Asymmetrie reduction of activated alkenes by pentaerythritol tetranitrate reductase: Specificity and control of stereochemical outcome by reaction optimisation

Article abstract of DOI:10.1002/adsc.200900574

We show that pentaerythritol tetranitrate reductase (PETNR), a member of the 'ene' reductase old yellow enzyme family, catalyses the asymmetric reduction of a variety of industrially relevant activated α,β-unsaturated alkenes including enones, enals, maleimides and nitroalkenes. We have rationalised the broad substrate specificity and stereochemical outcome of these reductions by reference to molecular models of enzyme-substrate complexes based on the crystal complex of the PETNR with 2cyclohexenone 4a. The optical purity of products is variable (49-99% ee), depending on the substrate type and nature of substituents. Generally, high enantioselectivity was observed for reaction products with stereogenic centres at Cβ (>99% ee). However, for the substrates existing in two isomeric forms (e.g., citral 11a or nitroalkenes 18-19a), an enantio-divergent course of the reduction of E/Z-forms may lead to lower enantiopurities of the products. We also demonstrate that the poor optical purity obtained for products with stereogenic centres at Ca is due to non-enzymatic racemisation. In reactions with ketoisophorone 3a we show that product racemisation is prevented through reaction optimisation, specifically by shortening reaction time and through control of solution pH. We suggest this as a general strategy for improved recovery of optically pure products with other biocatalytic conversions where there is potential for product racemisation.

Full text of DOI:10.1002/adsc.200900574

FULL PAPERS
DOI: 10.1002/adsc.200900574
Asymmetric Reduction of Activated Alkenes by Pentaerythritol
Tetranitrate Reductase: Specificity and Control of Stereochemical
Outcome by Reaction Optimisation
a
b
b
a
Anna Fryszkowska, Helen Toogood, Michiyo Sakuma, John M. Gardiner,
c,
b,
Gill M. Stephens, * and Nigel S. Scrutton *
Manchester Interdisciplinary Biocentre, The School of Chemistry, The University of Manchester, 131 Princess Street,
Manchester M1 7DN, U.K.
Manchester Interdisciplinary Biocentre, Faculty of Life Sciences, The University of Manchester, 131 Princess Street,
Manchester M1 7DN, U.K.
Fax : (+44)-161-306-8918; phone: (+44)-161-306-5152; e-mail: nigel.scrutton@manchester.ac.uk
Manchester Interdisciplinary Biocentre, The School of Chemical Engineering and Analytical Science, The University of
Manchester, 131 Princess Street, Manchester M1 7DN, U.K.
a
b
c
Fax : (+44)-161-306-8918; phone: (+44)-161-306-4377; e-mail: gill.stephens@manchester.ac.uk
Received: August 19, 2009; Published online: November 18, 2009
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/adsc.200900574.
Abstract: We show that pentaerythritol tetranitrate divergent course of the reduction of E/Z-forms may
reductase (PETNR), a member of the ꢀeneꢁ reduc- lead to lower enantiopurities of the products. We
tase old yellow enzyme family, catalyses the asym- also demonstrate that the poor optical purity ob-
metric reduction of a variety of industrially relevant tained for products with stereogenic centres at Ca is
activated a,b-unsaturated alkenes including enones, due to non-enzymatic racemisation. In reactions with
enals, maleimides and nitroalkenes. We have ration- ketoisophorone 3a we show that product racemisa-
alised the broad substrate specificity and stereo- tion is prevented through reaction optimisation, spe-
chemical outcome of these reductions by reference cifically by shortening reaction time and through
to molecular models of enzyme-substrate complexes control of solution pH. We suggest this as a general
based on the crystal complex of the PETNR with 2- strategy for improved recovery of optically pure
cyclohexenone 4a. The optical purity of products is products with other biocatalytic conversions where
variable (49–99% ee), depending on the substrate there is potential for product racemisation.
type and nature of substituents. Generally, high
enantioselectivity was observed for reaction products
with stereogenic centres at Cb (>99% ee). However, Keywords: asymmetric synthesis; biocatalysis; ene
for the substrates existing in two isomeric forms reductase; old yellow enzyme; pentaerythritol tetra-
(
e.g., citral 11a or nitroalkenes 18–19a), an enantio- nitrate reductase; stereochemistry
Introduction
the expense of NAD(P)H is known to proceed as a
[
11]
net trans-addition of 2H. These flavoenzymes, ob-
[1–10]
In recent years there has been an intensive search for tained primarily from plants and microorganisms,
enzymatic methods in organic synthesis given the in- have proved to be useful biocatalysts as they reduce a
creasing demand for optically pure compounds. The wide spectrum of a,b-unsaturated ketones, aldehydes,
biocatalytic asymmetric reduction of activated C=C nitroalkenes, carboxylic acids and their derivatives,
bonds has received much attention because it can yielding products with a variety of biotechnological
generate up to two new stereogenic centres in a single and pharmaceutical applications. However, the stereo-
[1–10]
step.
Unlike transition metal-catalyzed cis-hydro- selectivity and optical purity of the respective prod-
genation, biocatalytic reduction of activated a,b-unsat- ucts varies widely between these enzymes and is de-
urated alkenes catalyzed by flavin-dependent ꢀeneꢁ re- pendent on the nature of the cofactor or cofactor-re-
[
2]
ductases from the old yellow enzyme (OYE) family at generation system used.
2976
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 2976 – 2990

Asymmetric Reduction of Activated Alkenes by Pentaerythritol Tetranitrate Reductase
FULL PAPERS
Pentaerythritol tetranitrate reductase (PETNR) ically (<5 ppm oxygen). Compared to anaerobic
from the anaerobic microorganism Enterobacter cloa- assays, aerobic reduction of 2-cyclohexenone 4a and
[12]
cae st. PB2 is related to the prototypical OYE from 2-methylpentenal 10a by PETNR showed a significant
[
13]
Saccharomyces pastorianus.
PETNR was isolated reduction in the overall steady-state reaction rate
via its ability to degrade a variety of high explosives, (18% and 48%, respectively), showing the importance
including reduction of nitroaromatics to Meisenheim- of excluding oxygen from the reactions. We therefore
er hydride complexes and hydrolytic cleavage of ni- ran all reactions under anaerobic conditions.
[14,15]
trate esters.
More recently we reported that
Steady-state parameters (Table 1) indicated a high
PETNR catalyses the enantioselective reduction of a specificity (k /K ) of PETNR for maleimides 1–2a
cat
m
variety of (E)- and (Z)-a,b-unsaturated nitroalkenes and ketoisophorone 3a. The biological functions and
[
16]
under biphasic reaction conditions. As the enzyme physiological substrates of PETNR and many other
exhibited some unusual behaviour in comparison to OYEs remain unknown, despite the various hypothe-
[
16]
[17,18]
other OYEs, including by-product formation,
we ses that have been made.
It is worth noting that
explored the ability of PETNR to act as a general the K for 1a was within the range typical for natural
m
ꢀeneꢁ reductase. In this paper, we describe the stereo- substrates of enzymes. Interestingly, substrates 1–3a
selective reduction of a,b-unsaturated alkenes includ- possess two activating groups (carbonyl groups) and
ing enals, enones, enamides and nitroolefins using they were also readily reduced by members of OYE
[
1]
PETNR. We propose a rationale for the substrate family. By contrast, other cyclohexanone derivatives
specificity and stereoselectivity of PETNR, based on 4–7a, which possess only one carbonyl group, had kcat/
kinetic and stereochemical data in conjunction with Km two orders of magnitude lower, despite being
modelling approaches. We also demonstrate that the structurally similar to ketoisophorone 3a (Table 1, en-
optimisation of reaction conditions (e.g., enzyme con- tries 4–8 vs. entry 3). In general, ketones 4–8a and 13a
centration, pH and reaction time) can improve sub- were relatively poor substrates under steady-state
stantially the enantiopurity of the products. We show conditions (Table 1), whereas aldehydes 9–13a were
that this optimisation is required to minimise ꢀoff- reduced readily by PETNR.
enzymeꢁ racemisation, which may account for the low PETNR did not reduce the following a,b-unsaturat-
enantiopurity of some products of biocatalytic reac- ed carboxylic acid derivatives: trans-cinnamide, trans-
tions reported in the literature.
cinnamic acid, 3-phenylmaleic acid, 1-cyclohexene-1-
carboxylic acids (and their respective esters), methyl
trans-cinnamate, and methyl cyclohexenecarboxylate.
We also found no activity towards substrates contain-
ing activating groups such as nitrile or halogens (cro-
tonitrile, 3-bromocyclohexene, trans-cinnamonitrile,
Results and Discussion
Substrate Specificity of PETNR for a,b-Unsaturated
Alkenes
cyclohexene-1-carbonitrile
either in spectrophotometric assays or measured by
PETNR is known to reduce the C=C double bonds of GC analysis.
and
b-bromostyrene)
[16]
[13]
nitroolefins,
prednisone
and 2-cyclohexenone
We noted that PETNR has a restricted substrate
[14]
(
4a), but the substrate range has not been explored specificity, which possibly results from the difficulty in
further. Therefore, we benchmarked the biocatalytic accommodating the large substituents at Ca and Cb
ability of PETNR against other enzymes from the in the active site. The selectivity for the alkene acti-
[1–6,9]
OYE family
by studying anaerobic bioreduction vating group is also restricted. The kinetic data
of a variety of a,b-unsaturated activated olefins. Ini- (Table 1) demonstrate a general preference for small
tially, we determined the kinetic constants (apparent cyclic molecules (1–9a). For the aldehyde series,
K , k , K) of PETNR of anaerobic reduction of the smaller substitutents at Cb were preferred (10a vs.
m
cat
i
substrates under steady-state turnover conditions 11–13a), while the presence of a methyl substituent at
Table 1, Supporting Information Figure S1a–i) prior Ca increased the activity in comparison to the unsub-
to analysing reaction products and the stereochemical stituted parent molecule. For cinnamaldehyde deriva-
outcome of the reaction. tives (E)-12a and (E)-13a, the introduction of an a-
(
Although the biocatalytic reductions of alkenes by methyl group led to a decrease in k , but this was
cat
OYEs are routinely conducted under aerobic condi- compensated for by the 21-fold decrease in K , result-
m
[1–10]
tions,
molecular oxygen can compete with the ing in an overall 3-fold improvement in specificity
alkene substrates for the oxidation of enzyme-bound (k /K ). Similar trends were previously observed for
cat
m
[17]
[16]
FMNH2. This reduces the overall reaction rate, and nitroalkenes. In contrast, the activity with 2-methyl-
also yields undesirable reactive products (hydrogen cyclohexenone 5a was lower than with the respective
[
17]
peroxide and superoxide).
We confirmed that parent ketone (4a; Table 1). The presence of a methyl
oxygen affects the reaction rate of PETNR by meas- substituent at Cb (3-methylcyclohexanone 15a and 3-
uring the reaction rates both aerobically and anaerob- methylcyclopentanone 16a) leads to a significant re-
Adv. Synth. Catal. 2009, 351, 2976 – 2990
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2977

