Kinetic resolution of glyceraldehyde using an aldehyde dehydrogenase from Deinococcus geothermalis DSM 11300 combined with electrochemical cofactor recycling

Article abstract of DOI:10.1016/j.molcatb.2011.09.011

Glyceraldehyde and glyceric acid are both valuable chiral starting materials. Aldehyde dehydrogenases (ALDHs) accept a broad scope of endo- and exogenous aldehydes, such as glyceraldehyde, and convert them into the corresponding carboxylic acid. Here we present cloning, overexpression and kinetic data on two ALDHs from Escherichia coli BL21 and Deinococcus geothermalis. The two ALDHs have a similar substrate scope and favor short to medium chain aldehydes, both oxidize glyceraldehyde to glyceric acid. The ALDH variant of D. geothermalis shows the higher specific activity towards glyceraldehyde and has an elevated activity optimum compared with the BL21 enzyme. The ALDH of G. geothermalis was also applied to conduct a kinetic resolution of glyceraldehyde with electrochemical cofactor recycling. .

Full text of DOI:10.1016/j.molcatb.2011.09.011

Journal of Molecular Catalysis B: Enzymatic 74 (2012) 144–150
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Kinetic resolution of glyceraldehyde using an aldehyde dehydrogenase from
Deinococcus geothermalis DSM 11300 combined with electrochemical cofactor
recycling
H. Wulfa, M. Perzborna, G. Sieversb, F. Scholzb, U.T. Bornscheuera,∗
a Dept. of Biotechnology & Enzyme Catalysis, Institute of Biochemistry, Greifswald University, Felix-Hausdorff-Str 4, D-17487 Greifswald, Germany
b Dept. of Analytical & Environmental Chemistry, Institute of Biochemistry, Greifswald University, Felix-Hausdorff-Str 4, D-17487 Greifswald, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Glyceraldehyde and glyceric acid are both valuable chiral starting materials. Aldehyde dehydrogenases
(ALDHs) accept a broad scope of endo- and exogenous aldehydes, such as glyceraldehyde, and convert
them into the corresponding carboxylic acid. Here we present cloning, overexpression and kinetic data
on two ALDHs from Escherichia coli BL21 and Deinococcus geothermalis. The two ALDHs have a similar
substrate scope and favor short to medium chain aldehydes, both oxidize glyceraldehyde to glyceric acid.
The ALDH variant of D. geothermalis shows the higher specific activity towards glyceraldehyde and has
an elevated activity optimum compared with the BL21 enzyme. The ALDH of G. geothermalis was also
applied to conduct a kinetic resolution of glyceraldehyde with electrochemical cofactor recycling.
© 2011 Elsevier B.V. All rights reserved.
Received 28 March 2011
Received in revised form
11 September 2011
Accepted 17 September 2011
Available online 23 September 2011
Keywords:
Aldehyde dehydrogenase
Glyceraldehyde
Glyceric acid
Electrochemical cofactor regeneration
Kinetic resolution
1. Introduction
binds to the enzyme. Secondly, an aldehyde is bound followed by
a nucleophilic attack by the sulfur of a catalytic cysteine residue.
Aldehyde dehydrogenases belong to an enzyme class widely
spread over every kingdom of life, catalyzing the irreversible oxi-
dation of a wide variety of endo- and exogenous aldehydes to their
corresponding carboxylic acids. The protein YdcW of Escherichia coli
has been described belonging to the superfamily of aldehyde
dehydrogenases (ALDHs), and was annotated as betaine aldehyde
dehydrogenase (EC 1.2.1.8). Kinetic studies revealed the enzyme
should be more specifically described as medium chain aldehyde
dehydrogenase (EC 1.2.1.3) [1,2]. Like many other ALDHs the crys-
tal structure of the protein reveals a homotetrameric quarternary
structure with a catalytic site and NAD+ bound to each subunit.
The subunits consist of a catalytic domain, a cofactor binding
domain, and an arm-like oligomerisation domain including the C-
terminus. The enzyme Dgeo 1120 represents an ALDH from the
radiation resistant and thermophilic organism Deinococcus geother-
malis (type strain DSM 11300) [3,4] and has been to the best of our
knowledge not yet characterized. The enzyme was annotated as
succinic-semialdehyde dehydrogenase.
The carbonyl hydride of the tetrahedral intermediate is then trans-
ferred to the C4 of the nicotinamide moiety of the NAD+. The
conformational change of the NADH then favors the activation of a
water molecule by a glutamate residue. The water molecule attacks
the substrate–enzyme thioester bond, and the corresponding car-
boxylic acid is released [5–7]. Studies on YdcW regarding conserved
residues and NADH positioning, derived from crystal structures,
suggest that a similar mechanism in bacterial ALDH can be assumed
[2].
