Comparison of three enoate reductases and their potential use for biotransformations

Article abstract of DOI:10.1002/adsc.200700074

Enoate reductases (ERs) selectively reduce carbon-carbon double bonds in α,β-unsaturated carbonyl compounds and thus can be employed to prepare enantiomerically pure aldehydes, ketones, and esters. Most known ERs, most notably Old Yellow Enzyme (OYE), are biochemically very well characterized. Some ERs have only been used in whole-cell systems, with endogenous ketoreductases often interfering with the ER activity. Not many ERs are biocatalytically characterized as to specificity and stability. Here, we cloned the genes and expressed three non-related ERs, two of them novel, in E. coli: XenA from Pseudomonas putida, KYE1 from Kluyveromyces lactis, and Yers-ER from Yersinia bercovieri. All three proteins showed broad ER specificity and broad temperature and pH optima but different specificity patterns. All three proteins prefer NADPH as cofactor over NADH and are stable up to 40°C. By coupling Yers-ER with glucose dehydrogenase (GDH) to recycle NADP(H), conversion of > 99 % within one hour was obtained for the reduction of 2-cyclohexenone. Upon lowering the loadings of Yers-ER and GDH, we discovered rapid deactivation of either enzyme, especially of the thermostable GDH. We found that the presence of enone substrate, rather than oxygen or elevated temperature, is responsible for deactivation. In summary, we successfully demonstrate the wide specificity of enoate reductases for a range of α,β- unsaturated carbonyl compounds as well as coupling to glucose dehydrogenase for recycling of NAD(P)(H); however, the stability limitations we found need to be overcome to envision large-scale use of ERs in synthesis.

Full text of DOI:10.1002/adsc.200700074

FULL PAPERS
DOI: 10.1002/adsc.200700074
Comparison of Three Enoate Reductases and their PotentialUse
for Biotransformations
Javier F. Chaparro-Riggers,a Thomas A. Rogers,a Eduardo Vazquez-Figueroa,a
Karen M. Polizzi,a and Andreas S. Bommariusa,b,*
a
School of Chemical & Biomolecular Engineering, Parker H. Petit Institute of Bioengineering and Bioscience, Georgia
Institute of Technology, 315 Ferst Drive, Atlanta, GA 30332-0363, USA
Phone: (+1)-404-385-1334; fax: (+1)-404-894-2291; e-mail: andreas.bommarius@chbe.gatech.edu
School of Chemistry and Biochemistry, Georgia Institute of Technology, 315 Ferst Drive, Atlanta, GA 30332-0363, USA
b
Received: January 30, 2007
Abstract: Enoate reductases (ERs) selectively to 408C. By coupling Yers-ER with glucose dehydro-
reduce carbon-carbon double bonds in a,b-unsaturat- genase (GDH) to recycle NADP(H), conversion of
ed carbonyl compounds and thus can be employed
>99% within one hour was obtained for the reduc-
to prepare enantiomerically pure aldehydes, ketones, tion of 2-cyclohexenone. Upon lowering the loadings
and esters. Most known ERs, most notably Old of Yers-ER and GDH, we discovered rapid deactiva-
Yellow Enzyme (OYE), are biochemically very well tion of either enzyme, especially of the thermostable
characterized. Some ERs have only been used in GDH. We found that the presence of enone sub-
whole-cell systems, with endogenous ketoreductases strate, rather than oxygen or elevated temperature, is
often interfering with the ER activity. Not many ERs responsible for deactivation. In summary, we success-
are biocatalytically characterized as to specificity and fully demonstrate the wide specificity of enoate re-
stability. Here, we cloned the genes and expressed ductases for a range of a,b-unsaturated carbonyl
three non-related ERs, two of them novel, in E. coli: compounds as well as coupling to glucose dehydro-
XenA from Pseudomonas putida, KYE1 from Kluy- genase for recycling of NAD(P)(H); however, the
veromyces lactis, and Yers-ER from Yersinia berco- stability limitations we found need to be overcome
vieri. All three proteins showed broad ER specificity to envision large-scale use of ERs in synthesis.
and broad temperature and pH optima but different
specificity patterns. All three proteins prefer Keywords: biotransformations; enoate reductases;
NADPH as cofactor over NADH and are stable up enzymes; Old Yellow Enzyme; redox chemistry
Introduction
been characterized thoroughly,[9] its physiological role
has only recently begun to emerge.[10–12]
Biocatalysts are increasingly used during the synthesis
The OYE family has grown in recent years and a
of enantiomerically pure compounds of industrial in- variety of different enzymes have been characterized,
terest (pharmaceuticals, food, and crop protection), including different yeast OYEs,[13,14] XenA/XenB re-
due to their exquisite regioselectivity and stereoselec- ductase,[15] bacterial morphine reductase,[16] bacterial
tivity. These enable them to perform difficult chemi- nitrate ester reductases (pentaerythritol tetranitrate
cal synthesis, while circumventing the use of compli- reductase),[17] YqjM from Bacillus subtilis,[18] plant ox-
cated traditional synthetic routes.[1–4] For reductive re- ophytodienoic acid reductases,[19,20] and an enzyme in-
actions, ketoreductases are already an established volved in prostaglandin synthesis in Trypanosoma
part of the synthetic toolbox. One of the emerging re- cruzi.[21] Despite this variety in substrate specificity,
ductive chemistries is the asymmetric reduction of the family shares a common a8b8 barrel (TIM barrel)
carbon double bonds. Enoate reductases (ERs) per- fold with the flavin mononucleotide (FMN) binding
form such reactions and have been recognized for within the barrel near the carboxyl-terminal ends of
decades, but the stereoselective application was only the b-strands, however, there are significant variations
recently published.[5–7] Enoate reductase from Saccha- in the capping subdomain.[22] Members of the OYE
romyces carlsbergensis, also known as ꢀOld Yellow family also share several common functional charac-
Enzymeꢁ (OYE), was the first isolated flavin-contain- teristics. Although clear preferences are observed for
ing enzyme in 1933.[8] Despite the fact that OYE has every enzyme, they are generally able to reduce nitro
Adv. Synth. Catal. 2007, 349, 1521 – 1531
ꢂ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1521

Javier F. Chaparro-Riggers et al.
