Asymmetric Ketone Reduction by Imine Reductases

Article abstract of DOI:10.1002/cbic.201600647

The rapidly growing area of asymmetric imine reduction by imine reductases (IREDs) has provided alternative routes to chiral amines. Here we report the expansion of the reaction scope of IREDs by showing the stereoselective reduction of 2,2,2-trifluoroacetophenone. Assisted by an in silico analysis of energy barriers, we evaluated asymmetric hydrogenations of carbonyls and imines while considering the influence of substrate reactivity on the chemoselectivity of this novel class of reductases. We report the asymmetric reduction of C=N as well as C=O bonds catalysed by members of the IRED enzyme family.

Full text of DOI:10.1002/cbic.201600647

Accepted Article
Title: Asymmetric Ketone Reduction by Imine Reductases
Authors: Maike Lenz, Jan Meisner, Leann Quertinmont, Stefan Lutz,
Johannes Kästner, and Bettina Nestl
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: ChemBioChem 10.1002/cbic.201600647
Link to VoR: http://dx.doi.org/10.1002/cbic.201600647
A Journal of
www.chembiochem.org

10.1002/cbic.201600647
ChemBioChem
COMMUNICATION
Asymmetric Ketone Reduction by Imine Reductases
Maike Lenz[a], Jan Meisner[b], Leann Quertinmont[c], Stefan Lutz[c], Johannes Kästner[b] and Bettina M.
Nestl*[a]
Abstract: The rapidly growing area of asymmetric imine reduction by
imine reductases (IREDs) has provided alternative routes to chiral
amines. Here we report the expansion of the reaction scope of IREDs
by
showing
the
stereoselective
reduction
of
2,2,2-
trifluoroacetophenone (1). Assisted by an in silico analysis of energy
barriers, we evaluated asymmetric hydrogenations of carbonyls and
imines considering the influence of substrate reactivity on the
chemoselectivity of this novel class of reductases. We report on the
asymmetric reduction of C=N as well as C=O bonds catalyzed by
members of the IRED enzyme family.
The biocatalytic, stereocontrolled addition of hydrogen from
NAD(P)H to -unsaturated carbonyl compounds, cyclic/acyclic
imines and aldehydes/ketones during asymmetric catalytic
transformation is a highly efficient and competitive alternative for
the synthesis of chiral products[1]. In this respect, the enzymatic
asymmetric reduction of imines by NADPH-dependent imine
reductases (IREDs) has been intensively developed in recent
years. To date more than 30 IREDs are characterized as
catalyzing intramolecular asymmetric reductions of various cyclic
imines[2,3]. In addition, IREDs were recently utilized as chiral
catalysts for reductive aminations, demonstrating high
chemoselectivities for reducing the in situ formed imine
intermediate while leaving the C=O bond unaffected[4,5]. In the
course of the establishment of the IRED database (Imine
Reductase Engineering Database; http://ired.biocatnet.de/), we
identified that most imine reductases functionally are annotated
as dehydrogenases. However, the examination of the putative
dehydrogenase activity for IREDs from Streptosporangium
roseum, Streptomyces turgidiscabies and Paenibacillus elgii
using glycerol-3-phosphate and 6-phosphogluconate as
substrates showed no dehydrogenase activity[6]. We and others
have recently demonstrated that IREDs appear to be incapable of
reducing C=O bonds[7,8]. Conversely, to this day no asymmetric
reduction of imines by dehydrogenases or carbonyl/keto
reductases (KREDs) has been reported[2]. This led us to explore
the relationship between IREDs and their highly related carbonyl-
reducing counterparts like hydroxylisobutyrate and ß-hydroxyacid
dehydrogenases[6,9]. The question of how proteins in nature
evolve new functions is of fundamental significance. Herein, we
describe the first example of the asymmetric reduction of
activated ketones by two enantiocomplementary IREDs (Scheme
1).
Scheme 1. Imine reductase-catalyzed asymmetric reduction of imines and
ketones.
