Finding the Selectivity Switch - A Rational Approach towards Stereocomplementary Variants of the Ene Reductase YqjM

Article abstract of DOI:10.1002/adsc.201500149

Ene reductases from the Old Yellow Enzyme family are versatile biocatalysts useful for the synthesis of optically active compounds. One disadvantage of biocatalysts when compared to competing catalysts in chemical syntheses is that often only one stereoisomer of the product is available. Another drawback can be the lack of activity in certain enzyme-substrate combinations. We were able to approach both of these challenges rationally in the case of the enzymatic synthesis of methyl 3-hydroxy-2-methylpropanoate (commonly denoted as the Roche ester) and derivatives thereof using the ene reductase YqjM. By a highly efficient, concept-based approach of designing mutant variants of YqjM and engineering substrates we could alter both the rate constant and the enantioselectivity of the reaction. Preparative scale reactions have been performed with successful mutants. In addition, the iterative modification of the substrate gave experiment-based insights into the binding mode of the Roche ester precursor and its derivatives.

Full text of DOI:10.1002/adsc.201500149

FULL PAPERS
DOI: 10.1002/adsc.201500149
Finding the Selectivity Switch – A Rational Approach towards
Stereocomplementary Variants of the Ene Reductase YqjM
Elisabeth Rüthlein,a Thomas Classen,b Lina Dobnikar,a Melanie Schçlzel,a
and Jçrg Pietruszkaa,b,*
a
Institut für Bioorganische Chemie, Heinrich-Heine-Universität Düsseldorf im Forschungszentrum Jülich, 52426 Jülich,
Germany
Fax: (+49)-(0)2461-616-196; phone: (+49)-(0)2461-614-158; e-mail: J.Pietruszka@fz-juelich.de
Institute of Bio- and Geosciences IBG-1: Biotechnology, Forschungszentrum Jülich GmbH, 52426 Jülich, Germany
b
Received: February 12, 2015; Revised: March 17, 2015; Published online: April 29, 2015
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201500149.
Abstract: Ene reductases from the Old Yellow tase YqjM. By a highly efficient, concept-based ap-
Enzyme family are versatile biocatalysts useful for proach of designing mutant variants of YqjM and en-
the synthesis of optically active compounds. One dis- gineering substrates we could alter both the rate con-
advantage of biocatalysts when compared to compet- stant and the enantioselectivity of the reaction. Prep-
ing catalysts in chemical syntheses is that often only arative scale reactions have been performed with
one stereoisomer of the product is available. Anoth- successful mutants. In addition, the iterative modifi-
er drawback can be the lack of activity in certain cation of the substrate gave experiment-based in-
enzyme–substrate combinations. We were able to ap- sights into the binding mode of the Roche ester pre-
proach both of these challenges rationally in the case cursor and its derivatives.
of the enzymatic synthesis of methyl 3-hydroxy-2-
methylpropanoate (commonly denoted as the Roche Keywords: enantioselectivity; enzyme catalysis; pro-
ester) and derivatives thereof using the ene reduc- tein design; protein engineering; reduction
Introduction
OYEs is well understood and includes the binding of
the substrate above a flavin mononucleotide (FMN)
The number of industrial applications of biocatalysts via two hydrogen bonds to histidine or asparagine res-
has grown rapidly since protein engineering has first idues. The hydride from N-5 of the reduced flavin co-
become viable in the 1990s.[1] Possible drawbacks of factor then attacks the electronically activated Cb po-
known enzymes such as low stability, regio- or stereo-
selectivity can now be addressed and biocatalysts are
becoming applicable to new reactions and pro-
cesses.[1a] Oxidoreductases from the Old Yellow
Enzyme (OYE) family (EC 1.6.99.1)[2] are of great in-
terest for stereoselective synthesis as they catalyse the
asymmetric reduction of alkenes, creating up to two
stereogenic centres in one step. Additionally these fla-
voproteins show a broad substrate scope, converting
C=C double bonds conjugated to an electron-with-
drawing group. Examples of substrates are a,b-unsa-
turated carbonyl compounds (enals and enones), car-
boxylic acids and derivatives thereof (esters, cyclic
imides, nitriles, lactones), as well as nitroalkenes.[3]
Scheme 1 shows a generic bioreduction catalysed by
an ꢀeneꢁ reductase from the OYE family using the
conversion of cyclohexenone (1a) to cyclohexanone
Scheme 1. Bioreduction of the standard substrate cyclohex-
enone (1a) to cyclohexanone (1b) which is efficiently cata-
lysed by most ene reductases. Parts of the biocatalyst that
(1b) as an example. The biocatalytic mechanism of are relevant for mechanistic discussions are shown in grey.
Adv. Synth. Catal. 2015, 357, 1775 – 1786
ꢂ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1775