FULL PAPERS
esters, nitro aromatic substrates, and simple and com- cloned into pET28a (+) vector (Novagen, San Diego,
plex aldehydes and ketones.[23] Kinetics proceed by a CA) at the following restriction sites: NdeI and XhoI
ping-pong mechanism, i.e., consisting of an oxidative for Kluyveromyces lactis (KYE1), and NdeI and Hin-
and reductive half reaction.[13,14,17] The proposed dIII for both XenA and Yersinia bercovieri (Yers-
mechanism for OYE-catalyzed asymmetric reduction ER). These constructs enabled the expression of
proceeds according to an anti-addition hydrogenation mature ER with an N-terminal polyhistidine tag for
manner (trans-hydrogenation) and is consistent with convenient purification.
experimental results.[27–29]
Sequence analysis of Yers-ER and XenA showed
The asymmetric reduction of a,b-unsaturated car- the same nucleotide sequence as derived from the
bonyl compounds creates up to two new chiral centers NCBI databank (accession files ZP 00823209 and
and is an interesting approach towards enantiomeri- AAF02538), while KYE1 (P40952) revealed one dif-
cally pure compounds. Despite the success of the ho- ference at position 768 (t768g), leading to an amino
mogeneous catalysis of enantioselective alkene reduc- acid change from valine to glycine (V257G).
tion, high stereoselectivity nearly always depends on
XenA is a xenobiotic reductase from the Gram-
olefin proximity to highly polar functional groups negative, aerobic bacterium Pseudomonas putida.
such as amides, acids, and alcohols. Attempts to gen- KYE1 was isolated from yeast and shares a high
eralize these procedures to alkenes conjugated with amino acid identity with OYE1 from the yeast Sac-
less polar groups such as aldehydes or ketones have charomyces cerevisiae (71% aa identity, 84% aa simi-
been much less successful,[6] although some exceptions larity). KYE1 had already been mentioned in the lit-
have been reported.[24–26] Therefore, ERs are interest- erature,[31] but has never been characterized for its
ing for the hydrogenation of unsaturated aldehyde or ER function. Finally, Yers-ER is from the Gram-neg-
ketone substrates, rounding off the chemistꢁs toolbox. ative, aerobic bacterium Yersinia bercovieri and
The stereoselective application of these enzymes has shares high identity with the pentaerythritol tetrani-
been reported recently.[6]
trate reductase from Enterobacter cloacae (76% aa
Currently, the use of ERs is especially limited by identity, 86% aa similarity).[17] As already expected
two main factors: i) the full biocatalytic potential of from the wide variety of possible functions and hosts,
biochemically characterized ERs has not been ex- the amino acid identities and similarities between the
plored, especially with respect to their specificity on three cloned ER are low: KYE1 to XenA (19.0 and
enones beyond 2-cyclohexenone, and ii) until now, 35.5%, respectively); KYE1 to Yers-ER (37.0 and
whole cells were generally used and these are limited 47.1%, respectively); and XenA to Yers-ER (24.3
by the presence of endogenous ketoreductases in the and 39.9%, respectively). The sequence alignment re-
host cells, which also employ NAD(P)H to reduce the veals only a few conserved residues, among them the
carbonyl function.
active site residues histidine and tyrosine (Table 1 and
In the first part of this paper, we address the first Figure 1). Yers-ER (365 aa) and XenA (363 aa) share
limitation by cloning three ERs from very different a similar size, in contrast to KYE1 (398 aa), but all
sources and characterizing their substrate spectra, pH share similar isoelectric points (calculated pI=6.06,
and temperature profiles. We overcome the second 6.01, and 5.99, respectively). The his-tagged ERs
limitation by using an isolated ER in combination
with an isolated cofactor regeneration system, here
Table 1. Amino acid sequence identities and similarities be-
glucose dehydrogenase (GDH). For straightforward
tween the three enoate reductases with some other pub-
purification, we cloned a polyhistidine tag at the N-
lished proteins.[a]
terminus of each enzyme. In summary, all these fac-
tors and findings are important for successful and eco-
nomically feasible biotransformations in the future.
KYE1
Yers-ER
XenA
CaMa
KYE1
OYE1
OYE2
OYE3
Yers-ER
PETN
ZYMM
XenA
Yqjm
75/85
-
38/54
36/53
39/54
38/54
39/54
-
76/85
41/57
31/47
32/48
44/58
38/54
27/45
28/48
27/44
29/44
28/45
31/47
27/41
32/48
-
70/82
70/81
67/78
36/53
38/51
32/49
28/48
30/48
36/51
29/47
Results
Cloning, Overexpression, and Purification
The gene of xenobiotic reductase A (XenA) from
Pseudomonas putida, which shows ER activity toward
2-cyclohexen-1-one,[15] was kindly provided by Brian
Fox (University of Wisconsin, Madison, WI). Addi-
tionally, ER genes were isolated from the genomic
DNA of Kluyveromyces lactis ATCC 8585 and Yersi-
nia bercovieri ATCC 43970. The gene products were
39/52
28/43
30/45
OPR1
PrSa
[a]
Amino acid identity/similarity levels (in%); for descrip-
tion of ER source organisms, see caption to Figure 1.
1522
asc.wiley-vch.de
ꢂ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1521 – 1531

Three Enoate Reductases and their Potential Use
FULL PAPERS
share a slightly lower pI (Yers-ER: 5.80, KYE1: 5.61,
XenA: 5.76).
The heterologous expression in E. coli BL21(DE3)
G
(Novagen) was induced at an OD600 of 0.5 with
0.1 mM IPTG and incubated overnight at 378C. All
three ERs showed activity in the supernatant. The
his-tagged enzymes were purified according to the
standard Ni2+-NTA bead protocol from Qiagen (Va-
lencia, CA), and purification was monitored by SDS-
PAGE (Figure 2). The eluate did not contain any
other bands, even while overloading the gel. The ac-
tivity of a non-tagged XenA (in lysate) toward 2-cy-
clohexenone was measured, and these results indicat-
ed that his-tagging is not detrimental to ER activity.
