Characterization of three novel enzymes with imine reductase activity

Article abstract of DOI:10.1016/j.molcatb.2014.09.017

Imine reductases (IRED) are promising catalysts for the synthesis of optically pure secondary cyclic amines. Three novel IREDs from Paenibacillus elgii B69, Streptomyces ipomoeae 91-03 and Pseudomonas putida KT2440 were identified by amino acid or structural similarity search, cloned and recombinantly expressed in E. coli and their substrate scope was investigated. Besides the acceptance of cyclic amines, also acyclic amines could be identified as substrates for all IREDs. For the IRED from P. putida, a crystal structure (PDB-code 3L6D) is available in the database, but the function of the protein was not investigated so far. This enzyme showed the highest apparent E-value of approximately E_app = 52 for (R)-methylpyrrolidine of the IREDs investigated in this study. Thus, an excellent enantiomeric purity of >99% and 97% conversion was reached in a biocatalytic reaction using resting cells after 24 h. Interestingly, a histidine residue could be confirmed as a catalytic residue by mutagenesis, but the residue is placed one turn aside compared to the formally known position of the catalytic Asp187 of Streptomyces kanamyceticus IRED.

Full text of DOI:10.1016/j.molcatb.2014.09.017

Journal of Molecular Catalysis B: Enzymatic 110 (2014) 126–132
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Characterization of three novel enzymes with imine reductase activity
M. Ganda, H. Müllera, R. Wardengab, M. Höhnea,∗
a Institute of Biochemistry, Greifswald University, Felix-Hausdorff-Str. 4, 17487 Greifswald, Germany
b Enzymicals AG, Walther-Rathenau-Straße 49a, 17489 Greifswald, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 1 July 2014
Received in revised form
23 September 2014
Accepted 28 September 2014
Available online 7 October 2014
Imine reductases (IRED) are promising catalysts for the synthesis of optically pure secondary cyclic
amines. Three novel IREDs from Paenibacillus elgii B69, Streptomyces ipomoeae 91-03 and Pseudomonas
putida KT2440 were identified by amino acid or structural similarity search, cloned and recombinantly
expressed in E. coli and their substrate scope was investigated. Besides the acceptance of cyclic amines,
also acyclic amines could be identified as substrates for all IREDs. For the IRED from P. putida, a crys-
tal structure (PDB-code 3L6D) is available in the database, but the function of the protein was not
investigated so far. This enzyme showed the highest apparent E-value of approximately Eapp = 52 for
(R)-methylpyrrolidine of the IREDs investigated in this study. Thus, an excellent enantiomeric purity of
>99% and 97% conversion was reached in a biocatalytic reaction using resting cells after 24 h. Interest-
ingly, a histidine residue could be confirmed as a catalytic residue by mutagenesis, but the residue is
placed one turn aside compared to the formally known position of the catalytic Asp187 of Streptomyces
kanamyceticus IRED.
Keywords:
Imine reductases
Enzyme catalysis
Chiral secondary amines
Protein function assignment
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
enantiomer to the imine, which in turn is reduced non-selectively
in situ by chemical means to the amine racemate [11,12]. This
A large number of secondary metabolites are composed of chi-
ral secondary amines [1]. Their various physiological activities are
explored and utilized by pharmaceutical companies [2]. Hence,
many active ingredients of pharmaceuticals represent molecules
having one or more stereogenic centres connected to a nitrogen
atom. As a consequence, a variety of chemical methods for the
enantioselective synthesis of optically pure secondary amines were
developed during the last decades, e.g. organocatalytic or metal
catalyzed hydrogenation of imines [3,4]. Biocatalysis is more and
more accepted as an efficient synthesis strategy for regio- and
stereospecific conversions, because it eliminates the use of envi-
ronmentally toxic heavy metals, enables mild conditions and offers
high selectivities [5]. Furthermore, protein engineering allows the
fine-tuning of the employed enzymes to broaden their applicabil-
ity [6]. In contrary to the preparation of primary amines, which can
be accessed by various enzymes such as lipases, monoamine oxi-
dases and especially by asymmetric synthesis with transaminases
[7–10], for secondary amines there are fewer options: monoamine
oxidases allow the deracemization of an amine by oxidizing one
leads to the enrichment of the non-reactive enantiomer up to
100%. In a second approach, cyclic sec-amines can be obtained
in a two-step, one pot reaction if the substrate is carefully cho-
sen: the presence of a halogen, ester or ketone moiety facilitates
a spontaneous, intramolecular cyclization via nucleophilic substi-
tution or imine formation [10,13–15]. Dependent on the strategy,
a follow-up reductive step is necessary to yield the desired amine.
A third option discovered recently are NADPH-dependent imine
reductases (IRED) (Fig. 1). So far, only few IREDs were described as
recombinant proteins and investigated for the asymmetric synthe-
sis of secondary amines from the corresponding prochiral imines:
the (S)-selective IRED originating from Streptomyces sp. GF3546
(SIR-Sgf3546) [16] and Streptomyces aurantiacus (SIR-Sau) [17], and
the (R)-selective enzymes obtained from Streptomyces sp. GF3587
(RIR-Sgf3587) [18] and Streptomyces kanamyceticus (RIR-Ska). Crys-
tal structures were solved for SIR-Sgf3546, SIR-Sau, and RIR-Ska
[17,19]. During the revision of this manuscript, three additional
enzymes were characterized originating from Streptosporangium
roseum DSM 43021 (RIR-Sro), Streptomyces turgidiscabies (RIR-Stu)
and Paenibacillus elgii B69 (SIR-Pel) [20]. The latter enzyme is
one of the proteins that we characterized in detail in this study.
Furthermore, the group of Thomas Ward designed artificial met-
alloenzymes that facilitate the reduction of imines by transfer
hydrogenation [21,22].