FULL PAPERS
Elisabeth Rüthlein et al.
sition of the substrate. Subsequently, a tyrosine resi- stated in the title, how subtle changes control the ori-
due delivers a proton to Ca from the opposite side to entation of substrate binding in an alkene reductase.[13]
give the desired product with defined absolute config- Nevertheless, the impact of W116 mutations on the
uration. Alternatively the protonation may occur di- facial selectivity for the carvone substrate extended to
rectly from the solvent which consequently leads to certain Morita–Baylis–Hillman (MBH) adducts[6] and
undetermined configuration at Ca. For regeneration partly also to a homologous enzyme. The analogous
of the flavin an appropriate cofactor recycling system position in the ene reductase OYE 2.6 from Pichia
can be used to avoid the need for stoichiometric stipitis (Ile113) showed an effect on stereoselectivity
amounts of NAD(P)H.[4]
for the before-mentioned Morita–Baylis–Hillman ad-
There are several synthetically useful enzyme–sub- ducts, although full inversion could only be achieved
strate combinations that show excellent enantioselec- in combination with two other mutations.[11b]
tivity for either (R)- or (S)-product. Enantiocomple-
A bioinformatical analysis of wild type stereocom-
mentary ene reductases, however, are rarely available plementary ene reductases revealed sequence motifs
[exceptions being, for example, OPR 1 and OPR 3 that might indicate the stereoselectivity and the oligo-
for (E)-1-nitro-2-phenylpropene and the aldehyde an- meric state.[14] Experiments supported the prediction
alogue[5] or OYE 1 and OYE 2.6 for certain Morita– of the quaternary structure. The stereoselectivity pre-
Baylis–Hillman adducts[6]]. Thus a large number of diction, however, could not be validated.[14,15]
strategies have been developed to gain access to both
Amongst all of these methods, directed evolution is
enantiomers of each substrate that are either focused to date the most broadly applicable and although ad-
on the enzyme or the substrate.[7] Substrate-based ap- vanced strategies like ISM reduce the size of a particu-
proaches applied to ene reductases, e.g., rely on lar mutant library drastically the screening effort still
a switch in the (E)/(Z) geometry of the substrate,[7,8] can be significant. We tried to identify synthetically
or on variations in the protection group.[9] Recent relevant variants by a rational, concept-based ap-
enzyme-based approaches towards stereocomplemen- proach using a very limited library. We redesigned the
tary biocatalytic reactions often rely on site saturation active site of the bacterial OYE YqjM to alter the
mutagenesis endeavours.[1a,10] Those can be rationally binding mode of the substrate consequently reversing
guided by choosing specific amino acids near the the stereopreference. The presented procedure aims
active site and additionally include various cycles of for a stereocomplementary reaction altering, among
optimisation within a directed evolution study.[6,11] All others, the mechanistically essential amino acid
those strategies have proven successful for individual His167 rather than focusing only on the surroundings
cases but generalisation of found mutations to of the catalytic process. The hydrogen bonding histi-
a broader substrate scope or homologous enzymes is dine or asparagine residues have so far either been
rarely possible.
excluded from studies on stereoselectivity of ene re-
One example of obtained mutants showing some ductases (YqjM[11c]), have not shown any improved
generality regarding the substrate is a directed evolu- variants (OYE 2.6[16]) or showed stereoselectivity en-
tion study on the ene reductase YqjM based on itera- hancing instead of complementary behaviour [pen-
tive site saturation mutagenesis (ISM)[10a] and combi- taerythritol tetranitrate (PETN) reductase[17]]. As
natorial active site saturation (CASTing).[12] It result- a model reaction, the reduction of a-exo-methylene
ed in protein variants [(R): C26D/I69T, (S): C26G or carbonyl compounds (2a–6a) shown in Scheme 2 has
C26G/A60V) of stereocomplementary selectivity for been chosen. This includes the precursor of the indus-
3-substituted cyclohexenones with aliphatic groups trially relevant Roche ester[18] (2b) which can be re-
(methyl, ethyl, butyl, isopropyl) as well as methyl car- duced using six different wild type ene reductases.[19]
boxylate as the side chain.[11c]
Only one of them – OYE 2.6 – was found to show
A site saturation mutagenesis study on W116 of (S)-selectivity and high conversion rates. For OYE 2.6
OYE 1 from Saccharomyces pastorianus found a possi- and OYE 1 stereocomplementary variants could be
ble explanation for the findings on stereoselectivity. It found by site saturation mutagenesis and will be dis-
yielded six variants of OYE 1 (I, A, N, M, Q, V) cussed in relation to our results.[6,11b] To the best of
which show partly or fully stereocomplementary be- our knowledge no preparative scale bioreductions of
haviour compared to the wild type for the conversion the chosen substrates (2a–6a) have been reported to
of (S)- or (R)-carvone.[11a] The crystallographic struc- date.
tures of wild type OYE 1 and those key variants
soaked with (S)- or (R)-carvone, respectively, were
solved, which allowed the substrate binding modes to
be studied in detail. A dependence of the stereoselec-
tivity on the distance between residue 116 and the
side chain of (S)-carvone was found. However, the
same was not true for (R)-carvone which shows, as
1776
asc.wiley-vch.de
ꢂ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2015, 357, 1775 – 1786

FULL PAPERS
Finding the Selectivity Switch – A Rational Approach
Table 1. Steady-state kinetic parameters for the conversion
of substrates 2a and 3a catalysed by YqjM wt and reference
values measured for the standard substrate cyclohexenone
(1a).
Substrate
M [mM]
2a
3a
1a
K
92.77Æ6.48
0.68Æ0.02
0.52Æ0.12 0.22Æ0.01
2.51Æ0.18 2.64Æ0.04
a)
kcat [sÀ1]
kcatKMÀ1 [sÀ1 mmÀ1] 0.007Æ0.0006 4.8Æ1.45
12Æ0.08
[a]
For enzyme concentration measurement the following
specific extinction coefficient of YqjM was used:
[e]280nm =1.015·4.53(A432nm AÀ1
) +
0.8 mLmgÀ1 cmÀ1]
280nm
see Ref.[20]
mononucleotide and the active site. To achieve this,
the free hydroxy group of the Roche ester precursor
2a was protected with a benzyl (5a) or allyl (6a)
group (Scheme 2).
Scheme 2. A) General reaction scheme showing the sub-
strates tested for stereopreference of the designed YqjM
mutants including the glucose dehydrogenase (GDH) cofac-
tor recycling system used to regenerate NADPH. B) Engi-
neered substrates 3a to 6a show increased conversions com-
pared to Roche ester precursor 2a. (Initial concentration of
substrates: 5-10 mM).
As shown previously,[19] protection increased con-
versions by YqjM wild type about 2.5-fold to almost
90% (Scheme 2B). Additionally, the protection im-
proved (R)-selectivity to >99% in both cases. Sub-
strates 2a–4a are accessible in one step via Morita–
Baylis–Hillman reaction from formaldehyde and the
appropriate unsaturated carbonyl compound, protec-
tion of 2a with benzyl or allyl groups to access 5a and
6a required an additional reaction step, respectively.
Results and Discussion
Substrate Engineering
Catalytic efficiency tends to decrease as mutations are
introduced in rational design mutagenesis studies.
Therefore, targeted compounds should be readily con- Enzyme Design and Enzymatic Conversions
verted by the wild type enzyme aiming for syntheti-
cally useful rational mutants. As reported previous- In order to reverse the stereoselectivity of the reac-
ly,[19] bioreductions of the Roche ester precursor tion, we aimed to induce a flipped binding of the sub-
methyl 2-hydroxymethylacrylate (2a) by the ene re- strate compared to its native position based on a con-
ductase YqjM show good (R)-selectivity (96% ee) but ceptually new approach that we will refer to as the
low conversions (34%). Hence, we aimed to find “rotational flip”. Figure 1A shows the binding mode
more active substrates of equal synthetic value.
of the Roche ester precursor 2a in YqjM wild type
Two disadvantageous properties of the Roche ester that was assumed from in silico docking. For a produc-
needed to be addressed: its small inherent polarisa- tive binding two atom positions of the substrate are
tion and its hydrophilicity. The inherent polarisation supposed to be crucial. The carbonyl oxygen should
of the C=C double bond is small due to the relatively be placed in hydrogen bonding distance to two histi-
low electron-withdrawing effect of the ester group. dine or asparagine residues. Cb should be as close as
Thus we aimed to design a more activated substrate possible to N-5 of FMN. Therefore two point muta-
by inserting a ketone group instead of the methyl tions were strategically incorporated into the active
ester. Kinetic evaluation of both molecules as sub- site:
strates for YqjM wild type showed a 180-fold lower
The role of the first mutation, H167A, is to loosen
Michaelis–Menten constant KM for the ketone (3a). the substrate from its original position by exchanging
Its optimised value of 0.52 mM is in the range of the one of the essential hydrogen bonding histidines for
readily converted reference substrate cyclohexenone an alanine. To reactivate and flip the substrate a hy-
(1a). The catalytic efficiency increased 680-fold from drogen bond donating histidine or asparagine residue
ester 2a to ketone 3a (Table 1). Thus the methyl was installed in place of Cys26. This concept of “rota-
ketone 3a as well as the more sterically demanding tional flip” differs from previous attempts to achieve
ethyl ketone 4a were included in the study a flipped binding mode in ene reductases. Usually
(Scheme 2).
a flip around the fixed axis along carbonyl–oxygen
Furthermore, we increased the hydrophobicity of and Cb (referred to as “static flip”) was targeted and
the substrate to strengthen its interaction with flavin found in crystal structures.[13] For the “rotational flip”
Adv. Synth. Catal. 2015, 357, 1775 – 1786
ꢂ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1777