Anna Fryszkowska et al.
FULL PAPERS
[a]
Table 1. Steady state kinetics of PETNR with a,b-unsaturated activated alkenes.
À1
À1
À1
Entry
Substrate
kcat (s )
K (mM)
m
K (mM)
i
k /K (s mM )
cat m
1
1a
2a
3a
7.65Æ0.13
3.1Æ0.2
–
2468
2
3
8.00Æ0.35
5.52Æ0.10
112.1Æ10.3
13.7Æ0.5
3.20Æ0.70
71
–
404
5
4
5
4a
5a
6.12Æ0.17
1239Æ81
23.3Æ1.7
[
b]
nd
nd
nd
4
2
[b]
6
7
6a
7a
nd
nd
nd
nd
nd
nd
[b]
1
[
b]
8
9
8a
nd
nd
nd
<0.5
81
9a
4.10Æ0.10
6.23Æ0.33
50.4Æ3.0
102.3Æ10.5
3.13Æ0.63
1.93Æ0.46
1
1
1
0
1
2
10a
61
11a
12a
13a
0.240Æ0.004
0.83Æ0.02
4.70Æ0.4
26.7Æ1.3
27.9Æ1.5
589Æ103
–
9
2.50Æ0.54
30
8
[
[
c,d]
d]
1
1
3
–
[b]
4
14a
nd
nd
nd
<0.5
[
a]
Reactions (0.3 mL) were performed in potassium phosphate buffer (50 mM, KH PO /K HPO , pH 7.0) containing
2
4
2
4
NADPH (100 mM), oxidising substrate (2.5 mM to 50 mM) PETNR (12–2000 mM) and 5% ethanol to solubilise the sub-
strates. The reactions were followed continuously by monitoring NADPH oxidation at 340 nm for 2 min at 258C under
oxygen free atmosphere (<5 ppm).
[
b]
k /K was determined from the slope of the progress curve as is standard when Km values are high compared to sub-
cat
m
strate concentration.
[
[
c]
Reactions with oxidizing substrates measured at 365 nm to avoid substrate spectral overlap.
Calculated from a partial curve due to substrate absorbance at 365 nm, >1.2 mM substrate; nd=not determined due to
the poor solubility of the substrate and/or weak activity with PETNR.
d]
duction in enzyme activity (data not shown). The recent studies show that directed evolution can be
enzyme was inactive for more sterically hindered sub- used to extend the substrate specificity of OYEs to-
[
5,19]
strates such as 3-phenylcyclohexene-1-one.
wards more sterically hindered substrates.
Similar trends have also been seen with other
We noted that substrate inhibition occurred with
[2–4]
OYEs,
which is consistent with the structural simi- many substrates, with K values in the low mM range
i
larity of the enzymes within the family. However, (Table 1; Supporting Information Figure S1a–i). This
2978
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 2976 – 2990

Asymmetric Reduction of Activated Alkenes by Pentaerythritol Tetranitrate Reductase
FULL PAPERS
is a common feature of catalysis by PETNR with both methylcyclohexenone 15a was reduced to its respec-
[
12]
reducing and oxidising substrates.
However, this tive (3S)-enantiomer of 15b, while all 2-methylcyclo-
problem can be solved by using a continuous feed of hexanone derivatives 3–7a were reduced to their re-
substrate, to maintain the substrate concentration at a spective (2R)-enantiomers (Table 2, entries 3–6). The
non-inhibitory level whilst still ensuring the desired presence of the bulky substituent remote to the reac-
productivity. Furthermore, the enzyme concentration tion site did not affect the stereochemical outcome of
can be increased if necessary to compensate for sub- the reaction and the reduction of carvones (5R)-6a
strate inhibition.
and (5S)-7a gave the diastereoisomeric products with
the same absolute configuration on carbon C2 (2R) in
95% and 88% de, respectively. The enantioselectivity
Stereoselectivity of PETNR-Catalyzed Reduction of
a,b-Unsaturated Alkenes
obtained in the reactions with 2-methyl-substituted
cyclohexenones 3a and 5–7a varied from 49 to 98%
ee, whereas 3-methylcyclohexanone 15b was obtained
The mechanism of alkene reduction catalyzed by in optically pure form. Similarly, 3-methylcyclopente-
OYEs is known to proceed via hydride transfer (from none 16a yielded the (3S)-enantiomer of 16b in high
reduced flavin, FMNH ) to the Cb atom of the sub- optical purity (>99% ee), while the reduction of 2-
2
strate and proton transfer (from a conserved tyrosine methylcylopentenone 17a furnished the product 17b
residue, Y186 in PETNR, or from water) to the Ca in moderate ee (70%). Interestingly, the (S)-enantio-
[
11]
position of the substrate.
Hydrogenation of the meric product, rather than the expected (R)-enantio-
C=C bond occurs in a trans-fashion, consistent with mer of 17b was formed. This unexpected change of
the crystal structure of PETNR in complex with cy- enantioselectivity for formation of 2-methylcyclopen-
[
14]
clohexenone 4a (Scheme 1).
The binding of the tanone 17b has also been reported for other OYEs, al-
substrates is dictated by two basic amino acid resi- though no explanation for this observation was
[
2–4]
[5]
dues, His181 and His184 for PETNR, which interact given.
In light of recent work by Stewart et al. it
with the electron-withdrawing group of the substrates is possible that substrate 17a may bind in a ꢀflippedꢁ
and orientate them to allow the hydride transfer to orientation, while maintaining an optimal Cb-FMN
[
21]
proceed. To obtain stereochemical insight into N5 distance and angle (1058), similar to that seen
[
5]
PETNR-catalyzed reactions, we studied the reduction for OYE1 mutant W116I.
of compounds with different activating groups and PETNR proved to be highly efficient and highly
substitution patterns. The assays were performed stereoselective in the reduction of 2-methylmaleimide
anaerobically against a variety of substituted a,b-un- derivatives 1a–b giving the respective (2R)-products
saturated substrates (5 mM substrate; 2 mM PETNR) in quantitative yields and high enantioselectivities
using a glucose 6-phosphate dehydrogenase/NADPH (>99% ee, Table 2 and Scheme 1c). Similar results
[
20]
[2,3]
(
G6PDH) cofactor recycling system
of reducing equivalents (Table 2).
The stereochemical outcome of the bioreductions
can be interpreted by reference to the mechanism of PETNR, although only with low or moderate ee
as the source were also seen with other OYEs,
resulted only in the product (R)-2b at 75% ee.
Aldehydes 10–11a were reduced effectively by
although OYE2
[3]
[
14]
reduction of cyclohexenone 4a (Scheme 1). Thus, 3- (Table 2, entries 7 and 8). It should be noted that for
the aldehydes tested in this study (9–11a) the yields
obtained were not quantitative (50–60%) despite high
observed substrate depletion rates (>95%). We were
unable to identify the side product(s) formed from
these substrates. The corresponding enol products
could not be detected suggesting the side reactions
did not involve carbonyl bond reduction, so non-enzy-
matic substrate and/or product decomposition is
likely.
The low enantiopurity of the aldehydes, as well as
nitroalkenes with a stereogenic centre at Ca can be
attributed to the racemisation of the products in
[
4,22]
aqueous media.
The interpretation of the stereo-
chemical outcome of the reduction of citral 11a [the
citral 11a used in this study was a mixture of geranial
(
E)-11a and neral (Z)-11a in approximately 65:35
Scheme 1. Model of the postulated binding mode and mech-
anism of alkene reduction by PETNR. Panel a: substrates
3
1
ratio (by GC)] and nitroalkenes 18a and 19a is more
–6a and 15a; panel b: substrates 16–17a; panel c: substrates complex, because these substrates isomerise between
–2a; panel d: substrates 10–11a; panel e: substrates 18–19a. their E/Z forms under the reaction conditions.
[23,24]
Adv. Synth. Catal. 2009, 351, 2976 – 2990
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2979

