of 12000 psi. After lysis, cell debris were
removed by centrifugation (20000 rpm, 90 min and
FDH activity was determined by measuring the
+
reduction of NAD to NADH in 1 mL reaction mixture
at 22 °C using
a
Hach-Lange DR-3900
4
°C) and the clear lysate was used for protein
spectrophotometer (Hach Lange GmbH, Düsseldorf,
Germany) at a wavelength of 340 nm. The standard
purification by immobilised metal affinity
chromatography (IMAC). IMAC involved a HisTrap
FF Crude 5 mL column (GE Healthcare Life Sciences,
Chicago, USA) in an ÄKTA FPLC system (GE
Healthcare Life Sciences, Chicago, USA). The column
was equilibrated with 25 mL of washing buffer (500
-1
assay was composed of 100 mmol L phosphate buffer
-1
(pH 7.0), 100 mmol L sodium formate, and 2.5 mmol
-1
+
L NAD . The reaction was started by adding RjFDH
to the assay. For kinetic measurements, the enzyme
-1
concentration was set to 3.5 µg mL . K and Vmax
m
-1
-1
mmol L NaCl, 50 mmol L potassium phosphate at
pH 7.0) and 4% (v/v) elution buffer (500 mmol L
were calculated from the measurements and various
-1
+
concentrations of NAD and formate were analysed
-1
NaCl, 500 mmol
L
imidazole, 50 mmol
potassium phosphate, pH 7.0). The lysate was
applied to the ÄKTA system, the flow rate was set to
using hyperbolic regression with Hyper32 software
(https://hyper32.software.informer.com/). One unit
-1
L
+
was defined as the reduction of 1 µmol NAD per
-1
1
mL min , and the step gradient was set until the
minute. To measure the pH dependency of enzyme
activity, different buffers (pH 4.0-5.0: sodium acetate,
pH 6.0-8.0: potassium phosphate, pH 9.0:
-1
concentration of imidazole reached 500 mmol L . The
whole flow-through was fractionated and the fractions
containing the recombinant FDH or LeuDH were
tested for their respective activities. Fractions with the
highest activity were pooled and dialysed against
triethanolamine
(TEA),
pH
10.0:
3-
(Cyclohexylamino)-1-propanesulfonic acid) were
-1
used at a final concentration of 100 mmol L . LeuDH
activity was determined by measuring the oxidation of
-1
potassium phosphate buffer (50 mmol L , pH 7.0) for
0 h at 6 °C. ADH-A was purified by heat precipitation
at 70 °C for 50 min followed by centrifugation at
0.000 rpm for 45 min at 4°C. Protein concentration
+
2
NADH to NAD in 1 mL of reaction mixture at 22 °C
using a Hach-Lange DR-3900 spectrophotometer
(Hach Lange GmbH, Düsseldorf, Germany) at a
wavelength of 340 nm. The standard assay was
2
was determined using a NanoDrop ND-1000
spectrophotometer (Thermo Fisher Scientific,
Waltham, USA) at a wavelength of 280 nm and a
Pierce BCA Protein Assay Kit (Thermo Fisher
Scientific, Waltham, USA) according to the
manufacturer’s instructions.
-1
composed of 100 mmol L TEA buffer (pH 8.5), 200
-1
-1
mmol L ammonium formate, 20 mmol L
trimethylpyruvic acid (TMP), and 0.5 mmol L
-1
NADH. All activities were calculated as the average of
three independent experiments.
Polyacrylamide gel electrophoresis (PAGE)
Determination of optimum temperature and thermal
stability
9
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.202000536
To determine thermal stability, 30 µL of purified
RjFDH was incubated at different temperatures for 20
min in a thermocycler (Mastercycler personal;
Eppendorf AG, Hamburg). The residual activity after
excess. Each set-up was stirred on a magnetic stirrer at
22 °C. Samples were taken at defined time-points and
analysed for TMP consumption using a KNAUER
Smartline HPLC (Berlin, Germany) system equipped
with a LiChrosorb RP-18 column (250 mm × 4 mm, 5
µm) at 30 °C. An isocratic mobile phase consisting of
25 mmol L ammonium dihydrogen phosphate with
10% (v/v) acetonitrile (pH 6.0) was applied. UV
detection was carried out at 205 nm for 14 min, and
TMP was detected at a retention time of 6.60 min.
20 min of incubation was measured photometrically,
as described previously. RjFDH incubated on ice was
used as a reference. Thermal unfolding experiments
were performed on a Tycho spectrofluorometer
-1
(
Nanotemper
Technologies,
Munich)
with
concentrated enzyme (as obtained after purification) as
well as 1:4 diluted enzyme. The optimum temperature
was investigated in the range of 30–75 °C using a Cary
Acknowledgements
60
UV-Vis spectrophotometer. Cuvettes were
preheated at the respective temperatures and the
reaction was carried out under standard conditions.