FULL PAPERS
Elisabeth Rüthlein et al.
Figure 2. Stereopreference of YqjM single and double mu-
tants designed on the basis of the introduced concept re-
garding the conversion of substrates 2a to 6a to products 2b
to 6b. Only assays showing 5% conversion or more were
considered which was not feasible for all reactions catalysed
by YqjM mutants C26N and C26N/H167A.
Figure 3A and Figure 3B show the outcome of in
silico docking that might contribute to understanding
the achieved selectivities. The single mutation
H167A, where the native binding site has been cor-
rupted, apparently influences the binding mode for
substrates 2a–5a as is indicated by decreased enantio-
meric excesses. The C26N/H single mutations which
should offer an alternative binding site to the sub-
strate show less impact on stereoselectivity in the con-
version of the three ester substrates and the methyl
ketone 3a as was expected in the presence of the
intact binding site His167 and His164, respectively.
The ethyl ketone 4a, however, exhibits complete ste-
reocomplementary behaviour for YqjM C26H com-
pared to YqjM wt [wt: 60% ee (R)-; C26H: 60% ee
(S)-selectivity]. This effect decreases to an ee of 20%
(R) upon combination with the H167A mutation. The
above-mentioned concept of “rotational flip”, which
would predict additivity of the two mutations, is not
applicable here. In silico docking suggested a possible
binding. As shown in Figure 3A the substrate 4a
seems to form an intramolecular hydrogen bond
within YqjM C26H due to the steric demand of the
newly established histidine. The resulting six-mem-
bered ring is coordinated via hydrogen bonds to
His167 and His164, which causes the activation in
a flipped mode and the inverted configuration of the
Figure 1. A) Structure of YqjM active site. The essential
amino acids, FMN, and the positions for the mutagenesis
study are shown explicitly. The binding mode of the Roche
ester precursor (2a) resulting from in silico docking is shown
in grey. Hydrogen bonds are drawn as dashed lines. B) Illus-
tration of different approaches to a flipped substrate binding
demonstrated by the native and envisioned positions of sub-
strate 2a in s-trans conformation relative to the highly con-
served hydrogen bonding amino acids His167 and His164.
Yellow circles indicate fixed atom sites. In this work the “ro-
tational flip” was targeted.
an additional degree of rotational freedom around Cb products. This flipped binding was not observed as
exists. The different concepts and targeted changes in a docking result once His167 is mutated to alanine.
binding patterns are shown in Figure 1B. The illustra- The reduction of the ester substrate 2a shows, in line
tions (Figure 1) show the substrate in the s-trans con- with the assumed “rotational flip”, stereoinversion
formation. Alternatively, the substrate could be (86% ee) for the double mutant C26N/H167A. The
trapped as its s-cis-rotamer in both the wild type and possible explanatory illustration from docking shows
the altered, active site. However, that does not that the newly installed asparagine and the partial
change the general argument.[21]
loss of the former binding site in terms of His167 lead
In order to investigate the stereoselectivity of the to a flipped binding mode which depends on the abili-
generated mutants, the above-mentioned exo-methyl- ty of the substrate to form hydrogen bonds (Fig-
ene substrates 2a–6a have been screened. The ach- ure 3B). The allyl-protected substrate 6a does not
ieved enantiomeric excesses are displayed in Figure 2. show any influence on its binding mode; for the
1778
asc.wiley-vch.de
ꢂ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2015, 357, 1775 – 1786

FULL PAPERS
Finding the Selectivity Switch – A Rational Approach
Figure 3. Binding modes for selected mutant–substrate combinations, which are in accordance with the experimentally deter-
mined stereoselectivities [preferably (S)-configuration]. Important residues are displayed explicitly, while the mutated sites
are highlighted in orange. The substrates are coloured in grey. Hydrogen bonds are drawn as dashed lines. A) YqjM C26H
with docked ethyl ketone 4a. B) The Roche ester precursor 2a docked into the active site of YqjM C26N/H167A. The newly
implemented binding site enables the pre-(S)-binding mode. C)/D) Similar docking results for the binding mode of substrate
5a in the active centre of mutants YqjM I69A/H167A and YqjM I69Y/H167A.
benzyl protection significantly less (R)-product 5b can YqjM variants based on mutation sites from a directed
be found in the case of the double mutant C26H/ evolution approach carried out by the Reetz group
H167A [30% ee (R)]. These experimental results (Cys26, Tyr28, Ile69, Ala60).[11c] Of those variants
were promising, but not applicable to preparative showing conversion towards the screening substrate
scale reactions as the most promising mutant–sub- 2a only the Ile69 mutations had an impact on stereo-
strate combination (YqjM C26N/H167A–Roche ester chemistry (results shown in the Supporting Informa-
precursor 2a) showed only conversions of up to 10%. tion). The desired effect was observed with all sub-
Therefore, further positions influencing the stereo- strates except 6a. Ile69, situated in a loop region very
chemistry of the bioreductions needed to be identi- close to the active centre, was chosen to be replaced
fied. We screened a small library of seven promising by amino acids with different steric demands to inves-
Adv. Synth. Catal. 2015, 357, 1775 – 1786
ꢂ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1779

FULL PAPERS
Elisabeth Rüthlein et al.
Figure 4. Stereopreference of the most promising YqjM mutants regarding the bioreduction of substrates 2a to 6a to prod-
ucts 2b to 6b and respective conversion values for the assay reactions.
tigate if Ile69 could be blocking putative binding lected. Tyrosine additionally increases the overall
modes of the polar substrate side chains. Alanine – as acidity of the active site above the substrate which
a very small amino acid – and tyrosine have been se- might improve the conversion rate for novel substrate
1780
asc.wiley-vch.de
ꢂ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2015, 357, 1775 – 1786