Characterization
The three cloned ERs are derived from a broad spec-
trum of possible enzyme functions. XenA is a xenobi-
otic reductase, while KYE1 is a homologue of OYE1
and Yers-ER is a homologue of pentaerythritol tetra-
nitrate reductase from Enterobacter cloacae. Never-
theless, the comparison of the substrate specificity on
different a,b-unsaturated carbonyl compounds reveals
that all three enzymes showed broad substrate specif-
icity and were active on almost all tested substrates
(Table 2). For KYE1, the specific activities are gener-
ally low (ꢀ2.5 U/mg) and did not vary much between
the different substrates. The specific activity of XenA
varied the most: (R)-carvone and 3-methyl-2-butenal
were not substrates for XenA, but-3-en-2-one and
maleimide yielded the highest activity levels mea-
Figure 1. Amino acid alignment (with ClustalW) of some
published ERs with the three species characterized in this
paper. The bold amino acids are histidine (important for
substrate/ligand binding) and tyrosine (important for proton
donation during the reduction of a,b-unsaturated carbonyl
compounds). If the enzyme possesses two histidines in the
active site (bold amino acids solid) for substrate binding, it
can reduce trinitrotoluene (TNT) to the H-Meisenheimer
complex, whereas the enzymes with only one histidine
cannot.[32] Aligned from top to bottom are: CaMa, old
yellow enzyme from Candida macedoniensis (BAD24850);
KYE1 old yellow enzyme from Kluyveromyces lactis
(P40952); OYE1, old yellow enzyme from Saccharomyces
carlbergensis (Q02899); OYE2, old yellow enzyme from Sac-
charomyces cerevisiae (NP 012049); OYE3, old yellow
enzyme from Saccharomyces cerevisiae (CAA97878); Yers-
ER, ER from Yersinia bercovieri (ZP 00823209); PETN,
pentaerythritol tetranitrate reductase from Enterobacter clo-
acae (2 ABA); ZYMM, ER from Zymomonas mobilis
(Q5NLA1); XenA, xenobiotic reductase A from Pseudomo-
nas putida (AAF02538); Xqjm from Bacillus subtilis (1Z48);
PrSa from Trypanosoma cruzi (BAC24024); OPR1, 12-oxo-
phytodienoate reductase from Arabidopsis thaliana
(AAC78440).
Adv. Synth. Catal. 2007, 349, 1521 – 1531
ꢂ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1523

Javier F. Chaparro-Riggers et al.
FULL PAPERS
strates which still contain double bonds (phenylpropa-
nal and dihydrocarvone), though in lower amounts
compared to the a,b-unsaturated carbonyl substrates.
Surprisingly, XenA oxidized cofactor in the presence
of dihydrocarvone, but did not react with the enoate
(R)-carvone. The tendency of these products toward
further reductions is under investigation. This finding
highlights the importance of controls for the biocata-
lytic use of ERs.
Cofactor Specificity
After testing substrate specificity, the cofactor prefer-
ence of the three ERs was checked with four different
enone substrates (Table 3). NADPH was always the
preferred cofactor, and the enzymes showed very
little or no activity with NADH. Interestingly, the
choice of enone substrate seems to influence the
NADPH/NADH specificity ratio. While NADPH did
not react at all in the presence of 2-cyclohexen-1-one
and rarely in presence of maleimide, all three ERs
showed activity with NADH in presence of methyl
vinyl ketone and ethyl-3-(2-furyl)propanoate. This
phenomenon is all the more interesting, as ER is
known to react according to a ping-pong mecha-
nism,[31] i.e., the reductive (substrate) and oxidative
(FMN by NADPH) half-reactions occur independent-
ly, thus, the different enone substrates should not in-
fluence cofactor specificity.
Figure 2. SDS-PAGE of each step in the Ni2+-NTA purifica-
tion of the ER from Yersinia bercovieri. The standard proto-
col was used from QIAGEN with Ni2+-NTA beads. Lane
1: ladder; lane 2: cell lysate, lane 3: flow through from the
binding solution; lane 4: flow through from the wash buffer;
lane 5: eluted fraction.
sured in our study, 22.69 and 25.82 U/mg, respectively.
The Yers-ER was most consistent, showing the high-
est specific activity levels on eight of the twelve sub- Enzyme Stability
strates.
The substrates can be divided into four groups: i) The lack of stability of enzymes against heat and or-
substrates which reacted equally well for all three en- ganic solvents can hamper their economical applica-
zymes, 2-cyclohexen-1-one, trans,trans-2,4-hexadienal, tion. We investigated the thermal tolerance of all
citral, and cinnamaldehyde; ii) substrates which did three ERs by taking samples after various incubation
not react with XenA: (R)-carvone and 3-methyl-2-bu- times at three temperatures (30, 37, and 458C) and
tenal; iii) substrates which show strong preference for measuring residual activity at 258C. We always found
Yers-ER, ketoisophorone and maleic anhydride, and, deactivation to fit a first-order rate law. Although
finally iv) substrates which show strong preference for none of the three proteins is particularly thermosta-
both XenA and Yers-ER, maleimide, N-ethylmalei- ble, Yers-ER at 378C was 25-fold and 7-fold more
mide, but-3-en-2-one, and ethyl 3-(tetrahydrofuryl)- stable than XenA and KYE1, respectively (Table 4).
propanoate.
Yers-ER at 458C in degassed solution deactivates
Importantly, the broad spectrum of native functions with a half-life of 129 min, compared with 117 min in
of the three proteins investigated was reflected in the air-saturated solution, which demonstrates that
broad spectrum of substrate preference and range of oxygen has no influence on stability. The half-life of
specific activity.
