Appl Microbiol Biotechnol (2013) 97:5815–5824

DOI 10.1007/s00253-012-4474-5

BIOTECHNOLOGICALLY RELEVANT ENZYMES AND PROTEINS

Novel CAD-like enzymes from Escherichia coli K-12

as additional tools in chemical production
André Pick & Broder Rühmann & Jochen Schmid &
Volker Sieber

Received: 7 August 2012 / Revised: 21 September 2012 / Accepted: 24 September 2012 / Published online: 24 October 2012
# The Author(s) 2012. This article is published with open access at Springerlink.com

Abstract In analyzing the reductive power of Escherichia Keyword Alcohol dehydrogenase . Escherichia coli .
coli K-12 for metabolic engineering approaches, we identi- Metabolic engineering . YqhD . YahK . YjgB . Bulk chemicals
fied YahK and YjgB, two medium-chain dehydrogenases/
reductases subgrouped to the cinnamyl alcohol dehydroge-
nase family, as being important. Identification was achieved Introduction
using a stepwise purification protocol starting with crude
extract. For exact characterization, the genes were cloned The production of fuels and chemicals from biomass and
into pET28a vector and expressed with N-terminal His tag. such “greening” the chemical industry is an important
Substrate specificity studies revealed that a large variety of issue of today. Within this context, the microbial produc-
aldehydes but no ketones are converted by both enzymes. tion of industrially relevant alcohols has been the subject
YahK and and YjgB strongly preferred NADPH as cofactor. of many studies in recent years. Especially methods of
The structure of YjgB was modeled using YahK as template metabolic engineering and synthetic pathway design have
for a comparison of the active center giving a first insight to been applied to enable different microorganisms to pro-
the different substrate preferences. The enzyme activity for duce molecules like ethanol, n-butanol (Berezina et al.
YahK, YjgB, and YqhD was determined on the basis of the 2010), isobutanol (Liao et al. 2010), different propane
temperature. YahK showed a constant increase in activity diols (Emptage et al. 2003; Nakamura and Whited
until 60 °C, whereas YjgB was most active between 37 and 2003; Berríos-Rivera et al. 2003) and butane diols (Yim
50 °C. YqhD achieved the highest activity at 50 °C. Com- et al. 2011; Carothers et al. 2009; Nielsen et al. 2010), as
paring YjgB and Yahk referring to the catalytic efficiency, well as aromatic alcohols like furfuryl alcohol (Heer et
YjgB achieved for almost all substrates higher rates (butyr- al. 2009). The final step of all these syntheses is the
aldehyde 221 s−1 mM−1, benzaldehyde 1,305 s−1 mM−1). reduction of an aldehyde to an alcohol. One enzyme
Exceptions are the two substrates glyceraldehydes (no ac- stands out in its utilization for this step: the Escherichia
tivity for YjgB) and isobutyraldehyde (YjgB 0.26 s−1 mM−1) coli alcohol dehydrogenase YqhD (Jarboe 2010; Tang et
which are more efficiently converted by YahK (glyceralde- al. 2009; Liao et al. 2010). Accordingly, this enzyme was
hyde 2.8 s−1 mM−1, isobutyraldehyde 14.6 s−1 mM−1). YahK discovered within a project directed by DuPont for the
and even more so YjgB are good candidates for the reduc- production of 1,3-propanediol using an engineered E. coli
tion of aldehydes in metabolic engineering approaches and strain (Emptage et al. 2003; Nakamura and Whited
could replace the currently used YqhD. 2003). YqhD showed to be the better candidate for the
reduction of 3-hydroxypropionaldehyde compared to the
designated DhaT from Klebsiella pneumoniae (Wang et
A. Pick : B. Rühmann : J. Schmid : V. Sieber (*) al. 2005; Skraly et al. 1998). In our case, investigations of the
Wissenschaftszentrum Straubing, Lehrstuhl für Chemie Biogener reductive power of E. coli revealed additional activities that
Rohstoffe, Technische Universität München,
can be important for the microbial production of alcohols and
Schulgasse 16,
94315 Straubing, Germany that are especially suited for the reduction of multifunctional
e-mail: [email protected] aldehyde compounds with partially even higher activity than
5816 Appl Microbiol Biotechnol (2013) 97:5815–5824

YqhD. Here, we report the identification and characterization Protein identification

of two such enzymes from E. coli, the zinc-dependent alcohol
dehydrogenases YahK and YjgB. All purification steps of alcohol dehydrogenase activities
(ADHs) from E. coli W3110 were performed using an
ÄKTA UPC-900 FPLC-system (GE Healthcare, Munich,
Materials and methods Germany) at room temperature. All buffers were filtered
with 0.2-μm regenerated cellulose membranes. Fractions
Reagents were stored at −20 °C.

