Inversion of cpADH5 Enantiopreference and Altered Chain Length Specificity for Methyl 3-Hydroxyalkanoates

Article abstract of DOI:10.1002/chem.201702581

Expanding the substrate scope of enzymes opens up new routes for synthesis of valuable chemicals. Ketone-functionalized fatty acid derivatives and corresponding chiral alcohols are valuable building blocks for the synthesis of a variety of chemicals including pharmaceuticals. The alcohol dehydrogenase from Candida parapsilosis (cpADH5) catalyzes the reversible oxidations of chiral alcohols and has a broad substrate range; a challenge for cpADH5 is to convert alcohols with small substituents (methyl or ethyl) next to the oxidized alcohol moiety. Molecular docking studies revealed that W286 is located in the small binding pocket and limits the access to substrates that contain aliphatic chains longer than ethyl substituent. In the current manuscript, we report that positions L119 and W286 are key residues to boost oxidation of medium chain methyl 3-hydroxy fatty acids; interestingly the enantiopreference toward methyl 3-hydroxybutyrate was inverted. Kinetic characterization of W286A showed a 5.5 fold increase of V_max and a 9.6 fold decrease of K_m values toward methyl 3-hydroxyhexanoate (V_max: 2.48 U mg^? and K_m: 4.76 mm). Simultaneous saturation at positions 119 and 286 library yielded a double mutant (L119M/W286S) with more than 30-fold improved activity toward methyl 3-hydroxyoctanoate (WT: no conversion; L119M/W286S: 30 %) and inverted enantiopreference (S-enantiomer ≥99 % activity decrease and R-enantiomer >20-fold activity improvement) toward methyl 3-hydroxybutyrate.

Full text of DOI:10.1002/chem.201702581

A Journal of
Accepted Article
Title: Inversion of cpADH5 Enantiopreference and Altered Chain
Length Specificity for methyl 3-hydroxyalkanoates
Authors: Yunus Ensari, Gaurao V. Dhoke, Mehdi D. Davari, Marco
Bocola, Anna Joelle Ruff, and Ulrich Schwaneberg
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201702581
Link to VoR: http://dx.doi.org/10.1002/chem.201702581
Supported by

Chemistry - A European Journal
10.1002/chem.201702581
Submission to: Chemistry – A European Journal
Manuscript type: Full Paper
Inversion of cpADH5 Enantiopreference and
Altered Chain Length Specificity for methyl 3-
hydroxyalkanoates
1,2§
1§
1
1
1
Yunus Ensari, Gaurao V. Dhoke, Mehdi D. Davari, Marco Bocola, Anna Joëlle Ruff,
1
,3
Ulrich Schwaneberg, *
1
RWTH Aachen University, Lehrstuhl für Biotechnologie, Worringerweg 3, 52074 Aachen,
Germany, Tel.: +49 241 80 24170, Fax: +49 241 80 22387
2
Kafkas University, Faculty of Engineering and Architecture, Department of Bioengineering,
Kars, Turkey.
3
DWI-Leibniz Institut für Interaktive Materialien, Forckenbeckstraße 50, 52056 Aachen,
Germany
§
Y.E. and G.V.D. are joint first authors.
*
Corresponding author: u.schwaneberg@biotec.rwth-aachen.de
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.2017xxxxx
1
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
Abstract
Expanding the substrate scope of enzymes opens up new routes for synthesis of valuable
chemicals. Ketone-functionalized fatty acid derivatives and corresponding chiral alcohols are
valuable building blocks for the synthesis of a variety of chemicals including pharmaceuticals.
The alcohol dehydrogenase from Candida parapsilosis (cpADH5) catalyzes the reversible
oxidations of chiral alcohols and has a broad substrate range; a challenge for cpADH5 is to
convert alcohols with small substituents (methyl or ethyl) next to the oxidized alcohol moiety.
Molecular docking studies revealed that W286 is located in the small binding pocket and
limits the access to substrates that contain aliphatic chains longer than ethyl substituent. In the
current manuscript, we report that positions L119 and W286 are key residues to boost
oxidation of medium chain methyl 3-hydroxy fatty acids; interestingly the enantiopreference
toward methyl 3-hydroxybutyrate was inverted. Kinetic characterization of W286A showed a
5
.5 fold increase of Vmax and a 9.6 fold decrease of K values toward methyl 3-
m
hydroxyhexanoate (Vmax: 2.48 U/mg and K : 4.76 mM). Simultaneous saturation at positions
m
1
19 and 286 library yielded a double mutant (L119M/W286S) with more than 30 fold
improved activity toward methyl 3-hydroxyoctanoate (WT: no conversion; L119M/W286S:
0 % ) and inverted enantiopreference (S-enantiomer ≥ 99 % activity decrease and R-
3
enantiomer > 20 fold activity improvement) toward methyl 3-hydroxybutyrate.
Keywords: protein engineering, directed evolution, alcohol dehydrogenase, CPCR, fatty acid
methyl ester, methyl 3-hydroxyalkanoate
Running title: Inversion of cpADH5 Enantiopreference and Altered Chain Length Specificity
2
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
Introduction
Selective oxidation of primary or secondary alcohols to carbonyl compounds (aldehydes and
[
1]
ketones) is a key reaction in organic synthesis and industry. Ketones can be converted into
various chemical compounds via redox reactions, chain elongation, trans-amination or -
[
2]
acylation, and aldol condensation. Furthermore, ketone groups are used for the synthesis of
[
3]
asymmetric amines. Ketone-functionalized fatty acid derivatives are versatile compounds in
organic chemistry, are used as intermediates and building blocks in the production of
[
1a, 4]
detergents, plasticizer, multipurpose grease, cosmetics, and lubricants.
α-keto fatty acids
[
5]
or their methyl esters are used for L-amino acid and building block synthesis and β-keto
[
4a]
fatty acids and esters for biologically active heterocycles .
β-Keto esters, such as methyl 3-ketoalkanoates, are important synthons in organic synthesis
[
4a, 4c, 6]
for the production of a number of natural products.
Methyl 3-oxobutyrate has been
[
7]
used for the synthesis of pharmaceuticals, dye compounds and agricultural chemicals.
Methyl 3-oxohexanoate is utilized as building block and intermediate in the synthesis of
pharmaceutically active compounds such as mitochondrial branched-chain aminotransferase
[
8]
inhibitor and G-lactam derivatives. Methyl 3-oxooctanoate is used as intermediate in
synthesis of vitamin D derivatives and therapeutic agents used in treating high blood pressure,
[
9]
cardiac disease, and arteriosclerosis. Furthermore, enantiomerically pure 3-hydroxyalkanoic
acid monomers are of high interest as building block for the synthesis of new biodegradable
polymers, antibiotics, vitamins, flavor additive, pheromones. Corresponding methyl esters are
[10]
in discussion as biofuel .