Anna Fryszkowska et al.
FULL PAPERS
+
[a]
Table 2. Reduction of various activated alkenes by PETNR using NADP /G6PDH co-factor regeneration system.
[b]
[b]
[c]
Entry
Substrate
Conf.
Product
t [h]
Conv. [%]
Yield [%]
ee [%]
1
1a
2a
(R)-1b
24
>99
98
99
>99
>99
2
(R)-2b
24
>99
3
4
3a
5a
(R)-3b
(R)-5b
24
24
>99
>99
95
93
49
98
[
[
d,e]
d,e]
5
6
(5R)-6a
(5S)-7a
A
T
N
T
E
N
N
(2R,5R)-6b
24
24
>99
>99
82
95
95%
88%
AHCTUNGTRENNUNG( 2R, 5S)-7b
7
8
10a
11a
(S)-10b
(S)-11b
24
24
>99
>99
85
56
66
87
9
15a
16a
17a
(S)-15b
(S)-16b
(S)-17b
24
24
24
25
14
58
18
14
56
>99
>99
70
1
1
0
1
12
13
14
(E)-18a
(Z)-18a
(E)-19a
(S)-18b
(S)-18b
(S)-19b
24
24
24
96
70
87
40
69
95
64
>99
64
[
a]
Conditions: Reactions contained potassium phosphate buffer (50 mM KH PO /K HPO pH 7.0, 1.0 mL), alkene (5 mM),
PETN reductase (2 mM), glucose 6-phosphate dehydrogenase (10 U), glucose 6-phosphate (15 mM) and NADP (10 mM).
Reactions were agitated at 308C at 130 rpm for 24 h. Alkenes 1a, 3a and 18–19a were added as a DMF solution [2%
2
4
2
4
+
ACHTUNGTRENNUNG( v/v) final DMF concentration].
by GC using DB-Wax column.
By GC or HPLC.
[
[
b]
c]
We previously published data for the reduction of a (2%) to solubilise the substrates in a single phase
library of nitroalkenes by PETNR under biphasic re- aqueous system (Table 2, entries 12–14). Similarly to
[
16]
[16]
action conditions (isooctane/phosphate buffer).
To our previous study, under monophasic reaction con-
obtain a like-for-like comparison with other substrates ditions, the (Z)-isomer of 18a gave the respective (S)-
reported in this paper, we have also repeated the re- product quantitatively in high optical purity (95% ee,
ductions of 2-aryl-1-nitropropenes 18–19a, using DMF compared to 96% in the biphasic system). The reduc-
2980
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 2976 – 2990

Asymmetric Reduction of Activated Alkenes by Pentaerythritol Tetranitrate Reductase
FULL PAPERS
tion of (E)-isomers of nitroalkenes 18a and 19a, fur- erence for the (Z)-isomers of b,b-disubstituted nitro-
nished the same (S)-enantiomer of the products in alkenes 18a and 19a compared with their respective
[
16]
moderate enantioselectivity (69 and 64% ee, com- (E)-counterparts.
Since the reduction of both (E)
pared with 89 and 72% in the biphasic system). More- and (Z)-isomers of nitroalkene 18a lead to the forma-
over, for (E)-nitroolefins 18–19a the yields were not tion of (S)-enantiomer (Table 2, entries 12 and 13),
quantitative and the formation of by-products was ob- one cannot exclude an enzyme-driven shift in equilib-
[
16]
served, as described previously.
The respective 1- rium between the (E) and (Z)-isomers towards the
aryl-2-nitropropanes were obtained under the reac- isomer that binds more strongly to the active site,
tion conditions in almost racemic form (data not leading to an apparent enantioconvergent course of
shown), compared with up to 54% ee reported for the the reduction of both isomers. The kinetic data ob-
[
16]
[16]
biphasic reductions.
tained for nitroolefins
and modelling of both iso-
As the configuration of the newly formed stereo- mers of citral 11a [(E)-geranial and (Z)-neral] into
genic centre at Cb depends on highly stereoselective the active site of PETNR (Supporting Information,
hydride transfer from the flavin (Scheme 1d–e), the Figure S2a–b) suggests that the enzyme prefers a
moderate ees obtained for 11b and 18–19b are likely small substituent at Cb in the position trans to the al-
to be the result of an enantiodivergent course for the dehyde/nitro group (Me vs. Ph for nitroalkenes and
[
23]
reduction of (E) and (Z)-isomers of the substrates.
The kinetic data for nitroalkenes showed a clear pref-
Me vs. alkyl for citral).
+
Table 3. Reduction of activated alkenes by PETNR over 24–72 h hours with NADPH and an NADP /G6PDH co-factor re-
[a]
generation system, respectively.
2
4 h
48 h
ee [%]
72 h
ee [%]
[b]
[b]
[c]
[c]
[c]
Entry
Conf.
Product
Cofactor
NADPH
Conv. [%]
Yield [%]
ee [%]
1
2
>99
>99
>99
>99
93
99
–
>99
–
>99
+
NADP /G6PDH
(
(
R)-1b
R)-3b
3
4
NADPH
>99
>99
>99
>99
21
57
–
26
–
13
+
NADP /G6PDH
5
6
NADPH
NADP /G6PDH
>99
>99
>99
>99
94
96
–
94
–
91
+
(R)-5b
[
b]
7
8
NADPH
NADP /G6PDH
>99
>99
88
81
93% de
94% de
–
–
+
[b]
[b]
[b]
92% de
89% de
ACHTUNGTRENNUNG( 2R,5R)-6b
b
9
1
NADPH
NADP /G6PDH
>99
>99
97
91
81% de
87% de
—
80% de
—
75% de
+
[b]
[b]
[b]
0
ACHTUNGTRENNUNG( 2R,5S)-7b
1
1
1
2
NADPH
NADP /G6PDH
54
53
53
53
>99
>99
–
>99
–
>99
+
(S)-15b
S)-16b
1
1
3
4
NADPH
44
41
41
40
>99
>99
–
>99
–
>99
(
+
NADP /G6PDH
1
1
5
6
NADPH
NADP /G6PDH
>99
>99
>99
>99
65
71
–
65
–
58
+
(S)-17b
[
a]
Conditions for assays with NADPH: alkene (5 mM), PETN reductase (2 mM), NADPH (15 mM) in phosphate buffer
+
(
50 mM, 1 mL total volume), 308C at 130 rpm, 24 h; Conditions for assays with NADP /G6PDH: alkene (5 mM), PETN
+
reductase (2 mM), glucose 6-phosphate dehydrogenase (10 U), glucose 6-phosphate (15 mM) and NADP (10 mM) in
phosphate buffer (50 mM, 1 mL total volume), 308C at 130 rpm, 24 h. Alkenes 1a and 3a were added as a DMF solution
(
2% final DMF solution).
[
[
b]
By GC using a DB-Wax column.
By GC or HPLC. Conv.=% conversion.
c]
Adv. Synth. Catal. 2009, 351, 2976 – 2990
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2981