Restriction enzymes, alkaline phosphatase, phusionTM Crude extract

polymerase, and T4 ligase are from New England Biol-
abs (Frankfurt, Germany). Taq polymerase was obtained E. coli W3110 was cultivated in Luria–Bertoni (LB) medi-
from Rapidozym (Berlin, Germany). Oligonucleotides were um at 150 rpm and 37 °C. The cells were harvested by
from biomers.net (Ulm, Germany). DNase was obtained centrifugation at 4,580×g for 25 min at 4 °C. Pellets were
from Serva (Heidelberg, Germany). All chemicals were of stored at −20 °C. Ten grams of frozen cells were dissolved
analytical grade or higher quality and purchased from in 40 ml of 20 mM Tris-HCl (pH 9), 20 μL of 2 M MgCl2
Sigma-Aldrich, Merck, or Carl Roth. All columns used for and DNase in a final concentration of 50 μg/ml. Cell lysate
protein purification were from GE Healthcare (Munich, was prepared with a cell disruptor (IUL constant systems) at
Germany). 1.33 kbar. Removal of cell debris was performed by centri-
fugation (4,580×g, 15 min, 4 °C). Purification of cell lysate
Strains and plasmids was performed by denaturation for 15 min at 60 °C in a
water bath. Finally, the suspension was centrifuged at
The following strains were used during this work: E. coli 21,100×g for 15 min at 4 °C and filtrated via 0.45-μm
K-12 W3110, KeioCollection BW25113 (Baba et al. cellulose acetate membranes.
2006), E. coli XL1 Blue, and E. coli BL21(DE3). For
cloning of the genes yjgB (GenBankTM U14003.1), yahK Anion exchange chromatography
(GenBank T M U00096.2), and yqhD (Genbank T M
GQ478251.1), genomic DNA of E. coli K12 W3110 The cell lysate was applied to a HiTrap Sepharose Q XL 1 ml
was used as PCR template. For cloning of yjgB, the column equilibrated with 20 mM Tris-HCl buffer (pH 9.0,
primers F-NdeI-yjgB-E.c.- CGACAGCATATGTCGAT 1 ml/min). After washing with 20 bed volume of equilibra-
GATAAAAAGCTATGCCGC and R-XhoI-yjgB-E.c.- tion buffer, elution was executed with eluent B (20 mM Tris-
GACGATCTCGAGTCAAAAATCGGCTTTCAACAC HCl, 1 M NaCl and pH 9.0), starting with 10 % for 20 min,
CACGC, for yahK the primers F-NheI-yahK-E.c- followed by a linear gradient up to 60 % within 90 min, and
GACAGGCTAGCATGAAGATCAAAGCTGTTGGTGC column wash with 100 % for 10 min. The eluted fraction
and R-XhoI-yahK-E.c.-GACGATCTCGAGTCAGTCTGT was monitored at 280 nm and sampled (every 1.5 ml) into a
TAGTGTGCGATTATCG, and for yqhD the primers F- 96-deep-well microtiter plate for further analysis.
N h e I - y q h D - E . c . - G A C A G G C TA G C AT G G C G A A
CAACTTTAATCTGCACAC and R-XhoI-yqhD-E.c.- Hydrophobic interaction chromatography
G A C G A C T C G A G T TA G C G G G C G G C T T C G TATA
TACGG were used. PCR products were digested with NdeI or Different fractions from the first purification, with the de-
NheI and XhoI and cloned into pET28a(+) (Novagen), cut sired activity, were pooled. Ammonium sulfate was added
with the same enzymes, creating the plasmids pET28a-NH- slowly up to a saturation of 15 % and stirred on ice for 1 h.
yjgB-E.c, pET28a-NH-yahK-E.c., and pET28a-NH-yqhD- After filtration by a 0.45-μm cellulose acetate membrane,
E.c. Multiplication of the plasmids was performed by E. coli the solution was loaded in 1 ml/min onto a HiTrap Phenyl
XL1 Blue (Stratagene) in Luria–Bertoni medium containing HP column (1-ml column). The linear gradient was applied
30 μg/ml kanamycin. The E. coli strain BL21(DE3) (Novagen) after 20 min washing with sodium phosphate buffer (pH 7.0)
was used for expression. and 2 M NaCl. After 30 min, the washing was stopped with
100 % sodium phosphate buffer (pH 7.0).
Isolation of genomic DNA from E. coli K-12 W3110
Protein digestion
The genomic DNA from E. coli K-12 W3110 was isolated
from cells of an overnight culture using the protocol of Chen An aliquot of the fraction was treated with a finale concen-
and Kuo (1993). tration of 10 mM dithiothreitol for 15 min in a water bath at
Appl Microbiol Biotechnol (2013) 97:5815–5824 5817

60 °C. Free cysteine residues were then alkylated with a Enzyme expression and purification
finale concentration of 60 mM iodoacetamide in the dark for
15 min at room temperature. Reduced and alkylated frac- Protein expression is exemplarily described for one enzyme
tions were either loaded on the SDS gel or directly digested and was performed for other proteins by the same proce-
by trypsin (1.6 μg/ml) overnight at 37 °C. In-gel digestion dure. E. coli BL21(DE3) containing the plasmid of interest
was preformed with the protocol from OMX system (Pro- was grown in 50 ml autoinduction media for efficient pro-
teome X Solution, Germany). tein expression (Studier 2005). In the case of YjgB, YahK,
and YqhD, additional ZnCl2 was added to the media with a
HPLC analysis of digested proteins final concentration of 0.1 mM. The preculture was incubated
in 4 ml of LB medium with 100 μg/ml kanamycin at 37 °C
The HPLC system (Ultimate 3000RS Dionex, Germany) overnight on a rotary shaker (180 rpm). Expression culture
used consisted of a degasser (SRD 3400), a pump module was inoculated with a 1:100 dilution of overnight culture.
(HPG 3400RS), an auto sampler (WPS 3000TRS), a column Incubation was performed foremost 3 h at 37 °C followed
compartment (TCC 3000RS), a diode array detector (DAD by incubation for 21 h at 16 °C. Cells were harvested by
3000RS), and an ESI-ion-trap (HCT Bruker, Germany). centrifugation and resuspended in 50 mM sodium phosphate
Data were collected and analyzed with Bruker HyStar and buffer (pH 8.0, 20 mM imidazol, 500 mM NaCl, and 10 %
Data Analysis software. The Gravity column (100 mm glycerol). Crude extracts were prepared by use of a cell
length, 2 mm i.d., and 1.8 μm particle size, Macherey- disrupter (IUL Instruments) and subsequent addition of
Nagel, Germany) was tempered at 40 °C. Flow rate was MgCl2 to a final concentration of 2.5 mM in combination
set to 0.2 ml/min and the gradient was programmed as with DNase (1 μg/ml) and a following incubation for 20 min
follows: mobile phase B started at 10 % for 5 min, increas- at room temperature for DNA degradation. The insoluble
ing to 75 % over 45 min, and returning in 0.2 min to starting fraction of the lysate was removed by centrifugation at
conditions for 4.8 min (mobile phase A, 0.1 % formic acid 20,000 rpm for 40 min at 4 °C. The supernatant was
in water; mobile phase B, 0.1 % formic acid in acetonitrile). filtered through a 0.45-μm syringe filter and applied to an
Temperature of the auto sampler was 10 °C and injection affinity resin column, 5 ml HisTrapTM FF, equilibrated with
volume was set to 10 μl. the resuspension buffer using the ÄKTA UPC-900 FPLC-
system. The enzyme was washed with 20 ml of resuspen-
ESI-ion trap parameter sion buffer and eluted with 50 mM sodium phosphate
buffer (pH 8.0, 500 mM imidazol, 500 mM NaCl, and
The ion trap was operated in the ultra standard enhanced 10 % glycerol). Aliquots of each eluted fraction were
mode (8,100m/z/s) from m/z 300 to 1,500 (m/z 100 to 2,300 subjected to 12 % SDS-PAGE. The fractions containing
for MS/MS). The ICC target was set to 200,000 with a the eluted protein were pooled and the protein was desalted
maximum accumulation time of 150 ms and five averages using a HiPrepTM 26/10 Desalting column which prelimi-
(three for MS/MS). The ion source parameters were set as nary equilibrated with 50 mM Tris-HCl, pH 8.0. Protein
follows: capillary voltage 4 kV, dry temperature 365 °C, concentrations were determined using a Bradford assay
nebulizer pressure 40 psi, and a dry gas flow of 9 l/min Auto Roti®-nanoquant (Carl Roth).
MS mode with a smart target mass of 800m/z and a MS/MS
fragmentation amplitude of 0.5 was used. Determination of kinetic parameters