FULL PAPERS
Finding the Selectivity Switch – A Rational Approach
binding modes, (e.g., in the case of C26N/H) where enables the substrates to adopt a flipped binding
Tyr169 is not available as a protonating residue. mode in proximity to the cavity. Binding energy re-
Screening of these Ile69-mutants regarding biore- sults from hydrophobic interactions. Although tyro-
ductions of the optimised substrates 3a–6a yielded in- sine is much larger than alanine, the tyrosine residue
creased activity compared to the first mutant library in YqjM I69Y/H167A creates a cavity because of an
in all cases and an influence on enantioselectivity in adopted kinked conformation (Figure 3D). As
most cases. Only for the allylic product 6b could an H167A/I69A and H167A/I69Y show similar effects
almost unchanged enantiomeric excess of 94–95% (R) on the stereoselectivity in all cases this explanation
be observed (Figure 4). Encouraged by these results seems more probable than the originally made as-
we generated five of the six possible double mutants sumption for the effect of I69Y that it could improve
containing the previous point mutations at Cys26 and the rate and selectivity of protonation because of
His167 combined with those at Ile69. (YqjM C26N/ overall increased acidity above the substrate.
I69Y could not be obtained despite several attempts).
Thereby, for substrates 2a, 4a, and 5a variants of
ꢀ90% ee opposite to the wild type could be found Preparative Scale Reactions
(Figure 4). For compound 2a even three double mu-
tants (C26N/I69A, H167A/I69Y and H167A/I69A) To demonstrate the synthetic applicability of the ob-
fulfilled this criterion and almost all combined single tained enzyme–substrate combinations, we performed
mutations exhibited additive or more than additive ef- preparative scale bioreductions for the best pairs. In
fects.[22] For the methyl ketone 3a only one of these case of the Roche ester precursor 2a the (R)-selective
variants (C26N/I69A) showed moderate (S)-selectivi- wild type and the most active mutant of >90% ee
ty of 68% ee. This YqjM mutant also showed the best (S)-selectivity (C26N/I69A) were chosen. For the ke-
screening result of 94% ee (S) for the ethyl ketone tones and the allyl only one combination, respectively,
(4a). These findings support the interpretation that seemed promising with respect to the screening (wt–
these mutations only rely on steric interactions to 3a; C26N/I69A–4a; wt–6a). Preparative scale biore-
change the substrate binding pattern. In contrast, only ductions of the benzylic substrate 5a were attempted
the mutants H167A/I69A and H167A/I69Y had using mutant H167A/I69A and the other highly (S)-
a large impact on stereochemistry for the hydroxy- selective mutant H167A/I69Y. The results are sum-
protected ester substrates 5a and 6a. Those mutants marised in Table 2. Although the stereoselectivities
yielded racemic product 6b and 11% ee (S)-6b, re- from assay reactions were not fully reproducible
spectively, for the allylic substrate. When the benzyl- (which could result from uncontrolled pH conditions
substituted substrate 5a was converted almost perfect in assay scale as small changes in cofactor recycling
complementary selectivity of 90% ee (S) was ob- have been shown to have an impact on stereoselecti-
tained. Docking studies revealed that the truncation vity[5a,17a]) and the results for the Roche ester (2b)
of Ile69 creates a cavity, which can be accessed by the itself regarding enantiomeric excess as well as turn-
allyl and preferably by the benzyl group (Figure 3C/ over numbers leave room for improvement, we could
D). The loss of the former hydrogen bond to His167 obtain both configurations efficiently and in good to
Table 2. Preparative bioreductions of substrates 2a to 6a.
Entry
Substrate[a]
YqjM
Variant
ee[b]
[%]
Yield[c]
[%]
Ratio[d]
Product:Substrate
Biocatalyst
[mM]
Biocatalyst
[mg]
TTN[e]
1
2a
2a
3a
4a
5a
6a
wt
90 (R)
85 (S)
93 (R)
93 (S)
77 (S)
>99 (R)
47
44
86
95
10
50
100:0
85:15
100:0
100:0
95:5[g]
100:0
37
403
7
50
252
3
30
160
6
40
100
3
1336
248
6941
878
400
13895
2[f]
3
C26N/I69A
wt
C26N/I69A
H167A/I69A
wt
4
5f)
6
[a]
1 mmol substrate was converted.
[b]
Determined via GC according to references as shown in the Supporting Information. The stereodescriptor of the prefer-
entially formed enantiomer is shown in parenthesis.
[c]
In case of incomplete conversion yields refer to mixtures of product and remaining starting material as depicted in the
ratio column.
[d]
[e]
[f]
1
Determined via H NMR.
Molecular weight of YqjM: 39.7 kg/mol.
Reaction conducted at pH 6.5 instead of 7.5 and in only 10 mL total reaction volume.
Additionally unknown side product occurred.
[g]
Adv. Synth. Catal. 2015, 357, 1775 – 1786
ꢂ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1781

FULL PAPERS
Elisabeth Rüthlein et al.
excellent selectivity for the protected esters or the
keto derivatives. The protected ester substrates gave
the products in >99% ee (R)-selectivity (wt–allyl 6a)
and 77% ee (S)-selectivity (H167A/I69A–benzyl 5a).
The keto derivatives 3a and 4a delivered both stereo-
isomers in 93% ee and high yields (86%, 95%) utilis-
ing the wild type enzyme or the mutant C26N/I69A.
All the obtained products are versatile synthetic inter-
mediates of numerous (also industrial) applications
[selected examples: (S)-2b,[23] (R)-2b,[24] 3b,[25] (S)-
4b,[26] (S)-5b,[27] (R)-6b[28]]. The mutant derived prod-
uct (S)-4b has been used, for example, in the patented
Figure 5. Stereopreference of the most promising YqjM mu-
total synthesis of the polyketide discodermolide tants catalysing the bioreduction of substrates 7a to 9a.
where it is being synthesised in two steps from enan-
tiomerically pure Roche ester (2a).[26] With our ap-
proach (S)-4b could be obtained in a two-step process Stewart already stated[13] – completely reverse the
from the non-chiral starting materials ethyl vinyl newly established interactions. This is especially strik-
ketone and formaldehyde.
ing for the allyl protection compared to the benzyl
protection (5a, 6a) in our study. In which way also
structurally more distinct substrates such as, for exam-
ple, cyclic structures or alkenes activated by other
electron-withdrawing groups such as aldehyde, nitro
General Utility of the Mutations
To test the general utility of the most promising or nitrile moieties behave, remains to be examined.
YqjM variants, three other substrates were screened Also whether analogous mutations in other ene re-
regarding any impact on their stereopreference: an- ductases from the OYE family would have a similar
other terminal alkene 7a as well as two known YqjM impact on stereochemistry has yet to be determined.
substrates with triple-substituted double bonds (8a, To analyse the available data we compared the muta-
9a, Scheme 3).[20,29] The outcome is illustrated graphi- tion sites used in our study to previously mutated
cally in Figure 5. No impact on the binding mode of sites in OYE 2.6, OYE 1, PETN reductase and YqjM
phenyl-substituted alkene 7a could be found. The yielding stereocomplementary enzyme variants
triple-substituted olefins 8a and 9a show – although to (Table 3).[6,11b] This showed that mutations at an anal-
a much lesser extent – a behaviour similar to methyl ogous amino acid position to Cys26 and I69 have pre-
ester 2a. The same three mutants, namely C26N/I69A, viously yielded enzyme variants of inverted stereopre-
H167A/I69A and H167A/I69Y provided reduction ference for 3-substituted cyclohexenones,[11c] (S)-car-
products of significantly altered ee values. Although vone,[13] (E)-2-phenyl-1-nitropropene[17b] or ester
that did not lead to inverted selectivity compared to 2a,[11a,b] respectively. These sites have now been dem-
YqjM wild type, the achieved alteration can be an ad- onstrated to be also applicable in the context of cer-
vanced starting point for further optimisation. Espe- tain herein tested Baylis–Hillman adducts. However,
cially in case of the a-methyl-substituted ester 8a this His167-analogous mutation sites have not yet been
could be beneficial as – other than for 9a – the stereo- found to affect stereoinversion in these ene reductas-
complementary bioreduction cannot be realised by an es.
(E)/(Z) switch.[30]
Generally these findings indicate that the found set
of mutations is worthwhile to be tested when enantio- Conclusions
complementary YqjM variants for bioreduction of
structurally related substrates are required. Yet, Altogether, we have succeeded in finding synthetical-
subtle changes in the substrate structure can – as Jon ly useful YqjM variants which are highly selective for
the non-wt stereopreference in the bioreduction of
the targeted Baylis–Hillman adducts 2a to 5a using
a rational mutagenesis approach.
By the concept-based proceeding the mutagenesis
effort has been significantly reduced compared to sat-
uration mutagenesis (19 mutants per site) or directed
evolution endeavours (>1000 clones[34]). Only 17
single or double mutants had to be generated and
screened for their stereopreference to find enzyme
variants of inverted enantioselectivity (90–99% ee)
Scheme 3. Test substrates to examine the more general utili-
ty of the best performing YqjM variants. Substrate 7a,[29]
substrate 8a and 9a.[20]
1782
asc.wiley-vch.de
ꢂ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2015, 357, 1775 – 1786