Anna Fryszkowska et al.
FULL PAPERS
Enantioselectivity in Relation to Reaction Conditions
Table 4. pH dependence for the reduction of ketoisophorone
[a]
3
a with 0.02 mM PETNR.
The bioreduction catalysed by PETNR gave products
with variable degrees of optical purity (49–>99% ee;
Table 2), which was also observed with other
Entry pH Conv.
Yield
(1.5 h)
[%]
ee
(1.5 h)
[%]
ee
(24 h)
[%]
[b]
[b]
[c]
[c]
(
1.5 h)
[%]
[2–4]
OYEs.
As previous studies have shown that the
1
2
3
4
5
6
7
8
9
5.0 2
5.5 5
1
3
>99
>99
>99
>99
99
98
97
95
92
93
93
91
87
79
62
41
13
0
nature of the cofactor or cofactor regeneration system
affects the enantiopurity and even the absolute con-
[
2–4]
6.0 13
6.5 24
7.0 27
7.5 28
8.0 33
8.5 32
9.0 31
12
22
26
27
31
31
30
figuration of the product,
we studied the effect of
+
using NADP /G6PDH as cofactor recycling system
vs. reactions with excess NADPH (Table 3). Both re-
action conditions proved to be equally efficient (infer-
red from similar yields and conversion) and the prod-
ucts had similar optical purities after 24 h. When we
+
extended the reaction time for the NADP /G6PDH
[a]
Conditions: The reactions contained Bis-Tris/HEPPS/
CHES buffer (75 mM/38 mM/38 mM, respectively, 1 mL
total volume), alkene 3a (5 mM), PETNR (0.02 mM), and
NADPH (6 mM). Reactions were agitated at 308C at
system (48–72 h), we observed high optical purity of
the reduction products possessing a stereogenic centre
at Cb (>99% ee for 15–16b). This is in contrast to all
the 2-methyl-substituted cyclohexenones 3b and 5–7b
and 2-methylcyclopenenone 17b, which lost enantio-
purity over time. A similar time-dependent loss of
enantiopurity for these compounds was seen with
other OYEs, although it was not investigated in
1
30 rpm for 1.5 and 24 h. Alkene 3a was added as a
DMF solution (2% final DMF concentration).
By GC using a DB-Wax column.
By GC using Chirasil-DEX-CB column.
[
[
b]
c]
[2,3]
detail.
Since the enzymatic reaction with these sub-
Table 5. Enzyme concentration dependence for the reduc-
tion of ketoisophorone 3a after 2 h.
strates was complete after 24 h, we concluded that the
loss of optical purity was due to racemisation of prod-
ucts with a stereogenic centre at Ca. The apparent Entry Enzyme concentra- Conv.
rate of racemisation varied depending on the nature
of the product. For example, while levodione 3b
racemised quickly, other cyclohexanone derivatives 5–
[a]
[
b]
[b]
[c]
Yield
[%]
ee
tion
[%]
[%]
1
2
3
4
2 mM
>99
>99
37
95
95
33
2
93
96
95
0.2 mM
0.02 mM
0.002 mM
7b racemised at a slower rate. Maleimide 1a gave vir-
tually optically pure products, which were stable on
prolonged incubation (Table 3, entry 2), as it is not
prone to racemisation.
2
>99
[a]
Conditions: alkene 3a (5 mM), PETN reductase,
NADPH (6 mM) in phosphate buffer (50 mM, pH 7.0,
Therefore, we investigated the conditions influenc-
ing the enantiopurity of the product of PETNR-cata-
lysed reductions. As a model compound we chose ke-
toisophorone 7a to levodione (R)-7b, which is an im-
1
mL total volume), 308C at 130 rpm, 2 h. Alkene 3a was
added as a DMF solution (2% final DMF solution).
By GC using a DB-Wax column.
By GC using Chirasil-DEX-CB.
[b]
[c]
[
25]
portant synthon in the synthesis of zeaxanthin. Ini-
tially, we investigated the effect of pH and enzyme
concentration on the optical purity of product (R)-3b assays at pH 7.0 to minimise racemisation whilst ob-
(
Table 4 and Table 5).
taining reasonable reaction rates.
PETNR had a broad pH optimum at basic pH
Reactions carried out at higher enzyme concentra-
values (pH 7.0–9.0), with only low conversion ob- tions yielded a slightly lower optical purity of the
served below pH 6.0 (Table 4). The enzyme possesses product, suggesting an apparent enzymatic effect
up to two histidine residues, which are involved in (Table 5). However, the racemisation process could
[
13,14]
substrate binding (Scheme 1).
ring of histidine is mostly protonated below pH 6.0 emisation within the active site;
pK of imidazole ring 5.97), this may explain the lysed racemisation specific to basic amino acid groups
Since the imidazole be attributed potentially to: (i) enzyme-catalysed rac-
[
23,26]
(ii) enzyme-cata-
(
a
poor activity at lower pH. As expected, reductions located on the enzyme surface (which may involve
catalysed at acidic or neutral pH (5.0–7.0) yielded the either PETNR or G6PDH or both); (iii) influence of
product (R)-3b with high enantiopurity, with lower glucose 6-phosphate or the 6-phosphogluconate
optical purity obtained at pH 7.5–9.0. The pH effect formed during cofactor regeneration, which may de-
was more pronounced if reactions were carried out crease the pH, if buffer capacity is exceeded; (iv)
for 24 h, reflecting a time-dependent loss of enantio- non-enzymatic racemisation. To determine whether
purity (Table 4). Therefore, we conducted further the observed racemisation was purely non-enzymatic
or enzyme-driven, we determined the rate of racemi-
2982
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 2976 – 2990

Asymmetric Reduction of Activated Alkenes by Pentaerythritol Tetranitrate Reductase
FULL PAPERS
sation of product at pH 7.0 over 24 h (Table 6 and mising the reaction time was crucial to obtain high
Figure S3 in Supporting Information). Control reac- enantiopurity of the product. Indeed, when we stud-
tions were performed where after 8 h of incubation ied the time effect on the optical purity of levodione
either the PETNR reductive activity was inhibited by 3b at higher enzyme concentration (2 mM, Supporting
the addition of 1,4,5,6-tetrahydroNADPH (NADPH ; Information, Table S1 and Figure S4), we observed
4
[
27]
an excellent inhibitor of PETNR activity ), or total the same rate of racemisation (3% ee/h) and the high
protein was removed from the reactions by filtration ees (ꢀ95%) were obtained only for the assays con-
and the reactions were incubated for further 16 h. ducted for less then 1 h.
This enabled us to distinguish between the potential
mechanisms of racemisation (Table 6, entries 10–12).
These studies showed that a combination of (i) opti-
mising the solution pH and (ii) reducing the time the
The data showed a decline in enantiopurity of (R)- product spends in aqueous solution (by shortening the
b with time at a rate of approximately 3% ee/h reaction time or using biphasic reaction systems, as
Table 6). The enantiomeric excess of 3b obtained shown before ) is critical to improving the enantio-
3
(
[
16]
after 24 h (30% ee) was virtually the same as those purity of products subject to racemisation. We
obtained in the control experiments (entry 10 vs. 11 achieved complete conversion and obtained optically
and 12), what clearly demonstrates that the racemisa- pure levodione 3b (97% ee) using 2 mM enzyme con-
tion is primarily a non-enzymatic, chemical process. centration at pH 7.0 when conducting the reduction
This suggests that the effect of enzyme concentration for only 30 min. This contrasts with the relatively
on product enantiopurity (Table 5) might be due to poor ees obtained from longer reaction times (Table 1
the balance between the rates of reduction and race- and Table 3). Optimisation in this way may be benefi-
misation processes. At high enzyme concentration, cial to other biocatalytic processes catalysed by diffe
the reaction is completed rapidly, and the product ent enzymes where product racemisation is an issue.
R)-3b is left in the aqueous reaction mixture, allow- For reactions yielding products where racemisation is
ACHTUNGTRENNUNGr -
(
ing it to racemise (Table 5, entry 1). At low enzyme negligible, optimisation studies should be more fo-
concentrations (entry 4), the substrate is reduced cused on optimising enzyme activity and/or increasing
much more slowly, and the freshly formed (R)-enan- substrate/product solubility and media engineering.
tiomer was still accumulating when the reaction was
terminated after 2 h. Thus, the continued synthesis of
(
R)-3b would compensate for the racemisation only Models of Substrate-Bound PETNR and NMR
to a small extent. Therefore we concluded that mini- Studies on the Stereochemistry
Kinetic data for the compounds possessing two acti-
vating groups 1–3a suggest that they are especially
good substrates for PETNR-mediated reduction
Table 6. The effect of time on the reduction of ketoisophor-
one 3a and product racemisation catalysed by PETNR
[a]
(
0.02 mM).
(
Table 1). To gain a greater understanding of the ste-
[b]
[b]
[c]
Entry
Time [h]
Conv. [%]
Yield [%]
ee [%]
reochemical outcome of the reactions, we modelled
the potential binding modes of maleimide 2a and ke-
toisophorone 3a in the active site of PETNR
1
2
3
4
5
6
7
8
9
0.5
1
1.5
3
4
6
8
12
22
14
27
37
53
69
85
94
98
12
23
35
53
65
80
87
88
91
95
nd
nd
>99
99
98
93
90
82
76
69
33
30
30
29
(
Figure 1 a and b). We maintained a similar relative
position and orientation of the C=C and C=O bonds
(in red) as in the structure of PETNR containing
bound 2-cyclohexenone 4a as well as a near optimal
[
14,16]
Cb-FMN N5 distance and orientation.
Both models predict that the substrates bind to the
active site of PETNR on the si-face of the isoalloxa-
zine ring of the FMN, with the carbonyl group O2
within hydrogen bonding distance of both H181 NE2
and H184 ND1 as seen in the 2-cyclohexenone and
>99
>99
99
10
11
12
24[
d]
24[
e]
24
99
[
a]
[
13,14]
Conditions: alkene 3a (5 mM), PETNR (0.02 mM),
progesterone-bound structures.
In the case of the
NADPH (6 mM) in phosphate buffer (50 mM, pH 7.0,
model for maleimide 2a binding (Figure 1b), the
second carbonyl oxygen O5 is within hydrogen bond-
ing distance of the OG1 atom of Y351. This residue
(Y375 in OYE1) is known to be involved in binding
the inhibitor, p-hydroxybenzaldehyde, via a hydrogen
bond between atom OG1 and the inhibitor carbonyl
1
mL total volume), 308C at 130 rpm, 2 h. Alkene 3a was
added as a DMF solution (2% final DMF solution).
By GC using a DB-Wax column.
By GC using a Chirasil-DEX-CB column.
Filtered using 10 K millipore filter after 8 h to remove all
enzyme content.
[
[
[
b]
c]
d]
[
17,28]
[
e]
PETNR inhibitor NADPH (1.2 mg) added after 8 h to atom O1’.
This suggests a possible role of Y351 in
4
inhibit enzymatic activity.
the binding and correct orientation of some substrates
Adv. Synth. Catal. 2009, 351, 2976 – 2990
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2983