The aim of this study was to investigate the promiscuous behavior
of IREDs in respect of ketone reductions by elucidating the
electronic effects of substituents on substrate molecules and their
influence on catalysis. IRED-catalyzed asymmetric hydrogenation
reactions using 2,2,2-trifluoroacetophenone (1a), benzoyl cyanide
(1b), acetophenone (1c) as substrates were examined. The
fluorinated aromatic carbonyl compound 1a was transformed with
57.4% and 23.8% conversion into the corresponding
α-(trifluoromethyl)benzyl alcohol (2a) by the purified R- and
S-selective IREDs, respectively (Table 1). Both IREDs formed the
S-enantiomer of 2a with excellent enantioselectivities up to 96%
(Table 1). No product formation was detected with benzoyl
cyanide (1b) and acetophenone (1c) as substrates (Table 2, entry
10 and 11). We assume the size of the cyano group of benzoyl
cyanide to cause a non-productive orientation of the substrate in
the binding pocket due to steric hindrance. Despite the fact that
acetophenone is one of the model substrates for alcohol
dehydrogenases and KREDs[10–12]
, however, only a few
dehydrogenases are described for the synthesis of enantiopure
fluorinated alcohols from 1a[12–15]. Activities between 21 and 99%
were observed in the reduction of 1a with alcohol
dehydrogenases[13–15]
.
In order to exclude the contribution of E. coli-based
dehydrogenases, a cell-free protein expression system was
applied by decoupling protein expression from the host and its
metabolic pathways. An example for the application of the in vitro
translation/transcription (IVTT) was lately shown by the group of
Lutz, who prepared a cell-free mutant library of the ene reductase
OYE1[16,17]. In biotransformations with the in vitro expressed R-
selective IRED, 5% of 2a was formed (Table 2). We used the
PUREexpress® In Vitro Protein Synthesis KIT from New England
Biolabs. The reduced product formation of 2a can be explained
by a significantly reduced enzyme concentration in the reaction.
Successful expression of the IRED was verified in
[a]
M. Lenz, Dr. B. M. Nestl
Institute of Technical Biochemistry
Universitaet Stuttgart
Allmandring 31, 70569 Stuttgart (Germany)
E-mail: Bettina.Nestl@itb.uni-stuttgart.de
J. Meisner, Prof. Dr. J. Kästner
Institute for Theoretical Chemistry
Universitaet Stuttgart
Pfaffenwaldring 55, 70569 Stuttgart (Germany)
L. Quertinmont, Prof. Dr. S. Lutz
Department of Chemistry
[b]
[c]
biotransformations
using
the
model
substrate
3,4-
dihydroisoquinoline (Figure S5, supporting information). Due to its
high substrate specificity, imine reducing dihydrofolate reductase
served as template in the negative control. As no reduction of 1a
Emory University
1515 Dickey Drive, Atlanta, Georgia 30322 (USA)
Supporting information for this article is given via a link at the end of
the document.((Please delete this text if not appropriate))
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600647
ChemBioChem
COMMUNICATION
was observed in the subsequent biotransformations, reduction by
components of the IVTT system can be excluded.
studies of Turner and Xu that demonstrated the reduction of
iminium salts and the associated faster conversion of iminium
ions in comparison with the corresponding imines[8,18]. According
to our calculations, energy barriers for selected ketones are
higher than for iminium ions of the chosen model substrates,
however, lower than for imine compounds (Table 2). As expected
the ketone reactivity is mostly influenced by the CF3-group with
the highest negative inducible effect.
Table 1. Activities and selectivities of IREDs in the conversion of
2,2,2-trifluoroacetophenone
1a into the corresponding alcohol
α-(trifluoromethyl)benzylalcohol 2a in comparison to selected controls.[a]
Table 2. In silico calculated energy barriers in kJ/mol for the
hydride transfer from the nicotinamide subunit of NAD(P)H to
imine, iminium ion and ketone substrates.
Entry
Compound
Energy barrier [kJ/mol][a]
1
171.5
Catalyst
Cofactor
NADPH
NADPH
Formed product [%]
57.4 ± 0.7
ee [%]
96 (S)
87 (S)
2
3
174.6
153.4
R-IRED-Sr
S-IRED-Pe
23.8 ± 0.2
inactivated
R-IRED-Sr
NADPH
˂ 1.0
-
BSA
NADPH
NADH
˂ 0.5
-
-
4
152.2
R-IRED-Sr
1.4 ± 0.3
R-IRED-Sr
(expressed in vitro)
NADPH
4.5 ± 0.2 [a]
n.d.