FULL PAPERS
Finding the Selectivity Switch – A Rational Approach
Table 3. Amino acid positions that influence stereoselectivity in certain ene reductases regarding the bioreduction of 2a[6,11b]
(bold type) or other substrates and their analogue positions in homologue enzymes. PDB-Codes of the aligned protein struc-
tures: YqjM: 1z42,[31] OYE 1: 3TX9,[13] OYE 2.6: 3TJL,[32] PETN reductase: 1H50[33]
.
Entry
YqjM
OYE 1
OYE 2.6
PETN reductase
Specific mutations[a] – Substrate
1
2
3
4
C26
H167
I69
A104
–
T37
T35
T26
C26D/G – CH,[b,11c] T26S-(E)-2-phenyl-1-nitropropene[17b]
N194
Y82[c]
W116
F250
G72
H191
Y78[c]
I113
F247
A68
H184
Y68[c]
W102
F240
A58
–
Y78W – 2a[11b]
W116H/Q – 2a,[6] W116I – (S)-carvone,[11a] I113V/C – 2a[11b]
F247A/H – 2a[11b]
6
A60
A60V – CH[b,11c]
[a]
Mutations resulting in stereoinversion for the shown substrates either alone or in combination.
Results for certain 3-substituted cyclohexenones indicated as CH.
Alignment according to 3D structure instead of primary sequence.
[b]
[c]
for three out of five structurally related substrates Table 4. Oligonucleotides.[a]
(2a, 4a, 5a). The proposed strategic single mutations
Name
Sequence 5’!3’
affected stereoselectivity in five out of nine cases.
Two of the three most selective stereocomplementary
YqjM variants from screening stage (C26N/I69A,
H167A/I69A and H167A/I69Y) actually included the
His167 mutation that assumingly corrupts the active
centre. Suggested binding modes from in silico dock-
ing showed the proposed additional rotational degree
of freedom that makes an entirely different orienta-
tion of the substrate within the active site possible.
Screening regarding a broader substrate scope of
our best specific mutants yielded no bioreductions of
inverted stereoselectivity, but still similar trends of ee
values in the bioreductions of triple-substituted struc-
turally related substrates 8a and 9a.
Furthermore, we have demonstrated the synthetic
utility of the promising bioreductions from screening
stage in preparative scale using YqjM wild type or
mutants. Thereby all five structurally related products
(2a–6a) could be obtained with (R)- or (S)-configured
stereogenic centres in 77–99% ee and up to 95%
yield. All of the accessed products are widely used
chiral synthetic intermediates and can be produced by
the presented chemoenzymatic approach in two (2b
to 4b) or four (5b and 6b) reaction steps starting from
simple non-chiral starting material.
C26N fw
C26H fw
I69Y fw
I69 A fw
TGTCATGTCGCCAATGAACATGTATTCTTCTCATG
TGTCATGTCGCCAATGCAATGTATTCTTCTCATG
CCTCAAGGACGATATACTGACCAAGACTTA
CCTCAAGGACGAGCGACTGACCAAGAC
H167 A fw AAATTCATGCGGCGGCGGGATATTTAAT
[a]
The second QuikChange primer is the reversed comple-
ment of each.
tion, a FRENCH Press pressure cell from Thermo Fisher
Scientific (Schwerte, Germany) was used. NADP+ and
NADPH were a generous gift of Codexis (Redwood City,
CA). GDH was a kind gift from Prof. Dr. W. Hummel. Glu-
cose oxidase (from Aspergillus niger Type VII) was bought
at Sigma Aldrich (St. Louis, USA).
The yqjM gene was obtained as described before[20] and
the assembled vector contains the coding sequence for
YqjM with an N-terminally fused 6 His-tag/TEV site. For
site-direct mutagenesis QuikChange PCR (Stratagene, La-
Jolla CA) was performed using the oligonucleotides shown
in Table 4.
For double mutations two consecutive PCRs were per-
formed. The insert was checked by sequencing using T7 oli-
gonucleotides. Expression and purification of each ene re-
ductase was conducted as previously reported.[20]
Activity Assay for Steady-State Kinetics and
Evaluation of Enzymes
Experimental Section
The activity was measured photometrically by the consump-
tion of NADPH [Abs (340 nm), e340nm =6.2 mMÀ1 cmÀ1] in-
cluding an oxygen consuming system.
Molecular Biology and Catalyst Preparation
For cloning, restriction enzymes were bought from Fermen-
tas (St. Leon-Roth, Germany) the vectors were purchased
from Merck (Darmstadt, Germany), and oligonucleotides
were obtained as desalted lyophilisates from Sigma–Aldrich
(Steinheim, Germany). Chromatographic columns and ma-
trices were purchased from GE Healthcare (München, Ger-
many) and Qiagen (Hilden, Germany), respectively. Chro-
matographic steps were carried out on an ¾KTA purifier
from GE Healthcare (München, Germany). For cell disrup-
Assay conditions (500 mL): KPi (20 mM, pH 6.5); sub-
strate in KPi (1 mM); glucose (20 mM); glucose oxidase (ca.
20 U); NADPH (0.15 mM); YqjM wt (10 mL). Every mea-
surement was conducted in triplicate and corrected by
a blank not containing any substrate. Protein concentrations
to assign specific activities were derived as described
before.[20] Graphs were fitted using Origin 9.0.
To get an idea of the relative activity of enzymes in [U]
the assay was conducted using the substrate cyclohexenone.
Adv. Synth. Catal. 2015, 357, 1775 – 1786
ꢂ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1783