XenA and KYE1 at that temperature was less than
We also investigated whether reaction products of 5 min. Since sufficient stability is an important factor
a,b-enone reduction further reacted to undesirable for biocatalysis, we chose the Yers-ER for further
by-products. We chose products with two different characterization.
structural features. The first group did not contain
The activity-temperature profile between 108C and
any further unsaturated double bonds (succinimide 558C is depicted in Figure 3. Since NADPH degrades
and butanone) and, as expected, showed no further at a significant rate at ꢁ 458C, each sample beyond
reductive activity. On the other hand, KYE1 and 458C was prepared as a separate tube of assay solu-
Yers-ER still showed reductive activity on two sub- tion and NADPH was added last before starting the
1524
asc.wiley-vch.de
ꢂ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1521 – 1531

Three Enoate Reductases and their Potential Use
Table 2. The specific activity of the three enoate reductases on different a,b-unsaturated carbonyl compounds.[a]
FULL PAPERS
Molecule
Structure
Specific activity [U/mg]
KYE1 XenA Yers-ER
2-cyclohexen-1-one
1.54
2.74
4.22
trans,trans-2,4-hexadienal&
1.51
0.67
3.18
1.75
2.19
1.97
citral*
cinnamaldehyde&
2.30
0.73
1.56
nd
2.91
2.54
(R)-carvone*
3-methyl-2-butenal
maleic anhydride
0.45
0.86
-0.43
1.91
0.39
5.20
ketoisophorone
maleimide
1.33
2.04
1.84
10.11
18.88
25.82
N-ethylmaleimide&
2.46
2.14
2.36
14.87
22.69
9.96
15.81
14.85
12.87
but-3-en-2-one (methyl vinyl ketone)
ethyl 3-(tetrahydrofuran)propanoate [ethyl-3-(2-furyl)propanoate]*
[a]
Assay conditions: 200 mM sodium phosphate, pH 7.5, 10 mM substrate, 0.2 mM NADPH, 258C. Extinction coefficient of
NADPH of 6.22 mMÀ1 cmÀ1 at 340 nm or 3.51 mMÀ1 cmÀ1 at 365 nm (substrates are marked with &) was used in calcula-
tion of specific acitivity.[33] Compounds with an asterisk did not appear completely soluble at 10 mM, so the listed specific
activities may correspond to the solubility limit. Alternate names for some of these compounds are shown in parentheses.
nd=not detected.
reaction. Thermal deactivation dominates over activa- activation energy Ea of 25.11 kJmolÀ1 (r2 =0.9937; for
tion at temperatures beyond 558C, the highest tem- the range of 10–408C: 24.06 kJmolÀ1, r2 =0.9964).
perature investigated. A maximum specific activity of
6.15 U/mg was obtained for 2-cyclohexen-1-one at
458C, almost 2.5-fold higher than at room tempera- Activity-pH Profile
ture (208C). The activation energy was calculated
from an Arrhenius plot of specific activity over 1/T The activity-pH profile was measured for the purified
from data in the range of 10 to 458C; we obtained an Yers-ER (Figure 4). The enzyme showed a broad pH
Adv. Synth. Catal. 2007, 349, 1521 – 1531
ꢂ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1525

Javier F. Chaparro-Riggers et al.
FULL PAPERS
Table 3. Cofactor preference for KYE1, XenA, and Yers-
ER.[a]
Substrate
Specific Ac- Specific
tivity [U/
mg]
NADPH
Activity
[U/mg]
NADH
Kluyveromyces 2-cyclohexen-1- 1.54
nd
one
maleimide
methyl vinyl
ketone
2.04
2.14
nd
0.30
ethyl 3-(2-furyl)- 2.36
propanoate
2-cyclohexen-1- 2.74
one
0.56
nd
Figure 3. Activity-temperature profile of Yers-ER. The sam-
ples were heated immediately before the assay and con-
tained extra NADPH (250 mM) to ensure saturation with re-
duced enzyme at the start of the run. The buffer contained
200 mM sodium phosphate buffer at pH 7.5, 5 mM 2-cyclo-
hexen-1-one and 200 mM NADPH.
XenA
maleimide
methyl vinyl
ketone
25.82
22.69
1.50
1.43
ethyl 3-(2-furyl)- 9.96
propanoate
2-cyclohexen-1- 4.22
one
1.08
nd
Yersinia
maleimide
methyl vinyl
ketone
18.88
14.85
nd
0.69
ethyl 3-(2-furyl)- 12.87
propanoate
0.56
[a]
Assay conditions: 200 mM phosphate, pH 7.5, 10 mM
substrate, 0.2 mM cofactor, 258C. Detection limit: 0.1 U/
mg. nd=not detected.
Table 4. Half-lives of enoate reductase (Yers-ER) and glu-
cose dehydrogenase (GDH-103).
Figure 4. Activity-pH profile of Yers-ER. The measurements
were conducted in buffered solution (see Experimental Sec-
tion) at 258C with 5 mM 2-cyclohexen-1-one and 200 mM
NADPH.
Enzyme
Temperature [8C]
Half-life [min]
XenA
XenA
XenA
KYE1
KYE1
KYE1
Yers-ER
Yers-ER
Yers-ER
Yers-ER[a]
Yers-ER[b]
GDH-103[c]
GDH-103[a]
GDH-103[b]
30
37
45
30
37
45
30
37
45
30
30
30
30
30
600
99
<5
1146
210
around pH 5.5 and 8.8, possibly corresponding to the
pKa values of the catalytic histidine182 (pH 6.04) but
far from the typical pKa value of tyrosines of 10.13
(tyrosine 188 is crucial for binding).
<5
>2160
1488
117
78
Inhibition Patterns
2880
~7800
1.6
Substrate and product inhibition can significantly
reduce the biocatalytic usefulness of enzymes. There-
fore, the specific activity of Yers-ER was measured at
258C in 200 mM phosphate (pH 7.5) with 200 mM
NADPH and various concentrations of 2-cyclohexe-
none (substrate) and 2-cyclohexanone (product). For
the substrate inhibition study, the concentration of 2-
cyclohexenone was varied from 5–20 mM and no ac-
564
[a]
In the presence of 25 gLÀ1 of 2-cyclohexen-1-one.
In the presence of 25 gLÀ1 of 2-cyclohexanone.
Literature value from Pollard et al.[37]
[b]
[c]
optimum between pH 6.0–8.5 with a maximum specif- tivity loss was detected. To test for product inhibition,
ic activity (3.46 U/mg) around pH 7.0. The activity the concentration of 2-cyclohexanone was varied from
dropped off sharply between pH 5–6 and pH 8.5–9.5, 0–100 mM in a reaction solution containing 10 mM
with inflection points of the activity-pH profile substrate. No decrease in activity was observed and
1526
asc.wiley-vch.de
ꢂ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1521 – 1531

Three Enoate Reductases and their Potential Use
FULL PAPERS
product inhibition was ruled out for the tested condi-
tions.
Finally, kinetic data were measured for the Yers-
ER at 258C in 200 mM phosphate buffer (pH 7.5) and
varying 2-cyclohexenone (substrate) concentrations.