Anna Fryszkowska et al.
FULL PAPERS
and Cb, is predicted to be less favoured due to a sig-
nificant clash between the substrate C12 atom and
W102 (results not shown).
The model of ketoisophorone 3a (Figure 1a) shows
a closer match in both the position and orientation to
[
14]
the known position of cyclohexene 4a, resulting in
a change in orientation of the substrate with respect
to the model of 4a. This results in the absence of an
interaction between the second carbonyl group (O7)
and Y351 OH. A model of 2a in the same relative ori-
entation as in 4a would be less favoured as there
would be severe clashes between the 2’- and 6’-methyl
groups with residues W102 and Q241. In this position
the OH atom of the putative proton donor Y186 is
about 3.2 ꢃ from both the Ca and Cb atoms of the
substrate.
Interestingly, both the binding modes of the sub-
strates 1–3a should result in the same stereochemical
outcome for reduction, leading to the formation of
the (R)-enantiomer, as depicted in Scheme 2 for mal-
eimides 1–2a. This makes it more difficult to predict
which of the different possible binding mode/s of
these substrates are catalytically competent and which
catalytic step determines the stereochemistry: proto-
nation or hydride addition.
Scheme 2. Stereochemical course of the reduction of male-
AHCTUNGTRENNUGNi mides 1–2a by PETNR (R=Ph or H). Two possible binding
modes of 1–2a give the same stereochemical outcome: a)
hydrogen at new stereogenic centre (C2) is derived from
protonation step; b) from hydride addition step.
Figure 1. Models of the active site of PETNR containing (a)
ketoisophorone 3a and (b) 2-methylmaleimide 2a overlaid
with the known position of 2-cyclohexenone 4a. The posi-
tion of 4a is derived from a superimposition of the 2CH-
PETNR structure onto the model (pdb code 1GVQ). All
residues are shown as atom-coloured sticks with green,
[14]
To investigate further the binding modes of the sub-
yellow, magenta and blue carbons for amino acids, FMN, strates in PETNR, we envisaged using NMR to deter-
substrate 3a or 2a and 4a, respectively. Predicted hydrogen
mine whether the hydrogen at the newly formed ste-
reogenic centre of the products came from either the
FMN by hydride transfer or from a water-mediated
protonation step. Unfortunately, the reduction of the
substrate 3a by PETNR in deuterated phosphate
bonds are indicated by red dotted lines. The figures were
[29]
generated in Pymol.
within the active site. A minor clash between the C12 buffer (pD 7.0), with NADPH as a cofactor lead to a
atom (methyl group at Cb) and T26 OG1 was also complete exchange of all the ring protons to deuteri-
present in this model. A second model where the sub- um. This clearly established significant proton ex-
strate was rotated 1808, swapping the positions of Ca change for levodione 3b, which is easily enolisable.
2984
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 2976 – 2990

Asymmetric Reduction of Activated Alkenes by Pentaerythritol Tetranitrate Reductase
FULL PAPERS
1
2
Figure 2. H NMR (400 MHz) analysis of N-phenyl-2-methylsuccinamide (1b): a) reduction in the presence of (R)-[4- H]-
NADPH (NADPD) in phosphate buffer; b) reduction in the presence of NADPH in deuterated phosphate buffer; c) a
parent non-deuterated compound.
Fortunately, the racemisation of succinimides is negli-
These results suggest that, while both of the postu-
gible so we performed two parallel bioreductions of lated binding modes are possible, the orientation de-
maleimide 1a with PETNR using either NADPH in picted in Scheme 2b is favoured and the proton at the
deuterated phosphate buffer pD 7.0, or pro-R newly formed stereogenic centre C2 is derived mostly
NADPD in phosphate buffer pH 7. In both experi- from the hydride transfer step. This agrees with our
ments deuterium was found at C2 and C3 positions, modelling studies with substrate 2a, which predicts
but the level of incorporation biased either toward that hydride transfer occurs predominately on the C2
HA or H (Figure 2). The retention of geminal cou- atom (Figure 1b).
C
pling constant JAB (about 18 Hz) and the complete
loss of the vicinal coupling constant JAC (about 4 Hz)
present in the spectrum of a parent, non-deuterated Conclusions
compound (Figure 2, bottom spectrum), suggests that
the total yield deuterium incorporation between these We have shown that PETNR has excellent potential
2
positions was 100%, however, unequally distributed. as a biocatalyst due to its ability to catalyse the asym-
The NMR analysis of the product 1b obtained in deu- metric bioreduction of a variety of industrially rele-
terated buffer showed approximately 65% incorpora- vant a,b-unsaturated activated alkenes in a stereose-
tion of deuterium at the C3 position (D ) and approx- lective manner. It is an efficient catalyst for the reduc-
A
imately 35% at the position C2 (D , by NMR integra- tion of a variety of enones, enals, maleimides and ni-
C
tion, Figure 2 middle spectrum). In contrast, the reac- troalkenes. It terms of stereoselectivity it is similar to
[
4]
tion where NADPD served as reducing cofactor led other OYEs such as OPR3 from tomato and NCR-
[
3]
primarily to deuterium incorporation at the C2 posi- reductase from Zymomonas mobilis. This provides
tion (70% of D ), whereas 30% was incorporated at further evidence for the catalytic promiscuity of
C
C3 (Figure 2 top spectrum).
PETNR, which was originally isolated by its ability to
Adv. Synth. Catal. 2009, 351, 2976 – 2990
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2985

Anna Fryszkowska et al.
FULL PAPERS
reduce nitroaromatic explosives via the formation of dising substrate reduction. All the solutions were made
[
15,30]
anaerobic and the reactions were set-up in the oxygen free
the Meisenheimer-hydride complex.
environment (<5 ppm of O ).
Enzymes obtained from natural sources rarely pos-
sess the combined properties necessary for industrial
applications such as high activity with non-natural
substrates, and more-importantly high stereoselectiv-
ity. Our study demonstrates that optimisation of the
reaction conditions such as pH, reaction time and mosphere (<5 ppm oxygen). The following extinction coef-
2
Steady-State Kinetic Analysis
All kinetics measurements were performed within an anae-
robic glove box (Belle Technology Ltd) under a nitrogen at-
[
31]
enzyme concentration appears to be critical in signifi- ficients were used to calculate the concentration of NADPH
À1
3
À1
cantly increasing the enantiopurity of products which and enzyme: NADPH
(e340 =6220M cm ;
e =
365
À1
À1
À1
À1
3
256M cm ) and PETNR (e =11.3ꢄ10 M cm ). The
are subject to racemisation under the reaction condi-
tions. We were able to dramatically improve the enan-
tiopurity of the product (R)-levodione 3b from 49%
to 97% ee by simply increasing the enzyme concentra-
tion and reducing the reaction time. Such optimisa-
tion studies could be broadly applied to biotransfor-
mations of other substrates by OYEs, and to other re-
464
following extinction coefficients were used to calculate the
concentration of the alkene substrates 1a (e221
=
=
=
À1
À1
À1
À1
À1
À1
À1
À1
À1
À1
2
2
1
8
1200M cm ); 2a (e221 =19100M cm ); 3a (e246
1100M cm ); 4a (e231 =11000M cm ); 5a (e
242
À1
À1
À1
0450M cm ); (5R)-6a and (5S)-7a (e =7600M cm );
2
42
À1
À1
À1
À1
À1
a (e223 =14100M cm ); 9a (e =13800M cm ); 10a
237
À1
À1
À1
À1
(
e234 =14600M cm ); 11a (e244 =15100M cm ); 12a
À1
À1
action types where product racemisation is a problem. (e288 =24800M cm ); 13a (e =24100M cm ); and 14a
292
À1
À1
À1
À1
Our work emphasises the importance of using a (e292 =23200M cm ); 15a (e241 =14700M cm ); 16a
À1
À1
À1
À1
hybrid approach, combining studies of both substrate (e231 =16200M cm ); 17a (e232 =10900M cm ).
Steady-state kinetic measurements were performed anae-
specificity/enantioselectivity with analysis of models
robically using a BioTek Synergy HT microtiter plate reader
for the target enzyme-substrate complex whilst identi-
with a 0.84 cm path length. Reactions (0.3 mL) were per-
fying and screening industrially useful biocatalysts.
formed in buffer (50 mM KH PO /K HPO pH 7.0) contain-
ing NADPH (100 mM), oxidising substrate (2.5 mM to
0 mM) PETNR (12–2000 mM) and 5% ethanol to solubilise
the substrates. The reactions were followed continuously by
monitoring NADPH oxidation at 340 nm for 2 min at 258C.
2
4
2
4
Our approach has allowed us to rationalise the sub-
strate specificity of PETNR and to predict the impor-
tance of Y351 in substrate binding of double-activated
molecules such as 1–3a. Modelling approaches should
5
reduce the burden of performing extensive substrate Reactions with oxidising substrates that absorbed at 340 nm
screens to identify compounds that are accommodat- were monitored at 365 nm. To determine the kinetic param-
ed in the enzyme active site as well as introducing eters with each substrate (apparent K , kcat and K), the re-
m i
actions were performed with variable oxidising substrate
concentrations. For substrates that were reduced by PETNR
at very low rates, only the k /K parameter could be deter-
new functional groups into target substrates. Structu-
ACHTUNGTRENNUNGr al modelling approaches also suggest strategies to
cat
m
generate new improved biocatalysts by directed muta-
genesis. Structure-based identification and design of
biocatalysts complements other powerful laboratory
evolution approaches for the creation of new biocata-
lysts that are not necessarily dependent on structural
mined using substrate concentrations (10–40 mM) well below
the estimated K . Initial velocity data as a function of oxi-
m
dising substrate were analysed by fitting to either the stan-
dard Michaelis–Menten rate equation, or to an equation
that incorporates substrate inhibition [Eq. (1)].
[
5,19,32]
knowledge of the enzyme.
Experimental Section
General Enzymatic Procedures
Aerobic reactions were performed as above, except all
solutions were prepared, and all data collected outside the
glove box. For all microtiter plate kinetic assays, a minimum
of 6 replicates was generated per data point to ensure data
accuracy and reproducibility.
Glucose 6-phosphate and glucose 6-phosphate dehydrogen-
ase (G6PDH) from Leuconostoc mesenteroides, KH PO ,
2
4
K HPO ,
CHES
acid], Bis-Tris {1,3-bis
propane} and HEPPS [N-(2-hydroxyethy1)piperazine-N’-
{2-(N-cyclohexylamino)ethanesulfonic
2
4
A
H
U
G
R
N
U
G
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
propanesulfonic acid] were obtained from Sigma Aldrich.
+
NADP and NADPH were obtained from Melford. (R)-
General Procedures for the Reduction of Alkenes by
PETNR
2
[
4- H]-NADPH (NADPD; >95% deuterium incorporation
1
by H NMR) and the inhibitor 1,4,5,6-tetrahydroNADPH
(
NADPH ) were kindly supplied by Dr. Christopher R.
General procedure 1: Standard reactions (1.0 mL) were per-
4
[13]
Pudney. PETNR was prepared as described previously.
formed in buffer (50 mM KH PO /K HPO pH 7.0) contain-
2
4
2
4
Prior to use, PETNR was deoxygenated by passage through
a BioRad 10DG column equilibrated in anaerobic reaction
buffer to avoid unproductive flavin reoxidation during oxi-
ing alkene (5 mM), NADPH (15 mM) and PETNR (2 mM).
The reactions were shaken at 308C at 130 rpm followed by
reaction termination by extraction with ethyl acetate
2986
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 2976 – 2990