∗
Corresponding author. Tel.: +49 3834 86 22832; fax: +49 3834 86 794367.
E-mail addresses: rainer.wardenga@enzymicals.com (R. Wardenga),
Matthias.Hoehne@uni-greifswald.de (M. Höhne).
http://dx.doi.org/10.1016/j.molcatb.2014.09.017
1381-1177/© 2014 Elsevier B.V. All rights reserved.

M. Gand et al. / Journal of Molecular Catalysis B: Enzymatic 110 (2014) 126–132
127
Fig. 1. Enzymatic imine reduction. ꢀ1-Pyrrolin-carboxylate reductase (ꢀ1-PCR) catalyzes the reduction of cyclic imines (reaction A) or acyclic imines (reaction B), R1 COOH,
n = 1–2. Isolated imine reductases (IRED) reduce the analogous cyclic substrates where R1 is an alkyl substituent, e.g. a methyl group, and n = 1–3. The question, whether
IREDs also act on acyclic substrates (reaction B) is addressed in this study. As the equilibrium favors the hydrolysis of acyclic imines in water (compared to the fairly stable
cyclic imines), oxidative deamination of acyclic secondary amines is chosen in initial screening reactions (reverse reaction).
The main challenges for applying IREDs in biocatalysis are (i)
only few enzymes are available, (ii) they display a low activity
and (iii) some enzymes are reported to show a narrow substrate
specificity, especially the (R)-selective imine reductase from Strep-
tomyces GF3587 (RIR-Sgf3587) [18]. It was suggested earlier that
the ‘true’ natural substrate has not yet been identified [19], or that
imine reduction might be a promiscuous activity.
hypothesized that IREDs could also be active towards the analo-
gous acyclic substrates 3b and 4b (Fig. 1B). Therefore, a second aim
of this study was to investigate, whether IREDs are able to perform
this reaction either as kinetic resolution, or preferentially, in an
asymmetric synthesis by generating a stereocenter via reductive
amination.
To address the above-mentioned problems, we characterized
three proteins similar in sequence or structure to RIR-Ska or SIR-
Sgf3546. For their characterization, we screened a broad panel
of possible substrates. Besides the substrates of the hydrox-
yisobutyrate dehydrogenase superfamily (3-hydroxyisobutyrate,
6-phosphogluconate) and cyclic amines, we also examined various
alcohols, ketones, aldehydes, amino acids, prim- and sec-amines.
An additional question is whether IREDs are restricted towards
cyclic imine substrates. In principle, imine formation from a ketone
and a primary amine of choice would allow the synthesis of acyclic
secondary amines. This would be a desirable reaction. During the
preparation of this manuscript, Huber et al. published a proof of
principle for the reductive amination of three methyl alkyl ketones
with methylamine using SIR-Sgf3546 and SIR-Sau: product forma-
tion could be detected with conversions ranging from 0.1% to 9%.
However, the enzymes showed a very low catalytic turnover of
1.7 molecules/week for the best substrate. Thus, these enzymes
are not yet applicable for preparative reactions at larger scale [17].
Another example is described where whole cells were applied to
reduce acyclic imines [23], but a huge cell mass (18 g for 0.18 mmol
substrate) was needed and the amino acid sequence of the acting
catalyst is unknown. In nature, this synthetic strategy is applied
for the synthesis of various secondary amines, such as opines. In
this case an ␣-keto acid reacts with an amino acid and the formed
imine is reduced stereoselectively to yield an opine [24]. These
products usually contain two or more carboxylate functionalities,
but these polar groups are often not desired in pharmaceuti-
cally active sec-amines. Another reaction yielding secondary amino
acids is catalyzed by ꢀ1-pyrroline-carboxylate-reductase (ꢀ1-
PCR), which is involved in proline/lysine metabolism: the cyclic
substrate ꢀ1-pyrroline-carboxylate 1a is reduced to proline 2a
(Fig. 1A). Interestingly, ꢀ1-PCR also reduces acyclic imines like
3a formed by in situ condensation of ␣-keto acids like pyruvate
with methyl- and ethylamine to the N-alkylated ␣-amino acid
4a as a side reaction [25] (Fig. 1B). As the model substrate of
IREDs, 2-methylpyrroline 1b, differs from 1a only by the replace-
ment of the carboxylic acid group by an alkyl substituent, we
2. Methods and materials
2.1. General
All chemicals were purchased from Fluka (Buchs, Switzerland),
Sigma (Steinheim, Germany), Carl-Roth (Karlsruhe, Germany) and
Merck (Darmstadt, Germany). The pSGX3 plasmid bearing an insert
of a putative reductase with unknown function, (NCBI Gene iden-
tification number GI: 283807276, PDB-accession code 3L6D) was
kindly provided by the New York Structural Genomics Research
Consortium, Department of Biochemistry of Albert Einstein College
of Medicine, New York, USA). The genes of the previously described
imine reductases SIR-Sgf3546 (GI: 460838082) and RIR-Sgf3587
(GI: 505423772) and two putative reductases identified by BLAST
search (GIs: 498183793 and 496688866) were ordered as synthetic
genes (GenScript, Piscataway, USA) as subcloned constructs in the
pET28b(+) vector, which creates a N-terminal His6-tag for purifi-
cation purposes. 2-Methylpiperideine was synthesized according
to Leipold et al. [26] and for the synthesis of 2-propylpiperideine,
the same synthesis procedure was adapted (compare Section 1.1,
supplementary information).
2.2. Protein biosynthesis and purification
Kanamycin was used in all cultivations as antibiotic. In a typical
procedure, E. coli BL21 (DE3) cells transformed with pET28b IRED
were inoculated in LB medium from an overnight culture (1:100,
v/v) and were grown at 30 ◦C until an OD600 of 0.2–0.5 was reached.