Glyceraldehyde (GLA) and glyceric acid (GA) are possible sub-
strate and product of ALDH catalyzed reactions and both can serve
as chiral building blocks. GLA for example can serve as struc-
tural motif in 4-(dihydroxyalkyl)-␤-lactams [8,9]. GA is known
as a liver stimulant in dogs and has effects on the acetaldehyde
detoxification in rats [10,11]. GA is a building block in glycerate
phospholipid analogues, e.g. acyl glycerate derivatives bearing an
“inverse ester” function compared to natural phospholipids, at the
carboxylic group of the GA that circumvents the hydrolysis by phos-
pholipase PLA1 [12].
The reaction mechanism of human ALDHs has been identified
as an ordered sequential mechanism [5]. First, the NAD+ cofactor
The enzymes YdcW of E. coli BL21 (DE3) (ALDH-BL21) and
dgeo 1120 of D. geothermalis 11300 (ALDH-11300) were cloned
and examined during this work. Since ALDHs depend on expensive
cofactors like NAD+ or NADP+ as electron acceptors, prepara-
tive scale biotransformations using these dehydrogenases require
∗
Corresponding author. Tel.: +49 3834 86 4367; fax: +49 3834 86 794367.
E-mail address: uwe.bornscheuer@uni-greifswald.de (U.T. Bornscheuer).
1381-1177/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2011.09.011

H. Wulf et al. / Journal of Molecular Catalysis B: Enzymatic 74 (2012) 144–150
145
Canada), PCR-purification and gel extraction (Analytik Jena, Jena,
Germany) were performed according to the manufacturer proto-
cols. Standard procedures such as DNA cloning and manipulations
were performed as described by Sambrook and Russell [20].
DNA sequencing was carried out by GATC Biotech AG (Konstanz,
Germany).
The gene ydcW (aldh-BL21, GenBank accession code:
GI:313848522, locus tag B21 01412, protein: YP 604587 (NCBI))
was amplified from gDNA of E. coli BL21(DE3) by PCR using the
primer pair forward: 5ꢀ-CAT ATG CAA CAT AAG TTA CTG ATT AAC
GGA G-3ꢀ and reverse: 5ꢀ-GGA TCC TTA ATG TTT AAC CAT GAC GTG-
3ꢀ. The gene aldh-11300 (GenBank accession code: GI:94554390,
locus tag Dgeo 1120, protein: Q1IZB6 DEIGD (UniProt)) was
amplified from gDNA of D. geothermalis DSM 11300 using the
primer pair forward: 5ꢀ-CAT ATG ACC CCT GAC CCC CAG CAC CCT
GAG AAG AC-3ꢀ and reverse: 5ꢀ-GGA TCC TCA GCC CAC CAC CTG
GGC AGC CTG GCC CCC GCT TCT GGA T-3ꢀ. PCR products were
separated in a 1.2% agarose gel from which the desired fragments
were purified, cloned into the vector pCR®II-TOPO® and E. coli One
Shot® TOP10 cells were transformed with the plasmids bearing
the respective insert. YdcW (ALDH-BL21) and ALDH-11300 in
pCR®II-TOPO® were sequenced and cloned into pET-22b(+) and
pET-28b(+) via NdeI and BamHI restrictions sites (underlined).
Expression strain E. coli BL21 (DE3) was transformed with the four
different originated vectors.
Fig. 1. Reaction scheme for the biocatalysis using ALDH and electrochemical cofac-
tor recycling.
cofactor recycling systems like enzyme-coupled or substrate-
coupled systems to reoxidize the cofactors [13]. In order to employ
low concentrations of NAD+ in the biotransformation of glycer-
aldehyde to glyceric acid, an electrochemical cofactor recycling
system can be used. ALDHs in immobilized form have already been
used in aldehyde oxidations with electrochemical cofactor regen-
eration [14]. Also glyceraldehyde was oxidized this way applying
immobilized PQQ dependent oxidoreductases at modified glassy
carbon electrodes [15]. NADH can be oxidized at carbon elec-
trodes at neutral to alkaline pH but high overpotentials (above
1 V vs. Ag/AgCl) have to be applied [16,17]. The high positive
potentials may lead to unspecific side products and enzyme deacti-
vation. To circumvent the high potentials, NADH oxidation at glassy
carbon could successfully be carried out using the 2,2-azinobis(3-
ethylbenzothiazoline-6-sulfonate (ABTS) as mediator in a horse
liver alcohol dehydrogenase based biotransformation approach
[18].
2.4. Expression of ALDH-BL21 and ALDH-11300 and enzyme
purification
In this paper we describe the cloning, overexpression and prop-
erties of two ALDHs such as kinetic data and oxidative kinetic
resolution of glyceraldehyde (GLA) combined with electrochemi-
cal cofactor recycling using ALDH-11300 to yield glyceric acid (GA)
(Fig. 1).