Enzyme assay For the characterization of recombinant YahK and YjgB,

enzyme activity was assayed for the reduction of carbonyl
The ADH activity was determined photometrically by mon- compounds and the oxidation of alcohols. For the reduction
itoring the increase/decrease of NADP+/NADPH at 340 nm of carbonyl compounds, the assays were conducted at pH
in a Mulitskan® spectrum spectrophotometer (Thermo Fisher 7.5, 50 mM Tris-HCl, and 37 °C. The oxidation of alcohols
Scientific). The reaction mixture contained 50 mM Tris-HCl, was assayed with a slightly higher pH of 8.5, 50 mM Tris-
pH 7.5, 0.3 mM NAD(P)H a defined aldehyde as sub- HCl, and 37 °C. For every substrate, Km and kcat values were
strate, and the purified enzyme at 37 °C. One unit of determined as well as for the corresponding cofactor NAD
enzyme activity was defined as the amount of protein (P)+/NAD(P)H. Additionally, the activity of YqhD, YahK,
that oxidizes 1 μmol of NAD(P)H/min at 37 °C. Cal- and YjgB was determined at different temperatures using
culation of Michaelis–Menten kinetics for determination 50 mM Tris-HCl, pH 7.5, and 0.3 mM NADPH with butyr-
of Km and Vmax was done with SigmaPlot 11.0 (Systat aldehyde as the substrate. The decrease of NADPH at
Software). Alcohols with alkyl chains longer than C6 340 nm was determined using a Shimadzu UV-1800 UV-
were not tested as substrates due to solubility problems. spectrophotometer (Duisburg, Germany)
5818 Appl Microbiol Biotechnol (2013) 97:5815–5824

Structure modeling activity peak was detected (Fig. 1). All active fractions were
collected and subjected to a hydrophobic interaction chroma-
The structure of YjgB was modeled with “One to one tography (Fig. 2). Analysis of the active fractions on an SDS-
threading” using YahK as template with Phyre2 (Kelley and PAGE showed one single band corresponding to the NADPH-
Sternberg 2009). YahK was chosen as starting point due to the dependent reductase activity (Fig. 3a, c). BLAST analysis of
highest similarity (32 %) in a previous global modeling ap- sequencing results of the protein band from the SDS gel by
proach of Phyre2. In the crystal structure of YahK, the section MS revealed it to be YahK (37,954 Da), a so far uncharac-
between amino acid 269 and 277 was not resolved. As this terized oxidoreductase.
part does not seem to be involved in the catalytic mechanism, To confirm this result, a deletion of yahK in E. coli was
it was neglected. Using 3DLigandSite, the cofactor NADPH examined. We therefore analyzed ΔyahK strain (JW0317)
was integrated into the structure for both proteins (Wass et al. from Keio collection, in the same way as wild type of
2010). Finally the substrate butyraldehyde was docked into E. coli W3110. Anion exchange chromatography of the
the active site using YASARA (www.yasara.org) and energy knockout variant showed a new enzyme activity eluting ear-
minimization was performed using the force field AMBER99. lier as the YahK activity of the wild type E. coli (Fig. 1). No
corresponding band to YahK (∼38 kDa) could be detected by
the SDS gel (Fig. 3b).
Results In the hydrophobic interaction chromatography (HIC)
purification (Fig. 2), the newly identified activity showed
Analysis of reductive activity of E. coli longer retention time (RT) than YahK with a low UV signal
at 280 nm. By SDS gel electrophoresis no detectable band
Since 1997, the complete genome of E. coli is accessible and was identified (data not shown). For that reason, the com-
mostly annotated; however, nearly 10 % of its genes with their plete active fraction was subjected to a tryptic digestion and
potentially encoded proteins are still unidentified (Blattner et analyzed via MS. Sequencing of the fraction revealed the
al. 1997; Riley et al. 2006; Feist et al. 2007). In addition, the enzyme to be YjgB, another so far uncharacterized zinc-type
exact activity of many of the annotated genes is not known. We alcohol dehydrogenase-like protein.
were interested to find which enzymes in E. coli are responsible
for the major NADPH-dependent aldehyde reductase activity. Homologue expression
Within the genome of E. coli, more than 20 enzymes can be
identified that could be potentially important for this reaction. The genes of YjgB and YahK as well as of the previously
We therefore purified respective enzymes from cell lysate characterized YqhD were amplified by PCR, cloned, and
and identified them by protein sequencing. In short, lysed cells expressed in E. coli BL21 (DE3). Comparison between the
of an E. coli culture grown on LB medium were prepared and cell pellet and the cell-free soluble extracts revealed the for-
enzymes were partially purified by anion exchange chroma- mation of inclusion bodies. The ratio of insoluble and soluble
tography. All eluted fractions were analyzed for their enzyme was 50:50 (data not shown). The enzymes were
NADPH-dependent aldehyde reductase activity. One major purified via an N-terminal His tag for the determination of
Fig. 1 AXC purification with
normalized RT because of
different loading volumes
(gradient starts at 0 min),
negative retention time shows
loading and wash out unbound
sample: UV280nm signal from
purification of E. coli wild type
(thick solid line), analyzed
fraction (1.5 ml) activity of E.
coli wild type (thin solid line),
UV280nm signal from
purification of E. coli Δyahk
(thick broken line), analyzed
fraction (1.5 ml) activity of E.
coli Δyahk (thin broken line)
Appl Microbiol Biotechnol (2013) 97:5815–5824 5819