Oxidation of alcohol to ketones can be performed efficiently by a great number of chemical
oxidation methods. Most commonly used processes comprise a variety of organic or inorganic
oxidants, i.e. Corey oxidation with pyridinium chlorochromate, Dess-Martin oxidation with
3
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
periodinane, Oppenauer oxidations with aluminium isopropoxide in toxic and organic
solvents such as methanol, tetrahydrofuran (THF), acetone, and acetonitrile at elevated
[
1a, 1c, 11]
temperature.
Oxidation catalysts have achieved higher conversion rates, unfortunately
[
1a,
often large amount of organic or inorganic wastes and undesired by-products are produced.
1
c, 11b, 12]
Aerobic oxidation of alcohols in aqueous solutions using sodium, iron and gold based
[
11d, 11e,
catalysts for eco-friendly and efficient catalysis has recently attracted a lot of attention.
1
3]
Selectivity, undesired by-product formation, substrate scope and conversion rates are still
[
1c, 12, 14]
major challenges in oxidation of alcohols.
Alcohol dehydrogenases (ADHs) are
promising alternatives to metal catalysts and perform reversible oxidations of primary or
+
secondary alcohols to their corresponding aldehydes or ketones, respectively using NAD or
NADH as a cofactor (Figure 1) . Several alcohol dehydrogenases and their engineered
mutants from yeast, Lactobacillus brevis, Rhodococcus erythropolis and Ralstonia eutropha
were identified. Oxidation capability of aforementioned ADHs toward primary and secondary
alcohols, cyclic alkanes, bulky alcohols such as cyclohexanol, cyclooctanol, 2-
[
15]
phenylenthanol, 2-octanol was described in literature.
Candida parapsilosis alcohol dehydrogenase 5 (cpADH5), formerly also named as Candida
parapsilosis carbonyl reductase 2 (CPCR2), belongs to the zinc-dependent medium chain
[
16]
alcohol dehydrogenases (MDRs) family.
cpADH5 has a broad substrate range toward
aliphatic ketones, keto esters, cyclic ketones, and aryl or aromatic substituted secondary
[
16a, 16c, 17]
alcohols.
secondary alcohols such as (S)-2-butanol, (S)-1-phenylethanol, and achiral 2-propanol.
cpADH5 has mainly been used for asymmetric reduction of prochiral aldehydes or ketones to
Yamamoto and coworkers reported that cpADH5 is capable of oxidizing
[
18]
[
19]
the enantiomerically pure (S)-enantiomeric alcohols
as the reduction rate of cpADH5 is
Recently, cpADH5 was engineered to
[
16a]
more than 200 times higher than its oxidation rate.
extend its substrate scope toward methylated cyclic in order to generate new synthesis routes
4
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
through enantiomeric alcohols with a high added value, such as 2-methyl cyclohexanone and
[
20]
2
-chloroacetophenone. Furthermore, cpADH5 has been investigated in detail by employing
computational approaches. Carbonyl reduction mechanism as well as structural determinants
of substrate spectrum was determined by using quantum mechanichs/molecular mechanics
[
17, 21]
(
QM/MM) techniques.
predicted through virtual screening and QM/MM calculations.
binding pocket was engineered through structure guided design for reduction of aryl ketones
Natural substrates and thus, function of the cpADH5 were
[
16a]
In addition, its substrate
[
22]
such as 2-bromo-4`-chloroacetophenone.
Recently, enantioselectivity inversion of
Thermoanaerobacter brockii secondary alcohol dehydrogenase (TbSADH) was explained
[
23]
through Molecular Dynamics simulations.
Engineering of proteins through rational or semi-rational approaches enables to tailor enzyme
[
24]
properties such as substrate acceptance, thermostability and enantiopreference. This study
reports the first reengineered cpADH5 for the efficient oxidation of methyl 3-
hydroxyhexanoate and methyl 3-hydroxyoctanoate (Figure 1), which were not converted by
the cpADH5 WT, to their corresponding ketones. In this study, we mutated and redesigned
the substrate binding pocket of cpADH5 in order to extend its substrate scope toward
medium-chain 3-hydroxy fatty acid methyl esters (FAME) which contain 6-12 carbon atoms
in their alkyl chains. Furthermore, we describe the inverted enantiopreference for the
oxidation of methyl 3-hydroxybutyrate which was observed for our engineered cpADH5
variant.
5
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
ADH
R
R
NAD+
NADH
R
R
Name
Structure
Name
Structure
1
1
1
1
1
a
b
c
d
e
CH3
CH3
methyl (S)-3-hydroxybutyrate
methyl (R)-3-hydroxybutyrate
methyl 3-hydroxypentanoate
methyl 3-hydroxyhexanoate
methyl 3-hydroxyoctanoate
2a CH3
methyl 3-oxobutyrate
C H
2c C H5
methyl 3-oxopentanoate
methyl 3-oxohexanoate
methyl 3-oxooctanoate
2
5
2
C H
2d C H8
3
8
3
C H
2e C H
11
5
11
5
+
Figure 1. Alcohol oxidation reaction catalyzed by cpADH5 by using NAD as cofactor and
chemical structures of alcohol and ketone substrates considered in this study.
Results and Discussion
cpADH5 reengineering was performed to oxidize 3-hydroxy fatty acid methyl esters to
corresponding ketones. In a first step, amino acid positions that govern oxidation of Me3-
OHC (1d) were identified by computational assisted methods, single and double site
6
saturation libraries of two identified positions were generated and screened in order to obtain
variants that oxidize Me3-OHC (1d). Afterwards kinetic characterization of four improved
6
cpADH5 variants and WT were shown. Finally, substrate specificity, selectivity and
enantiopreference of improved variants were identified and structure-function relationship of
the selective oxidation of 3-hydroxy fatty acid methyl esters of Me3-OHC oxidation were
4
discovered .
Identification of Relevant Amino Acid Positions for Site Saturation Mutagenesis
[
20]
cpADH5 WT and L119M, which converts bulky substrates,
were investigated for Me3-
OHC (1d) oxidation. The cpADH5 WT had no activity toward Me3-OHC (1d) and the
6
6
6
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
L119M variant had moderate activity toward Me3-OHC (1d) (Vmax: 0.37 U/mg). In order to
6
improve cpADH5 activity toward Me3-OHC a mechanism-based substrate docking study
6
was performed as described in Experimental Section to identify amino acid positions for Site
[16b]
Saturation Mutagenesis (SSM).