ALDH-BL21 and ALDH-11300, both with and without N-
terminal His6-tag, were expressed in E. coli BL21 (DE3) using 5 ml
LB medium supplemented with 100 ␮g/ml ampicillin or 50 ␮g/ml
kanamycin for expression control and activity measurements. Puri-
fied enzyme was obtained by the expression of His6-tagged variants
of the two enzymes in 500 ml LB medium containing 50 ␮g/ml
kanamycin. Cells were grown at 37 ◦C to an optical density of 0.6
at 600 nm. Enzyme expression was induced with 0.4 mM IPTG and
proceeded for 5–24 h at 30 ◦C. Cells were centrifuged (4000 × g, 4 ◦C,
15 min), the cell pellet was washed twice with 20 mM Tris–HCl
buffer (pH 7.5) and disrupted in the same buffer by sonication
(2–4 min, 30% power, 50% pulse). For biotransformation experi-
ments cells were washed and resuspended with 50 mM sodium
phosphate buffer (pH 7.5) (PB) and subsequently lysed in 50 mM
sodium phosphate, 0.5 M NaCl (buffer A) using a French® Press
(Thermo Fisher, Waltham, MA). The crude cell extract was cen-
trifuged for 15 min at 16,000 × g and 4 ◦C and the supernatant
was diluted 1:10 with buffer B (50 mM sodium phosphate, 0.5 M
NaCl, 0.3 M imidazole, pH 7.5). For the purification of ALDH-BL21
and ALDH-11300 with the hexahistidine affinity tag (His6-tag) the
diluted crude extract was injected into a 5 ml HisTrapTM FF crude
column for immobilized metal ion affinity chromatography (IMAC)
using the ÄKTApuriferTM (GE-Healthcare, Munich, Germany). The
column with bound enzyme was washed with 20% buffer B and
fractions were collected while bound enzyme was eluted by apply-
ing 100% buffer B. For the storage of purified enzyme in solution
or lyophilization, the purified enzyme was subjected to gel filtra-
tion for the removal of imidazole and high salt concentration using
ÄKTApuriferTM equipped with Sephadex G-25 Superfine column
(26 × 110 mm).
2. Experimental
2.1. Chemicals
All chemicals were purchased from Sigma–Aldrich (St. Louis,
USA) and Carl-Roth (Karlsruhe, Germany). Polymerases were
obtained from Roboklon (Berlin, Germany), Restrictions enzymes
and DNA ligases were obtained from Fermentas (Burlington,
Canada).
2.2. Bacterial strains, plasmids and growth conditions
E. coli BL21 (DE3) was purchased from Novagen (Darmstadt,
Germany) and used as expression strain and for the isolation of
genomic DNA. D. geothermalis DSM 11300 was obtained from DSMZ
(Braunschweig, Germany). E. coli One Shot® TOP10 cells (F− mcrA
ꢀ(mrr-hsdRMS-mcrBC) ꢁ80lacZꢀM15 ꢀlacX74 recA1 araD139
ꢀ(araleu)7697 galU galK rpsL (StrR) endA1 nupG) and vector pCR®II-
TOPO® for subcloning were purchased from Invitrogen (Karlsruhe,
Germany). E. coli strains were routinely cultured in lysogeny broth
[19] without glucose at 37 ◦C, when necessary supplemented with
ampicillin (100 ␮g/ml) or kanamycin (50 ␮g/ml). D. geothermalis
was cultured in LB medium at 45 ◦C for four days. The expression
plasmids pET-22b(+) and pET-28b(+) were obtained from Novagen
(Darmstadt, Germany), the latter was used for enzyme expression
with N-terminal His6-tag.
2.5. Determination of protein concentration and SDS-PAGE
2.3. Cloning of ydcW and aldh-11300
Polyacrylamide gel electrophoresis was carried out using 4%
stacking gel and 12% resolving gel according to the method
described by Laemmli [21]. Roti-Mark® STANDARD (Roth, Karl-
sruhe, MW: 14–200 kDa) was used as protein standard. The
Genomic DNA (gDNA) from E. coli BL21 (DE3) and D. geother-
malis was isolated using the innuPREP Bacteria DNA Kit (Analytik
Jena, Jena, Germany). Plasmid isolations (Fermentas, Burlington,

146
H. Wulf et al. / Journal of Molecular Catalysis B: Enzymatic 74 (2012) 144–150
protein concentration was determined either with the BCA-assay
(Uptima, Montluc¸ on, France) or with Coomassie Brilliant-Blue
using Roti®-Nanoquant (Carl-Roth, Karlsruhe, Germany) according
to manufacturer protocols using BSA as standard.