Protein expression was initiated by the addition of IPTG to a
final concentration of 0.2 mM. Eight hours after induction, the
cells were harvested (4000 g, 4 ◦C, 10 min) and resuspended in
sodium phosphate buffer (50 mM, pH 7.5). The cells were lysed
with a French press (Thermo Fisher Scientific, Waltheim, MA,
USA) and centrifuged to separate the cell debris from the solu-
ble protein. The His6-tagged IREDs were purified by metal affinity
chromatography using an ÄKTA purifier (GE Healthcare, Chalfont
St Giles, UK) with a HisTrapTM FF 5 mL column (GE Healthcare). In a

128
M. Gand et al. / Journal of Molecular Catalysis B: Enzymatic 110 (2014) 126–132
second step, the IREDs were desalted using Centricon filters (15 mL)
with a molecular weight limit of 30 kDa (Merck Millipore, Billerica,
MA, USA). The purified enzymes were washed three times with
7.5 mL sodium phosphate buffer (50 mM, pH 7.5) to remove imid-
azole and NaCl. Finally, the purified proteins were concentrated by
reducing the volume to ∼2 mL. The purity was evaluated with SDS-
PAGE, while the protein content was measured with the Bradford
assay [27].
GC–MS analysis was performed by the same equipment men-
tioned above. Hydrogen was the carrier gas. The injection volumes
were 1.5 ␮L.
Analysis of 1b and 2b was performed starting at 60 ◦C for
15 min followed by an increment of 24 ◦C/min to 120 ◦C followed
by a 2 min hold. The retention times were assigned using com-
mercial standards: (S)-2b, 7.5 min; (R)-2b 9.0 min; 1b, 16.5 min;
4-methylacetophenone (standard), 16.5 min. The column flow rate
was of 1.96 mL/min.
2.3. Determination of enzyme activity
For 1c and 2c the following temperature programme was
applied: 50 ◦C for 15 min followed by an increment of 24 ◦C/min
to 120 ◦C followed by a 2 min hold. The retention times were as
following: (R)-2c, 10.5 min; (S)-2c, 13.5 min. The later eluting peak
was assigned to the (S)-enantiomer of 2c as it is known that SIR-
Sgf3546 produces 2c with (S)-selectivity [26]; 2-methylpiperdeine
(1c), 18 min; 4-methylacetophenone (standard), 20 min. The flow
rate was the same as mentioned above.
The activities of the purified IREDs were determined from a
liquid-phase spectrophotometric assay (Tecan infinite M200 pro,
Männedorf, Swiss) monitoring the change of NADPH concentra-
tion at 340 nm (ε = 6.22 mM−1 cm−1). Reaction mixtures contained
buffer (100 mM, sodium phosphate pH 7.5 for the reduction or
glycine–NaOH pH 10.5 for the oxidation), 0.2 mM NADPH or 0.2 mM
NADP+ for the reduction or oxidation, respectively, in the pres-
ence of 5% (v/v) DMSO as cosolvent, and the substrate at the
desired concentration. The reaction was started by adding the
purified enzyme to the mixture. The reactions were performed at
room temperature. One unit is defined as the amount of protein
that oxidizes 1 ␮mol NADPH/min. Controls were done using the
same reaction composition but by substitution of enzyme or sub-
strate with buffer. The reported activities are the mean of three
purifications (from independent cultivations), each measured in
triplicates.
Biotransformations: oxidative deamination of rac-N-methyl-1-
phenylethylamine was performed at 30 ◦C and 1000 rpm (Eppen-
dorf thermomixer, Hamburg, Germany) in sodium phosphate
buffer (100 mM, pH 8) using purified IRED RIR-Sip (4.5 mg/mL).
The reaction mixtures typically contained 5 mM substrate, 5% (v/v)
DMSO and 5 mM NADP+. 200 ␮L aliquots were removed in defined
time intervals for a period of 24 h. 4-Methylacetophenone was
added at 2.5 mM final concentration as internal standard dur-
ing extraction with two volumes of ethyl acetate followed by
drying over anhydrous MgSO4. No derivatization had to be per-
formed. As a negative control, biotransformations were performed
in the same way without enzyme. Product analysis (acetophe-
none, 1-phenylethylamine, N-methyl-1-phenylethylamine) was
performed by GC–MS analysis with a Shimadzu GC-2010 and
GCMS-QP2010 device (Tokyo, Japan), equipped with a HYDRODEX
␤-TBDAc column (25 m × 0.25 ␮m, SGE Analytical Science, Mil-
ton Keynes, UK). Hydrogen was the carrier gas with a column
flow rate of 1.95 mL/min. The analysis was performed with a gra-
dient starting at 65 ◦C for 40 min and then 1 ◦C/min to 100 ◦C
followed by 20 ◦C/min to 180 ◦C for 5 min. The retention times
were as following: 4-methylacetophenone (standard), 75.25 min;
acetophenone, 54.5 min; (S)-(+)-N-methyl-1-phenylethylamine,
32.5 min; (R)-(−)-N-methyl-1-phenylethylamine 35.5 min, (R)-1-
phenylethylamine 60.2 min, (S)-1-phenylethylamine 56 min.
Whole cell biotransformations were performed at 30 ◦C and
1000 rpm in sodium phosphate buffer (100 mM, pH 7.0) by using
freshly harvested resting cells of E. coli BL21 (DE3) expressing an
IRED at a final OD600nm of 90. The reaction mixtures contained 2.5
or 5 mM imine and 50 mM of glucose.