Fig. 2 HIC purification with

normalized RT because of
different loading volumes
(gradient starts at 0 min),
negative retention time shows
loading and wash out unbound
sample: UV280nm signal from
purification of E. coli wild type
(thick solid line), analyzed
fraction (1.5 ml) activity of E.
coli wild type (thin solid line),
UV280nm signal from
purification of E. coli Δyahk
(thick broken line), analyzed
fraction (1.5 ml) activity of E.
coli Δyahk (thin broken line)

the kinetic parameters. The purified proteins appeared as aldehydes and alcohols were used as substrates, and also the
single band on SDS polyacrylamide gels. The molecular NAD(P)+/NAD(P)H cofactor concentration was varied to de-
weight was calculated to be 40.43 kDa for YahK and termine the corresponding Km (Table 1). The Km values to-
38.66 kDa for YjgB (including the additional amino acids of wards the different substrates varied between 0.22 and
the His tag). After 14 days at 8 °C in desalting buffer (50 mM 193.7 mM for YjgB and 0.135 and 52.6 mM for YahK. Both
Tris-HCl pH 8.0), YahK, YjgB, and YqhD still exhibited 85 % enzymes strongly prefer aldehyde compounds and no activity
of the initial activity. For long-term storage at −20 °C, glycerol was measured using a ketone as substrate. NADPH cannot be
(25 % v/v) was added and no loss of activity was observed substituted by NADH for both enzymes without losing >99 %
after 4 months for all three enzymes. of the activity. For both enzymes, hexanal represents the best
substrate concerning the turnover number.
Substrate and cofactor specificity of YahK and YjgB YahK shows lower Km values than YjgB for linear aliphatic
aldehydes (factor of 10 for acetaldehyde, 2 for propionalde-
The purified enzymes were used to determine their kinetic hyde, or 1.5 for hexanal). This difference is much more
parameters kcat and Km for different substrates. Various pronounced for the branched aldehyde isobutanal. Here the

kDa M 1 2 3 4 5 M kDa M 1 2 3 M kDa M 1 2 3 4 M

200
120 a 200 b 200
120
c
150
100
85 120 100
100 85
70 85
60 70
60
50 70
60
50
40 50
40
40
30
30
25 30

25
25
20

20

15

Fig. 3 SDS gel (12 %) with M PageRuler unstained protein ladder from E. coli Δyahk, 1–3 fractions without YahK band (no activity)
(Fermentas): a fractions of AXC purification from E. coli wild type, 1 from RT 46.8–49.8 min; c fraction of HIC purification from E. coli
crude extract; 2–5 1 ml active fraction from RT 45.8–49.8 min; b AXC wild type, 1–4 1 ml active fraction from RT 26.8–30.8 min
5820 Appl Microbiol Biotechnol (2013) 97:5815–5824

Table 1 Kinetic parameters

determined for YjgB, YahK, and Enzyme Cofactor Substrate Km (mM) kcat (s−1) kcat/Km (s−1 mM−1)
YqhD* corresponds to data from
Jarboe (2010) for different YjgB NADPH Acetaldehyde 73.4±5.4 332.6±26.1 4.5
substrates NADPH Propionaldehyde 19.6±0.8 207.2±2.9 10.6
NADPH Glyceraldehyde n.c.a. n.c.a. n.c.a.
NADPH Butyraldehyde 2.1±0.09 464.4±24.2 221
NADPH Isobutyraldehyde 193.7±27.5 50.8±9.4 0.26
NADPH Crotonaldehyde 2.4±0.2 77.6±3.3 32.3
NADPH 2-Butanone n.c.a. n.c.a. n.c.a.
NADPH Glutaraldehyde 5.8±0.5 311.9±14.8 53.8
NADPH 5-Hydroxyvalerate 59.7±3.5 2.6±0.04 0.04
NADPH Hexanaldehyde 0.34±0.03 418.9±15.7 1,232
NADPH Benzaldehyde 0.24±0.02 313.0±7.4 1,305
NADPH Furfural 0.22±0.02 224.13±26.9 1,018
NADP+ Butanol 3.5±0.28 13.79±0.6 3.94
NADP+ 1,4-Butanediol 24.1±1.5 12.9±0.4 0.53
Butanol NADP+ 0.076±0.01 14.37±0.2 189
Butyraldehyde NADPH 0.06±0.003 224.1±36.5 3,734
YahK NADPH Acetaldehyde 13.3±1.0 11.18±0.7 0.84
NADPH Propionaldehyde 10.9±1.0 11.6±0.6 1.1
NADPH Glyceraldehyde 4.4±0.4 12.3±1.3 2.8
NADPH Butyraldehyde 2.1±0.1 41.6±0.6 19.8
NADPH Isobutyraldehyde 2.2±0.2 32.1±1.2 14.6
NADPH Crotonaldehyde 3.6±0.4 32.6±2.1 9.1
NADPH 2-Butanone n.c.a. n.c.a. n.c.a.
NADPH Glutaraldehyde 4.1±0.4 13.4±1.5 3.3
NADPH 5-Hydroxyvalerat 52.6±13.2 0.18±0.02 0.003
NADPH Hexanaldehyde 0.37±0.02 18.3±0.6 49.5
NADPH Benzaldehyde 0.29±0.02 7.75±0.5 26.7
NADPH Furfural 0.135±0.03 12.5±1.9 92.3
NADP+ Butanol 6.6±0.3 4.7±0.07 0.71
NADP+ 1,4-Butanediol 38.5±2.9 6.7±0.3 0.18
Butanol NADP+ 0.012±0.001 7.47±0.07 622.0
Butyraldehyde NADPH 0.011±0.001 9.8±0.4 894.4
YqhD* NADPH Acetaldehyde 30 1.1 0.033
NADPH Propanaldehyde 3.3 45 14
NADPH Butyraldehyde 0.67 60 87
NADPH Isobutyraldehyde 2 1 0.5
NADPH Glyceraldehyde 1.4 3.4 2.4
NADPH Furfural 9 n/a n/a
n.c.a. no catalytic activity, n/a NADPH 0.008 n/a n/a
not available