The assay solution contained 200 mM NADPH, a con-
centration at which the enzyme was very close to sat-
uration with respect to NADPH (Figure 5). At
Scheme 1. Enzyme-coupled cofactor regeneration system
with enoate reductase from Yersinia bercovieri and glucose
dehydrogenase (GDH-103).
Figure 5. Determining the saturation concentration of
NADPH for Yers-ER. Assay conditions: 258C, pH 7.5,
10 mM 2-cyclohexen-1-one.
10 mM substrate concentration, the KM value for
NADPH is calculated to be 40Æ5 mM. The catalytic
constant kcat (118.2Æ0.6 minÀ1), a Michaelis constant
KM (1.24Æ0.01 mM) and specificity ratio kcat/KM
(95.32Æ0.71 mMÀ1 minÀ1) values were determined.
Coupling of the Enoate Reductase with the Cofactor
Recycling System
Figure 6. Enzymatic oxidation of 25 g/L of 2-cyclohexen-1-
one to 2-cyclohexanone at 308C in 200 mM phosphate
The coupling of Yers-ER with GDH was performed pH 7.5.
in the batch mode (Scheme 1).
All reactions were run at 308C in 200 mM phos-
phate buffer (pH 7.5) containing 25 g/L (260 mM) of
the substrate 2-cyclohexen-1-one, 53.8 g/L (299 mM, >99% and >16% conversion were measured after
15% molar excess) of glucose, 1 g/L (1.2 mM) of co- 29 h, respectively. Such a dramatic difference in reac-
factor NADP+ and both enzymes at varying ratios of tion rate between the 8:16 kU/L and the 4:8 kU/L
kU/L. Four reactions were run at these conditions at enzyme loadings for Yers-ER and GDH-103 suggests
8:16, 8:8, 4:8, and 2:4 kU/L of Yers-ER to GDH-103, significant enzyme deactivation at the operating con-
respectively (Figure 6).
ditions. To test this hypothesis, the residual activities
It is evident that enzyme loading had a significant of Yers-ER and GDH-103 were monitored separately
effect on reaction rate. At 8:16 kU/L of Yers-ER to in batch experiments in the presence of reactant or
GDH-103, the reaction reached >99% conversion product at the operating conditions (308C, 200 mM
within an hour (217 turnovers), while at a loading of phosphate pH 7.5 and 25 g/L of 2-cyclohexen-1-one or
8:8 kU/L the reaction required 6 h for the same 2-cyclohexanone) in the absence of cofactor. The re-
degree of conversion (>99%) (Figure 6). This differ- sidual activity was measured by monitoring
ence in reaction rate might be explained by the in- NADP(H) depletion (Yers-ER) or formation (GDH-
creased supply of reduced cofactor achieved at a 103). Assuming first-order deactivation (corroborated
higher loading of GDH-103. At more diluted levels of above for ER and independently verified for GDH
Yers-ER and GDH-103, 4:8 kU/L and 2:4 kU/L, from various organisms[40]), half-lives were estimated
Adv. Synth. Catal. 2007, 349, 1521 – 1531
ꢂ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1527

Javier F. Chaparro-Riggers et al.
FULL PAPERS
for Yers-ER to be 78 min and 48 h in the presence of function, which can lower selectivity and produce ad-
reactant (2-cyclohexen-1-one) and product (2-cyclo- ditional impurities.[36,37] The use of isolated enzymes
hexanone), respectively. For GDH-103, the half-lives can circumvent this problem but still requires a cofac-
were estimated to be 1.6 min and 9.4 h in the presence tor regeneration system. Importantly, both ER and
of reactant and product, respectively. The addition of the regenerating enzyme have to be active at the
260 mM 2-cyclohexen-1-one leads to a more than 30- same pH range. As maximum activity was measured
fold decrease of the half-life at 308C from >2160 min for Yers-ER in the pH range of 6.0–8.5, we chose a
to 78 min for Yers-ER. A high degree of precipitation commercially available and stabilized glucose dehy-
was observed in both enzyme samples which con- drogenase (GDH-103); GDH-103 has been reported
tained 260 mM 2-cyclohexen-1-one.
to lose its activity only above pH 9.[39,41] The coupled
system revealed that conversions of >99% can be
easily obtained when using high amount of enzymes
and short incubation times. This confirms the advant-
age of employing isolated enzymes. Another advant-
Discussion
For the synthesis of enantiomerically pure com- age is that isolated enzyme systems show higher pro-
pounds, the reduction of a,b-unsaturated carbonyl ductivity and are less tedious to optimize than whole-
compounds is of great interest. Homogeneous chemi- cell systems.[36,37] Lastly, the expression and purifica-
cal catalysts have dominated cis-hydrogenation,[44] tion of his-tagged, isolated enzymes is straightfor-
while trans-hydrogenation seems to be more challeng- ward, does not seem to influence activity very much,
ing for this type of catalyst. The asymmetric reduction and allows easy assessment of new activities and reac-
with OYE was found to be consistent with an anti-ad- tions without interference from side reactions.
dition mechanism (trans-hydrogenation).[27–30] This
Conversion dropped off more rapidly than expected
means that ERs are an important addition to the upon longer incubation time. Therefore, we measured
chemistꢁs toolbox at many process scales, but their ap- the residual activity at different incubation times in
plication has thus far encountered two major limita- the presence of 25 g/L of the substrate 2-cyclohexe-
tions.
none or the product 2-cyclohexanone. We obtained
First, isolated enzymes have not been characterized the following two unanticipated results.