2.6. Enzyme assays
The aldehyde dehydrogenase activity was determined spec-
trophotometrically at 30 ◦C by monitoring the rate of reduction
of NAD+ at 340 nm. The standard reaction mixture (1 ml) con-
tained 80 mM Tris–HCl buffer (pH 9.0), 0.35 mM NAD+, 2.5 mM
d,l-glyceraldehyde and an appropriate amount of the enzyme. The
reaction was started by adding d,l-glyceraldehyde to the mixture.
One unit of aldehyde dehydrogenase activity was defined as the
amount of enzyme that catalyzes the formation of 1 ␮mol NADH
per minute.
In order to determine the apparent KM and vmax values for
different aldehydes the described activity test was carried out
with varying substrate concentrations at a fixed concentration of
0.35 mM NAD+. Initial velocities were recorded during the first 11 s
of catalytic turnover. Apparent KM and vmax-values were obtained
by linear regression fitting of a Hanes plot. These values were
used for non-linear regression, which effectively fits the data
directly to the best hyperbola according to the method of Wilkin-
son [22] and Duggleby [23] using the program HYPER (J.S. Easterby,
http://www.liv.ac.uk/∼jse/software.html).
Fig. 2. Scheme of the biotransformation reactor vessel. The carbon foam working
electrode is cylindrically surrounded by a platinum net shaped auxiliary electrode.
The Ag/AgCl reference electrode is contacted via a KCl solution saturated agar tube.
The electrodes are located inside a stirred 100 ml beaker.
equipped with a Chiralpak® QN-AX precolumn (4 × 10 mm) and
column (4.6 × 150 mm, both Chiral Technologies Europe, Illkirch,
France) at 23 ◦C with methanol/acetic acid/ammonia acetate
98/2/0.5 (v/v/w) as eluent at 1 ml/min flow rate and evaporative
light scattering detection. The retention time of l-GA was 5.9 min
and the d-GA retention time was 7.1 min.
The effect of temperature on the enzyme activity was deter-
mined by incubating the standard reaction mixture at different
temperatures ranging from 12 to 59 ◦C for 10 min at pH 9.0. For
determining the effect of the pH towards enzyme activity, purified
enzyme was incubated at 30 ◦C for 5 min in the following buffers at
100 mM: sodium acetate (pH 4.0–5.5), Bis–Tris–HCl (pH 5.5–7.0),
sodium phosphate (pH 6.0–8.0) and Tris–HCl (pH 6.5–9.7) Activity
was measured as described above.
3. Results and discussion
3.1. Amplification, cloning and protein expression of ALDH-BL21
and ALDH-11300
2.7. Electrochemical measurements and biotransformations
Electrochemical measurements were carried out with an Auto-
lab PSTAT 10 (Eco Chemie, Utrecht, Netherlands). All potentials
refer to the Ag/AgCl electrode (3 mol l−1 KCl, E = 0.208 V vs. SHE). In
cyclic voltammetry the second scan was used. The sweep rate was
5 mV s−1 and the potential range between 0 and 1 V vs. Ag/AgCl.
The buffer system consisted of 100 mM sodium phosphate pH 8.0
(buffer PB).
Biotransformation experiments were carried out using the
electrochemical cofactor recycling system. A foam glassy-carbon
electrode (30 PPI RVC Duocel®, ERG Materials and Aerospace, Oak-
land, USA) was used as working electrode. A Pt-net surrounded the
working electrode and served as auxiliary electrode. The above-
mentioned Ag/AgCl reference electrode contacted the reactor via a
KCl saturated agar tube (Fig. 2).
The genes encoding proteins ALDH-BL21 and ALDH-11300 could
be successfully amplified and cloned into pET-22b(+) without His6-
tag and pET-28b(+) bearing the genes with N-terminal His6-tag in
order to enhance and simplify the protein purification procedure.
The cloned DNA sequence of the ydcW differed in one nucleotide
(A994G) to the database entry B21 01412. This difference leads
to the mutation S332G at protein level in ALDH-BL21. Both genes
could be overexpressed in the host strain E. coli BL21 (DE3). The
His6-tag variants of ALDH-BL21 and ALDH-11300 were purified to
apparent homogeneity by IMAC (as determined by SDS-PAGE, data
not shown) at yields of 39% (ALDH-BL21) and 66% (ALDH-11300)
of the total activity.
The buffer reaction solution consisted of 0.5 mM ABTS and 2 mM
NADH in 20 ml of 100 mM sodium phosphate (pH 8.0). Experi-
ments were performed at room temperature and the potential of
the working electrode was hold at 0.7 V vs. Ag/AgCl. As biocatalyst a
preparation of 5–10 U of purified and lyophilized ALDH-11300 was
used. Racemic d,l-glyceraldehyde was added up to 40–50 mM.