Analysis of 1d and 2d was performed starting with 50 ◦C for
10 min followed by an increment of 4 ◦C/min to 110 ◦C followed
by a 10 min hold and a second increment of 24 ◦C/min to 180 ◦C
followed by a 2 min hold. The retention times were as following:
propylpiperdines, 20.9 min and 21.2 min (near baseline separa-
tion); 2-propylpiperideine (1d), 26 min; 4-methylacetophenone
(standard), 31.3 min. The column flow rate was of 1.00 mL/min.
For site-directed mutagenesis a modified QuikChange® PCR
protocol was used according to literature [28], and the product
transformed in E. coli TOP10 cells to repair the nicked ends. Finally
the plasmids were isolated and used to transform BL21 (DE3) for
protein expression. A list of the primers used in the present study
can be found in the supplementary information (Table S1).
3. Results and discussion
3.1. Identification and expression of novel IREDs
Two proteins from P. elgii B69 (IRED-Pel) and S. ipomoeae 91-
03 (IRED-Sip) were identified from a BLAST search. Both proteins
were annotated as 6-phosphogluconate dehydrogenases. IRED-Pel
showed 65% sequence identity to the known SIR-Sgf3546 [16]
while the IRED-Sip has 55% sequence identity towards the RIR-
Sgf3587 [18]. The third protein, a putative reductase from P. putida,
was chosen from a structural alignment of all available structures
of the 3-hydroxyisobutyrate dehydrogenase superfamily contain-
ing for example the structures of S. kanamyceticus (R)-selective
imine reductase (RIR-Ska) [PDB-code 3ZHB], ␤-hydroxyl acid dehy-
drogenase [PDB-code 3CKY], tartronate semialdehyde reductase
[PDB-code 1VPD], and 6-phosphogluconate dehydrogenase [PDB-
code 4GWP] [29–31]. A remarkable structural difference of RIR-Ska
to proteins with other enzyme activities of this superfamily is the
dimer formation by reciprocal domain swapping between two sub-
units [19]. Interestingly, the phenomenon of domain sharing is also
observed in the crystal structure of the putative oxidoreductase
from P. putida (PDB-code 3L6D, compare Fig. S1, supplementary
information). This enzyme has not yet been experimentally char-
acterized. It is annotated as 6-phosophogluconate dehydrogenase
(Uniprot accession number Q88J51, www.uniprot.org). We have
chosen this enzyme for further studies, although the sequence iden-
tity to RIR-Ska is low (29%) and instead of the catalytic Asp187 as
observed in RIR-Ska, an alanine residue is placed at the structural
equivalent position 176 in this oxidoreductase.
Samples were taken after different time points, basified with
20 vol% 10 M NaOH, three times extracted by using 0.5 reaction vol-
umes dichloromethane, containing 2.5 mM 4-methylacetophenone
as internal standard, and used directly for GC analysis. Two mod-
ified biotransformations were performed as negative controls in
the same way as described above using (i) BL21 cells lacking the
expression plasmid or (ii) performing the reaction without cells.
For calibrating the GC analysis and proof of product stability under
reaction conditions, buffered solutions with different concentra-
tions of imines and amines were prepared and measured.
The corresponding genes were expressed in E. coli and purified
by metal affinity chromatography for biochemical characterization
and determination of the substrate profile. We found nearly the
same amount of protein in the soluble and insoluble fractions (a
representative result from a SDS-PAGE and a summary of protein

M. Gand et al. / Journal of Molecular Catalysis B: Enzymatic 110 (2014) 126–132
129
Table 1
Activity ratio of the used cofactors NADPH/NADH or NADP+/NAD+. The oxidation was
performed with 20 mM (R)- or (S)-2b at pH 10.5 in 100 mM glycine–NaOH buffer.
The reduction was done with 20 mM 1b at pH 7.5 in 100 mM sodium phosphate
buffer.
Reaction direction
RIR-Sip
SIR-Pel
RIR-Ppu (3L6D)
Oxidation
Reduction
6.5
85
42
4.7
7
17.5
Fig. 2. Intended kinetic resolution of N-methyl-1-phenylethylamine catalyzed by
IRED.
rather than imines as substrates in our first screening, because
acyclic imines are not stable in water and rapidly hydrolyze into the
amine and ketone (Fig. 1). Notably, nine secondary acyclic amines
were converted, but to a different extent (Table 3). The enantiose-
lectivity was first analyzed by comparing the oxidative activities
towards (R)- and (S)-2b: RIR-Ppu showed the highest apparent E-
value of approximately Eapp = 52 for the (R)-enantiomer, followed
by RIR-Sgf3587 with Eapp = 19. The RIR-Sip showed only a 3 fold
higher activity towards the (R)-enantiomer. The SIR-Sgf3546 and
SIR-Pel have nearly the same preference for the (S)-enantiomer
with an Eapp = 6–7.
To confirm the reactions and enantioselectivities observed in
the photometric assays, we performed the asymmetric reduction
of three imines using BL21 resting cells having expressed the
IREDs (Table 4). In contrast to the oxidation reaction, perfect enan-
tiomeric excesses were obtained for 2b and 2c in the reduction
with all enzymes. A similar trend (different enantioselectivities
in the oxidation and reduction reactions) was observed earlier
for SIR-Sgf3546 [26]. As a third substrate, 2-n-propylpiperideine
1d was investigated, yielding the amine 2d. The obtained enan-
tiomeric excess was low except for the reaction employing SIR-Pel,
where an excellent enantioselectivity, but a very low conversion
was detected.
A cyclic secondary amino acid, 2-piperidinecarboxylic acid,
could be identified as a substrate for four IREDs with 1–15% activ-
ity compared to 2c. RIR-Ppu is the only enzyme which converts the
amino acids l-proline and l-glutamic acid with 32% and 8% activity
relative to (R)-2b. This indicates that the active sites can also accept
charged molecules, but with less efficiency.