Based upon our previous experimentation with cpADH5,
the substrate binding pocket is known to be composed of two parts (a large and a small
binding pocket, Figure 2). The smaller binding pocket is close to the catalytic zinc ion and,
and can accommodate “only” methyl or ethyl groups adjacent to hydroxyl or carbonyl
[
16a, 16b, 22]
groups.
Our molecular docking studies revealed that there is insufficient space for an
alkyl chain longer than two carbon atoms next to 3-hydroxyl group of Me3-OHC (1d) in the
6
small binding pocket due to the residue W286 (Figure 3). Previous study of cpADH5 has
[
4b]
shown that butryaldehydes are natural substrates of cpADH5.
The catalytic zinc ion
activates the hydroxyl group of the substrate Me3-OHC .Thus the distance between the
6
catalytic zinc ion and the hydroxyl group of substrates (termed D1 in Figure. 3), which should
[
16b, 25]
be shorter than 2.5 Å, plays a crucial role in initiation of the catalytic activity.
Figure 3
shows that there is no interaction between the catalytic zinc ion and 3-hydroxyl group of Me3-
OHC (1d) due to a steric hindrance in the presence of W286 in WT (D1; 4.626 Å and
6
binding energy; -5.25 kcal/mol). The distance D1 between the catalytic zinc ion and 3-
hydroxyl group of Me3-OHC (1d) is shorter when tryptophan is located at position 286 is
6
substituted with alanine (D1; 1.713 Å and binding energy; -4.74 kcal/mol). The alkyl branch
(
propyl- substituent adjacent to β-hydroxyl moiety) of Me3-OHC fits well into the small
6
binding pocket of cpADH5. Molecular docking results showed that the enlarged binding
pocket of cpADH5 becomes beneficial for the oxidation of Me3-OHC (1d) to its
6
corresponding ketone.
7
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
Figure 2. Catalytic binding pocket of cpADH5; binding pocket consists of a small (red) and a
large (blue) binding pocket. The oxidized hydroxyl group of substrate adopts an orientation
toward the catalytic zinc and the side chains next to carbonyl functionality of the substrate
occupies the small binding pocket (red color) while its alkyl chain with the methyl ester group
adjusted into the large binding pocket (blue color). The catalytic zinc ions (orange sphere) and
+
their coordinating residues (C44, H65 and D154) and the NAD cofactors are shown as sticks,
while the substrates are shown in a ball-and-stick representation. The substitution W268 is
shown in red and L119 in blue.
In summary, preliminary experimental results and computational analysis identified the
positions L119 and W286 for SSM studies to improve oxidation performance of cpADH5
toward Me3-OHC (1d).
6
8
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
Figure 3. Catalytic binding poses of methyl 3-hydroxyhexanoate in the active site of cpADH5
A; WT and B; W286A variant. The catalytic zinc ions (orange sphere) and their coordinating
residues (C44, H65 and D154) along with S46 (involved in proton relay mechanism) and the
+
NAD cofactors are shown as sticks, while the substrates (Me3-OHC (1d)) are shown in a
6
ball-and-stick representation. The substitution W268 is shown in magenta. Non-polar
hydrogen atoms are not shown for clarity. D1 distance is shown as reciprocal arrow.
Screening of SSM Libraries toward Methyl 3-hydroxyhexanoate and Identification of
Activity Improved Variants
The compound Me3-OHC (1d) was selected as substrate among 3-hydroxy fatty acid methyl
6
esters because the corresponding ketone is an important intermediate for the production of
[
8]
pharmaceutically active compounds. In order to obtain improved cpADH5 variants for
Me3-OHC (1d) oxidation, NNK single site saturation mutagenesis libraries at the positions
6
L119 and W286 were generated as described in Experimental Section. Both SSM libraries
were screened toward Me3-OHC (1d) oxidation by spectrophotometrically monitoring of
6
9
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
[
22]
NADH production as previously reported. One hundred eighty clones were screened per
[
26]
SSM positions (95 % coverage ) and 1080 clones for SSM double positions (83 %
[
26]
coverage; calculated with GLUE-IT ). Twenty variants with high activity toward Me3-
OHC (1d) were rescreened and the 5 most active variants were sequenced for each library. In
6
all five cases an exchange from leucine to methionine was obtained at position L119, which is
in agreement with previous reports about the conversion of bulky substrates like 2-methyl
[
20]
cyclohexanone. At position 286, the two variants, with the amino acid exchanges W286A
and W286S showed a significantly higher activity toward Me3-OHC (1d). Moreover, these
6
activity results obtained from variants W286A and W286S are in accordance to our docking
prediction which implies that enlarging the cavity size enables conversion of Me3-OHC (1d).
6
Sequencing of the 5 most promising variants from SSM double library yielded in all cases the
variant L119M/W286S.
Kinetic Characterization of Improved cpADH5 Variants
The kinetic characterization of purified (> 98 % purity, see Figure S2 in Supporting
Information) cpADH5 WT, L119M, W286A, W286S, and L119M/W286S was performed by
determining Vmax, K , and k values toward Me-(S)-3-OHC (1a), Me3-OHC (1c), and
m
cat
4
5
Me3-OHC (1d) and showed in Table 1. As already mentioned, cpADH5 WT has almost no
6
activity toward Me3-OHC (1d). Thus, kinetic values of cpADH5 toward Me3-OHC (1d)
6
6
could not be calculated even with increasing the substrate concentration up to 100 mM. Out of
four improved variants, L119M/W286S showed the highest Vmax value for Me3-OHC (1d)
6
which is approximately 6-fold higher when compared to L119M. Furthermore, the
L119M/W286S variant displays a cooperative effect on turnover number (k ) for the
cat
oxidation of Me3-OHC (1d) compared to single substituted variants L119M and W286S
6
(
L119M/W286S: 160,03 1/s; L119M: 16,80 1/s; W286S: 87,99 1/s). Among the single
substituted variants, the W286A variant is the “best-performing” variant toward Me3-OHC6
1
0
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
(
1d) in respect to turnover number. Moreover, W286A showed a 9.6-fold decreased K value
m
and a 5.5-fold higher Vmax value toward Me3-OHC (1d) when compared to the L119M
6
variant. These results suggest that by enlarging the small binding pocket of cpADH5 instead
of modifying the large binding pocket, a more beneficial effect for the conversion of the target
substrate Me3-OHC (1d) can be achieved. As general trend all improved variants showed
6
compared to cpADH5 WT improved activity and decreased K toward target substrate Me3-
m
OHC (1d).