3.2. Temperature- and pH-optimum
Since both enzymes could be expressed in active form with and
without His6-tag, subsequent experiments were performed with
the His6-tagged enzyme variants using glyceraldehyde as substrate.
The temperature optimum of ALDH-BL21 from E. coli was between
30 and 35 ◦C (Fig. 3). The maximum activity for the ALDH-11300
from the mesothermophilic organism D. geothermalis was deter-
mined to be more than 10 ◦C higher, between 43 and 47 ◦C (Fig. 3),
making this enzyme more useful for applications at elevated tem-
peratures around 45 ◦C. The temperature range corresponds to the
optimum growth temperature of D. geothermalis of 45–50 ◦C [3]
since chaperones and compatible solutes may additionally stabi-
lize the enzyme at temperatures of over 45 ◦C in its physiological
environment.
2.8. Product determination by HPLC
Concentrations of glyceraldehyde (GLA) and glyceric acid (GA)
were determined using a Shimadzu LC 10 HPLC equipped with a
Eurokat-H precolumn (30 × 8 mm) and column (300 × 8 mm, both
Knauer, Berlin, Germany) at 80 ◦C with 0.01 N sulfuric acid at
0.8 ml/min flow rate and refractive index detection. Chiral analysis
of glyceric acid was conducted using a Merck Hitachi LaChrom HPLC

H. Wulf et al. / Journal of Molecular Catalysis B: Enzymatic 74 (2012) 144–150
147
The biocatalytic reduction of GLA and GA with NADH was inves-
tigated too. However no activity was observed with both enzymes
towards GLA or GA. Eriksson et al. [10] showed that treatment
of rats with GA enhances ethanol and acetaldehyde clearance.
The authors suggest that GA enhances acetaldehyde detoxification
by enhanced cofactor regeneration through ALDH catalyzed sub-
strate reductions. Since the BRENDA-Database [24] solely lists one
example of a ALDH converting an acid to an aldehyde, namely 3-
hydroxypropionic acid to 3-hydroxypropionaldehyde with a poor
activity of 4% of the oxidation capacity [25] and data of this work
show no action of ALDH against glyceric acid this hypothesis seems
to be unlikely. Because of the high difference in the redox poten-
tial, acids need to be phosphorylated prior to be metabolized and
reduced by aldehyde dehydrogenases at reasonable reaction veloc-
ities.
Both aldehyde dehydrogenases showed no activity towards 1-
propanol, 1,2-propandiol and glycerol as substrates. These findings
indicate the chemoselectivity of these enzymes towards the oxida-
tion of a terminal carbonyl group.
The activity of ALDH-11300 against several substrates was
determined, and the kinetic parameters of the Michaelis–Menten
equation were determined for six substrates at various concen-
trations (Table 1). The highest turnover rates were found for the
oxidation of isobutyraldehyde. The slowest turnover was observed
with the aromatic substrate phenylacetaldehyde. However with
3.76 s−1 mM−1 the kcat/KM value was still better than that of glycer-
aldehyde, which had the second best turnover but also the second
highest KM. The best kcat/KM was determined for butyraldehyde.
Since no constants could be determined for valeraldehyde (C5) and
caproaldehyde (C6), butyraldehyde represents the best kinetic data
for the longest unbranched carbon chain. Valeraldehyde showed
a significant inactivation above 150 ␮M (Fig. 5). The plots for
caproaldehyde and valeraldehyde showed a non Michaelis–Menten
like, sigmoidal slope.
Kinetic data were also determined for the ALDH-BL21. Similar
to the findings of Gruez et al. our data show an increase in the
turnover as the chain length of the substrate increases. In contrast
to the findings for the K12-ALDH [2], the best substrate according
to vmax values is the C4 substrate butyraldehyde. The best kcat/KM
value has isobutyraldehyde before caproaldehyde (Table 2).
The enzyme activity after storing at 0 ◦C and at room temper-
ature (RT) was determined during more than 40 days. ALDH-BL21
was more stable than ALD-11300 during both storing conditions.
Activity of ALDH-BL21 after storing at 0 ◦C for 20 days showed 95%
Fig. 3. Temperature profiles of ALDH-BL21 (ꢀ) and ALDH-11300 (᭹).
Both enzymes showed the highest activity at pH 8.5 in Tris–HCl.
At pH 5.5 the enzymes showed residual activities of 10% and a
complete loss of activity took place at pH 4.0 (Fig. 4).
3.3. Activity and steady-state kinetics
For ALDH-11300 from D. geothermalis, the oxidation of glycer-
aldehyde and other substrates was measured for the first time in
this work. For the homolog ALDH enzyme from E. coli K12 activ-
ity against glyceraldehyde was previously described by Gruez et al.