Km of YjgB is 64 times higher. For tested substrates carrying aldehydes is similar for both enzymes. Turnover numbers
additional functional groups and such being more polar are show a more uniform behavior; generally they are higher with
mostly better recognized by YjgB. Crotonaldehyde and 1,4- YjgB than YahK. This is more pronounced with aliphatic
butanediol have lower Km values with YjgB, whereas glutar- aldehydes as substrates (factors between 5 and 20 between
aldehyde and 5-hydroxyvalerate have almost the same Km the two enzymes) than for bi- and trifunctional molecules
values with YjgB and YahK. Glyceraldehyde shows an inter- (factors between 3 and 10). Aromatic molecules are converted
mediate behavior, which is reasonable as it is carrying addi- faster with YjgB as well. In summary, YjgB has the higher
tional functions but can be considered branched as well and is catalytic efficiency for most substrates, isobutyraldehyde be-
only accepted by YahK as substrate. Recognition of aromatic ing the one exception. YahK and YjgB show inhibition by
Appl Microbiol Biotechnol (2013) 97:5815–5824 5821

some substrates (hexanal, benzaldehyde, and furfural at con- which it declined. YqhD showed an increase in activity until
centrations higher than ca. 2 mM). We compared YjgB and 50 °C with 12 U/mg. At higher temperatures the activity
YahK with kinetic data published for YqhD (Table 1). This has decreases.
to be considered with caution, as test conditions for YqhD
varied. But in general, YqhD has the lower Km values but also Sequence and structural comparison
much lower turnover numbers leading to an overall reduced
catalytic efficiency. The size of YahK is 349 amino acids and for YjgB 339
Activity of YqhD, YahK, and YjgB was determined in the amino acids in accordance to Jörnvall et al. (1999a), both are
temperature range from 15 to 60 °C using butyraldehyde as classified as medium-chain dehydrogenases (Nordling et al.
substrate (Fig. 4). Every enzyme showed a unique temper- 2002). An alignment for all 17 possible medium-chain
ature profile. YahK appears to be stable up to 60 °C; there is dehydrogenases/reductases (MDR) enzymes of E. coli
a constant increase in activity up to 65 U/mg at 60 °C. revealed the closest relationship of YahK and YjgB to each
Interestingly, the specific activity of YjgB remained con- other without mentioning YqhD (Jörnvall et al. 1999b). All
stant at around 460 U/mg between 37 and 50 °C, above three enzymes are grouped to the MDR superfamily, YahK
and YjgB belong to the cinnamyl alcohol dehydrogenase
family, whereas YqhD to the polyol dehydrogenase family
(Cambillau et al. 2004; Persson et al. 2008). YahK as well as
YjgB possess the GHEX2GX5(G,A)X2(I,V,A,C,S) protein
pattern that can be found in Zn-containing MDRs and the
GX1-3GX1-3G pattern located in the nucleotide-binding re-
gion. There exists an entry in the protein data bank (PDB
entry 1UUF) for YahK without any further information. The
structure was solved like that of YqhD in a structural
genomics program determining the crystal structures of E.
coli open reading frame (ORF) products of unknown func-
tion (Sulzenbacher et al. 2002; Vincentelli et al. 2003). We
modeled the structure of YjgB based on the YahK structure
to compare the active sites of both enzymes (Fig. 5). From
this perspective in both cases, the substrate is embedded
through the cofactor on one side and coordinated through
the Zn2+ ion. There is a pronounced difference between both
enzymes in the active center. YahK exhibits more space
through a cysteine at position 88, whereas in YjgB, the
corresponding residue is a tryptophane (AS 91) that limits
the available space. This substantially reduces the size of the
substrate binding pocket. Such in YjgB, the space between
the cofactor and the catalytic zinc represents a short hollow
tube. The differences in shape of the substrate binding
pocket could explain the different substrate preference of
the two enzymes. As shown above, YjgB strongly prefers
unbranched substrates, which is well demonstrated by the
marked difference between n-butyraldehyde and isobutyral-
dehyde. In contrast, both substrates react very similar in
YahK. When tryptophane 91 was removed, the cavity in
YigB would be similar to that of YahK or even slightly
larger.