FULL PAPERS
Elisabeth Rüthlein et al.
Conditions for Small Scale Bioreductions
Representative Procedure for Scale-Up Bioreductions
For assay reactions in small scale the following conditions
were used: substrate (1 mL), MeTHF (1% v/v), NADP+
(0.2 mM), 20 mL glucose (20 mM), GDH (ca. 1 U), YqjM wt
(0.3–1 U) or an appropriate amount of YqjM variant. Add
KPi (20 mM, pH 7.5) to 1 mL.
The chosen substrate (ca. 1 mmol) was placed under argon
in a two-necked flask and KPi (100 mM, pH 7.5 degassed),
NADP+ (0.3 mM), glucose (250 mM), 100 mL 2-methyltetra-
hydrofuran (1% v/v) and GDH (~2.5 U in KPi buffer pH 7)
were added and premixed. Then a pH electrode and a bu-
rette were installed for on-line pH control and the pH was
adjusted. Finally purified YqjM was added (either as lyophi-
lisate or as a solution in KPi, pH 6.5). Total starting volume
was 20 mL. After 24 h the reaction was either stopped or an-
other aliquot of enzyme, coenzyme and GDH were added
to drive the reaction to completion. After complete conver-
sion the reaction was quenched using (NH4)2SO4 (0.8 g/mL)
and stirred for 1 h. The reaction mixture was filtered
through a pad of celite and the aqueous phase was extracted
with 2-methoxy-2-methylpropane (MTBE), dried over
MgSO4 and the solvent was removed under reduced pres-
sure. The products were analysed by 1H NMR and chiral
GC and compared to the synthesised reference material.
(R)-Methyl 2-(hydroxymethylene)propionoate [Roche
ester, (R)-2b]: The product was obtained as yellow oil;
yield: 68 mg (0.58 mmol, 58%). 1H NMR (CDCl3,
600 MHz): d=3.72 (s, 3H, OCH3), 3.68–3.77 (m, 1H,
CH2OH), 2.69 (ddq, 3J=7.2 Hz, 3J1’,3b =4.7 Hz, 3J=4.7 Hz,
In silico Docking
For the purpose of illustration, docking studies have been
performed and might serve as explanation, but these dock-
ing studies are not intended to provide a proof of molecular
binding modes. However, wild type structure and mutants
have been generated in silico using UCSF Chimera 1.8[35]
starting from YqjM wt structure pdb:1z42[31] bearing para-
hydroxybenzaldehyde as inhibitor. All non-amino acid
atoms but FMN have been deleted in order to obtain an
open state active site. Mutants have been introduced accord-
ing to the Dunbrack rotamer library[36] with respect to clash-
es. If clashes occurred, structure minimisation has been car-
ried out. After protonation, structures were finalised for
docking in AutoDock Tools 4.2[37] in PMV. Docking algo-
rithms have been chosen in their competence to reproduce
the inhibitor binding mode from pdb:1z42. Thus, the genetic
algorithm in AutoDock 4.2h[38] has been chosen. Necessary
maps have been generated for the search room having its
centre on top on one FMN and sulphate [(x, y, z)=(17.314,
54.744, 26.091)] and a cube length of 22.5 Š (grid spacing
0.375 Š). Generic files for map generation and docking can
be found in the Supporting Information. The ligands have
been prepared using Chem 3D v 12.0 3D from Cambridge-
Soft as well as ADT 4.2. As an empirical result, the bond be-
tween the exo-methylene group and the respective carbonyl
moiety has been set non-rotatable in its s-trans conforma-
tion. A liberated rotation did not yield any productive bind-
ing modes. The virtual screening has been set-up using Rac-
coon.[39]
3
1H,CHCH3), 1.19 (d, J=7.2 Hz, 3H, CHCH3); GC [column
FS-Lipodex E; 608C (5’ iso), 58CminÀ1 to 658C (30’ iso),
208CminÀ1 to 1508C (5’ iso)]: (R): 32.4 min, (S): 31.1 min,
90% ee (R), [a]2D0: À13.5 (c 0.96, CHCl3).
(S)-Methyl 2-(hydroxymethylene)propionoate [Roche
ester, (S)-2b]: The product was obtained as a yellow oil;
yield: 52 mg (0.44 mmol, 44%). GC [column FS-Lipodex G;
508C (40ꢁ iso), 208CminÀ1 to 1508C (5’ iso)]: tR (R):
25.3 min, (S): 27.8 min, 85% ee (S), [a]2D0: +10.4 (c 0.79,
CHCl3).
(R)-4-Hydroxy-3-methylbutan-2-one [(R)-3b]: The prod-
uct was obtained as
a colourless oil; yield: 88 mg
(0.86 mmol, 86%). 1H NMR (CDCl3, 600 MHz): d=3.72
(ddd, 3J3a,OH =6.0 Hz, 3J3a,2 =7.3 Hz, 2J3a,3b =11.2 Hz, 1H, 3-
3
3
2
Chemical Reactions
Ha), 3.67 (ddd, J3b,OH =6.8 Hz, J3b,2 =4.2 Hz, J3b,3a =11.2 Hz,
1H, 3-Hb), 2.74 (ddq, 3J2,1’ =7.3 Hz, 3J2,3b =4.2 Hz, J2,3a
=
3
All chemicals were purchased from Carl Roth (Karlsruhe,
Germany) and Sigma–Aldrich (Steinheim, Germany) in an-
alytical grade. Solvents for inert reactions were dried using
the solvent purification system MB SPS-800 from MBraun
(Garching, Germany). The enantiomer analytics were per-
formed on a gas chromatograph and an HPLC from Thermo
Scientific (Waltham; USA) using columns from Macherey
and Nagel (Düren, Germany) or Daicel (Tokyo, Japan), re-
spectively. pH control in preparative scale reactions was
achieved using the titration system Titrino 747 from Büchi
(Essen, Germany).
3
3
7.3 Hz, 1H, 2-H), 2.35 (dd, J3a,OH =6.0 Hz, J3b,OH =6.8 Hz,
1H, OH), 2.19 (s, 3H, COCH3), 1.14 (d, J1’,2 =7.3 Hz, 3H,
3
CHCH3); GC [column: FS-Lipodex E, 608C (5’ iso),
58CminÀ1 to 1508C (5’ iso)]: tR (R): 15.8 min; (S): 15.5 min,
93% ee (R), [a]2D0: +24.1 (c 1.05, CHCl3).
(S)-1-Hydroxy-2-methylpentan-3-one [(S)-4b]: The prod-
uct was obtained as a yellow oil; yield: 98 mg(0.84 mmol,
95%). 1H NMR (CDCl3, 600 MHz): d=3.77–3.62 (m, 2H,
3
CH2OH), 2.76 (ddq, 3J4,1’ =7.3 Hz, 3J4,5a =4.3 Hz, J4,5b
=
=
=
=
3
4.3 Hz, 1H, CHCH3), 2.57 (dq, 3J2b,2a =17.9 Hz, J2b,1
3
7.3 Hz, 1H, CH2CH3), 2.47 (dq, 3J2a,2b =17.9 Hz, J2a,1
Substrates were synthesised according to the literature
(2a–4a,[40] 5a–6a[41]). Reference material methyl 2-(hydroxy-
methylene)propionoate [(R)-2b, (S)-2b] was purchased at
Sigma Aldrich and the other references were synthesised
from the ester according to known procedures.[42] All the
3
7.3 Hz, 1H, CH2CH3), 2.23 (s, 1H, OH), 1.13 (d, J1’,4
3
7.3 Hz, 3H, CHCH3), 1.06 (t, J1,2 =7.3 Hz, 3H, CH2CH3);
GC [column: FS-Lipodex E, 608C (5’ iso), 58CminÀ1 to
1508C (5’ iso)]: tR (R): 16.6 min, (S): 16.3 min, 93% ee (S),
[a]20: À15.3 (c 1.2, CHCl3).
1
substances were analysed by H NMR, 13C NMR, MS (EI),
(DS)-Methyl 3-(benzyloxy)-2-methylpropionate [(S)-5b]:
The product was obtained as a yellow oil; yield: 20 mg
GC and IR.
1
(0.1 mmol, 10%). H NMR (CDCl3, 600 MHz): d=7.39–7.25
4
(m, 5H, Har), 4.52 (d, J1’’,3’’ =2.6 Hz, 2H, 1’’-H), 3.70 (s, 3H,
OCH3), 3.66 (dd, 2J3a,3b =9.1 Hz, 3J3a,2 =7.2 Hz, 1H, 3-Ha),
2
3
3.50 (dd, J3b.3a =9.1 Hz, J3b,2 =6.0 Hz, 1H, 3-Hb), 2.79 (ddq,
1784
asc.wiley-vch.de
ꢂ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2015, 357, 1775 – 1786