6
Kinetic performance of cpADH5 variants changed with length of the alkyl branch (ethyl- to
propyl- substituent) from Me-3-OH-C to Me-3-OH-C .The cpADH5 L119M variant has a
4
6
more than 2 fold increase in Vmax value for Me-(S)-3-OHC (1a) in comparison to the
4
cpADH5 WT without a significant change in substrate affinity. Substitutions at position 286
proved not to be beneficial or even decreased in case of the variant (W286A, W286A, and
L119M/W286S) the activity toward Me-(S)-3-OHC (1a). None of the cpADH5 variants (see
4
Table. 1) had significant Vmax improvements for Me3-OHC (1c) when compared to the
5
cpADH5 WT. The variant W286A had a 1.6 fold increase in turnover number and a 2.7 fold
decreased K value.
m
1
1
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
+
Table 1. Kinetic parameters of cpADH5 variants for selected substrate oxidation. NAD (0.25
mM) reduction was monitored spectrophotometrically for oxidation of selected substrates at
3
40 nm. Obtained data were fitted with the GraphPad Prism6 software for Windows using the
Michaelis-Menten model. All conversions were performed at least in triplicates. Substrates
were used as racemate unless stated otherwise. ND: not determined.
Variant
Substrate
Vmax [U/mg]
4.82 ± 0.23
1.84 ± 0.03
ND
K [mM]
kcat [1/s]
232.18
88.49
ND
k /K [1/mM s]
m
cat
m
methyl (S)-3-hydroxybutyrate
methyl 3-hydroxypentanoate
methyl 3-hydroxyhexanoate
methyl (S)-3-hydroxybutyrate
methyl 3-hydroxypentanoate
methyl 3-hydroxyhexanoate
methyl (S)-3-hydroxybutyrate
methyl 3-hydroxypentanoate
methyl 3-hydroxyhexanoate
methyl (S)-3-hydroxybutyrate
methyl 3-hydroxypentanoate
methyl 3-hydroxyhexanoate
methyl (S)-3-hydroxybutyrate
methyl 3-hydroxypentanoate
methyl 3-hydroxyhexanoate
1.43 ± 0.22
12.94 ± 0.48
ND
162.36
cpADH5
WT
6.84
ND
10.89 ± 0.49
0.28 ± 0.01
0.37 ± 0.01
0.48 ± 0.02
2.48 ± 0.06
2.02 ± 0.09
0.77 ± 0.02
1.36 ± 0.03
1.79 ± 0.06
0.44 ± 0.01
0.27 ± 0.01
2.17 ± 0.08
1.59 ± 0.19
0.60 ± 0.05
29.76 ± 2.3
5.32 ± 0.50
4.76 ± 0.38
3.09 ± 0.39
4.47 ± 0.25
13.63 ± 0.64
3.14 ± 0.29
3.50 ± 0.22
8.30 ± 0.72
5.34 ± 0.49
981.08
12.16
16.80
27.03
139.86
113.63
37.99
66.62
87.99
32.45
19.76
160.03
618.59
20.24
0.56
cpADH5
L119M
5.08
cpADH5
W286A
29.38
36.89
8.50
cpADH5
W286S
4.89
28.02
9.26
cpADH5
L119M /
W286S
2.38
29.97
General Trends in Substrate Specificity, Selectivity, and Enantiopreference of Improved
cpADH5 Variants
Chain length specificity
The targeted 3-hydroxy FAME substrate for cpADH5 engineering studies was Me3-OHC6
(1d). In order to investigate the chain length specificity of improved cpADH5 variants, three
shorter (Me-(S)-3-OHC (1a), Me-(R)-3-OHC (1b), and Me3-OHC (1c)) and one longer
4
4
5
(
Me3-OHC (1e)) 3-hydroxy FAME substrates were investigated for oxidation to their
8
corresponding ketones. Investigations also comprised reduction of 3-keto FAME substrates
Me3-OC (2a), Me3-OC (2c), Me3-OC (2d), and Me3-OC (2e)). As a general trend,
(
4
5
6
8
1
2
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
cpADH5 variants with substitutions in the small binding pocket (W286A, W286S, and
L119M/W286S) showed a chain length specificity shift from small chain substrates (Me-(S)-
3
-OHC (1a), Me-(R)-3-OHC (1b), and Me3-OHC (1c)) to medium chain substrates (Me3-
4 4 5
OHC (1d) and Me3-OHC (1e)). Moreover, W286A, W286S, and L119M/W286S showed a
6
8
significant decrease in the reduction rate of Me3-OC (2a), Me3-OC (2c), and a moderate
4
5
decrease in the reduction of both Me3-OC (2d) and Me3-OC (2e). Subsequently, molecular
6
8
docking studies were performed in order to elucidate the structure-function relationship
underlying the activity improvement toward Me3-OHC (1d) and Me3-OHC (1e) as well as
6
8
inversion of the enantiopreference for methyl 3-hydroxybutayrate.
Figure 4. Specific activities of cpADH5 WT and improved variants for oxidation of 3-
hydroxy FAME substrates (Me-(S)-3-OHC , Me-(R)-3-OHC , Me-3-OHC , Me-3-OHC , Me-
4
4
5
6
3
-OHC ). All activity measurements were performed at 10 mM substrate concentration and
8
NADH production was monitored at 340 nm.; WT, red; L119M, grey; W286A, blue; W286S,
green and L119M/W286S, black. Measurements were performed in triplicate and specific
activities were normalized to the activity of cpADH5 WT toward methyl (S)-3-
hydroxybutyrate.
1
3
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
Chain selectivity
The cpADH5 WT has the highest oxidation activity toward Me-(S)-3-OHC (1a) among the
4
five investigated 3-hydroxy FAMEs (Figure 4). cpADH5 WT activity is in agreement with
previous studies reporting a poor conversion of substrates that carry bulky side chains next to
[
15a, 20]
carbonyl functionality (e.g. 3-hexanone, 2-methylcyclohexanone, and propiophenone).