Interestingly, our attempts to detect activities for various sub-
strates of the following substance classes failed: amino acids
(except of l-proline and l-glutamic acid), amines, substrates of
structurally similar enzymes (3-hydroxy-2-methyl-propanoic acid,
6-phosphogluconic acid), dihydrofolic acid, aldehydes, ketones or
keto acids.
yields can be found in Fig. S2 and Table S2, supplementary infor-
mation).
3.2. Optimization of reaction conditions and cofactor preference
In our first attempts, 2-methylpyrroline 1b and (R)- or (S)-2-
methylpyrrolidine 2b were used as model substrates in different
buffers using both NADPH and NADH as cofactors (or their oxi-
dized forms). We were pleased to detect imine reduction and amine
oxidation for all three purified enzymes. The reductase from S. ipo-
moeae and P. putida (3L6D) preferred the (R)-enantiomer of 2b
in the oxidative reaction and thus are referred to as (R)-selective
imine reductases (RIR-Sip and RIR-Ppu) throughout this study. The
specific oxidation activities towards (R)-2b were 16 mU/mg and
23 mU/mg under the selected reaction conditions, respectively. The
reductase of P. elgii preferred (S)-2b with an activity of 20 mU/mg
and is therefore designated as SIR-Pel.
Prior to the screening of a large panel of different substrates,
the pH profiles of the enzymes were investigated to elucidate the
optimal reaction conditions to guarantee a maximal sensitivity of
the assay: for the oxidation reaction, higher pH-values lead to an
increased activity. The most suitable buffer for the SIR-Pel, RIR-Sip
and RIR-Ppu was sodium phosphate at pH 8, glycine–NaOH at pH
10.5 or pH 10, respectively (pH profiles see Figs. S3–S5, A, supple-
mentary information). On the contrary, the optimal buffer system
for the reduction of imines was sodium phosphate at pH 6–7 (see
Figs. S3–S5, B, supplementary information). These results are in line
with the previously described RIR-Sgf3587 and SIR-Sgf3546 (see
Figs. S3 and S4, C and D, supplementary information). All enzymes
showed a varying, but significant preference for the phosphorylated
cofactors NADP(H) over NAD(H) (Table 1).
3.3. Substrate profiles and enantioselectivities
For the reduction, 2-methylpyrroline, 2-methylpiperideine and
1-methyl-3,4-dihydroisoquinoline were investigated (Table 2).
Towards theses substrates, the identified IREDs of this study
showed lower activities compared to the already known enzymes
and prefer the six-membered ring substrate 1c over 1b.
3.4. Towards the kinetic resolution of
N-methyl-1-phenylethylamine
Six secondary cyclic amines were oxidized by the new enzymes.
Compared to SIR-Sgf3546 and RIR-Sgf3587, the new IREDs showed
similar substrate profiles, but with lower activities towards most
substrates (Table 3). To elucidate the possibility of the enzymes to
act on acyclic substrates, we investigated acyclic secondary amines
The slow, but selective activity towards the (S)-enantiomer of
N-methyl-1-phenylethylamine 5 encouraged us to investigate a
kinetic resolution (Fig. 2). In biotransformations with rac-5, we
found that acetophenone 6 was formed as main product of the
deamination, as confirmed via GC-MS. 1-Phenylethylamine could
Table 2
Specific activities (mU mg−1) in the NADPH-dependent reduction catalyzed by five different IREDs.b Each compound was used at a concentration of 20 mM.
Substance
RIR- Sgf3587
RIR-Sip
SIR-Sgf354
SIR-Pel
20
RIR-Ppu
2-methylpyrroline
2-methylpiperideine
1-methyl-3,4-dihydroisoquinoline
490 62
1241 563
27 3.4
254 31
50 20
57 7.4
10 7.4
3
66 36
n.m.
37 0.4
0.3
a
43
4
4
1.2
9
–
n.m., not measured.
a
Minimal activity detectable.
b
For the following compounds, activity was not detectable for any of the investigated IREDs: aldehydes and ketones (ethanol, butanal, 2-methylbutanal, 3-methylbutanal,
hexanal, heptanal, benzaldehyde, 4-pyridinecarboxaldehyde, ␣-methyl-benzylacetaldehyde, 4-nitrobenzaldehyde, 4-oxo-butanoic acid, 2-octanone, 1-phenylmethyl-3-
pyrrolidinone, 2-pyrrolidinone, 2-oxo-1,1-dimethylethyl-1-piperidinecarboxylic acid ester), keto acids (2-oxo-propanoic acid, 2-oxo-butyrate, 3-methyl-2-oxobutyrate,
2-oxo-hexanoate, 4-methyl-2-oxo-pentanoic acid, 2-oxo-pentanedioic acid, 2-oxo-benzylpropanoic acid), 2-[(5-nitro-2-furanyl)methylene]-hydrazinecarboxamide.

130
M. Gand et al. / Journal of Molecular Catalysis B: Enzymatic 110 (2014) 126–132
Table 3
Specific activities (mU mg−1) in the NADP+-dependent oxidation catalyzed by five different IREDs.a Each compound was used at a concentration of 20 mM.
Substance
RIR-Sgf3587
RIR-Sip
SIR-Sgf3546
SIR-Pel
RIR-Ppu
Cyclic amines and cyclic amino acids
(S)-(+)-2-methylpyrrolidine
(R)-(-)-2-methylpyrrolidine
2-phenylpyrrolidine
16
9
5
0.2
14 0.1
2.4 1.8
15 5.2
20 3.5
2.9 1.2
10 0.1
0.45 0.05
23 14
280 70
152 23
16 1.2
85 1.6
1.2 0.2
2-methylpiperidine
651 117
1412 234
224 84
n.d.