Asymmetric Reduction of Activated Alkenes by Pentaerythritol Tetranitrate Reductase
FULL PAPERS
(
0.9 mL) containing an internal standard (limonene, 5%
85% yields by silica gel-assisted reduction of their respective
v/v), and dried using MgSO . The extracts were analysed by
nitroalkenes by NaBH according to as described in the lit-
4
4
[33]
GC or HPLC to determine the yield, % conversion, and en-
antiomeric excess. Products were identified by comparison
with authentic reference material. Maleimide 1a ketoiso-
phorone 3a, and nitroalkenes 18--19a were added as a con-
centrated solution in 2% DMF.
erature. Model optically active nitroalkanes (R)-18b and
[24]
(R)-19b were synthesised as described before. 3-Phenyl-2-
cyclohexene-1-one was synthesised as described previous-
[34]
ly.
The rest of the compounds were synthesised as follows:
General Procedure 2: Reactions employing a NADPH re-
cycling system (1.0 mL) were performed in buffer (50 mM
KH PO /K HPO pH 7.0) containing alkene (5 mM; dis-
N-Phenyl-2-methylmaleimide (1a) was synthesised from
citraconic anhydride and aniline as pale yellow needles, ac-
[
35]
2
4
2
4
cording to the procedure described previously;
60%; mp 98–1008C (lit. mp 96–988C);
yield:
[35]
1
solved as above), glucose 6-phosphate dehydrogenase
H NMR
+
(
10 U), glucose 6-phosphate (15 mM), NADP (10 mM) and
PETNR (2 mM). The reactions were shaken at 308C at
30 rpm followed by extraction with ethyl acetate (0.9 mL)
(400 MHz, CDCl ): d=2.15 (d, J=1.8 Hz, 3H), 6.46 (q, J=
3
1
3
1.8 Hz, 1H), 7.16–7.31 (m, 5H); C NMR (100 MHz,
1
CDCl ): d=11.2., 126.0, 127.5, 129.12, 131.7, 145.8, 169.6,
3
containing an internal standard (limonene, 5% v/v) and ana-
lysed as described above.
General Procedure 3: To determine the effect of pH, all
reactions (1.0 mL) were performed in a broad pH range
buffer mixture (75 mM Bis-Tris, 38 mM HEPPS, 38 mM
CHES, pH 5.0–9.0) containing alkene (5 mM), NADPH
1
70.7; anal. calcd. for C H NO : C 54.05, H 4.54, N 12.61;
5 5 2
found: C 54.31, H 4.57, N 12.48. All spectral data were in
[35]
accordance with those reported in the literature.
R)-N-Phenyl-2-methylsuccinimide (1b) was synthesised
enzymatically at 50 mL scale according to general procedure
(
2
1
7
(36 h) as a white solid; yield: 98%; mp 133–1348C (lit. mp
(
6 mM) and PETNR (20 nM to 2 mM). The reactions were
[36] 1
32–1348C);
H NMR (500 MHz, CDCl ): d=1.45 (d, J=
3
agitated at 308C at 130 rpm followed by extraction with
ethyl acetate (0.9 mL) containing an internal standard and
analysed as described above.
General Procedure 4: This procedure is to determine if
the physical presence of the enzyme influences the rate of
racemisation of products (i.e., if basic amino acid residues
influence the racemisation rate independent of reductive ac-
tivity). After 8 h of biotransformations with ketoisophorone
.2 Hz, 3H), 2.50 (dd, J=4.3 Hz, J=17.6 Hz, 1H), 2.98–3.13
13
(
m, 2H), 7.26–7.50 (m, 5H); C NMR (100 MHz, CDCl ):
3
d=16.9, 34.8, 36.6, 126.4, 128.6, 129.1, 131.9, 175.4, 179.5;
[37]
[
>
a] : +9.90 (c 1.01, CHCl ) lit. +6.6 (c 0.56, CHCl );
D 3 3
99% ee (Chiralcel OD: hexane/i-PrOH, 9:1, t =53.5 min,
R
t =63.0 min); anal. calcd. for C H NO : C 69.83, H 5.86, N
s
11 11
2
7
.40; found: C 69.95, H 5.90, N, 7.28. A respective racemic
[2]
compound 1b was synthesised as described before as white
crystals; yield: 62%; mp 106–1078C (lit. mp 1078C, etha-
3a: a) the enzyme was removed from the reaction via centri-
fugation (5000 rpm for 5 min) in a 10 kDa milipore filter
[38]
nol). All spectral data were in accordance with those re-
spin column; b) PETNR inhibitor NADPH was added. The
[2]
4
ported in the literature.
reactions were further incubated and analysed after 24 h
2
-Methylmaleimide (2a) was synthesised according to the
(
total time) as described above.
[2,39]
procedure described previously; yield: 5%;
068C (lit. mp 1058C); H NMR (400 MHz, CDCl ): d=
mp 104–
[2] 1
1
2
1
3
.11 (d, J=1.6 Hz, 3H), 6.34 (q, J=1.8 Hz, 1H), 7.4 (bs,
Chemistry
H); anal. calcd. for C H NO : C 54.05, H 4.54, N 12.61;
5
5
2
found: C 54.31, H 4.57, N 12.48. All spectral data were in
NMR spectra were recorded on 400 MHz or 500 MHz spec-
trometers and referenced to the solvent, unless stated other-
wise. The chemical shifts are reported in ppm and coupling
constants (J) are given in hertz (Hz). Melting points were
determined using an electrothermal capillary apparatus and
are uncorrected. IR spectra were recorded neat on NaCl
plates. HPLC analysis was performed using an instrument
equipped with a UV detector. UV-visible data were record-
ed with a diode array spectrophotometer. The progress of
the reactions was monitored by GC or TLC on standard
silica gel plates. All chemicals obtained from commercial
sources and the solvents were of analytical grade. Citraconic
anhydride, substrates 3–12a, trans-cinnamide, trans-cinnamic
acid, methyl trans-cinnamate, trans-cinnamonitrile, 3-phenyl-
maleic acid, 1-cyclohexene-1-carboxylic acid, methyl 1-cyclo-
hexene-1-carboxylate, crotonitrile, 3-bromocyclohexene, 2-
cinnamonitrile, cyclohexene-1-carbonitrile, b-bromostyrene,
racemic products 3–12a, (R)-(+)-3-methylcyclopentanone
[35]
accordance with those reported in the literature.
R)-2-Methylsuccinimide (2b) was synthesised enzymati-
cally at 50 mL scale according to general procedure 2 (36 h)
(
[
40]
as a white solid; yield: 95%; mp 668C (lit. mp 628C);
1
H NMR (400 MHz, CDCl ): d=1.33 (d, J=7.2 Hz, 3H),
3
2
.33–2.39 (m, 1H), 2.90–2.98 (m, 2H), 8.9 (bs, 1H);
1
3
C NMR (100 MHz, CDCl ): d=16.4, 36.1, 37.5, 177.2,
3
[2]
1
81.4; >99% ee (by Chiral GC using Chiraldex B-TA);
[40]
D 3 D 3
[a] :+24.8 (c 1.60, CHCl ) lit. [a] : +29.4 (c 2.0, CHCl );
anal. calcd. for C H NO : C 53.09, H 6.24, N 12.38; found:
5
7
2
C 53.47, H 6.47, N 11.83. A respective racemic compound
2b was synthesised from maleimide 2a as a white powder;
[41]
yield: 98%; mp 68–708C (lit. mp 66–68.58C). All spectral
data were in accordance with those reported in the litera-
[2]
ture.
(R)-Levodione (3b) was synthesised from ketoisophorone
according to general procedure 2 (36 h) as a white solid;
yield: 90%; mp 68–708C [lit. mp 65–678C for racemate, 91–
[(R)-7b], (R)-(+)-3-methylcyclohexanone [(R)-6b], (S)-cit-
[25]
ronellal [(S)-11b] were from commercial suppliers (Aldrich,
Avocado). (+)-Dihydrocarvone was obtained from a com-
mercial supplier as a mixture of 2 isomers (2R,5R)-6b and
928C for (R)-enantiomer]; 27% ee (by Chiral GC on Chir-
[2] 1
asil-DEX-CB column);
H NMR (400 MHz, CDCl ): d=
3
1.12 (s, 3H), 1.14 (d, J=6.6 Hz, 3H), 1.16 (s, 3H), 2.34 (dd,
J=12.8 Hz, J=17.8 Hz, 1H), 2.52 (dd, J=0.7 Hz, J=
(
(
ly.
2R,5S)-6b in 77:20 ratio (>99% ee). Nitroalkenes (E)-18a,
Z)-18a and (E)-19a were synthesised as described previous-
15.4 Hz, 1H), 2.72–2.78 (m, 2H), 2.96–3.05 (m, 1H);
[
16,24]
13
Racemic nitroalkanes 18–19b were obtained in 80–
C NMR (100 MHz, CDCl ): d=14.6, 25.6, 26.5, 39.8, 44.2,
3
Adv. Synth. Catal. 2009, 351, 2976 – 2990
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2987