Docking simulations showed that the Me-(S)-3-OHC (1a) which only has a methyl
4
substitution on the small aliphatic chain that can easily occupy the smaller binding pocket of
cpADH5 WT as depicted in Figure 5. Therefore Me-(S)-3-OHC (1a) can easily be converted
4
by cpADH5 WT. The distance D1 between the hydroxyl oxygen and the catalytic zinc ion
was determined to be below 2 Å for Me-(S)-3-OHC (1a), which indicates that catalytic
4
activity should be initiated by the catalytic zinc ion. This is due to fact that substrates with a
single methyl substituent at their alkyl branch are sterically similar to the natural substrates
[
16a]
(butyraldehydes) of cpADH5.
All other variants showed a decrease of up to 95 % in
activity for the oxidation of Me-(S)-3-OHC (1a) with the exception of L119M. Variants with
4
a decreased activity toward Me-(S)-3-OHC (1a) showed an increase in binding energy and
4
catalytically important distances (see Supporting Information Table 1-5). As an example, the
W286S variant has a 0.28 kcal/mol increase in binding energy and in D3 distance to 4.07 Å.
The distance D3 between the terminal carbonyl oxygen of Me-(S)-3-OHC (1a) and the side
4
[
16b]
chain of catalytical S46 is shorter than 3 Å
which explains the decrease of the activity of
the W286S variant toward Me-(S)-3-OHC (1a). The increased size of the small binding
4
pocket also reduces the oxidation activity toward Me-(S)-3-OHC (1a). While the L119M
4
mutation decreases the total binding pocket size, mutations of residue 286 increased the total
size by enlarging the small binding pocket (Table S6 in Supporting Information, the binding
[
27]
site volume was calculated using HotSpot Wizard2.0). The latter is in accordance with a
previous study by Green et al. (1993) which reported that an enlargement of the binding
1
4
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
pocket has a negative effect on the conversion of small substrates by yeast alcohol
[
15a]
dehydrogenase.
Figure 5. A) Catalytic binding pose of methyl (S)-3-hydroxybutyrate (Binding energy; -4.44
kcal/mol) and B) Non-catalytic binding pose of methyl (R)-3-hydroxybutyrate (Binding
energy; -4.40 kcal/mol) in the active site of cpADH5 WT. Catalytic zinc ion (orange sphere)
and its coordinating residues (C44, H65 and D154) along with S46 (involved in proton relay
+
mechanism) and NAD cofactor are shown as sticks, however the substrates are shown in ball
and stick representation. Non-polar hydrogens were not shown for clarity. D1 distance is
shown as reciprocal arrow.
Enantiopreference
In order to investigate whether the L119M, W286A, W286S, and L119M/W286S substitution
alter the enantiopreference of the methyl 3-hydroxybutyrate, the activity toward Me-(R)-3-
OHC4 (1b) was determined. Surprisingly, the L119M/W286S variant displayed an
1
5
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
enantiopreference inversion toward methyl 3-hydroxybutyrate. L119M/W286S showed 99 %
decrease in relative activity toward Me-(S)-3-OHC (1a). Interestingly, L119M/W286S also
4
has an cooperative effect toward Me-(R)-3-OHC (1b) and showed a 23-fold activity increase
4
toward Me-(R)-3-OHC (1b) compared to cpADH5 WT. cpADH5 WT shows Prelog
4
[
28]
[16a]
specificity
and thus preferentially converts Me-(S)-3-OHC (1a)
and has no activity
4
toward Me-(R)-3-OHC (1b) (Figure 4).
4
Molecular docking revealed that no catalytic binding mode with a D1 <4 Å for Me-(R)-3-
OHC (1b) could be detected in the cpADH5 WT. The hydroxyl group on which the actual
4
oxidation takes places is far away from the catalytic zinc ion meaning that there is no
activation of reactive hydroxyl group of Me-(R)-3-OHC (1b) and consequently no oxidation
4
can take place. Molecular docking results were verified by experimental results, in which
cpADH5 WT shows higher activity toward Me-(S)-3-OHC (1a). This explains the fact that
4
cpADH5 WT is preferentially an S-selective ADH. A high enantiopreference is a key factor
for the production of pure enantiomers; therefore we studied in the detail the underlying
[
29]
structure-function relationships on the molecular level through molecular docking studies.
Molecular mechanisms underlying enantiomeric recognition of substrates by enzymes are not
well understood yet. Computational analysis were focused on the L119M/W286S variant. In
case of cpADH5 WT, there is not enough space for the suitable catalytic orientation of Me-
(
R)-3-OHC (1b) in the small binding pocket due to the blockage by large W286 residue; only
4
Me-(S)-3-OHC (1a) is capable of a catalytically active orientation (Figure 5). The latter also
4
explains the fact that cpADH5 WT is an S-selective enzyme and enables a conversion rate of
1
a of up to 99 %. To explain the shift in the enantiopreference from S to R in case of Me-(S)-
3
-OHC (1a), flexible docking has been carried out. In our docking protocol, all the residues
4
which are within 3 Å from the mutated positions were treated as flexible to take into account
conformational changes introduced by the mutation, while the rest of the protein was treated
1
6
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
as rigid. In case of the double variant L119M/W286S unreliable results were obtained with a
[
16b]
rigid ADH cavity (receptor) as reported in our previous publication.
The introduction of
flexibility to some of the residue sidechains which are in the vicinity of mutated amino acids
therefore is essential to further understand the effect of substitutions. In our docking
simulations, we found that after mutation of the small binding pocket of cpADH5, sufficient
3
3
space (588 Å for L119M and 636 Å for L119M/W286S) is available for the proper
orientation and catalytic binding of Me-(R)-3-OHC (1b), especially after replacing the W286
4
residue with serine along with L119M. The docking energy of -4.45 kcal/mol for the (R)-
enantiomer of methyl 3-hydroxybutyrate was found to be increased in comparison to its (S)-
enantiomer (-4.30 kcal/mol; Table S3 in Supporting Information), whereas in the WT no
productive docking orientation of the (R)-enatiomer could be found. Moreover, a
conformational change of the neighboring residues was observed in flexible sidechain
docking. C57 shows a slight flipping, whereas W116 displays a complete flipping which
indicates also indirect effects of the mutation L119M (Figure 6). Therefore, the binding of
Me-(R)-3-OHC (1b) in a catalytic binding mode is enabled by conformational changes
4
enlarging the small binding pocket in the cpADH5 double variant (L119M/W286S) which
leads to a faster conversion and inversion of enantiopreference from S to R. For a fully
quantitative prediction for enantiopreference of enzymatic reactions more sophisticated
[23]
computational protocols including MD simulations and QM/MM calculations are required
[
22]
to include also backbone rearrengments and free energy contributions. Wang et al.
observed a similar shift of enantiopreference as an effect of the modification of residues in the
small binding pocket of cpADH5. The increase in size of the small binding pocket (W285A,
W286A, and W285A/W286A) showed a shift in enantiopreference for reduction of aryl
ketones such as 3-chloroacetophenone, 4-bromoacetophenone, and 4-chloro-2-
[
22]
bromoacetophenone. Deng et al., inverted the enantiopreference of Gluconobacter oxydans
carbonyl reductase toward ethyl 2-oxo-4-phenylbutyrate by mutating the active site residues
1
7
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
C93 and Y149. Increased flexibility in the binding pocket resulted in case of the substitutions
[
30]
C93P/Y149A in an inversion of enantiopreference from R to S.