Discussion
Fig. 4 Specific activity of Yahk, YjgB, and YqhD at different temper-
atures; values were determined by photometric assay in 50 mM Tris-
In analyzing enzymes of E. coli responsible for reducing
HCl buffer at pH 7.5, 5 mM butyraldehyde, and 0.3 mM NADPH for aldehydes to primary alcohols, we found two so far uncharac-
different temperatures terized enzymes YahK and YjgB to be relevant. Interestingly,
5822 Appl Microbiol Biotechnol (2013) 97:5815–5824

this class have been found important in NADP/NADPH

homeostasis, lipid biosynthesis, amino acid metabolism,
or the formation of fusel alcohols and have been dis-
cussed for enzymes of this family (Larroy et al. 2002,
2003). The rather broad substrate range of YahK and
YjgB implies a more universal function like the men-
tioned NADP/NADPH homeostasis. This is an interest-
ing fact in combination to ongoing research regarding
the use of lignocellulosic hydrolysates as a cheap car-
bon source for fermentations. Depending on the type of
biomass, pretreatment hydrolysates can contain signifi-
cant amounts of furfural (app. 1 g l−1) which has a
tremendous effect on growth of E. coli (Miller et al.
2009a; Almeida et al. 2009). Different strategies for use
of hydrolysates are investigated whereas adapted strains
represent the most elegant way. A first attempt in an
adaptation process generated an E. coli mutant EMFR9
with downregulated expression of yqhD (Miller et al.
2010). The efficient detoxification of furfural leads to
an imbalance in the NADPH pool. This results from the
low Km of YqhD for NADPH (8 μM) (Miller et al.
2009b). This extreme low Km stands in contrast to the
eightfold higher Km of YjgB. The newly identified
dehydrogenases can be an interesting starting point for
Fig. 5 The Substrate Binding Pocket for a YahK (PDB 1UUF) and b further strain improvement in this way.
YjgB. Docking solutions for butyraldehyde (yellow) and the cofactor E. coli K-12 is supposed to contain 17 ORFs that are
NADPH (blue), respectively. The substrate is targeted through interac-
tion of the aldehyde function with the Zn2+ ion (purple) encoding MDR alcohol dehydrogenase but without iden-
tification of YqhD as a MDR (Jörnvall et al. 1999a, b).
So apparently, the well examined workhorse E. coli still
we could not detect YqhD, an enzyme which earlier has bears some biocatalysts which are promising candidates
been reported to be important (Atsumi et al. 2009; Jarboe for metabolic engineering approaches for the optimized
2010; Nakamura and Whited 2003). YjgB appears to be production of bulk chemicals and that might be better
the more active enzyme; however, its production seems to suited than currently applied enzymes. In this context,
be only induced when the gene coding for YahK is we suggest to rename all characterized Zn-dependent-
knocked out. Further studies on conditions that lead to and NADPH preferring alcohol dehydrogenases from E.
the induction of YjgB in comparison to YqhD and YahK coli YqhD, YahK, and YjgB to AdhZ1, AdhZ2, and
have to be performed. The in vivo functions of YahK as AdhZ3, respectively. The newly characterized enzymes
well as YjgB have yet to be determined. Due to their AdhZ2 and AdhZ3 should be considered, when knock-
similar substrate specificities compared to YqhD, a related out variants are prepared that are not supposed to re-
function is possible. First reports of the in vivo function of duce aldehydes, as for example shown in the production
YqhD describe involvement in a NADPH-dependent response of isobutyric acid (Zhang et al. 2011). AdhZ3 seems to
mechanism to lipid peroxidation (Perez et al. 2008). Addition- be especially useful for the conversion of bifunctional
ally, expression analysis concerning growth-limiting condi- substrates and AdhZ2 for branched chain substrates. So
tions using E. coli K-12 W3110 and comparison between wild far, the focus in metabolic engineering has been on
type and cold-sensitive deletion strains revealed an upre- AdhZ1 as it also accepts a wide range of carbonyl
gulation of YqhD (Phadtare and Inouye 2004; Hua et al. compounds rendering it an interesting candidate for
2004). Until now, no study dealing with the investiga- various mutation approaches though at lower activity
tion of global responses of E. coli focusing on an (Jarboe 2010; Tang et al. 2009; Liao et al. 2010). First
altered expression profile identified yahK and yjgB and engineering approaches for a further improvement of the
could connect it with a special enzyme activity. Cin- enzyme resulted in variants with increased catalytic
namyl alcohol dehydrogenases catalyze the last step in efficiency on 3-hydroxypropionaldehyde (Li et al. 2008).
the biosynthesis of monolignols in plants (Sibout et al. Using the more active AdhZ3 or AdhZ2 instead might be a
2005), which is hardly relevant in E. coli. Enzymes of better alternative.
Appl Microbiol Biotechnol (2013) 97:5815–5824 5823