Anna Fryszkowska et al.
FULL PAPERS
4
4.8, 52.7, 208.0, 214.1; [a] : À64.6 (c 0.61, MeOH) lit. À270
enantiomeric excess was determined using a Chiraldex B-
TA column (40 m, 0.25 mm), as described previously. The
D
[25]
[2]
(
c 0.4, MeOH); anal. calcd. for C H O : C 70.10, H 9.15;
9
14
2
found: C 70.31, H 9.43. All spectral data were in accordance
absolute configuration of levodione 3b was assigned by a
comparison of its optical rotation with literature data and
[42]
[25]
with those reported in the literature.
-Methyl-2-cylohexene-1-one (5a) was synthesised from
cyclohexenone 4a in a 3-step synthetic strategy, as described
2
the ee was determined using a CP-Chirasil-DEX-CB column
[2]
(25 m, 0.32 mm, 0.25 mm film), as described previously. En-
antiomeric excesses of ketones 5b and 15–17b were deter-
mined using a Chiraldex B-TA column (40 m, 0.25 mm), as
[43,44] 1
before.
H NMR (400 MHz, CDCl ): d=1.66–1.78 (m,
3
3
2
1
H), 1.95–2.02 (m, 2H), 2.29–2.33 (m, 2H), 2.40–2.44 (m,
1
3
[2]
H), 6.72–6.76 (m, 1H); C NMR (100 MHz, CDCl ): d=
described before, and the absolute configurations were as-
3
6.0, 23.3, 26.0, 38.3, 135.7, 145.6, 200.0.
signed by a comparison with authentic samples of (R)-5b
and (R)-17b obtained commercially. The absolute configura-
tions of ketones 15b and 16b were assigned by comparison
of the elution order on a Chiraldex B-TA column with the
ACHTUNGTRENNUNG( 2R,5R)- Dihydrocarvone (6b) was synthesised enzymati-
cally at 50 mL scale according to general procedure 2 (1.5
1
days) as a transparent oil; yield: 90%; H NMR (400 MHz,
[2]
CDCl ): d=1.04 (d, J=6.5 Hz, 3H), 1.36 (dq, J=3.5 Hz, J=
literature data. The absolute configurations of (2R,5R)-di-
hydrocarvone (2R,5R)-6b obtained from (5R)-carvone (5R)-
3
1
1
4
3
3.1 Hz, 1H), 1.57–1.70 (m, 1H), 1.71–1.73 (m. 3H),1.90–
.96 (m, 1H), 2.08–2.14 (m, 1H), 2.23–2.45 (m, 4H), 4.70–
6a and (2R,5S)-dihydrocarvone (2R,5S)-7b obtained from
13
[45]
.75 (m, 2H); C NMR (100 MHz, CDCl ): d=14.3, 20.5,
(S)-carvone (5S)-7a were assigned by NMR
and com-
3
0.7, 34.9, 44.7, 46.8, 47.0, 109.6, 147.6, 212.7; >99% ee by
pared with an authentic sample of (2RS,5R)-dihydrocarvone
6b obtained commercially as a mixture of trans:cis isomers
77:20 [by NMR and GC on DB-wax column: split 20, flow
1
GC, 96% de by H NMR All spectral data were in accord-
[45]
ance with those reported in the literature for
-Methylcyclohexanone (15b) was synthesised from the
respective 3-methyl-2-cylohexene-1-one 15a as described for
ACHTUNGTRENNUNG( 2S,5S)-6b.
À1
3
A
H
U
G
R
N
U
G
1 mLmin , injector: 2208C, detector: 2508C, temperature
À1
programme: 408C hold for 2 min, to 2108C at 158Cmin ,
1
17b as a transparent oil; yield: 77%; H NMR (400 MHz,
hold for 3 min; retention times: (2R,5R)-6b and (2S,5R)-6b
11.31 and 11.46 min, respectively]. Enantiomeric excesses of
dihydrocarvones 6–7b were determined using a Chiraldex B-
TA column (40 m, 0.25 mm column): split 100, injector
CDCl ): d=1.02 (dd, J=1.8 Hz, J=6.3 Hz, 3H), 1.28–1.38
3
1
3
(
m, 1H), 1.60–1.80 (m, 2H), 1.81–2.52 (m, 4H); C NMR
(
100 MHz, CDCl ): d=22.1, 25.3, 33.3, 34.2, 41.1, 50.0,
3
À1
2
12.1. All spectral data were in accordance with those re-
1808C, detector 2008C, flow 1 mLmin , 808C, hold 2 min,
[46,47]
À1
ported in the literature.
to 1108C at 108Cmin , hold 20 min, to 1208C at
À1
2
-Methylcyclopentanone
A
H
U
G
E
N
N
(17b) was synthesised from the
108Cmin , hold 10 min, to 1808C at 108C hold 1 min. Re-
respective 2-methyl-2-cylopetene-1-one 17a. Hydrogen was
purged through the solution of alkene (2 mmol) and Pd/C
tention times: (2R,5R)-6b 30.59 min, (2S,5R)-6b 31.71 min,
(2R,5S)-7b 32.88 min, (2S,5S)-7b 33.95 min. The enantiomer-
ic excess of 2-methylpentanal 10b was determined using a
Hydrodex-b-TBDAc column (25 m, 0.25 mm): split 100, flow
(
10%, 10 mg) in ethyl acetate or methanol (12 mL) until the
reaction was completed as detected by TLC. The reaction
mixture was filtered through a Celite plug and the solution
was evaporated under vacuum. Solid products were purified
by chromatography (hexane:ethyl acetate). Liquid products
À1
1 mLmin , injector: 1808C, detector: 2508C, temperature
À1
programme: 808C hold for 10 min, to 1208C at 48Cmin ,
À1
hold for 2 min, to 1808C at 108Cmin , hold for 2 min; re-
1
were vacuum-distilled (bulb-to-bulb). Yield: 20% : H NMR
tention times: (R)- and (S)-2-methylpentanal 10.86 and
11.36 min, respectively. The absolute configuration of (S)-
10b was assigned after reduction of aldehyde to primary al-
(
(
(
400 MHz, CDCl ): d=1.02 (d, J=7.0 Hz, 3H), 1.43–1.53
3
m, 1H), 1.73–1.83 (m, 1H), 1.92–2.05 (m, 1H), 2.07–2.17
1
3
m, 2H), 2.20–2.36 (m, 2H); C NMR (100 MHz, CDCl ):
cohol with NaBH and comparison of the chromatographic
3
4
d=14.2, 20.6, 31.9, 37.6, 44.0, 221.0. All spectral data were
in accordance with those reported in the literature.
data on
0.25 mm) as published previously.
excess of citronellal 11b was determined using a Hydrodex-
b-TBDAc column (25 m, 0.25 mm) and the absolute con-
figuration was assigned by comparison of the chromatogram
with an authentic sample of (S)-11b obtained commercially
a chiral column (Hydrodex-b-TBDM, 25 m,
[46,48]
[2,4]
The enantiomeric
[2]
Analytical Procedures
Determination of yield and % conversion: GC analysis of
conversion and yields was performed on a 30 m DB-Wax
(
Aldrich). The absolute configurations of nitroalkanes 18–
9b were determined by a comparison with literature data
1
(
0.32 mm, 0.25 mm) column, unless stated otherwise. Yield
[16,24]
and authentic samples obtained by us previously.
enantiomeric excess of nitroalkane 18b was determined
using CP-Chirasil-DEX-CB column (25 m, 0.32 mm,
.25 mm film) as described previously. The enantiomeric
excess of nitroalkane 19b was determined using a Chiralcel
The
and % conversion of the reduction of 2-methylmaleimide 2b
were determined using a Chiraldex B-TA column (40 m,
a
0.25 mm).
[2]
0
Determination of enantiomeric excess and absolute con-
figuration: The absolute configuration of N-phenyl-2-me-
thylsuccinimide 1b was determined by a comparison of its
[16,24]
OD column, respectively, as described previously.
The
ACHTUNGTRENNUNG
absolute configurations of nitroalkanes 18–19b were as-
[37]
optical rotation with literature data while the enantiomer-
ic excess was determined by HPLC using a Chiralcel OD
column [hexane/i-PrOH, 9:1, retention times: (R)- and (S)-
maleimide 20.0 min and 22.0 min, respectively] or a Chiral-
cel OJ column [hexane/i-PrOH, 9:1, (R)- and (S)-maleimide
[16,24]
signed by comparison with previously published data.
Molecular Modelling
5
3.4 min and 60.8 min, respectively]. The absolute configura-
The substrates 2a and 3a were modelled into the active site
[50]
[49]
tion of 2-methylsuccinimide 2b was determined by compar-
ing the optical rotation value with literature data and its
of PETNR using the methods of Sketcher
and Coot,
[40]
based on the position of the substrate in the 2-cyclohexe-
2988
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 2976 – 2990

