Combined Imine Reductase and Amine Oxidase Catalyzed Deracemization of Nitrogen Heterocycles

Article abstract of DOI:10.1002/cctc.201500822

A novel amine oxidase (AO)/imine reductase (IRED) system was developed for the deracemization of racemic amines. By combining (R)-6-hydroxy-d-nicotine oxidase (6-HDNO) with an (R)-IRED, a panel of racemic 2-substituted piperidines and pyrrolidines were deracemized to yield the (S)-amines in high yields and enantiomeric excess values. Other N-heterocycles were deracemized with monoamine oxidase (MAO-N) or 6-HDNO in combination with ammonia borane, which allowed comparison of the two enzyme deracemization approaches with that involving a chemical reducing agent.

Full text of DOI:10.1002/cctc.201500822

DOI: 10.1002/cctc.201500822
Communications
Combined Imine Reductase and Amine Oxidase Catalyzed
Deracemization of Nitrogen Heterocycles
[
a]
Rachel S. Heath, Marta Pontini, Shahed Hussain, and Nicholas J. Turner*
A novel amine oxidase (AO)/imine reductase (IRED) system was
developed for the deracemization of racemic amines. By com-
bining (R)-6-hydroxy-d-nicotine oxidase (6-HDNO) with an (R)-
IRED, a panel of racemic 2-substituted piperidines and pyrroli-
dines were deracemized to yield the (S)-amines in high yields
and enantiomeric excess values. Other N-heterocycles were de-
racemized with monoamine oxidase (MAO-N) or 6-HDNO in
combination with ammonia borane, which allowed comparison
of the two enzyme deracemization approaches with that
involving a chemical reducing agent.
Enantiomerically pure chiral N-heterocycles such as substituted
piperidines, morpholines, and piperazines constitute privileged
structures in medicinal chemistry and, hence, are frequently
encountered in pharmaceutical compounds that are currently
Scheme 1. Deracemization of chiral amines with amine oxidase and a) a
nonselective chemical reducing agent such as ammonia borane or b) an
imine reductase whole-cell biocatalyst.
in development. All of these motifs are challenging to prepare
by enantioselective catalysis and often require bespoke meth-
ods to address specific classes of compound. Even though clas-
sical resolution by diastereomeric salt formation and chroma-
tography on chiral supports still remains state of the art,
a number of asymmetric synthetic methodologies have been
[
5]
es, have been shown to catalyze enantioselective reduction
of cyclic imines. We envisaged that the nonselective chemical
reducing agent in Scheme 1a could be replaced with an IRED
and NADPH (Scheme 1b) to generate single-enantiomer
amines upon starting from racemic mixtures. Moreover, coex-
pression of the amine oxidase and imine reductase in whole
cells with the addition of glucose for NADPH recycling could
enable a one-pot deracemization process. The use of a highly
enantioselective (R)- or (S)-imine reductase should result in re-
action times that are reduced relative to those necessary for
chemical reducing agent, because there is no reduction of the
imine to the undesired enantiomer. Potential challenges for
this system included identifying reaction conditions suitable
for both enzyme activities, especially maintaining high IRED/
[
1]
reported. The widespread application of these strategies has
been hampered by restricted substrate scopes or the use of in-
convenient reaction procedures, for example, toxic or rare
metal catalysts, or the requirement for additional steps for
attachment and removal of (chiral) auxiliaries.
The recent emergence of technologies such as enzyme engi-
neering/evolution, metagenomics, and high-throughput
screening methodologies have provided powerful starting
points and tools to generate tunable biocatalysts for enantio-
selective synthesis. For example, enantiomerically pure cyclic
amines have been prepared by employing engineered amine
oxidases (AOs) in combination with a nonselective chemical re-
[2]
ducing agent (Scheme 1a). In this deracemization process,
the chemical reducing agent (typically ammonia borane) is
added in excess amount (at least fourfold) and reduces the
imine nonselectively back to the amine. Recently, imine reduc-
NADPH activity in the presence of H O2 generated by the
2
amine oxidase.
Previous applications of enantioselective amine oxidases
have primarily focused on the deracemization of pyrrolidines,
[
3,4]
[2]
tases (IREDs),
including artificial bioinspired imine reductas-
tetrahydroisoquinolines, and b-carbolines. Hence, for the new
IRED-based system we decided to focus primarily on the dera-
cemization of 2-substituted piperidines (Figure 1a). We also
report a comparison with the ammonia borane based system
and extension of the latter to encompass other N-heterocycles
(Figure 1b).
[
a] Dr. R. S. Heath, Dr. M. Pontini, S. Hussain, Prof. N. J. Turner
School of Chemistry
University of Manchester,
Manchester Institute of Biotechnology
1
31 Princess Street
We began by examining the combination of the (R)-selective