Acknowledgments While this paper was in review, complementary Hua Q, Yang C, Oshima T, Mori H, Shimizu K (2004) Analysis
data were published describing the role of all three alcohol dehydro- of gene expression in Escherichia coli in response to changes
genases (AdhZ1, AdhZ2 and AdhZ3) as candidates for the final reduction of growth-limiting nutrient in chemostat cultures. Appl Envi-
step in a metabolic engineering approach for production of aromatic ron Microbiol 70(4):2354–2366. doi:10.1128/AEM.70.4.2354-
compounds using E. coli. (Koma D, Yamanaka H, Moriyoshi K, Ohmoto 2366.2004
T, Sakai K (2012) Production of aromatic compounds by metabolically Jarboe LR (2010) YqhD: a broad-substrate range aldehyde reductase with
engineered escherichia coli with an expanded shikimate pathway. Appl various applications in production of biorenewable fuels and chem-
Environ Microbiol 78(17):6203. doi:10.1128/AEM.01148-12). icals. Appl Microbiol Biotechnol 89(2):249–257. doi:10.1007/
s00253-010-2912-9
Open Access This article is distributed under the terms of the Creative Jörnvall H, Höög J-O, Persson B (1999a) SDR and MDR: completed
Commons Attribution License which permits any use, distribution, and genome sequences show these protein families to be large, of old
reproduction in any medium, provided the original author(s) and the origin, and of complex nature. FEBS Lett 445:261–264.
source are credited. doi:10.1016/S0014-5793(99)00130-1
Jörnvall H, Shafqat J, Hoog JO, Hjelmqvist L, Oppermann UCT, Ibanez
C (1999b) An ethanol-inducible MDR ethanol dehydrogenase/acet-
aldehyde reductase in Escherichia coli—structural and enzymatic
References relationships to the eukaryotic protein forms. Eur J Biochem 263
(2):305–311. doi:10.1046/j.1432-1327.1999.00323.x
Kelley LA, Sternberg MJE (2009) Protein structure prediction on the
Almeida JRM, Bertilsson M, Gorwa-Grauslund MF, Gorsich S, Liden web: a case study using the Phyre server. Nat Protoc 4(3):363–
G (2009) Metabolic effects of furaldehydes and impacts on bio- 371. doi:10.1038/Nprot.2009.2
technological processes. Appl Microbiol Biotechnol 82(4):625– Larroy C, Pares X, Biosca JA (2002) Characterization of a Saccharomyces
638. doi:10.1007/s00253-009-1875-1 cerevisiae NADP(H)-dependent alcohol dehydrogenase (ADHVII), a
Atsumi S, Higashide W, Liao JC (2009) Direct photosynthetic recy- member of the cinnamyl alcohol dehydrogenase family. Eur J Bio-
cling of carbon dioxide to isobutyraldehyde. Nat Biotechnol 27 chem 269(22):5738–5745. doi:10.1046/j.1432-1033.2002.03296.x
(12):1177–1180 Larroy C, Fernandez MR, Gonzalez E, Pares X, Biosca JA (2003)
Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, Datsenko Properties and functional significance of Saccharomyces cerevi-
KA, Tomita M, Wanner BL, Mori H (2006) Construction of siae ADHVI. Chem-Biol Interact 143:229–238. doi:10.1016/
Escherichia coli K-12 in-frame, single-gene knockout mutants: S0009-2797(02)00166-7
the Keio collection. Mol Syst Biol 2. doi:10.1038/Msb4100050 Li HM, Chen J, Li YH (2008) Enhanced activity of yqhD oxidoreduc-
Berezina O, Zakharova N, Brandt A, Yarotsky S, Schwarz W, Zverlov tase in synthesis of 1,3-propanediol by error-prone PCR. Prog Nat
V (2010) Reconstructing the clostridial n-butanol metabolic path- Sci 18(12):1519–1524. doi:10.1016/j.pnsc.2008.04.014
way in Lactobacillus brevis. Appl Microbiol Biotechnol 87 Liao JC, Atsumi S, Wu TY, Eckl EM, Hawkins SD, Buelter T (2010)
(2):635–646. doi:10.1007/s00253-010-2480-z Engineering the isobutanol biosynthetic pathway in Escherichia
Berríos-Rivera S, San K-Y, Bennett G (2003) The effect of carbon coli by comparison of three aldehyde reductase/alcohol dehydro-
sources and lactate dehydrogenase deletion on 1,2-propanediol genase genes. Appl Microbiol Biotechnol 85(3):651–657.
production in Escherichia coli. J Ind Microbiol Biotechnol 30 doi:10.1007/s00253-009-2085-6
(1):34–40. doi:10.1007/s10295-002-0006-0 Miller EN, Jarboe LR, Turner PC, Pharkya P, Yomano LP, York SW,
Blattner FR, Plunkett G, Bloch CA, Perna NT, Burland V, Riley M, Nunn D, Shanmugam KT, Ingram LO (2009a) Furfural inhibits
ColladoVides J, Glasner JD, Rode CK, Mayhew GF, Gregor J, growth by limiting sulfur assimilation in ethanologenic Escher-
Davis NW, Kirkpatrick HA, Goeden MA, Rose DJ, Mau B, Shao ichia coli strain LY180. Appl Environ Microbiol 75(19):6132–
Y (1997) The complete genome sequence of Escherichia coli K- 6141. doi:10.1128/AEM.01187-09
12. Science 277(5331):1453–1462 Miller EN, Jarboe LR, Yomano LP, York SW, Shanmugam KT, Ingram
Cambillau C, Sulzenbacher G, Alvarez K, van den Heuvel RHH, LO (2009b) Silencing of NADPH-dependent oxidoreductase
Versluis C, Spinelli M, Campanacci V, Valencia C, Eklund H, genes (yqhD and dkgA) in furfural-resistant ethanologenic
Tegoni M (2004) Crystal structure of E. coli alcohol dehydroge- Escherichia coli. Appl Environ Microbiol 75(13):4315–4323.
nase YqhD: evidence of a covalently modified NADP coenzyme. doi:10.1128/AEM.00567-09
J Mol Biol 342(2):489–502. doi:10.1016/j.jmb.2004.07.034 Miller EN, Turner PC, Jarboe LR, Ingram LO (2010) Genetic changes
Carothers JM, Goler JA, Keasling JD (2009) Chemical synthesis using that increase 5-hydroxymethyl furfural resistance in ethanol-
synthetic biology. Curr Opin Biotechnol 20(4):498–503. producing Escherichia coli LY180. Biotechnol Lett 32(5):661–
doi:10.1016/j.copbio.2009.08.001 667. doi:10.1007/s10529-010-0209-9
Chen WP, Kuo TT (1993) A simple and rapid method for the prepa- Nakamura CE, Whited GM (2003) Metabolic engineering for the
ration of gram-negative bacterial genomic DNA. Nucleic Acids microbial production of 1,3-propanediol. Curr Opin Biotechnol
Res 21(9):2260–2260 14(5):454–459. doi:10.1016/j.copbio.2003.08.005
Emptage M, Haynie SL, Laffend LA, Pucci JP, GW (2003) Process for Nielsen DR, Yoon SH, Yuan CJ, Prather KLJ (2010) Metabolic engi-
the biological production of 1,3 propanediol with high titer. US neering of acetoin and meso-2,3-butanediol biosynthesis in E.
6514733 BI coli. Biotechnol J 5(3):274–284. doi:10.1002/biot.200900279
Feist AM, Henry CS, Reed JL, Krummenacker M, Joyce AR, Karp PD, Nordling E, Jornvall H, Persson B (2002) Medium-chain dehydro-
Broadbelt LJ, Hatzimanikatis V, Palsson BO (2007) A genome- genases/reductases (MDR)—family characterizations including
scale metabolic reconstruction for Escherichia coli K-12 MG1655 genome comparisons and active site modelling. Eur J Biochem
that accounts for 1260 ORFs and thermodynamic information. 269(17):4267–4276. doi:10.1046/j.1432-1033.2002.03114.x
Mol Syst Biol 3. doi:10.1038/Msb4100155 Perez JM, Arenas FA, Pradenas GA, Sandoval JM, Vasquez CC (2008)
Heer D, Heine D, Sauer U (2009) Resistance of Saccharomyces cer- Escherichia coli YqhD exhibits aldehyde reductase activity and
evisiae to high concentrations of furfural is based on NADPH- protects from the harmful effect of lipid peroxidation-derived
dependent reduction by at least two oxireductases. Appl Environ aldehydes. J Biol Chem 283(12):7346–7353. doi:10.1074/
Microbiol 75:7631–7638. doi:10.1128/AEM.01649-09 jbc.M708846200
5824 Appl Microbiol Biotechnol (2013) 97:5815–5824