5
6
82.6
69.2
[a] Product formation was calculated using GC areas from substrate and
product normalized to an internal standard after 48 h by GC-MS as detailed
in the supporting information. n.d. not determined
To ensure that the reduction of 2,2,2-trifluoroacetophenone (1a)
to α-(trifluoromethyl)benzyl alcohol (2a) was performed by the
IRED enzyme, further control experiments were performed. In
reactions with inactive R-IRED-Sr or bovine serum albumin (BSA)
with a nicotinamide cofactor regeneration system, only traces of
2a were obtained (Table 1). Therefore, reduction by the NADPH
cofactor can be excluded. In contrast to carbonyl reducing
enzymes IREDs are described to be exclusively dependent on
NADPH as cofactor[7]. It is thus not surprising that
biotransformations applying NADH showed 40 times less product
formation than those using NADPH (Table 1).
7
65.3
8
9
58.9
66.4
To support our experimental findings, we calculated energy
barriers for hydride transfers from the nicotinamide subunit of
NAD(P)H to a small panel of cyclic and acyclic imines and their
corresponding iminium ions as well as aromatic carbonyl
substrates (Table 2). We modelled the active center using
1-methylnicotinamide as hydride donor, substrate and one explicit
water molecule embedded in a dielectric environment (detailed
information in the supporting information). Due to the limitations
of the model (only one water molecule present, no protein-
substrate interactions, no protein environment), interpretation of
the results is restricted to relative differences in activation
energies. The calculated activation energies summarized in Table
2 show that two to three times lower energy barriers for iminium
ions (Table 2, entry 5-8) than for imines (Table 2, entry 1-4) were
obtained. These differences in reactivity are underlined by recent
10
11
80.6
112.6
[a] Energy barriers were determined as detailed in the
supporting information.
To compare our IRED activity for 1a with the activity for
characterized imine substrates, kinetic parameters were
determined. The results showed that the catalytic efficiency of
R-IRED-Sr for 2-methylpyrroline (2-MPN) being one of the model
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600647
ChemBioChem
COMMUNICATION
higher initial rates in transformations with iminium ions than with
the corresponding imines[8,18]. Regarding speculations of the role
of amino acid residues for C=N protonation in IREDs[9,23], there is
some supporting evidence based on mutations of putative
catalytically important aspartic acid residues displaying reduced
catalytic efficiencies over wild type enzyme[6]. Residual activities
can be attributed to the partly protonated imines in solution at pH
7. However, reductive amination reactions were conducted at pH
9, where protonated imines are not expected in solution. Product
formation up to 60% under these conditions[5] strongly suggests
imine protonation in the active site as part of the catalytic
mechanism. Efforts to determine crucial amino acid residues for
catalysis in IREDs were not yet successful. We hypothesize a
common orientation of substrate and cofactor for favourable
hydride and proton transfer in IREDs and carbonyl-reducing
enzymes. In this respect, fitting of imines into the binding pocket
of carbonyl reducing enzymes might result in steric clashes with
amino acid residues that normally interact with the free electron
pairs of oxygen. Furthermore, the substrate binding site of
dehydrogenases/KREDs might prevent productive hydride
transfer from the nicotinamide ring of NAD(P)H to alternate imine
substrates, because of unfavorable distances. Our theoretical and
experimental data have provided considerable information on the
size and electronic requirements for the asymmetric reduction of
imines and ketones with IRED catalysts. The substitution of the
keto functionality by a trifluoromethyl group increases the
substrate reactivity and thus, facilitates the asymmetric reduction
of ketones. The asymmetric hydrogenation of substituted ketones
highlights the difficulties in understanding the balance between
steric and electronic factors that govern the outcome of organic
transformations. We consequently conclude that by employing
chemical tools the chemoselectivity of IREDs can be driven.
substrates for IREDs was 600-fold higher than for 1a. The kcat/KM
values of 53.31 min-1 mM-1 for 2-MPN[6] and 0.085 min-1 mM-1 for
1a reveal the difference in the reduction of imines and carbonyls.
Moreover, the KM value of 8.78 mM for 1a is about eight times
higher compared to the KM of 2-MPN. Substrate inhibition has
been reported for IREDs[19,20] and has also been observed for 1a.
R-IRED-Sr showed decreased activities at substrate
concentrations above 7 mM as shown by the Michaelis-Menten
plot (Figure S7, supporting information). Even though Grogan and
Turner speculated in a recent publication that the reduction of
imines might represent a promiscuous activity of IREDs[19], the
determined kinetic constants indicate that IREDs have been
evolved in nature for the reduction of C=N bonds, if in other
substrates.