171 13.4
290 42
194 13.4
1.5 0.6
33 0.8
16 3.8
11 1.6
1.9 0.4
22 2.5
16 0.5
2-ethylpiperidine
7
3
2
37
4
2-propylpiperidine
0.1
28 1.5
n.d.
1,2,3,4-tetrahydroisoquinoline
0.9 0.2
0.7 0.2
2-piperidinecarboxylic acid
7.2 0.2
n.d.
1.4 0.6
n.d.
5
1.2
n.d.
l-Proline
n.d.
n.d.
7.4 0.6
Acyclic amines and amino acids with alkyl groups
N-methyl-2-propanamine
20 9.6
15 0.8
7
3.1
2
1.7 0.6
n.d.
N-methyl-2-butanamine
N-methyl-2-pentanamine
N-methyl-3-pentanamine
75 15.1
20 8.2
58
9
9
3
0.8
1.7 0.2
n.d.
93 3.6
n.d.
2
0.2 0.1
n.d.
480 140
309 13.9
1.8
n.d.
Diisopropylamine
n.d.
n.d.
n.d.
n.d.
13.5
27
2
6-methyl-N-(1-methylethyl)-2-heptanamine
n.d.
n.d.
93 3.4
n.d.
n.d.
n.d.
0.6 0.2
n.d.
2
l-Glutamic acid
1.9 0.3
Acyclic amines with aryl groups
(S)-(-)-N-methyl-1-Phenylethylamine
(R)-(+)-N-methyl-1-Phenylethylamine
N-methyl-1-phenylpropanamine
76 3.8
0.7 0.1
17 0.1
n.d.
1
0.3
0.7
0.9 0.4
n.d.
1
0.4
n.d.
8
n.d.
n.d.
7
4.2
17 0.8
0.6 0.1
n.d., no activity detectable over background (<0.5 mU).
a
For the following compounds, activity was not detectable for any of the investigated IREDs: amino acids (Gly, d- and l-Ala, l-Ser, l-Val, d-Pro, rac-Asp, l-Gln),
amines (R-NH2 with R = methyl–heptyl, benzyl, cyclohexyl, i-propyl, sec-butyl; 1- and 2-phenylethylamine, 3-phenylproylamine, 4-phenylbutylamine, N1,N1-dimethyl-
1,4-benzenediamine, 7-amino-4-methyl-2H-1-benzopyran-2-one, diisopropylamine, pyrrolidine, (S)-1-(2-pyrrolidinylmethyl)-pyrrolidine, (R)- and (S)-3-aminopyrrolidine,
(R)-and (S)-3-aminopiperidine), substrates of structurally similar enzymes (3-hydroxy-2-methyl-propanoic acid, 6-phosphogluconic acid), dihydrofolic acid (5 mM).
not be detected, indicating that the hydride was abstracted from
the chiral carbon of (S)-5 rather than from the N-methyl group.
Around 9 1.1% conversion was reached after 24 h. This leads to
a detectable enantiomeric excess of (R)-5 of 10%eeS, which repre-
sents the maximum value in case that the enzyme shows a perfect
enantioselectivity. In the control reaction the formation of ace-
tophenone was less than 0.01%, and no enantiomeric excess was
detectable. After a proof of the conversion of 5 as an acyclic sub-
strate, the reductive aminations of 6 with methylamine 7 and the
alternative reaction of formaldehyde with phenylethylamine were

M. Gand et al. / Journal of Molecular Catalysis B: Enzymatic 110 (2014) 126–132
131
Table 4
Reduction of three imines after 24 h using resting cells having expressed IRED.
Imine
RIR-Sgf3587
Conversion
SIR-Sgf3546
Conversion
RIR-Sip
SIR-Pel
RIR-Ppu
%ee
%ee
Conversion
%ee
Conversion
%ee
Conversion
%ee
1ba
1ca
1db
94 1%
76 3%
94 1%
(R)-99
(R)-99
35c
79 6%
80 1%
20 3%
(S)-99
(S)-99
37d
99 0.1%
57 5%
42 2%
(R)-99
(R)-99
67c
98 0.1%
68 12%
(S)-99
(S)-99
99d
97 0.6%
72 9%
(R)-99
(R)-99
69c
8
2%
7
1%
a
Substrate concentration: 5 mM.
Substrate concentration: 2.5 mM.
b
c, d
The absolute configuration could not be assigned. The enantiomer formed in excess elutes cat 20.9 or dat 21.2 min during GC analysis.
investigated in sodium phosphate buffer pH 7.5 and glycine–NaOH
buffer pH 10.5 using the photometric assay. Unfortunately, no
consumption of NADPH could be detected, indicating that these
reductive aminations are not possible under these conditions.
biocatalytic synthesis of secondary amines. Two of the IREDs inves-
tigated in this study, RIR-Sip and SIR-Pel, show a high sequence
similarity to RIR-Sgf3587 or SIR-Sgf3546, respectively, and their
enantiopreference matched that of the respective enzyme used as
search query. However, they showed lower activities towards the
investigated substrates. A crystal structure containing both NADP+
and an imine substrate is still missing, and because of this lack of
high quality structural information the molecular basis of activity
and substrate specificity is poorly understood. This limits the suc-
cess of in silico discovery of IREDs with preferably high activity
towards a desired substrate. Besides the already known subfam-
ilies of (R)- and (S)-selective IREDs [20], RIR-Ppu represents the
first characterized enzyme belonging to a third clade of sequences,
which might contain additional IREDs useful for biocatalysis (see
Fig. S7, supplementary information). Similar to the discovery and
characterization of four amine transaminases [33,34], the crystal
structure of RIR-Ppu was deposited in the PDB since 2010, but
owing to the lack of experimental characterization, its capability
of reducing imines was not recognized until now.