[2]. Activity could also be confirmed here for ALDH-BL21 from E. coli
BL21, which differs in one amino acid (V197I) from the K12 vari-
ant. For this enzyme a specific activity of 0.6 U/mg was determined
after IMAC purification. The newly characterized ALDH-11300 with
a specific activity of 3.5 U/mg is more active against glyceralde-
hyde compared to ALDH-BL21. The purified ALDH-BL21 had only
2% activity against GLA with NADP+ compared to NAD+ as cofactor.
ALDH-11300 showed 15% activity against GLA using NADP+ instead
of NAD+. Gruetz et al. found 6% activity using NADP+ in GLA oxida-
tion, whereas the vmax/KM ratio for NADP+ is still 25% of the NAD+
value [2].
Fig. 4. Effect of pH on the enzyme activity in the range from pH 4.0 to 9.7 with for different buffers: sodium acetate (ꢁ) Bis–Tris–HCl (᭹), sodium phosphate (ꢂ) und Tris–HCl
(ꢀ). (A) Activity of ALDH-BL21. (B) Activity of ALDH-11300.

148
H. Wulf et al. / Journal of Molecular Catalysis B: Enzymatic 74 (2012) 144–150
Table 1
Kinetic parameters of ALDH-11300 for six substrates, ordered according to increasing kcat-values. The error estimations represent a confidence interval of 95%.
Substrate
vmax [U mg−1
]
KM [mM]
kcat [1 s−1
]
kcat/KM [1 s−1 mM−1
]
Phenylacetaldehyde
Acetaldehyde
Propionaldehyde
Butyraldehyde
Glyceraldehyde
Isobutyraldehyde
1.74 0.24
1.97 0.14
2.76 0.16
2.9 0.4
3.05 0.35
3.22 0.15
0.13 0.07
0.21 0.06
0.14 0.03
0.07 0.04
2.52 0.95
2.67 0.4
1.56 0.22
1.78 0.13
2.49 0.14
2.61 0.36
2.75 0.32
2.9 0.13
12.4 3.35
8.59 2.06
17.39 4.65
37.74 9.9
1.09 0.33
1.09 0.33
Fig. 5. The activity of ALDH-11300 towards several substrates at different concentrations.
Table 2
Kinetic parameters of ALDH-BL21 for eight substrates, ordered according to increasing kcat-values. The error estimations represent a confidence interval of 95%.
Substrate
vmax [U mg−1
]
KM [mM]
kcat [1 s−1
]
kcat/KM [1 s−1 mM−1
]
Phenylacetaldehyde
Acetaldehyde
Isobutyraldehyde
Glyceraldehyde
Propionaldehyde
Valeraldehyde
0.08 0.03
0.42 0.01
0.55 0.03
0.95 0.1
1.08 0.04
1.24 0.11
1.41 0.11
1.64 0.25
4.87 3.98
0.56 0.06
0.02 0.004
8.44 2.66
0.05 0.01
0.07 0.02
0.05 0.02
0.08 0.04
0.07 0.03
0.38 0.01
0.49 0.03
0.86 0.09
0.97 0.04
1.12 0.1
1.27 0.1
1.48 0.23
0.01 0.01
0.67 0.14
26.47 5.94
0.1 0.03
21.05 4.21
16.28 4.73
25.21 6.04
18.98 5.8
Caproaldehyde
Butyraldehyde
of the initial value, whereas for ALD-11300 activity was reduced by
40% (0 ◦C) to 80% (room temperature) (Fig. 6). Storage of purified
enzymes at −20 ◦C resulted in complete loss of activity, but freeze
dried ALDH-11300 retained 80% of its activity towards GLA.
3.4. Biotransformation and determination of reaction products
Glyceraldehyde should be oxidized to GA in an ALDH cat-
alyzed kinetic resolution. For these biotransformation experiments
100 mM sodium phosphate pH 8.0 was used instead of Tris-buffer,
because tris (tris(hydroxymethyl)-aminomethane) is not appli-
cable in reactions with acyl-enzyme intermediates since it may
act as nucleophile itself [26]. Cyclic voltammetry measurements
show the two-electron transfer system of ABTS. The first redox
reaction is reversible and important for the NADH reoxidation.
The second (more positive) redox system is irreversible and must
be avoided as otherwise ABTS is eliminated from the equilib-
rium. The cyclic voltammogram of NADH shows an irreversible
system with a current peak at 580 mV vs. Ag/AgCl (Fig. 7, plot
d). A solution of 4 mM NADH and 1 mM ABTS shows an irre-
versible oxidation peak at 500 mV since ABTS•− is immediately
Fig. 6. Activity of the enzymes ALDH-BL21 und ALDH-11300 after storing at room
temperature and at 0 ◦C for 43 days.