We thank Dr. Tobias Wager (NanoTemper
Technologies GmbH) for providing the opportunity to
use Tycho spectrofluorometer for measurement.
Determination of solvent impact on enzyme activity
Impact of the water-miscible polar solvents DMSO,
acetonitrile, ethanol, acetone, and ethylene glycol on
the activity of RjFDH was measured by addition of the
respective solvents in different concentrations (1–50%
References
[
1] A. Weckbecker, H. Gröger, W. Hummel, in Biosystems
Engineering I (Eds.: C. Wittmann, R. Krull), Springer
Berlin Heidelberg, Berlin, Heidelberg, 2010, pp. 195.
[
v/v]) to the reaction mixture. The effect of a two-
phase system consisting of water and the immiscible
organic solvents methyl isobutyl ketone (MIBK),
cyclopentyl methyl ether (CPME), ethyl acetate, 2-
methyltetrahydrofuran, hexane, toluene, and n-butanol,
on RjFDH activity was investigated by mixing 500 µL
of 1:10 diluted RjFDH (50 mmol L phosphate buffer,
pH 7.0) and 500 µL of organic solvent at 22 °C in an
overhead shaker (60 rpm). Ten microlitres of the
aqueous phase was collected at defined time-points
and subjected to standard activity assay.
[2] X. Wang, T. Saba, H. H. P. Yiu, R. F. Howe, J. A.
Anderson, J. Shi, Chem 2017, 2, 621.
[
3] S. K. Spaans, R. A. Weusthuis, J. van der Oost, S. W.
M. Kengen, Front. Microbiol. 2015, 6: 742.
-
1
[
4] H. K. Chenault, G. M. Whitesides, Appl. Biochem.
Biotechnol. 1987, 14, 147.
[5] M. B. Mathews, B. Vennesland, J. Biol. Chem. 1950,
1
86, 667.
Determination of cofactor regeneration
[6] C. Vinals, E. Depiereux, E. Feytmans, Biochem.
Biophys. Res. Commun. 1993, 192, 182.
ADH-A from R. ruber [23] and different concentrations
-1
of RjFDH or CbFDH, were added to 25 mmol L
[
7] M. Jormakka, S. Törnroth, B. Byrne, S. Iwata, Science
-1
acetophenone, 100 mmol L potassium formate, and
2002, 295, 1863.
.25 mmol L-1 NAD in 100 mmol L phosphate
+
-1
0
[
1
8] V. I. Tishkov, V. O. Popov, Biochemistry 2004, 69,
252.
buffer (pH 7.0) to a total volume of 5 mL. ADH-A was
always added in excess. Samples were collected at
defined time-points and analysed for (S)-1-
phenylethanol using gas chromatography. Product
formation was measured using a Shimadzu GC2010-
Plus (Kyoto, Japan) equipped with a Hydrodex-γ-
DiMOM (0.25 mm, 50 m) column at 115 °C. (S)-1-
phenylethanol was detected at a retention time of 7.83
min. Likewise, LeuDH and RjFDH were added to 200
[
9] H. Nanba, Y. Takaoka, J. Hasegawa, Biosci.,
Biotechnol., Biochem. 2003, 67, 720.
[
10] W. A. van der Donk, H. Zhao, Curr. Opin.
Biotechnol. 2003, 14, 421.
[
8
11] V. I. Tishkov, V. O. Popov, Biomol. Eng. 2006, 23,
9.
-1
-1
mmol L ammonium formate, 10 mmol L TMP, and
[12] H. Schütte, J. Flossdorf, H. Sahm, M.-R. Kula, Eur. J.
Biochem. 1976, 62, 151.
-1
+
-1
0
8
.25 mmol L NAD in 100 mmol L TEA buffer (pH
.5) to a total volume of 5 mL. LeuDH was always in
1
0
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.202000536
[
13] A. S. Bommarius, M. Schwarm, K. Stingl, M.
Kottenhahn, K. Huthmacher, K. Drauz, Tetrahedron:
Asymmetry 1995, 6, 2851.
[29] V. I. Tishkov, A. D. Matorin, A. M. Rojkova, V. V.
Fedorchuk, P. A. Savitsky, L. A. Dementieva, V. S.
Lamzin, A. V. Mezentzev, V. O. Popov, FEBS Lett. 1996,
390, 104.
[
14] A. S. Bommarius, M. Schwarm, K. Drauz, J. Mol.
Catal. B: Enzym. 1998, 5, 1.
[
30] T. Watanabe, T. Hattori, S. Tengku, M. Shimada,
[
15] J. Z. Jin, D. L. Chang, J. Zhang, Appl. Biochem.
Enzyme Microb. Technol. 2005, 37, 68.
Biotechnol. 2011, 164, 376.
[
31] K. Schirwitz, A. Schmidt, V. S. Lamzin, Protein Sci.