Asymmetric Reduction of Activated Alkenes by Pentaerythritol Tetranitrate Reductase
FULL PAPERS
[
14]
none-bound structure (pdb code 1 gvq). The models were
refined using the structural idealisation functions of the pro-
[18] F. Amari, A. Fettouche, M. A. Samra, P. Kefalas, S. C.
Kampranis, A. M. Makris, J. Agric. Food Chem. 2008,
56, 11740–11751.
[19] S. Kille, D. J. Bougioukou, M. T. Reetz, in: Biotrans,
Berne, Switzerland, 2009.
[51]
gramme REFMAC to normalise the bond angles and dis-
tances of the substrates and minimise clashes.
[
[
[
[
[
[
[
20] Y. Meah, V. Massey, Proc. Natl. Acad. Sci. USA 2000,
7, 10733–10738.
21] M. W. Fraaije, A. Mattevi, Trends Biochem. Sci. 2000,
5, 126–132.
22] H. Ohta, N. Kobayashi, K. Ozaki, J. Org. Chem. 1989,
4, 1802–1804.
23] A. Muller, B. Hauer, B. Rosche, Biotechnol. Bioeng.
007, 98, 22–29.
9
Acknowledgements
2
This work was funded by the UK Biotechnology and Biologi-
cal Sciences Research Council. NSS is a BBSRC Professorial
Research Fellow and GS is a BBSRC Research Development
Fellow.
5
2
24] A. Fryszkowska, K. Fisher, J. M. Gardiner, G. M. Ste-
phens, J. Org. Chem. 2008, 73, 4295–4298.
25] H. G. W. Leuenberger, W. Boguth, E. Widmer, R. Zell,
Helv. Chim. Acta 1976, 59, 1832–1849.
26] D. A. Case, I. Cheatham, T. E. , T. Darden, H. Gohlke,
R. Luo, J. Merz, K. M. , A. Onufriev, C. Simmerling, B.
Wang, R. Woods, J. Comput. Chem. 2005, 26, 1668–
References
[
1] J. F. Chaparro-Riggers, T. A. Rogers, E. Vazquez-Fig-
ueroa, K. M. Polizzi, A. S. Bommarius, Adv. Synth.
Catal. 2007, 349, 1521–1531.
1
688.
27] C. R. Pudney, S. Hay, N. S. Scrutton, FEBS J. 2009, 276,
780–4789.
[
[
2] M. Hall, C. Stueckler, H. Ehammer, E. Pointner, G.
Oberdorfer, K. Gruber, B. Hauer, R. Stuermer, W.
Kroutil, P. Macheroux, K. Faber, Adv. Synth. Catal.
4
[
[
28] K. M. Fox, P. A. Karplus, Structure 1994, 2, 1089–1105.
29] W. L. DeLano, DeLano Scientific, Palo Alto, CA,
USA, 2002.
2
008, 350, 411–418.
[
[
3] M. Hall, C. Stueckler, B. Hauer, R. Stuermer, T. Frie-
drich, M. Breuer, W. Kroutil, K. Faber, Eur. J. Org.
Chem. 2008, 1511–1516.
4] M. Hall, C. Stueckler, W. Kroutil, P. Macheroux, K.
Faber, Angew. Chem. 2007, 119, 4008–4011; Angew.
Chem. Int. Ed. 2007, 46, 3934–3937.
[
30] H. Khan, T. Barna, R. J. Harris, N. C. Bruce, I. Barsu-
kov, A. W. Munro, P. C. Moody, N. S. Scrutton, J. Biol.
Chem. 2004, 279, 30563–30572.
[
31] L. G. J. L. Stephan Luetz, Biotechnol. Bioeng. 2008,
1
01, 647–653.
[
32] P. F. Mugford, U. G. Wagner, Y. Jiang, K. Faber, R. J.
Kazlauskas, Angew. Chem. 2008, 120, 8912–8923;
Angew. Chem. Int. Ed. 2008, 47, 8782–8793.
[5] S. K. Padhi, D. J. Bougioukou, J. D. Stewart, J. Am.
Chem. Soc. 2009, 131, 3271–3280.
6] C. Stueckler, M. Hall, H. Ehammer, E. Pointner, W.
Kroutil, P. Macheroux, K. Faber, Org. Lett. 2007, 9,
[
[
[
[
33] A. K. Sinhababu, R. T. Borchardt, Tetrahedron Lett.
1
983, 24, 227–230.
5
409–5411.
7] M. A. Swiderska, J. D. Stewart, Org. Lett. 2006, 8,
131–6133.
8] M. A. Swiderska, J. D. Stewart, J. Mol. Catal. B:
Enzym. 2006, 42, 52–54.
34] Z. Zhang, Z. Zha, C. Gan, C. Pan, Y. Zhou, Z. Wang,
[
[
[
M.-M. Zhou, J. Org. Chem. 2006, 71, 4339–4342.
6
35] D.-S. Choi, S. Huang, M. Huang, T. S. Barnard, R. D.
Adams, J. M. Seminario, J. M. Tour, J. Org. Chem.
1
998, 63, 2646–2655.
36] K. Balenovi, N. Bregant, J. Chem. Soc. 1965, 5131–
132.
37] T. Hirata, A. Takarada, A. Matsushima, Y. Kondo, H.
Hamada, Tetrahedron: Asymmetry 2004, 15, 15–16.
38] J. Kling, Ber. Dtsch. Chem. Ges. 1897, 30, 3041.
39] K. R. Shah, C. D. Blanton, J. Org. Chem. 1982, 47,
9] D. J. Bougioukou, J. D. Stewart, J. Am. Chem. Soc.
[
[
2
008, 130, 7655–7658.
5
[
[
[
[
[
10] B. Kosjek, F. J. Fleitz, P. G. Dormer, J. T. Kuethe, P. N.
Devine, Tetrahedron: Asymmetry 2008, 19, 1403–1406.
11] R. Stuermer, B. Hauer, M. Hall, K. Faber, Curr. Opin.
Chem. Biol. 2007, 11, 203–213.
[
[
12] C. E. French, S. Nicklin, N. C. Bruce, J. Bacteriol. 1996,
5
02–508.
1
78, 6623–6627.
[
40] T. Polonski, J. Chem. Soc. Perkin Trans. 1 1988, 629–
13] T. M. Barna, H. Khan, N. C. Bruce, I. Barsukov, N. S.
Scrutton, P. C. Moody, J. Mol. Biol. 2001, 310, 433–447.
14] H. Khan, R. J. Harris, T. Barna, D. H. Craig, N. C.
Bruce, A. W. Munro, P. C. Moody, N. S. Scrutton, J.
Biol. Chem. 2002, 277, 21906–21912.
637.
[41] R. A. Earl, F. W. Clough, L. B. Townsend, J. Heterocycl.
Chem. 1978, 15, 1479–1483.
[42] M. A. Sefton, G. K. Skouroumounis, R. A. Massywes-
tropp, P. J. Williams, Austr. J. Chem. 1989, 42, 2071–
2084.
[43] P. G. Baraldi, A. Barco, S. Benetti, G. P. Pollini, V. Za-
nirato, Tetrahedron Lett. 1984, 25, 4291–4294.
[44] P. N. Confalone, E. Baggiolini, B. Hennessy, G. Pizzola-
to, M. R. Uskokovic, J. Org. Chem. 1981, 46, 4923–
4927.
[
15] a) R. E. Williams, D. A. Rathbone, N. S. Scrutton, N. C.
Bruce, Appl. Environ. Microbiol. 2004, 70, 3566–3574;
b) J. Ye, A. Singh, O. P. Ward, World J. Microbiol. Bio-
technol. 2004, 20, 117–135.
[
16] H. S. Toogood, A. Fryszkowska, V. Hare, K. Fisher, A.
Roujeinikova, D. Leys, J. M. Gardiner, G. M. Stephens,
N. S. Scrutton, Adv. Synth. Catal. 2008, 350, 2789–2803.
17] R. E. Williams, N. C. Bruce, Microbiology 2002, 148,
[
[45] M. A. Maestro, L. Castedo, A. Mourino, J. Org. Chem.
1992, 57, 5208–5213.
1
607–1614.
Adv. Synth. Catal. 2009, 351, 2976 – 2990
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2989

Anna Fryszkowska et al.
FULL PAPERS
[
[
[
46] J. B. Stothers, C. T. Tan, Can. J. Chem. 1974, 52, 308–
[49] Collaborative Computational Project number 4, Acta
Crystallogr., Sect. D: Biol. Crystallogr. 1994, 50, 760–
763.
[50] P. Emsley, K. Cowtan, Acta Crystallogr., Sect. D: Biol.
Crystallogr. 2004, 60, 2126–2132.
3
14.
47] T. Tsuda, T. Hayashi, H. Satomi, T. Kawamoto, T. Sae-
gusa, J. Org. Chem. 1986, 51, 537–540.
48] Z. Jedlinski, A. Misiołek, W. Głowkowski, H. Janeczek,
A. Wolinska, Tetrahedron 1990, 46, 3547–3558.
[51] G. N. Murshudov, A. A. Vagin, E. J. Dodson, Acta
Crystallogr., Sect. D: Biol. Crystallogr. 1997, 53, 240–
2
55.
2990
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 2976 – 2990