H. Wulf et al. / Journal of Molecular Catalysis B: Enzymatic 74 (2012) 144–150
149
Fig. 7. The plot a displays the CV of ABTS and NAD, b displays ABTS, NAD and ALDH-11300, c displays ABTS, NAD, ALDH-11300 and GLA, d displays NAD (4 mM) and e displays
ABTS (1 mM). All CVs were measured in PB (100 mM, pH 8.0).
oxidized by NADH (Fig. 7 and [18]). The addition of enzyme and
GA shifts the current peak of the recycling system to higher over-
potentials, which may be due to blocking of catalytic active sites
at the electrode. The increase of the overpotential of the recy-
cling system requires a higher potential as it would be needed
for normal ABTS mediated NADH reoxidation. The potential for
the biotransformation should be at least 650 mV vs. Ag/AgCl
(Fig. 7) thus for biotransformation experiments 700 mV where
applied.
During the biotransformation experiments, the cofactor NADH
was indirectly recycled at constant potential of 700 mV at a glassy
carbon foam electrode. The foam electrode was used to generate
a greater contact surface, where ABTS and subsequently NADH
could be reoxidized (Fig. 1). The reaction vessel contained 20 ml
buffer solution with dissolved enzyme and substrates. Since NADH
was applied initially, the reaction started only after the current
was switched on. Samples were taken during the biocatalysis and
analyzed via ion-exclusion and chiral HPLC, and the transferred
electric charge was monitored (Fig. 8). In biotransformation A the
concentration of GA increased by 16.4 mM while the concentra-
tion of GLA decreased by 14.5 mM (Fig. 8). Biotransformation B
does not show such a high accumulation of D-GA, it increases by
7.9 mM while GLA drops in 12.7 mM (Fig. 8). The small experi-
mental discrepancies between drop of educts and gain of products
are most probably experimental errors. The measured transferred
charge (qm) apparently fits the theoretically transferred charge
(qt) best in the beginning of the biocatalysis (Fig. 8). The differ-
ence between qm and qt gains significance with progression of
the reaction. A subsequent oxidation of glyceric acid at the anode
is not possible, as proven in separate CV experiments (data not
shown), hence further oxidation of GA is not an explanation for
the differing qt and qm. A possible explanation could be the big
surface and the large number of pores and cavities of the car-
bon foam electrode. GLA and GA may diffuse into the pores, and
Fig. 8. Plots of biotransformation experiments A and B. The reactions differ solely
hence evade the chromatographic quantitation. Washing steps of
the working electrode after the biotransformation revealed that GA
and GLA where retained by the electrode and perhaps remained
in the cavities. This can lead to a systematic error in the deter-
mination of the extent of the biotransformation. The structure of
in the prolonged reaction time and greater number of sample points in biotransfor-
mation B. qm and qt refer to the left axis, concentrations of GLA, GA and %ee to the
two left axes.

150
H. Wulf et al. / Journal of Molecular Catalysis B: Enzymatic 74 (2012) 144–150
the porous working electrode will certainly lead to lower poten-
tials inside the pores, which will decrease the electrolysis efficiency
[27]. This feature must receive further attention when an upscal-
ing of the technique is intended. It might also be possible that
with reaction time the enzyme becomes oxidized at the electrode
as described by Manjón et al. for glucose dehydrogenase [28],
which can be an explanation for the uncompleted conversion of
the GLA.
The enantiomeric excess of the resulting D-GA decreased
during both biotransformations although no direct oxi-
dation of GLA at the electrode could be observed at the
applied potential. At the first measurement, the enan-
tiomeric excess of both reactions was still 98.5%ee and
100%ee but ended up in 91.8 and 87.9%ee, respectively
(Fig. 8).
References
[1] N.A. Sophos, V. Vasiliou, Chem. Biol. Interact. 143–144 (2003) 5–22.
[2] A. Gruez, V. Roig-Zamboni, S. Grisel, A. Salomoni, C. Valencia, V. Campanacci,
M. Tegoni, C. Cambillau, J. Mol. Biol. 343 (2004) 29–41.
[3] A.C. Ferreira, M.F. Nobre, F.A. Rainey, M.T. Silva, R. Wait, J. Burghardt, A.P. Chung,
M.S. Da Costa, Int. J. Syst. Bacteriol. 47 (1997) 939–947.
[4] K.S. Makarova, M.V. Omelchenko, E.K. Gaidamakova, V.Y. Matrosova, A.
Vasilenko, M. Zhai, A. Lapidus, A. Copeland, E. Kim, M. Land, K. Mavromatis, S.
Pitluck, P.M. Richardson, C. Detter, T. Brettin, E. Saunders, B. Lai, B. Ravel, K.M.