[
16] J. Li, J. Pan, J. Zhang, J. H. Xu, J. Mol. Catal. B:
2007, 16, 1146.
Enzym. 2014, 105, 11.
[
32] Y. Asano, T. Sekigawa, H. Inukai, A. Nakazawa, J.
[
17] W. Liu, H. Ma, J. Luo, W. Shen, X. Xu, S. Li, Y. Hu,
Bacteriol. 1988, 170, 3189.
H. Huang, Biochem. Eng. J. 2014, 91, 204.
[
33] S. Felber, PhD thesis 2001.
[
18] R. Kratzer, M. Pukl, S. Egger, B. Nidetzky, Microb.
Cell Fact. 2008, 7, 37.
[34] E. Andersson, B. Hahn-Hägerdal, Enzyme Microb.
Technol. 1990, 12, 242.
[
19] R. Jayabalan, M. Sathishkumar, E. S. Jeong, S. P.
Mun, S. E. Yun, Bioresour. Technol. 2012, 123, 686.
[35] H. J. Heipieper, G. Neumann, S. Cornelissen, F.
Meinhardt, Appl. Microbiol. Biotechnol. 2007, 74, 961.
[
20] H. Slusarczyk, S. Felber, M. R. Kula, M. Pohl, Eur. J.
Biochem. 2000, 267, 1280.
[36] A. M. Klibanov, Nature 2001, 409, 241.
[
21] V. Tishkov, A. Galkin, G. Marchenko, Y. Tsygankov,
[37] A. P. Demchenko, O. I. Rusyn, A. M. Egorov, V. I.
Tishkov, Biochim. Biophys. Acta, Protein Struct. Mol.
Enzymol. 1990, 1039, 290.
A. Egorov, Biotechnol. Appl. Biochem. 1993, 18, 201.
[
22] H. Nanba, Y. Takaoka, J. Hasegawa, Biosci.,
Biotechnol., Biochem. 2003, 67, 2145.
[38] H. T. Ding, D. F. Liu, Z. L. Li, Y. Q. Du, X. H. Xu,
Y. H. Zhao, J. Appl. Microbiol. 2011, 111, 1075.
[
23] V. V. Fedorchuk, A. G. Galkin, I. E. Yasny, L. B.
Kulakova, A. M. Rojkova, A. A. Filippova, V. I. Tishkov,
Biochemistry 2002, 67, 1145.
[39] I. Ahmed, A. Yokota, A. Yamazoe, T. Fujiwara, Int.
J. Syst. Evol. Microbiol. 2007, 57, 1117.
[
24] H. Choe, J. M. Ha, J. C. Joo, H. Kim, H. J. Yoon, S.
[40] L. Zhu, Z. Wu, J. M. Jin, S. Y. Tang, Appl. Microbiol.
Biotechnol. 2016, 100, 5805.
Kim, S. H. Son, R. M. Gengan, S. T. Jeon, R. Chang, et al.,
Acta Crystallogr., Sect. D: Biol. Crystallogr. 2015, 71,
3
13.
[41] B. Kosjek, W. Stampfer, M. Pogorevc, W. Goessler,
K. Faber, W. Kroutil, Biotechnol. Bioeng. 2004, 86, 55.
[
25] A. Waterhouse, M. Bertoni, S. Bienert, G. Studer, G.
Tauriello, R. Gumienny, F. T. Heer, T. A. P. de Beer, C.
Rempfer, L. Bordoli, et al., Nucleic Acids Res. 2018, 46,
[42] J. K. B. Cahn, C. A. Werlang, A. Baumschlager, S.
Brinkmann-Chen, S. L. Mayo, F. H. Arnold, ACS Synth.
Biol. 2017, 6, 326.
2
96.
[
26] R. Castillo, M. Oliva, S. Martí, V. Moliner, J. Phys.
[43] D. G. Gibson, L. Young, R. Y. Chuang, J. C. Venter,
Chem. B 2008, 112, 10012.
C. A. Hutchison III, H. O. Smith, Nat. Methods 2009, 6,
3
43.
[
27] M. Iida, K. Kitamura-Kimura, H. Maeda, S. Mineki,
Biosci., Biotechnol., Biochem. 1992, 56, 1966.
[44] U. K. Laemmli, Nature 1970, 227, 680.
[
28] W. Jiang, P. Lin, R. Yang, B. Fang, Appl. Microbiol.
Biotechnol. 2016, 100, 8425.
1
1
This article is protected by copyright. All rights reserved.

Advanced Synthesis & Catalysis
10.1002/adsc.202000536
FULL PAPER
Formate Dehydrogenase from Rhodococcus jostii
(
RjFDH) - A High-Performance Tool for NADH
Regeneration
Adv. Synth. Catal. 2020, 362, Page – Page
Alexander Boldt and Marion B. Ansorge-
Schumacher*
1
2
This article is protected by copyright. All rights reserved.