Reductases catalyzing asymmetric hydrogenations of ketones
and imines are closely related. For both enzyme families
NAD(P)H is described as cofactor providing the hydride ion for the
reduction of the appropriate substrate. The binding of the
nicotinamide cofactor not only influences protein dynamics but
also constitutes a key structural component in organizing all
catalytically important active site residues to form the binding
pocket[21]. In comparison to carbonyl reducing enzymes, catalytic
amino acid residues and the reduction mechanism of IREDs are
still under debate. A generic proton-donating catalysis including
hydride and proton transfer is expected for IREDs, closely related
to the mechanism of dehydrogenases/KREDs[22]. NADPH is
considered as the hydride donor, while the proton-donating
residue remains elusive[3]. Hydride transfer prior to protonation
seems unlikely due to the formation of a highly basic amide
intermediate, which is difficult to stabilize[2]. Therefore two
strategies for hydride- and proton-transfer can be envisaged: (i)
the formation of an iminium ion through protonation of the imine,
followed by addition of the hydride, or (ii) the concerted transfer of
hydride and proton[2]. Commonly found zinc or iron ions
functioning as Lewis acids in metal-containing alcohol
dehydrogenases have so far not been observed in IREDs. IREDs
are homodimeric proteins with each subunit binding one molecule
of NADPH via a Rossmann-fold motif. First insights into the
possible ligand-binding site and local residues of the IRED from
Amycolatopsis orientalis influencing the catalytic activity were
recently gained by the groups of Turner and Grogan. By co-
crystallization with NADPH and the amine product 1-methyl-
1,2,3,4-tetrahydrosioquinoline the ternary complex of the IRED
was examined[23]. Latest investigations from Tawfik and co-
workers on the evolution of the Rossmann-fold motif[24] and the
structural similarity of IREDs to hydroxylisobutyrate and ß-
hydroxyacid dehydrogenases[6,9] strongly suggest a common
ancestry. Despite these common elements, IREDs and KREDs
differ in their ability to reduce imines.
In summary, we describe the first example of the promiscuous
imine reductase-catalyzed asymmetric reduction of a highly
reactive carbonyl compound. Assisted by in silico calculations of
energy barriers for the hydride transfer from the nicotinamide
subunit of NAD(P)H to several imines and their corresponding
iminium ions, the present results contribute to a deeper
understanding of the reaction mechanism of imine reductases
and their evolution.
Acknowledgements
We express our cordial thanks to Prof. Dr. Bernhard Hauer and
Dr. Stephan Hammer for the fruitful discussions. We also
acknowledge financial support from the European Union and the
EFPIA companies’ in kind contribution for the Innovative Medicine
Initiative under Grant Agreement No. 115360 (Chemical
manufacturing methods for the 21st century pharmaceuticals
industries, CHEM21).
In the present study, we surveyed the hydride transfer to C=N and
C=O containing substrates. Theoretically calculated activation
energies encourage the speculation that imines are protonated
prior to the hydride transfer from NADPH. We demonstrated that
our selected iminium ions are more reactive than the
corresponding imines (Table 1). This is in accordance with the
work of Mayr describing the electrophilicity of benzaldehyde-
derived iminium ions and their higher reactivities compared to
substituted imines[25]. In pursuit of this trend, IREDs showed
Keywords: biocatalysis • asymmetric reduction • imine
reductase • promiscuity • carbonyl compounds
[1]
K. Drauz, H. Gröger, O. May, Eds. , Enzyme Catalysis in
Organic Synthesis: A Comprehensive Handbook, Wiley-
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600647
ChemBioChem
COMMUNICATION
VHC Verlag GmbH, 2012.
[2]
[3]
[4]
[5]
[6]
J. H. Schrittwieser, S. Velikogne, W. Kroutil, Adv. Synth.
Catal. 2015, 357, 1655–1685.
G. Grogan, N. J. Turner, Chem. - A Eur. J. 2016, 22, 1900-
1907.
T. Huber, L. Schneider, A. Präg, S. Gerhardt, O. Einsle, M.
Müller, ChemCatChem 2014, 6, 2248–2252.
P. N. Scheller, M. Lenz, S. C. Hammer, B. Hauer, B. M.
Nestl, ChemCatChem 2015, 7, 3239–3242.