FULL PAPERS
Finding the Selectivity Switch – A Rational Approach
3
2
3J2,1’ =7.1 Hz, J2,3b =6.0 Hz, J2,3a =7.2 Hz, 1H, 2-H), 1.18 (d,
3J1’,2 =7.1 Hz, 3H, 1’-H); HPLC [column: Chiracel OD-H
(Daicel); eluent n-heptane/iso-propyl alcohol (99.8:0.2);
flow rate: 0.5 mLminÀ1; detection: UV 205 nm]: tR (R):
28.3 min; (S): 36.1 min, 77% ee (S).
[6] A. Z. Walton, W. C. Conerly, Y. Pompeu, B. Sullivan,
J. D. Stewart, ACS Catal. 2011, 1, 989–993.
[7] G. Oberdorfer, K. Gruber, K. Faber, M. Hall, Synlett
2012, 1857–1864.
[8] G. Tasnadi, C. K. Winkler, D. Clay, N. Sultana, W. M.
Fabian, M. Hall, K. Ditrich, K. Faber, Chem. Eur. J.
2012, 18, 10362–10367.
[9] a) C. K. Winkler, C. Stueckler, N. J. Mueller, D. Press-
nitz, K. Faber, Eur. J. Org. Chem. 2010, 6354–6358;
b) C. Stueckler, C. K. Winkler, M. Hall, B. Hauer, M.
Bonnekessel, K. Zangger, K. Faber, Adv. Synth. Catal.
2011, 353, 1169–1173.
[10] a) M. T. Reetz, J. D. Carballeira, Nat. Protocols 2007, 2,
891–903; b) M. T. Reetz, Angew. Chem. 2011, 123, 144–
182; Angew. Chem. Int. Ed. 2011, 50, 138–174.
[11] a) S. K. Padhi, D. J. Bougioukou, J. D. Stewart, J. Am.
Chem. Soc. 2009, 131, 3271–3280; b) A. Z. Walton, B.
Sullivan, A. C. Patterson-Orazem, J. D. Stewart, ACS
Catal. 2014, 4, 2307–2318; c) D. J. Bougioukou, S. Kille,
A. Taglieber, M. T. Reetz, Adv. Synth. Catal. 2009, 351,
3287–3305.
(R)-Methyl 3-(allyloxy)-2-methylpropionate [(R)-6b]: The
product was obtained as a colourless oil; yield: 79 mg
1
(0.5 mmol, 50%). H NMR (CDCl3, 600 MHz): d=5.88 (ddt,
3J2’’,3’’b =17.2 Hz, 3J2’’,3’’a =10.5 Hz, 3J2’’,1’’ =5.5 Hz, 1H, 2’’-H),
5.26 (ddt, 3J3’’b,2’’ =17.2 Hz, 2J3’’b,3’’a =1.7 Hz, 4J3’’b,1’’ =1.6 Hz,
1H, 3’’-Hb), 5.17 (ddt, 3J3’’a,2’’ =10.5 Hz, 2J3’’a,3’’b =1.7 Hz,
4J3’’a,1’’ =1.3 Hz, 1H, 3’’-Ha), 3.98 (ddd, J1’’,2’’ =5.5 Hz, J1’’,3’’b
=
3
4
4
1.6 Hz, J1’’,3’’a =1.3 Hz, 2H, 1’’-H), 3.70 (s, 3H, OCH3), 3.63
(dd, 2J3a,3b =9.2 Hz, 3J3a,2 =7.2 Hz, 1H, 3-Ha), 3.46 (dd,
2J3b,3a =9.2 Hz, 3J3b,2 =5.9 Hz, 1H, 3-Hb), 2.76 (ddq, J2,1’
=
=
3
3
3
3
7.2 Hz, J2,3a =7.2 Hz, J2,3b =5.9 Hz, 1H, 2-H), 1.18 (d, J1’,2
7.2 Hz, 3H, 1’-H); GC [column: FS-Lipodex E, 608C (5’
iso), 58CminÀ1 to 1508C (5’ iso)]: tR (R): 9.6 min, (S):
9.9 min, >99% ee (R), [a]2D0: À14.8 (c 1.06, CHCl3).
[12] M. T. Reetz, L.-W. Wang, M. Bocola, Angew. Chem.
2006, 118, 1258–1263; Angew. Chem. Int. Ed. 2006, 45,
1236–1241.
Acknowledgements
[13] Y. A. Pompeu, B. Sullivan, J. D. Stewart, ACS Catal.
We gratefully acknowledge the Ministry of Innovation, Sci-
ence and Research of the German Federal State of North
Rhine-Westphalia (NRW) and the Heinrich-Heine-Universität
Düsseldorf (HHU; scholarship within the CLIB-Graduate
Cluster Industrial Biotechnology for M.S. and technology
platform “ExpressO” within the “Ziel 2-Programm 2007–
2013, NRW - EFRE”), the Deutsche Forschungsgemeinschaft,
the Deutsche Akademische Austauschdienst (scholarship for
L.D.) and the Forschungszentrum Jülich GmbH for the sup-
port of our projects. Generous gifts of enzymes from evocatal
GmbH and Prof. Dr. W. Hummel (IMET, HHU), were great-
ly appreciated. We would like to thank Birgit Henßen, Kiril
Lutsenko, Bea Paschold, Saskia Schuback, Sonja Meyer zu
Berstenhorst, for their support in GC analytics, syntheses,
cloning and mutagenesis and the entire IBOC staff for their
ongoing support.
2013, 3, 2376–2390.
[14] G. Oberdorfer, G. Steinkellner, C. Stueckler, K. Faber,
K. Gruber, ChemCatChem 2011, 3, 1562–1566.
[15] T. Classen, J. Pietruszka, S. M. Schuback, ChemCatCh-
em 2013, 5, 711–713.
[16] A. Patterson-Orazem, B. Sullivan, J. D. Stewart,
Bioorg. Med. Chem. 2014, 22, 5628–5632.
[17] a) H. S. Toogood, A. Fryszkowska, M. Hulley, M.
Sakuma, D. Mansell, G. M. Stephens, J. M. Gardiner,
N. S. Scrutton, ChemBioChem 2011, 12, 738–749; b) A.
Fryszkowska, H. Toogood, M. Sakuma, G. M. Stephens,
J. M. Gardiner, N. S. Scrutton, Catal. Sci. Technol. 2011,
1, 948–957.
[18] W. R. Roush, A. D. Palkowitz, M. J. Palmer, J. Org.
Chem. 1987, 52, 316–318.
[19] C. Stueckler, C. K. Winkler, M. Bonnekessel, K. Faber,
Adv. Synth. Catal. 2010, 352, 2663–2666.
[20] T. Classen, M. Korpak, M. Schçlzel, J. Pietruszka, ACS
Catal. 2014, 4, 1321–1331.
References
[21] Only a change in substrate–conformation from the wild
type to the mutated enzyme would probably leave the
side from which the proton addition originates the
same. The carbonyl group would move in its new posi-
tion only by the intramolecular rotation that changes
the substrate from s-trans to s-cis. That both conforma-
tions are generally possible has been reported previous-
ly (Ref.[7]).
[1] a) M. T. Reetz, J. Am. Chem. Soc. 2013, 135, 12480–
12496; b) U. T. Bornscheuer, G. W. Huisman, R. J. Ka-
zlauskas, S. Lutz, J. C. Moore, K. Robins, Nature 2012,
485, 185–194.
[2] H. S. Toogood, N. S. Scrutton, Curr. Opin. Chem. Biol.
2014, 19, 107–115.
[3] R. Stuermer, B. Hauer, M. Hall, K. Faber, Curr. Opin.
Chem. Biol. 2007, 11, 203–213.
[22] M. T. Reetz, Angew. Chem. 2013, 125, 2720–2729;
Angew. Chem. Int. Ed. 2013, 52, 2658–2666.
[4] a) P. A. Karplus, K. M. Fox, V. Massey, FASEB J. 1995,
9, 1518–1526; b) B. J. Brown, Z. Deng, P. A. Karplus, V.
Massey, J. Biol. Chem. 1998, 273, 32753–32762.
[5] a) 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.
2008, 350, 411–418; b) M. Hall, C. Stueckler, W. Krou-
til, P. Macheroux, K. Faber, Angew. Chem. 2007, 119,
4008–4011; Angew. Chem. Int. Ed. 2007, 46, 3934–3937.
[23] S. V. Ley, M. N. Tackett, M. L. Maddess, J. C. Ander-
son, P. E. Brennan, M. W. Cappi, J. P. Heer, C. Helgen,
M. Kori, C. Kouklovsky, S. P. Marsden, J. Norman, D. P.
Osborn, M. . Palomero, J. B. J. Pavey, C. Pinel, L. A.
Robinson, J. Schnaubelt, J. S. Scott, C. D. Spilling, H.
Watanabe, K. E. Wesson, M. C. Willis, Chem. Eur. J.
2009, 15, 2874–2914.
Adv. Synth. Catal. 2015, 357, 1775 – 1786
ꢂ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1785