A
B
C
C44
C44
Zn
C44
D154
NAD+
D154
S46
Zn
NAD+
H65
D1
Zn
D1
NAD+
D1
H65
D154
H65
S46
S46
Alcoholic
oxygen
L119
S286
M119
W286
S286
C57
C57
C57
W116
W116
W116
Figure 6. The binding poses of methyl (R)-3-hydroxybutyrate to the cpADH5 (A) WT, (B)
W286S, and (C) L119M/W286S (Binding energies for WT, W286S, and L119M/W286S are -
4
.40, -4.21, and -4.45 kcal/mol respectively). Catalytic zinc-ion (orange sphere) and its
coordinating residues (C44, H65, and D154) along with S46 (involved in proton relay
+
mechanism) and NAD cofactor are shown as sticks, while the substrate is shown in ball and
stick representation. Non-polar hydrogens were not shown for clarity. The substitutions are
shown in magenta; whereas residues showing conformational changes are indicated by green
letters. D1 distance is shown as reciprocal arrow.
In order to investigate the effect of an enlarged binding pocket on the conversion of longer
substrates, the activity of cpADH5 toward Me3-OHC (1e) was quantified. Me3-OHC (1e)
8
8
showed a similar trend to Me3-OHC (1d) for W286A and W286S. The L119M/W286S
6
variant showed the highest activity toward Me3-OHC (1e). While a cooperative effect of the
8
double substitution at the positions L119/W286 was observed for specific activity toward
Me3-OHC (1e), the same effect could not be observed for Me-(S)-3-OHC (1a), Me3-OHC
5
8
4
(
1c), and Me3-OHC (1d). Molecular docking of cpADH5 WT shows no catalytic binding
6
1
8
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
pose for Me3-OHC (1e). The same results were obtained for Me3-OHC (1d) . The resulting
8
6
best binding pose shows a non-catalytic orientation of the alcoholic group (D1; >4 Å away
from the catalytic zinc ion and binding energy -5.82 kcal/mol; Figure 7). The docking results
of cpADH5 L119M/W286S revealed a catalytic binding pose with a 1.68 Å D1 value for
Me3-OHC (1e) and a binding energy of -5.70 kcal/mol.
8
A
B
C44
S46
D154
C44
NAD+
Zn
S46
D1
H65
D1
H65
NAD+
Zn
Alcoholic
oxygen
D154
S286
M119
methyl 3-hydroxyoctanoate
methyl 3-hydroxyoctanoate
C57
W286
W116
Figure 7. Catalytic binding poses of methyl 3-hydroxyoctanate (1e) in the active site of
cpADH5 A; WT and B; L119M/W286S variant. Catalytic zinc-ion (orange sphere) and its
coordinating residues (C44, H65 and D154) along with S46 (involved in proton relay
+
mechanism) and the NAD cofactor are shown as sticks while the substrate is shown in ball
and stick representation. Substitutions L119M/W286S are displayed in magenta. Non-polar
hydrogens were not shown for clarity. Angle of L119M/W286S variant was changed for
better visualization. D1 distance is shown as reciprocal arrow.
1
9
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
Oxidation or reduction preference of cpADH5 variants
[
16a]
As previously reported
, cpADH5 mainly prefers carbonyl substrates instead of alcoholic
ones (Vmax values toward phenylacetaldehyde and corresponding alcohol 2-phenylethanol are
1.74 U/mg and 0.25 U/mg respectively). In terms of reduction reactions, the L119M variant
has an improved reduction activity toward bulky substrates (e.g. 2-methyl cyclohexanone, 3-
5
[
20]
methyl cyclohexanone, and 4-methyl cyclohexanone) and a slightly improved activity for
reduction of Me3-OC (2a), Me3-OC (2c), and Me3-OC (2d) (Figure S3 in Supporting
4
5
6
Information). Additionally, a similar reduction activity was found when compared to cpADH5
WT for the Me3-OC (2e) reduction. The substitution L119M is located in the large binding
8
pocket and the W286 substitutions (W286S, W286A) is located in the small binding pocket.
The variants W286A, W286S, and L119M/W286S show a decrease of up to 95 % in
reduction activity for Me3-OC4 (2a), Me3-OC (2c) and a decrease of up to 20 % in reduction
5
activity for Me3-OC (2d) and Me3-OC (2e) which can be attributed to the W286
6
8
substitutions in the small binding pocket.
Conclusions
In essence, enlarging the cavity size enabled cpADH5 to broaden the substrate scope by
oxidizing Me3-OHC (1d) and Me3-OHC (1e) to the corresponding ketones which cannot be
6
8
converted by the cpADH WT. Additionally, W286A, W286S, and L119M/W286S variants
have a chain length specificity shift from Me-(S)-3-OHC (1a), Me-(R)-3-OHC (1b), and
4
4
Me3-OHC (1c)) to the medium chain substrates (Me3-OHC (1d) and Me3-OHC (1e)).
5
6
8
Furthermore, the L119M/W286S variant shows an inversion of enantiopreference for methyl
-hydroxybutyrate. Molecular docking simulations revealed that an enlarged binding pocket
of cpADH5 leads to conformational changes and enantiopreference shift.
3
2
0
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
The latter studies revealed especially for the L119M/W286S variant how cpADH5’s
enantiopreference can be inverted for methyl 3-hydroxybutyrate and generated a first
understanding on the molecular level. Deepen these studies for 3-hydroxy FAMEs would
open novel synthetic routes to attractive building block in the synthesis of pheromones and β-
[
28]
lactam antibiotics.
Experimental Section
Chemicals
All chemicals were purchased from Sigma-Aldrich (Steinheim, Germany), Merck (Darmstadt,
+
Germany), and Roth (Karlsruhe, Germany) if not stated otherwise. NAD and NADH cofactor
was purchased from Sigma-Aldrich (Steinheim, Germany) and Julich Fine Chemicals (Julich,
Germany) respectively. PhuS polymerase was produced in-house. Salt-free oligonucleotides
were purchased from Eurofins MWG Operon (Ebersberg, Germany). Enzymes were
purchased from New England Biolabs (Frankfurt, Germany).