P. N. Scheller, S. Fademrecht, S. Hofelzer, J. Pleiss, F.
Leipold, N. J. Turner, B. M. Nestl, B. Hauer, ChemBioChem
2014, 15, 2201–2204.
[7]
[8]
M. Gand, H. Müller, R. Wardenga, M. Höhne, J. Mol. Catal.
B Enzym. 2014, 110, 126–132.
S. Hussain, F. Leipold, H. Man, E. Wells, S. P. France, K.
R. Mulholland, G. Grogan, N. J. Turner, ChemCatChem
2015, 7, 579–583.
[9]
S. Fademrecht, P. N. Scheller, B. M. Nestl, B. Hauer, J.
Pleiss, Proteins Struct. Funct. Bioinforma. 2016, 84, 600–
610.
[10]
[11]
[12]
[13]
N. H. Schlieben, K. Niefind, J. Mueller, B. Riebel, W.
Hummel, D. Schomburg, J. Mol. Biol. 2005, 349, 801–813.
H. G. Naik, B. Yeniad, C. E. Koning, A. Heise, Org. Biomol.
Chem. 2012, 10, 4961.
W. Borzęcka, I. Lavandera, V. Gotor, J. Org. Chem. 2013,
78, 7312–7317.
C. Rodríguez, W. Borzęcka, J. H. Sattler, W. Kroutil, I.
Lavandera, V. Gotor, Org. Biomol. Chem. 2014, 12, 673–
81.
[14]
[15]
[16]
[17]
[18]
[19]
K. Inoue, Y. Makino, N. Itoh, Tetrahedron Asymmetry 2005,
16, 2539–2549.
T. C. Rosen, R. Feldmann, P. Dünkelmann, T. Daußmann,
Tetrahedron Lett. 2006, 47, 4803–4806.
A. B. Daugherty, J. R. Horton, X. Cheng, S. Lutz, ACS
Catal. 2015, 5, 892–899.
L. T. Quertinmont, R. Orru, S. Lutz, Chem. Commun. 2015,
51, 122–124.
H. Li, Z.-J. Luan, G.-W. Zheng, J.-H. Xu, Adv. Synth. Catal.
2015, 357, 1692–1696.
M. Rodríguez-Mata, A. Frank, E. Wells, F. Leipold, N. J.
Turner, S. Hart, J. P. Turkenburg, G. Grogan,
ChemBioChem 2013, 14, 1372–1379.
[20]
F. Leipold, S. Hussain, D. Ghislieri, N. J. Turner,
ChemCatChem 2013, 5, 3505–3508.
[21]
[22]
G. Sanli, M. Blaber, J. Mol. Biol. 2001, 309, 1209–1218.
B. P. Schlegel, J. M. Jez, T. M. Penning, Biochemistry
1998, 37, 3538–3548.
[23]
G. A. Aleku, H. Man, S. P. France, F. Leipold, S. Hussain,
L. Toca-, R. Marchington, S. Hart, J. P. Turkenburg, G.
Grogan, et al., ACS Cata 2016, 6, 3880–3889.
P. Laurino, Á. Tóth-Petróczy, R. Meana-Pañeda, W. Li, D.
G. Truhlar, D. S. Tawfik, PLOS Biol. 2016, 1–23.
R. Appel, S. Chelli, T. Tokuyasu, K. Troshin, H. Mayr, J.
Am. Chem. Soc. 2013, 135, 6579–6587.
[24]
[25]
This article is protected by copyright. All rights reserved.

10.1002/cbic.201600647
ChemBioChem
COMMUNICATION
Entry for the Table of Contents (Please choose one layout)
Layout 1:
COMMUNICATION
IRED vs. KRED: Asymmetric reduction
of a highly reactive carbonyl compound
has been accomplished using an imine
reductase. Kinetics and in silico studies
have offered mechanistic insights,
allowing an understanding of this
promiscuous behaviour.
M. Lenz, J. Meisner, L. Quertinmont, S.
Lutz, J. Kästner, B.M. Nestl*
Page No. – Page No.
Chemoselectivity of Imine
Reductases Driven by the Inherent
Reactivity of Functional Groups
Layout 2:
COMMUNICATION
Author(s), Corresponding Author(s)*
((Insert TOC Graphic here))
Page No. – Page No.
Title
Text for Table of Contents
This article is protected by copyright. All rights reserved.