FULL PAPERS
Elisabeth Rüthlein et al.
[24] C. M. Neuhaus, M. Liniger, M. Stieger, K.-H. Altmann,
Angew. Chem. 2013, 125, 5978–5983; Angew. Chem.
Int. Ed. 2013, 52, 5866–5870.
[35] E. F. Pettersen, T. D. Goddard, C. C. Huang, G. S.
Couch, D. M. Greenblatt, E. C. Meng, T. E. Ferrin, J.
Comput. Chem. 2004, 25, 1605–1612.
[25] K. L. Rinehart, R. C. Pandey, V. F. German, Y. Nishika-
wa, J. Am. Chem. Soc. 1971, 93, 3738–3747.
[26] G. Florence, G. Koch, O. Loiseleur, S. J. Mickel, I. Pa-
terson (Novartis AG), WO Patent WO 2004009574A1,
2004.
[27] J. S. Yadav, A. S. Reddy, C. S. Reddy, B. V. S. Reddy, V.
Saddanapu, A. Addlagatta, Eur. J. Org. Chem. 2011,
696–706.
[28] D. Quintard, P. Bertrand, S. Vielle, E. Raimbaud, P.
Renard, B. Pfeiffer, J.-P. Gesson, Synlett 2003, 2033–
2036.
[29] J. Pietruszka, M. Schçlzel, Adv. Synth. Catal. 2012, 354,
751–756.
[36] R. L. Dunbrack Jr, Curr. Opin. Struct. Biol. 2002, 12,
431–440.
[37] M. F. Sanner, J. Mol. Graphics Modell. 1999, 17, 57–61.
[38] D. S. Goodsell, G. M. Morris, A. J. Olson, J. Mol. Rec-
ognit. 1996, 9, 1–5.
[39] S. Forli, Raccoon AutoDock VS: an automated tool for
preparing AutoDock virtual screenings, http://auto-
dock.scripps.edu/resources/raccoon (accessed 12/10/
2014).
[40] C. Yu, B. Liu, L. Hu, J. Org. Chem. 2001, 66, 5413–
5418.
[41] a) E. Sagot, D. S. Pickering, X. Pu, M. Umberti, T. B.
Stensbøl, B. Nielsen, M. Chapelet, J. Bolte, T. Gefflaut,
L. Bunch, J. Med. Chem. 2008, 51, 4093–4103; b) J. I.
Borrell, J. Teixidó, B. Martínez-Teipel, J. L. Matallana,
M. T. Copete, A. Llimargas, E. García, J. Med. Chem.
1998, 41, 3539–3545.
[42] a) A. Basha, M. Lipton, S. M. Weinreb, Tetrahedron
Lett. 1977, 18, 4171–4172; b) G. P. Luke, J. Morris, J.
Org. Chem. 1995, 60, 3013–3019; c) J. D. White, G. N.
Reddy, G. O. Spessard, J. Chem. Soc. Perkin Trans.
1 1993, 759–767; d) D. Quintard, P. Bertrand, C. Bach-
mann, J.-P. Gesson, Eur. J. Org. Chem. 2004, 4762–
4770.
[30] M. Korpak, J. Pietruszka, Adv. Synth. Catal. 2011, 353,
1420–1424.
[31] K. Kitzing, T. B. Fitzpatrick, C. Wilken, J. Sawa, G. P.
Bourenkov, P. Macheroux, T. Clausen, J. Biol. Chem.
2005, 280, 27904–27913.
[32] Y. A. Pompeu, B. Sullivan, A. Z. Walton, J. D. Stewart,
Adv. Synth. Catal. 2012, 354, 1949–1960.
[33] T. M. Barna, H. Khan, N. C. Bruce, I. Barsukov, N. S.
Scrutton, P. C. E. Moody, J. Mol. Biol. 2001, 310, 433–
447.
[34] S. Hoebenreich, F. E. Zilly, C. G. Acevedo-Rocha, M.
Zilly, M. T. Reetz, ACS Synth. Biol. 2014, 4, 317–331.
1786
asc.wiley-vch.de
ꢂ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2015, 357, 1775 – 1786