for their potential use in biotransformation. Our work
i) The kinetic stability of the thermostable GDH-
demonstrates that ERs from three different organisms 103[37] was lower than that for the thermolabile Yers-
showed, as expected, significant differences in their ER. In many cases, thermal stability correlates with
substrate preferences: i) xenobiotic reductase, XenA, the stability against other denaturing factors, such as
from Pseudomonas putida was known to react with organic solvents.[42,43] For GDH-103 compared to
cyclohexenone but no further substrates had been Yers-ER this is not the case, as we measured a half-
tested;[15] ii) KYE1 from Kluyveromyces lactis had life of 117 min for Yers-ER and around 4800 min for
been described in the literature but never tested for GDH-103 at pH 7.0 and 458C.[37]
its reductive activity;[31] finally, iii) the ER from Yersi-
ii) The presence of the substrate 2-cyclohexenone
nia bercovieri had not been described in the literature causes rapid loss of activity in both Yers-ER and
at all, but its kinetic parameters on 2-cyclohexenone GDH-103 (Table 4): half-life decreased about 30-fold
are similar to the ones of its homologue, pentaerythri- to 78 min for Yers-ER but a drastic 6300-fold to
tol tetranitrate reductase (PETN).[17] Based on the 1.7 min for GDH-103. In comparison, the presence of
differences in substrate spectrum, the isolation and the product cyclohexanone caused no deactivation at
characterization of additional enzymes would be ad- all in Yers-ER and an 18-fold reduction of half-life to
vantageous. Their development towards large-scale 564 min for GDH-103. This difference in deactivation
use would be improved with the establishment of a behavior is astonishing, given that one double bond
high-throughput screening system, which already constitutes the only structural difference between the
exists for ketoreductases.[36] Regarding the specificity two. This finding highlights the importance, but also
of these enzymes, tests with selected products of the the unpredictability, of measuring enzyme stability in
ER reaction revealed that, while the carbonyl group the presence of substrates and products, not just in
was not reduced, additionally present double bonds aqueous buffer at the same pH value and tempera-
could be further reduced and form unwanted side ture.
products.
Different strategies, such as chemical modification,
The second limitation results from the hitherto lyophilization, immobilization, additives, or protein
dominant use of whole cells for enoate reductases, engineering, can be used to stabilize enzymes, as low
since the specific reduction of the unsaturated double stability is a known problem.[34,35] The choice of strat-
bond is limited by the presence of endogenous ketor- egy is case-dependent, since the main target variable,
eductases in the host cells. The latter employ the e.g., costs or time of process development, can vary
same cofactor, NAD(P)H, to reduce the carbonyl substantially.
1528
asc.wiley-vch.de
ꢂ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1521 – 1531

Three Enoate Reductases and their Potential Use
FULL PAPERS
TGG CAA CAG CTT CCT CG-3’; Yersinia bercovieri: 5’-
ATA ATA ATA AAA GCT TTT ACA GCG TTG GGT
AAT CAG TGT AGC CCT TAG CAC CGC-3’).
Conclusions
In the present study we have characterized the sub-
strate spectrum of three ERs, XenA from Pseudomo-
nas putida, KYE1 from Kluyveromyces lactis, and ER
from Yersinia bercovieri, to a complex set of a,b-unsa-
turated carbonyl compounds. We demonstrated that
all three ERs feature broad substrate specificity but
different substrate preferences. These enzymes can be
part of a future screening set of enzymes, which can
be used to identify ERs for the synthesis of enantio-
merically pure compounds.
Furthermore, we demonstrate the feasibility of cell-
free biocatalytic reduction of a,b-unsaturated carbon-
yl compounds and have addressed the major limita-
tion of the in vitro application of ERs. By employing
glucose dehydrogenase as a cofactor regeneration
system, we eliminate the need for whole-cell catalysis,
which is often limited by side reactions and low activi-
ty. The coupled enzyme system Yers-ER/GDH-103 is
quickly deactivated by 2-cyclohexenone, our sub-
strate, much more so than by heat or by the presence
of oxygen or cyclohexanone, our reaction product.
Overall, we have shown that biocatalytic reduction
of a,b-unsaturated carbonyl compounds with re-
combinant, isolated ERs can straightforwardly be ap-
plied as a tool for biochemical synthesis from enones.
Amplification of the target DNA was performed using a
standard PCR protocol with Pfu polymerase from New Eng-
land Biolabs (Marlborough, MA) and PCR buffer from
Stratagene (La Jolla, CA). Setting up the PCR reactions in-
volved final DNA concentration of 100 ng (XenA, KYE1,
or Yers-ER), 200 mM of each dNTP, 10 mM of each primer
and 1 U of Pfu polymerase in a final volume of 45 mL. To
each of the reactions, 5 mL of the Pfu buffer were added.
DNA was amplified successfully for 30 cycles in an Eppen-
dorf Gradient Thermocyler (Eppendorf, Hamburg, Germa-
ny) using the following conditions: each cycle involved a de-
naturation step at 1 min 958C, an annealing step at 1 min at
508C, 55 8C or 608C, and an extension step at 3.5 min at
688C. Of the final reaction mixture of 50 mL aliquots were
analyzed on 1% agarose gels. Prior to any further use,
These PCR products were gel purified using a gel extraction
kit (Qiagen, Valencia, CA).
The PCR products were digested with the adequate re-
striction enzymes (XenA: NdeI/HindIII; KYE1: NdeI/XhoI;
Yers-ER: NdeI/HindIII), ligated into pET28a (+) (Novagen,
San Diego) and transformed into competent E. coli XL1-
Blue cells. The sequence of the ER genes was determined
by sequencing with an Applied Biosystems 3100 Genetic
Analyzer (Perkin–Elmer-AB, Boston, MA).
Expression and Purification of the Enoate Reductase
Genes
Heterologous expression of the ER genes in E. coli BL21
(DE3) was performed as follows: 5 mL starter LBKan cul-
tures were inoculated with aliquots from frozen stock cul-
ture and grown overnight at 378C. These starter cultures
were each used to inoculate a 500 mL culture (1% v/v),
which was vigorously aerated until OD600 reached 0.5–0.6, at
which point the cultures were incubated with 0.1 mm IPTG
(final concentration) and protein expression was performed
over night. Cells were harvest and pellets frozen away at
À808C or used directly for the protein purification.
ExperimentalSection
BacterialStrains, Media and Growth Conditions
The xenobiotic reductase A (XenA) from Pseudomonas
putida was kindly provided by Brian Fox.[15] The genomic
DNA of Kluyveromyces lactis (ATCC 8585D-5) and strain
Yersinia bercovieri (ATCC 43970) were obtained from
ATCC and grown in trypticase soy media (30 g/L trypticase
soy broth, BD 211768) under aerobic conditions at 268C.
The cell pellets were resuspended in lysis buffer (5 mL/g
wet weight; 50 mM NaH2PO4, 300 mM NaCl and 20 mM
imidazole) and sonicated on ice. The supernatant was sepa-
rated after centrifugation and the purification was per-
formed with Ni2+-NTA beads according to the standard pro-
tocol of Qiagen (Valencia, CA).