Construction of Single and Double Site-Saturation Mutagenesis (SSM) Libraries of
cpADH5
[
20]
pET22b-cpADH5, previously constructed , was used as template for saturation
[
20, 22]
mutagenesis. Positions 119 and 286 were selected according to a previous report
and
computational analysis respectively. For generation of SSM libraries, a modified two step
[
31]
QuikChange Mutagenesis (QCM) protocol was applied.
In a first step, two PCRs were
performed in separate tubes; one containing the forward primer (F286-NNK for residue 286
and F119-NNK for residue 119) and the other containing the reverse primer (R286-NNK for
residue 286 and R119-NNK for residue 119). After 3 PCR amplification cycles, the two
reactions were mixed and PCR was performed for 15 cycles. Each PCR reaction contained 15
ng plasmid template DNA, 1 U PhuS DNA polymerase, 1x PhuS DNA polymerase buffer, 0.2
2
1
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
mM dNTP mix and 0.4 µM primer (forward and reverse). The following NNK degenerated
primers were used for the construction of SSM libraries:
Position 286
F286-NNK; 5'-CTGGGTAGCTTTNNKGGCACCACCAATG-3' and
R286-NNK; 5'-CATTGGTGGTGCCMNNAAAGCTACCCAG-3',
Position 119
F119-NNK; 5'-GGCGACTGGTTTGGTNNKGGTTATGATGGTGG-3' and
R119-NNK; 5'-CCACCATCATAACCMNNACCAAACCAGTCGCC-3'.
Obtained PCR products were digested with 20 U DpnI (New England Biolabs, Frankfurt am
Main, Germany) to remove template DNA and were purified with PCR purification kit
®
according to manufacturer´s guideline (NucleoSpin Gel and PCR Clean-up kit, Macharey-
Nagel, Düren, Germany). Libraries were transformed into E. coli BL21 Gold (DE3) and
plated on LBAMP (100 µg/ml ampicillin) agar plates.
For the generation of a double Site Saturation Mutagenesis library, the first position (L119)
was mutated as described above, and transformed into E.coli DH5α and plated on LBAMP agar
plates. Resulting colonies were collected from each plate and plasmid purification was
®
performed (NucleoSpin Plasmid kit, Macharey-Nagel, Düren, Germany). The purified
plasmid was used as template for the mutagenesis at the second position (W286). DNA
concentrations were determined using the NanoDrop photometer from NanoDrop
Technologies (Wilmington, USA). DNA sequencing was performed by Eurofins MWG-
Operon (Ebersberg, Germany). Sequencing data was analyzed with the Clone Manager 9
Professional Edition Software (Scientific & Educational Software, Cary, NC, USA).
2
2
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
Libraries were constructed by transferring single colonies from each agar plate into 96-well
flat bottom microtiter plates (Greiner Bio-One GmbH, Frickenhausen, Germany) filled with
LB medium (150 µL; supplemented with 100 µg/ml ampicillin). Plates were tightly sealed
with insulation tape and incubated in a humidified microtiter plate shaker (Multitron II, Infors
GmbH, Einsbach, Germany; 16 h, 37 °C, 900 rpm, 70 % humidity). For long-term storage,
1
00 µl of 50 % (v/v) glycerol was added to each well and the libraries were stored at -80 °C as
master microtiter plate.
Protein Expression in 96-Well Plates
Expression in microtiter plates was performed by preparing a pre-culture by duplicating a
master plate with 96-pin replicator to 150 µL of LB medium supplied with ampicillin (100
µg/ml) in 96-well flat bottom microtiter plates. Plates were closed with lids, tightly sealed
with insulation tape and incubated for 16 h (37 °C, 900 rpm, 70 % relative humidity).
Subsequently, 10 µL of the pre-culture were transferred to V-bottom microtiter plates
(
transparent polystyrene plate, Corning GmbH, Kaiserslautern, Germany) containing 140 µl
terrific broth (TB) medium supplemented with ampicillin (100 µg/ml), trace element solution
3.4 µM CaCl , 0.6 µM ZnSO , 0.6 µM MnSO , 54.0 µM Na –EDTA, 61.8 µM FeCl , 0.6
(
2
4
4
2
3
µM CuSO and 0.8 µM CoCl in water), IPTG (0.1 mM), thiamine hydrochloride (0.3 mM)
4
2
and ZnCl (1 mM). Plates were closed with lids, tightly sealed and incubated in a microtiter
2
plate shaker (20 h, 15 °C, 900 rpm, 70 % relative humidity). Expression cultures were
harvested by centrifugation (4 °C, 3220 g, 15 min) using an Eppendorf 5810R centrifuge
(Eppendorf AG, Hamburg, Germany), the supernatant discarded and the pellets stored at -20
°
C until further use.
2
3
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
9
6-Well MTP Activity Assay for Screening of SSM Libraries.
The expressed libraries in 96-well microtiter plates were removed from cold storage (-20 °C)
and incubated at room temperature for 10 min. Cell lysates were prepared by resuspending
each cell pellet in lysis buffer (150 µL; 50 mM NaH PO , 300 mM NaCl, and 1 g/L
2
4
lysozyme). Cell lysis was performed by incubating the resuspended cells in a microtiter plate
shaker (1 h, 37 °C, 900 rpm, 70 % relative humidity) and followed by centrifugation (4 °C,
3
220 g, 20 min).
Screening of single and double SSM libraries was performed through a spectrophotometric
Tecan Infinite M200 PRO) monitoring of NADH generation as a result of oxidation of
methyl 3-hydroxyhexanoate to the corresponding ketone by cpADH5. Each reaction setup in
6 well MTP (Greiner Bio-One GmbH, Frickenhausen, Germany) contained 50 µl of crude
(
9
extract, 125 µL of 120 mM methyl 3-hydroxyhexanoate in triethanolamine buffer (TEA, 100
+
mM) at pH 8.0 and 25 µL of TEA buffer. The reaction was initiated by the addition of NAD
(50 µl of 1 mM) and absorption was monitored at 340 nm for 30 min with 30 sec. intervals at
3
0 °C (ɛ = 6200 1/(M cm)).
Flask Expression and Purification of cpADH5 Protein Variants.
cpADH5 WT and selected variants were expressed in shake flasks and purified through
[
16a]
streptavidin gravity column as previously described by Dhoke and Ensari et al.