Kemner, Y.I. Wolf, A. Sorokin, A.V. Gerasimova, M.S. Gelfand, J.K. Fredrickson,
E.V. Koonin, M.J. Daly, PLoS ONE 2 (2007) e955.
[5] R.I. Feldman, H. Weiner, J. Biol. Chem. 247 (1972) 267–272.
[6] P.K. Hammen, A. Allali-Hassani, K. Hallenga, T.D. Hurley, H. Weiner, Biochem-
istry 41 (2002) 7156–7168.
[7] S.J. Perez-Miller, T.D. Hurley, Biochemistry 42 (2003) 7100–7109.
[8] D.R. Wagle, C. Garai, J. Chiang, M.G. Monteleone, B.E. Kurys, T.W. Strohmeyer,
V.R. Hegde, M.S. Manhas, A.K. Bose, J. Org. Chem. 53 (1988) 4227–4236.
[9] P. Areces, E. Carrasco, M.E. Light, M. Santos, J. Plumet, Synlett 2007 (2007)
3180–3182.
[10] C.J.P. Eriksson, T.P.S. Saarenmaa, I.L. Bykov, P.U. Heino, Metabolism 56 (2007)
895–898.
4. Conclusion
[11] S.S. Handa, A. Sharma, K.K. Chakraborti, Fitoterapia 57 (1986) 307.
[12] R. Rosseto, C.M. Tcacenco, R. Ranganathan, J. Hajdu, Tetrahedron Lett. 49 (2008)
3500–3503.
[13] W. Kroutil, H. Mang, K. Edegger, K. Faber, Curr. Opin. Chem. Biol. 8 (2004)
120–126.
In this work we showed the cloning, purification, character-
ization and biocatalysis with two ALDH from E. coli BL21 and
D. geothermalis. Both enzymes could be overexpressed and puri-
fied with IMAC yielding 39–66% active enzyme. The ALDH-11300
with 3.5 U/mg had a better specific activity against glyceralde-
hyde then ALDH-BL21 with 0.6 U/mg. The pH profile was similar
for both enzymes and showed the highest activity at a pH of
8.5. The temperature optimum of activity at 45–50 ◦C was 10 ◦C
higher with ALDH-11300 then with ALDH-BL21. The best sub-
strate according to kcat/KM for ALDH-11300 and ALDH-BL21 were
butyraldehyde and isobutyraldehyde, respectively. Biotransfor-
mations of GLA to GA with ALDH-11300 and electrochemical
cofactor recycling led to concentrations up to 1.8 g/l d-GA with
88%ee.
[14] C.M. Moore, N.L. Akers, A.D. Hill, Z.C. Johnson, S.D. Minteer, Biomacromolecules
5 (2004) 1241–1247.
[15] R.L. Arechederra, S.D. Minteer, Electrochim. Acta 55 (2010) 7679–7682.
[16] S.B. Saidman, J.B. Bessone, Electrochim. Acta 45 (2000) 3151–3156.
[17] C. Kohlmann, W. Märkle, S. Lütz, J. Mol. Catal. B: Enzym. 51 (2008) 57–72.
[18] I. Schröder, E. Steckhan, A. Liese, J. Electroanal. Chem. 541 (2003) 109–115.
[19] G. Bertani, J. Bacteriol. 62 (1951) 293–300.
[20] J. Sambrook, D.W. Russell, Molecular Cloning: A Laboratory Manual, 3rd ed.,
Cold Spring Harbor Laboratory Press, New York, 2001.
[21] U.K. Laemmli, Nature 227 (1970) 680–685.
[22] G.N. Wilkinson, Biochem. J. 80 (1961) 324–332.
[23] R.G. Duggleby, Anal. Biochem. 110 (1981) 9–18.
[24] M. Scheer, A. Grote, A. Chang, I. Schomburg, C. Munaretto, M. Rother, C. Söhngen,
M. Stelzer, J. Thiele, D. Schomburg, Nucl. Acids Res. 39 (2010) 1–7.
[25] J.-E. Jo, S. Mohan Raj, C. Rathnasingh, E. Selvakumar, W.-C. Jung, S. Park, Appl.
Microbiol. Biotechnol. 81 (2008) 51–60.
[26] V. Kasche, R. Zöllner, Hoppe Seylers Z. Physiol. Chem. 363 (1982) 531–534.
[27] R.E. Sioda, Electrochim. Acta 16 (1971) 1569–1576.
Acknowledgements
[28] A. Manjón, J.M. Obón, P. Casanova, V.M. Fernández, J.L. Ilborra, Biotechnol. Lett.
24 (2002) 1227–1232.
The authors thank the “Fachagentur für Nachwachsende
Rohstoffe” (AZ06NR073, 22015906) for financial support.