The low activity towards most substrates indicates that the
real substrates might not have been identified so far. Although
we investigated the oxidation of different amino acids, primary
amines as well as the reduction of aldehydes, ketones, lactams, ␣-
keto carboxylic acids and 6-phosphogluconate, no activity could
be detected for these substrates (Table 3, footnote). Importantly,
we could show that IREDs are active on acyclic secondary amines,
although the activities and therefore the conversions are low.
Nevertheless, the obtained enantiomeric excess in the oxidation
reaction of rac-5 indicates that the enzyme selectively converts the
(R)-enantiomer. This is a step towards the envisioned synthesis
of secondary amines starting from a ketone and primary amine
of choice. Unfortunately, an asymmetric reductive amination of
ketones was not detectable in this study. However, this reaction
could be demonstrated with low conversion and different reac-
tion conditions by Huber et al. [17] very recently. This underlines
the potential of IREDs for biocatalysis. A successful reductive ami-
nation might be realized as the next step by protein engineering: an
optimized variant should promote imine formation and subsequent
reduction directly in the active site to overcome the unfavourable
equilibrium (hydrolysis).
3.5. Active site residue of the IRED Ppu
During IRED catalyzed imine reduction a hydride of NADPH is
transferred to the carbon atom of the C N double bond of the sub-
strate. Thereby, a proton has to be delivered to the nitrogen atom.
Asp187 located on an ␣-helix is likely to facilitate this protona-
tion step in RIR-Ska, as shown by crystallographic und mutagenesis
studies [19]. Two neighbouring leucines (Leu137 and Leu191) form
a hydrophobic sandwich and are hypothesized to increase the pKa
value of Asp187. Therefore, the pH-optimum of RIR-Ska is ≈3 pH
units above the pKa value of Asp in solution. In RIR-Ppu (3L6D),
however, an alanine replaces the catalytic aspartate at the corre-
sponding position 176. Interestingly, RIR-Ppu features a histidine
180 four amino acids downstream in the amino acid sequence.
Thus, this His180 is positioned one turn away, but its imidazole ring
occupies a similar direction like the side chain of Asp187 of RIR-Ska.
Therefore, it might take over the function as the proton shuttling
residue in this enzyme (maybe via a water molecule). Similar to
RIR-Ska, the neighbouring residues create a relatively hydrophobic
environment, composed by amino acids Met126, Val128, Ile142,
Ala176, Leu179, and Phe184. The next polar amino acid, His251,
is located at the entrance to the active site cleft of the protein.
˚
In between the distance of 9 A (C␤-carbon atoms), two water
molecules are modelled in the structure (for a comparison of the
active sites, see Fig. S6, supplementary information). Nevertheless,
the pH-optimum of this enzyme matches the pKa-value of histi-
dine of 6.0. In order to proof the catalytic function of His180, we
introduced a valine at this position. The purified H180V mutant
showed a tenfold reduced specific residual activity. This confirms
our hypothesis that H180 contributes to catalysis in the active site
of 3L6D.
3.6. Discussion and outlook
One disadvantage of the chemical transition metal-catalyzed
imine hydrogenation methods is the requirement to react the start-
ing ketone with nitrogen sources that contain activated groups, e.g.,
N-aryl, N-acetyl or N-phosphinoyl [32]. After the reduction step,
these activating groups are still present at the nitrogen atom and
usually must be removed as they are not part of the desired tar-
get structure of the natural product or drug. As stated in a recent
assessment of synthetic strategies for amine synthesis, a reaction
step efficient solution is required, but not available at the moment
[32]. Hence, the enzymatic enantioselective reduction of imines
is a desirable target reaction to prepare chiral secondary amines,
and a large toolbox of enzymes acting on different substrates is
desired. Consequently, this study aimed to identify novel imine
reductases in protein sequence databases, which can be applied for
4. Conclusions
We characterized three new IREDs and confirmed our hypoth-
esis that IREDs are not limited to cyclic amines, as conversion of
nine acyclic amines could be confirmed. The RIR-Sip converts one
of these acyclic substrates with a higher activity compared to the
already known cyclic amines. RIR-Ppu is the first representative
enzyme of a new subfamily. For this enzyme a crystal structure is
available. The structure reveals a different arrangement of active
site residues: a histidine contributes to catalysis, rather than the
formerly reported aspartic acid or tyrosine residues. This result
underlines the potential of in-depth characterization of enzymes

132
M. Gand et al. / Journal of Molecular Catalysis B: Enzymatic 110 (2014) 126–132
with unknown functions deposited in the PDB: novel biocatalysts
can be discovered and biochemical insights into their structure
function relationships can be gained.
[13] M. Girardin, S.G. Ouellet, D. Gauvreau, J.C. Moore, G. Hughes, P.N. Devine, P.D.
O’Shea, L.-C. Campeau, Org. Process. Res. Dev. 17 (2013) 61–68.
[14] C.K. Savile, J.M. Janey, E.C. Mundorff, J.C. Moore, S. Tam, W.R. Jarvis, J.C. Colbeck,
A. Krebber, F.J. Fleitz, J. Brands, P.N. Devine, G.W. Huisman, G.J. Hughes, Science
329 (2010) 305–309.
[15] R.C. Simon, F. Zepeck, W. Kroutil, Chem. Eur. J. 19 (2013) 2859–2865.
[16] K. Mitsukura, T. Kuramoto, T. Yoshida, N. Kimoto, H. Yamamoto, T. Nagasawa,
Appl. Microbiol. Biotechnol. 97 (2013) 8079–8086.
[17] T. Huber, L. Schneider, A. Präg, S. Gerhardt, O. Einsle, M. Müller, ChemCatChem
6 (2014) 2248–2252.