Purified
cpADH5 variants were analyzed with the Experion Automated Electrophoresis System (BIO-
RAD, Munich, Germany) and quantified with the Pierce BCA Protein assay Kit (Thermo
Scientific, Schwerte, Germany) according to manufacturer`s protocols.
2
4
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
cpADH5 Activity Assay and Determination of Kinetic Parameters
cpADH5 activity assays in cuvette-scale were performed in Cary 50 Bio UV/vis
spectrophotometer (Varian) with temperature-regulated cuvette chamber. Substrates
concentrations ranging from 0.25 to 50 mM (Me(S)3-OHC (1a), Me3-OHC (1c), and Me3-
4
5
+
OHC (1d)) were mixed with NAD (0.25 mM) in TEA buffer (100 mM) at pH 8.0. The
6
reaction was started by addition of 100 μL purified cpADH5. NADH absorption at 340 nm
was monitored for 10 min (with the shortest feasible intervals) at 30 °C (ε = 6200 1/(M cm).
Purified cpADH5 was diluted accordingly in order to obtain a linear absorbance change in the
monitored time scale. The linear part of the measured activity curve was used for the
calculation of cpADH5 enzymatic activity. For the determination of K and Vmax values, the
m
activity assay described above was used at different concentration ranges of substrates and all
measurements were performed in triplicate. Resulting data were fitted with GraphPad Prism6
software for Windows (GraphPad Software, La Jolla California USA, www.graphpad.com)
using Michaelis-Menten model.
In order to determine the substrate scope of obtained variants, the activity toward different
chain length alcohol (Me(S)3-OHC (1a), Me(R)3-OHC (1b), Me3-OHC (1c), Me3-OHC
4
4
5
6
6
(1d), and Me3-OHC (1e)) and ketone substrates (Me3-OC (2a), Me3-OC (2c), Me3-OC
8 4 5
(
2d), and Me3-OHC (2e)) were determined in microtiter plate through monitoring NADH
8
absorbance. 5 µL of purified cpADH5 variants were mixed with 25 µL of 100 mM substrates
(in TEA buffer pH 8.0) and 170 µL of TEA buffer. Each reaction was started by addition of
+
5
0 µL 1 mM cofactor. For oxidation reactions, NAD and reduction reactions NADH were
used.
2
5
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
Molecular Docking
Molecular docking was carried out by using the dock_run.mcr macro in YASARA Structure
[
32]
Version 13.9.8 with the optimized atomic charges and Van der Waals parameters for zinc
[33]
[16b]
ions and coordinating residues . Docking of selected 3-hydroxy fatty acid methyl ester
substrates to cpADH5 WT and resulting variants have been performed in order to generate a
first molecular understanding of the structure-function relationships of cpADH5 ability to
[16b, 21]
oxidize 3-hydroxy fatty acid methyl ester. In the substrate docking protocol
, the partial
charge of the catalytic zinc environment including the zinc ion and its coordinating residues
(C44, H65, D154 and water) have been re-parameterized. In addition, the oxidation
[
25]
mechanism-based distance criteria
(distances D1, D2 and D3 shown in Figure S1 in
Supporting Information) have been applied to extract the catalytically active pose from a pool
of docking poses. Distance D1 indicates the distance between alcoholic oxygen (O) of the
substrate and catalytic zinc (Zn) ion, distance D2 indicates the distance between the alcoholic
+
hydrogen (H) of the substrate to the carbon (C4) of the pyridine ring of NAD cofactor and
distance D3 specifies the distance between alcoholic oxygen (O) and hydrogen of Ser46
which is involved in the first proton transfer. In medium chain dehydrogenases, oxidation
reaction mechanism is a zinc-dependent process. The carbonyl group is first activated by the
[
16b, 25]
catalytic zinc and then the hydride transfer takes place.
Therefore, the distance between
the catalytic zinc and carbonyl group of substrate is of primary importance to identify the
catalytic poses. QM/MM calculations demonstrated that the hydride transfer is a linear
[
16b]
transfer process with C4-H-C angle is of ~150 degrees.
As we are using the non-bonded
zinc model, the variations in distance between zinc and hydroxyl group are descriptive for
[
16a]
catalysis.
based substrate docking protocol have been published before.
generated around the binding site of cpADH5 by centering on the catalytic zinc ion. The
The detailed procedure of the re-parameterization and our applied mechanism-
[
16b]
A grid box of 14 Å was
2
6
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
receptor (herein cpADH5 enzyme, X-ray structure, 4C4O, of carbonyl reductase from
[16c]
Candida parapsilosis)
was kept rigid and ligand (herein 3-hydroxy fatty acid methyl ester
substrates) was treated as flexible. Mutations were introduced using YASARA by swapping
the residue with requisite amino acid and optimization was performed by generating different
side chain rotamers. Finally, the lowest energy conformer of the generated mutant was
selected for further studies. However, in the case of the double mutant, all the residues within
3
Å proximity of the mutated position were treated flexible to account for the additional
[
34]
flexibility introduced by the mutations. Autodock4.2 was used for substrate docking along
[
35]
with Amber99 force field. Figures showing catalytic binding poses of different substrates
in active site of cpADH5 variants were generated with PyMOL.
2
7
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
ASSOCIATED CONTENT
Supporting Information
Electrophoresis analysis of purified cpADH5 variants, specific activities of cpADH5 WT and
improved variants on various chain length 3-oxo FAME substrates, figures showing the
catalytic binding modes of the substrates obtained by substrate docking simulations, binding
energies and catalytically important distances obtained from molecular docking and, substrate
binding pocket sizes of cpADH5 variants.
AUTHOR INFORMATION
Corresponding Author
u.schwaneberg@biotec.rwth-aachen.de
Notes:
The authors declare no competing financial interest.
AUTHOR CONTRIBUTIONS
ACKNOWLEDGMENT
Y.E. thanks Kafkas University, Kars, Turkey for funding. G.D., and M.D.D. would like to
acknowledge DFG SPP 1166 BioNoCo for financial support. Simulations were performed
with computing resources granted by JARA-HPC from RWTH Aachen University under
project RWTH0116.
2
8
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.201702581
ABBREVIATIONS
ADHs, Alcohol dehydrogenases; MDR, Medium chain dehydrogenase/reductase; FAME,
Fatty acid methyl esters; SSM, Site Saturation Mutagenesis; TEA, triethanolamine; NADH, β-
Nicotinamide adenine dinucleotide; SSM, Site-Saturation Mutagenesis.
2
9
This article is protected by copyright. All rights reserved.