Persson B, Hedlund J, Jornvall H (2008) The MDR superfamily. Cell Zenatti A, Campanacci V, Cambillau C (2002) A medium-
Mol Life Sci 65(24):3879–3894. doi:10.1007/s00018-008-8587-z throughput crystallization approach. Acta Crystallogr D 58
Phadtare S, Inouye M (2004) Genome-wide transcriptional analysis of the (12):2109–2115. doi:10.1107/S0907444902013938
cold shock response in wild-type and cold-sensitive, quadruple-csp- Tang XM, Tan YS, Zhu H, Zhao K, Shen W (2009) Microbial conver-
deletion strains of Escherichia coli. J Bacteriol 186(20):7007–7014. sion of glycerol to 1,3-propanediol by an engineered strain of
doi:10.1128/Jb.186.20.7007-7014.2004 Escherichia coli. Appl Environ Microbiol 75(6):1628–1634.
Riley M, Abe T, Arnaud MB, Berlyn MKB, Blattner FR, Chaudhuri doi:10.1128/Aem.02376-08
RR, Glasner JD, Horiuchi T, Keseler IM, Kosuge T, Mori H, Vincentelli R, Bignon C, Gruez A, Canaan S, Sulzenbacher G, Tegoni
Perna NT, Plunkett G, Rudd KE, Serres MH, Thomas GH, M, Campanacci V, Cambillau C (2003) Medium-scale structural
Thomson NR, Wishart D, Wanner BL (2006) Escherichia coli genomics: strategies for protein expression and crystallization.
K-12: a cooperatively developed annotation snapshot—2005. Accounts Chem Res 36(3):165–172. doi:10.1021/ar010130s
Nucleic Acids Res 34(1):1–9. doi:10.1093/Nar/Gkj405 Wang FH, Qu HJ, He H, Tan TW (2005) High-level expression of the
Sibout R, Eudes A, Mouille G, Pollet B, Lapierre C, Jouanin L, Seguin A 1,3-propanediol oxidoreductase from Klebsiella pneumoniae in
(2005) CINNAMYL ALCOHOL DEHYDROGENASE-C and -D Escherichia coli. Mol Biotechnol 31(3):211–219
are the primary genes involved in lignin biosynthesis in the floral Wass MN, Kelley LA, Sternberg MJE (2010) 3DLigandSite: predicting
stem of Arabidopsis. Plant Cell 17(7):2059–2076. doi:10.1105/ ligand-binding sites using similar structures. Nucleic Acids Res
tpc.105.030767 38(suppl 2):W469–W473. doi:10.1093/nar/gkq406
Skraly FA, Lytle BL, Cameron DC (1998) Construction and character- Yim H, Haselbeck R, Niu W, Pujol-Baxley C, Burgard A, Boldt J,
ization of a 1,3-propanediol operon. Appl Environ Microbiol 64 Khandurina J, Trawick JD, Osterhout RE, Stephen R, Estadilla J,
(1):98–105 Teisan S, Schreyer HB, Andrae S, Yang TH, Lee SY, Burk MJ,
Studier FW (2005) Protein production by auto-induction in high- Van Dien S (2011) Metabolic engineering of Escherichia coli for
density shaking cultures. Protein Expr Purif 41(1):207–234. direct production of 1,4-butanediol. Nat Chem Biol 7(7):445–452.
doi:10.1016/j.pep.2005.01.016 doi:10.1038/Nchembio.580
Sulzenbacher G, Gruez A, Roig-Zamboni V, Spinelli S, Valencia C, Zhang K, Woodruff AP, Xiong M, Zhou J, Dhande YK (2011) A
Pagot F, Vincentelli R, Bignon C, Salomoni A, Grisel S, Maurin synthetic metabolic pathway for production of the platform chemical
D, Huyghe C, Johansson K, Grassick A, Roussel A, Bourne Y, isobutyric acid. ChemSusChem 4(8):1068–1070. doi:10.1002/
Perrier S, Miallau L, Cantau P, Blanc E, Genevois M, Grossi A, cssc.201100045