[18] K. Mitsukura, M. Suzuki, S. Shinoda, T. Kuramoto, T. Yoshida, T. Nagasawa,
Biosci. Biotechnol. Biochem. 75 (2011) 1778–1782.
[19] M. Rodríguez-Mata, A. Frank, E. Wells, F. Leipold, N.J. Turner, S. Hart, J.P. Turken-
burg, G. Grogan, ChemBioChem 14 (2013) 1372–1379.
[20] P.N. Scheller, S. Fademrecht, S. Hofelzer, J. Pleiss, F. Leipold, N.J. Turner, B.M.
Nestl, B. Hauer, ChemBioChem (2014).
[21] M. Genz, V. Köhler, M. Krauss, D. Singer, R. Hoffmann, T.R. Ward, N. Sträter,
ChemCatChem 6 (2014) 736–740.
Acknowledgements
We thank Uwe T. Bornscheuer for his support and constructive
discussions. We thank the New York Structural Genomics Research
Consortium for providing the expression plasmid of RIR-Ppu.
Rainer Wardenga thanks the European Union (NewProt project,
KBBE-2011- 5, grant no. 289350) for financial support.
Appendix A. Supplementary data
[22] J.M. Zimbron, T. Heinisch, M. Schmid, D. Hamels, E.S. Nogueira, T. Schirmer, T.R.
Ward, J. Am. Chem. Soc. 135 (2013) 5384–5388.
[23] T. Vaijayanthi, A. Chadha, Tetrahedron: Asymmetry 19 (2008) 93–96.
[24] S.H. Smits, A. Mueller, L. Schmitt, M.K. Grieshaber, J. Mol. Biol. 381 (2008)
200–211.
Supplementary material related to this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.molcatb.
2014.09.017.
[25] H. Muramatsu, H. Mihara, R. Kakutani, M. Yasuda, M. Ueda, T. Kurihara, N. Esaki,
J. Biol. Chem. 280 (2005) 5329–5335.
References
[26] F. Leipold, S. Hussain, D. Ghislieri, N.J. Turner, ChemCatChem
3505–3508.
5 (2013)
[1] M. Wink, Phytochemistry 64 (2003) 3–19.
[27] M.M. Bradford, Anal. Biochem. 72 (1976) 248–254.
[28] A. Nobili, M.G. Gall, I.V. Pavlidis, M.L. Thompson, M. Schmidt, U.T. Bornscheuer,
FEBS J. 280 (2013) 3084–3093.
[2] B.W. Cue, J. Zhang, Green Chem. Lett. Rev. 2 (2009) 193–211.
[3] S.Y. Hsieh, M. Binanzer, I. Kreituss, J.W. Bode, Chem. Commun. 48 (2012)
8892–8894.
[29] S. Reitz, A. Alhapel, L.O. Essen, A.J. Pierik, J. Mol. Biol. 382 (2008) 802–811.
[30] J. Osipiuk, M. Zhou, S. Moy, F. Collart, A. Joachimiak, J. Struct. Funct. Genomics
10 (2009) 249–253.
[4] C. Li, J. Xiao, J. Am. Chem. Soc. 130 (2008) 13208–13209.
[5] K. Buchholz, V. Kasche, U.T. Bornscheuer, Biocatalysts and Enzyme Technology,
Wiley-VCH, Weinheim, 2005.
[31] T. Hitosugi, L. Zhou, S. Elf, J. Fan, H.B. Kang, J.H. Seo, C. Shan, Q. Dai, L. Zhang, J.
Xie, T.L. Gu, P. Jin, M. Aleckovic, G. LeRoy, Y. Kang, J.A. Sudderth, R.J. DeBerar-
dinis, C.H. Luan, G.Z. Chen, S. Muller, D.M. Shin, T.K. Owonikoko, S. Lonial, M.L.
Arellano, H.J. Khoury, F.R. Khuri, B.H. Lee, K. Ye, T.J. Boggon, S. Kang, C. He, J.
Chen, Cancer Cell 22 (2012) 585–600.
[32] T.C. Nugent, M. El-Shazly, Adv. Synth. Catal. 352 (2010) 753–819.
[33] F. Steffen-Munsberg, C. Vickers, A. Thontowi, S. Schätzle, T. Meinhardt, M.
Svedendahl Humble, H. Land, P. Berglund, U.T. Bornscheuer, M. Höhne, Chem-
CatChem 5 (2013) 154–157.
[6] U.T. Bornscheuer, G.W. Huisman, R.J. Kazlauskas, S. Lutz, J.C. Moore, K. Robins,
Nature 485 (2012) 185–194.
[7] M. Höhne, S. Kühl, R. Karen, U.T. Bornscheuer, ChemBioChem 9 (2008) 363–365.
[8] D. Ghislieri, N.J. Turner, Top. Catal. 57 (2014) 284–300.
[9] W. Kroutil, E.M. Fischereder, C.S. Fuchs, H. Lechner, F.G. Mutti, D. Pressnitz, A.
Rajagopalan, J.H. Sattler, R.C. Simon, E. Siirola, Org. Process. Res. Dev. 17 (2013)
751–759.
[10] H. Kohls, F. Steffen-Munsberg, M. Höhne, Curr. Opin. Chem. Biol. 19 (2014)
180–192.
[34] F. Steffen-Munsberg, C. Vickers, A. Thontowi, S. Schätzle, T. Tumlirsch, M.
Svedendahl Humble, H. Land, P. Berglund, U.T. Bornscheuer, M. Höhne, Chem-
CatChem 5 (2013) 150–153.
[11] M. Alexeeva, A. Enright, M.J. Dawson, M. Mahmoudian, N.J. Turner, Angew.
Chem. Int. Ed. 41 (2002) 3177–3180.
[12] M. Höhne, U.T. Bornscheuer, ChemCatChem 1 (2009) 42–51.