Primer Sequence and Cloning
The DNA sequences were identified using a search of the
NCBI Genebank (Accession files AAF02538 for XenA,
ZP 00823209 for Yersinia bercovieri and P40952 for Kluy-
veromyces lactis). The corresponding specific 5’- and 3’-pri-
mers were synthesized at MWG Biotech (High Point, NC).
The primers contain a gene-encoding segment (italic), a
restriction site (underlined), and an overhang. All 5’-primers
introduced a NdeI restriction site into the PCR fragment.
(XenA: 5’-AAC CAA CCA ACA TATG TCC GCA CTG
TTC GAA CCC TAC ACC C-3’; KYE1: 5’-AAC CAA CCA
ACA TAT GTC GTT TAT GAA CTT TGA ACC AAA
GCC ATT GGC-3’; Yersinia bercovieri: 5’-AAC CAA CCA
ACA TAT GAA GAC TGC TAA ACT GTT CTC TCC
TTT GAA GGT TGG-3’) The 3’-primers introduced a Hin-
dIII or a XhoI restriction site into the PCR fragment
(XenA: 5’-ATA ATA ATA AAA GCT TTT AGC GAT
AGC GCT CAA GCC AGT GCG C-3’; KYE1: 5’-ATA
ATA ATA ACT CGA GCT ATT TCT TGT AAC CCT
Enzyme Assay
Apparent steady-state kinetic constants were recorded by
monitoring the oxidation of NADPH at 340 nm using a
molar absorption coefficient of 6.22 mMÀ1 cmÀ1 or
3.51 mMÀ1 cmÀ1 for 365 nm, while systematically varying the
substrate concentration. Reactions were initiated by addi-
tion of enzyme and monitored for 2 min. All experiments
were conduced in triplicate and results were averaged. One
unit of enzyme activity was defined as the oxidation of
1 mmol NADPH per min at 258C in the assay buffer. Since
members of the OYE family catalyze the slow oxidation of
nicotinamide cofactor in the absence of substrate, we deter-
mined this background oxidase activity and subtracted it
from the specific activities. The enzyme assays were mea-
Adv. Synth. Catal. 2007, 349, 1521 – 1531
ꢂ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1529

Javier F. Chaparro-Riggers et al.
FULL PAPERS
sured at 258C in 200 mM phosphate, pH 7.5, 200 mM
NADPH, 42.9 nM Yers-ER), where 2-cyclohexen-1-one was
used as oxidative substrate.
Acknowledgements
We thank Prof. Charles Eckert, Prof. Charles Liotta and their
students, especially Elizabeth M. Hill, at Georgia Institute of
Technology for access and help with gas chromatography.
Additionally, we want to thank Merck & Co., Inc. (Rahway,
NJ) for providing us with GDH-103from Biocatalytics (Pa-
sadena, CA). DNA sequencing was provided by the Funda-
mental and Applied Molecular Evolution sequencing center
at Emory University (enabled by the National Science Foun-
dation grant MRI 0320786). Eduardo Vazquez-Figueroa
gratefully acknowledges funding by a NSF Graduate Re-
search Fellowship.
Activity-pH Profile
The pH profile was obtained for the Ni2+-NTA eluted frac-
tions. All samples were tested in triplicate in 1.5 mL methyl
acrylate disposable cuvettes. An appropriate buffer was
used for each different pH range: acetate (5.0–5.5), phos-
phate (6.0–7.5), Tris (8.0–9.0) and carbonate (9.5–10.0), all
at 200 mM. The rate of change of absorbance at 340 nm was
measured at 258C and an initial 2-cyclohexenone concentra-
tion of 5 mM.
Protein GelAnayl sis
References
Prior to SDS-PAGE, protein samples were diluted to the de-
sired concentration in deionized water if the initial concen-
tration was above 200 ng/mL. The 50 mL diluted samples
were then mixed with 50 mL of 2X sample buffer composed
of 125 mM Tris-HCl, pH 6.8, 4% SDS, 50% glycerols,
0.02% bromophenol blue, and 10% 2-mercaptoethanol.
Mixed samples were incubated at 1008C for 5 min and then
placed on ice. 10 to 20 mL of the samples were loaded onto
a 12% PAGETM Gold precast gel and run in a BioRad Mini
Protean 3 cell chamber (BioRad, Hercules, CA) at 125 V
for 45 min (running buffer: 25 mM Tris base, 192 mM gly-
cine, 0.1% SDS). Molecular weight standard BioRad Preci-
sion Plus protein ladder from BioRad was added to lanes
immediately adjacent to the sample lanes. Coomassie Blue
from Pierce (Rockford, IL) was used in staining.
[1] A. S. Bommarius, K. M. Polizzi, Chem. Eng. Sci. 2006,
61, 1004–1016.
[2] A. Zaks, Curr. Opin. Opin. Chem. Biol. 2001, 5, 130–
136.
[3] M. Mꢃller, Curr. Opin. Biotechnol. 2004, 15, 591–598.
[4] S. Panke, M. Wubbolts, Curr. Opin. Chem. Biol. 2005,
9, 188–194.
[5] M. Wada, A. Yoshizumi, Y. Noda, M. Kataoka, S. Shi-
mizu, H. Takagi, S. Nakamori, Appl. Environ. Micro-
biol. 2003, 69, 933–937.
[6] M. A. Swiderska, J. D. Stewart, J. Mol. Catal. B:
Enzym. 2006, 42, 52–54.
[7] A. Mꢃller, B. Hauer, B. Rosche J. Mol. Cat. B: Enzym.
2006, 38, 126–130.
[8] Warburg, O. , and Christian, W. , Biochem. Zeit 1933,
266, 377–411.
[9] A. S. Abramovitz, V. Massey, J. Biol. Chem. 1976, 251,
5327–5336.
[10] J. Lee, C. Godon, G. Lagniel, D. Spector, J. Garin, J.
Labarre, M. B. Toledano, J. Biol. Chem. 1999, 274,
6040–6046.
Determination of Protein Concentration
Protein concentrations were determined by the Bradford
method utilizing Coomassie Plus Protein assay reagent and
pre-diluted protein assay standards – BSA (Pierce Chemi-
cal) for the calibration curve.
[11] B. K. Haarer, D. C. Amberg, Mol. Biol. Cell 2004, 15,
4522–4531.
[12] R. Reekmans, K. De Smet, C. Chen, P. Van Hummelen,
R. Contreras, FEMS Yeast Res. 2005, 5, 711–725.
[13] K. Stott, K. Saito, D. J. Thiele, V. Massey, J. Biol.
Chem. 1993, 268, 6097–60106.
Coupling of the Enoate Reductase with the Cofactor
Recycling System
Enoate reductase from Yersinia bercovieri and GDH-103 (a
glucose dehydrogenase from Biocatalytics, Pasadena, CA)
were added in different ratios to a 200 mM phosphate buffer
of pH 7.5, 299 mM glucose and 1.2 mM of NADP+. The re-
action was performed in the batch mode at 308C and sub-
strate cyclohexen-1-one was added to a final concentration
of 260 mM. The work was carried out at 125 mg scale (5 mL
total volume), the temperature was maintained at 308C
using a water bath, pH was adjusted manually using 2M
NaOH and agitation was achieved using a magnetic stirrer
at 200 rpm. The conversion was determined via gas chroma-
tography (GC) by measuring the substrate and product con-
tent of a reaction sample extracted with ethyl acetate using
an Agilent GC-FID (model 6890) with a DB17 column
(Agilent, Palo Alto, CA).
[14] Y. S. Niino, S. Chakraborty, B. J. Brwon, V. Massey, J.
Biol. Chem. 1995, 270, 1983–1991.
[15] D. S. Blehert, B. G. Fox, G. H. Chambliss, J. Bacteriol.
1999, 181, 6254–6263.
[16] C. E. French, N. C. Bruce, Biochem. J. 1994, 301, 97–
103.
[17] C. E. French, S. Nicklin, N. C. Bruce, J. Bacteriol. 1996,
178, 6623–6627.
[18] T. B. Fitzpatrick, N. Amrhein, P. Macheroux, J. Biol.
Chem. 2003, 278, 19891–19897.
[19] F. Schaller, E. W. Weiler, J. Biol. Chem. 1997, 272,
28066–28072.
[20] J. Strassner, A. Fꢃrholz, P. Macheroux, X. X. Amrhein,
A. Schaller, J. Biol. Chem. 1999, 49, 35067–35073.
[21] B. K. Kubata, Z. Kubututu, T. Nozaki, C. J. Munday, S.
Fukuzumi, K. Ohkubo, M. Lazarus, T. Maruyama, S. K.
Martin, M. Duszenko, Y. Urade, J. Exp. Med. 2002,
196, 1241–1251.
1530
asc.wiley-vch.de
ꢂ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1521 – 1531

Three Enoate Reductases and their Potential Use
FULL PAPERS
[22] D. Van Den Hemel, A. Brigꢄ, S. N. Savvides, J. Van [34] A. S. Bommarius, J. Broering, J. F. Chaparro Riggers,
Beeumen, J. Biol. Chem. 2006, 281, 28152–28161.
[23] R. W. Williams, N. C. Bruce, Microbiology, 2002, 148,
1607–1614.
[24] D. A. Dobbs, K. P. M. Vanhessche, E. Brazi, V. Rauten-
strauch, J. Y. Lenoir, J. P. Genet, J. Wiles, S. H. Bergens,
Angew. Chem. Int. Ed. Engl. 2000, 39 1992–1995.
[25] B. H. Lipshutz, J. M. Servesko, Angew. Chem. Int. Ed.
Engl. 2003, 42, 4789–4792.
K. M. Polizzi, Curr. Opin. Biotechnol. 2006, 17,606–
610.
[35] K. M. Polizzi, A. S. Bommarius, J. Broering, J. F.Cha-
parro Riggers, Curr. Opin. Chem. Biol. 2007, in press.
[36] D. J. Pollard, K. Telari, J. Lane, G. Humphrey, C.
McWilliams, S. Nidositko, P. Salmon, J. Moore, Biotech-
nol Bioeng. 2006, 93, 674–686.
[37] D. J. Pollard, M. Truppo, J. Pollard, C. Chen, J. Moore,
Tetrahedron: Assymetry 2006, 17, 554–559.
[38] T. R. M. Barends, H. Yoshida, B. W. Dijkstra, Curr.
Opin. Biotechnol. 2004, 15:356–363.
[39] W. Hilt, G. Pfleiderer, P. Fortnagel, Biochim. Biophys.
Acta 1991, 1076, 298–304.
[40] Unpublished results.
[41] Description of GDH-103: mutant of GDH-102; excel-
lent thermostability and solvent stability; should be
screened in parallel with GDH-102 to identify optimal
enzyme for a specific application.
[42] V. G. H. Eijsink, S. Gaseidnes, T. V. Borchert, B.
van den Burg, Biomolecular Engineering 2005, 22, 21–
30.
[26] B. H. Lipshutz, J. M. Servesko, T. B. Petersen, P. P.
Papa, Org. Lett. 2004, 6, 1273–1275.
[27] R. M. Kohli, V. Massey, J. Biol. Chem. 1998, 273,
32763–32770.
[28] D. Xu, R. M. Kohli, V. Massey, Proc. Natl. Acad. Sci.
U.S.A. 1999, 96, 3556–3561.
[29] B. J. Brown, J. W. Hyun, S. Duvvuri, P. A. Karplus, V.
Massey, J. Biol. Chem. 2002, 277, 2138–2145.
[30] K. M. Fox, P. A. Karplus, Structure 1994, 2, 1089–1105.
[31] P. A. Karplus, K. M. Fox, V. Massey, The FASEB Jour-
nal 1995, 9, 1518–1526.
[32] H. L. Messiha, N. C. Bruce, B. M. Satelle, M. J. Sut-
cliffe, A. W. Munro, N. S. Scrutton, J. Biol. Chem. 2005,
310, 433–447.
[33] C. Larroy, M. R. Fernandez, E. Gonzalez, X. Pares,
J. A. Biosca, Biochem. J. 2002, 361, 163–172.
[43] D. A. Cowan, Physiol A. Physiol. 1997, 118, 429–438.
[44] R. Noyori, Angew. Chem. Int. Ed. 2002, 41, 2008–2022.
Adv. Synth. Catal. 2007, 349, 1521 – 1531
ꢂ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1531