Manchester, M1 7DN (UK)
Fax: (+44)161-275-1311
E-mail: nicholas.turner@manchester.ac.uk
[2f]
6
-hydroxy-d-nicotine oxidase (6-HDNO) E350L/E352D variant
and the (R)-IRED from Streptomyces sp. GF3587, for which the
[
4]
substrate scope has recently been reported. 2-(4-Methoxy-
Supporting Information (experimental methods) for this article is avail-
able on the WWW under http://dx.doi.org/10.1002/cctc.201500822.
phenyl)piperidine (1c) was selected as a test substrate in view
ChemCatChem 2016, 8, 117 – 120
117
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
3
4% conversion with 30%ee (Table 1, entry 1). A control reac-
tion under the same conditions but by using the analogous
imine as the substrate demonstrated 99% conversion with
9
9%ee (Table 1, entry 2); thus, the problem appeared to be
with the activity of 6-HDNO. In entry 3 (Table 1), the conditions
[
2f]
are those used previously with 6-HDNO that gave 99%ee.
The differences in these conditions compared to those of
entry 1 are therefore in pH, buffer molarity, and glucose addi-
tion. Entry 4 (Table 1) shows that the pH was not the issue. En-
tries 5–7 (Table 1) reveal that the combination of glucose and
lower buffer molarity attenuated the activity of 6-HDNO. We,
thus, confirmed that the IRED could work with less glucose
(
10 mm) and at a higher buffer concentration (Table 1, entries 8
and 9) so that under the optimized conditions of pH 7.4, 1m
buffer with 10 mm glucose we were able to combine 6-HDNO
and the IRED to achieve full conversion and 99%ee upon start-
ing from the racemic amine (Table 1, entry 10). We also noted
that if imine reductase biotransformations were spiked with
5
mm hydrogen peroxide (the maximum concentration that
should ever be present in the 6-HDNO/IRED system, assuming
Table 2. One-pot deracemization of amines by using 6-HDNO E350L/
[a]
E352D and (R)-IRED.
[b]
[c]
Entry
rac-Amine
Conversion
%]
ee
ee with (R)-IRED from
analogous imine [%]
[
[%]
1
2
3
4
5
6
7
8
9
10
1a
1b
1c
1d
1e
1 f
1g
1h
3a
3b
99
95
99
99
99
99
99
99
71
99
99 (S)
95 (S)
99 (S)
17 (S)
93 (S)
99 (S)
99 (S)
76 (n.a.)
72 (S)
99 (S)
37 (S)
89 (S)
>98 (S)
77 (S)
>98 (S)
n.t.
66 (S)
78 (n.a.)
8 (S)
Figure 1. Substrates and deracemization schemes for a) AO/IRED with piperi-
dines/pyrrolidines and b) AO/NH
lines, and thiomorpholines.
3 3
BH with piperidines, piperazines, morpho-
of the high reactivity of the amine with 6-HDNO and also the
corresponding imine with the IRED (full conversion and 99%
enantiomeric excess (ee) within 24 h). [Note that the (R)-IRED
generates the (S)-configured product for this substrate.] Initial
conditions for the deracemization were based on those previ-
ously used for imine reduction with the (R)-IRED (i.e., 100 mm
pH 7.4 potassium phosphate buffer with 50 mm glucose and
66 (S)
[a] Conditions: 5 mm amine, 1m pH 7.4 potassium phosphate buffer,
10 mm glucose, E. coli whole-cell biocatalysts, 308C, 24 h; n.t.=not
tested, n.a.=not assigned. [b] Based on remaining imine. [c] Absolute
configuration was determined by comparison with previous results with
(R)-IRED or 6-HDNO (see the Supporting Information).
[
4a]
[2f]
[
5]
5
mm amine). However, these conditions resulted in only
Table 1. Optimization of conditions for deracemization of rac-amine 1c by combining 6-HDNO E350L/E352D
with (R)-IRED.
a 1:1 ratio of oxidation of amine
to production of H O ), there
[a]
2
2
was no effect on the ee (>99%)
or conversion (>99%) upon
using 1m buffer; with 100 mm
buffer there was a slight de-
crease in conversion (95%),
though no influence on the ee.
The optimized conditions
were then applied to the IRED/6-
HDNO-catalyzed deracemization
of piperidines 1a–h as well as
pyrrolidines 3a and 3b (Table 2),
which resulted in both high con-
versions and high ee values in
Entry
Amine or imine 6-HDNO (R)-IRED
pH Buffer
Glucose
Conversion ee
[%]
concentration concentration [%]
[
m]
[mm]
1
2
3
4
5
6
7
8
9
amine
imine
amine
amine
amine
amine
amine
imine
imine
amine
yes
no
yes
yes
no
no
no
no
no
yes
yes
yes
7.4 0.1
7.4 0.1
50
50
0
0
0
50
50
50
10
10
34
>99
N/A
N/A
N/A
N/A
N/A
>99
>99
>99
30
>99
>99
>99
>99
30
>99
>99
>99
>99
yes
yes
yes
yes
yes
no
8
7.4
1
1
7.4 0.1
7.4 0.1
7.4
7.4
7.4
7.4
1
1
1
1
no
yes
1
0
[
a] Conditions: Potassium phosphate buffer, 5 mm substrate, E. coli whole-cell biocatalyst. N/A=not applicable.
ChemCatChem 2016, 8, 117 – 120
www.chemcatchem.org
118
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
many cases. For comparison, the right-hand column in Table 2
generate a single whole-cell biocatalyst capable of deracemiza-
tion; to date, we have only managed to obtain 84% conver-
sion and 84%ee for 1c with this type of system (see the Sup-
porting Information), as the two enzymes appear to require
different conditions for optimum expression. Further variation
in expression conditions (including promoters, media, etc.) and
biotransformation conditions may yet provide a solution to
this approach.
displays the ee obtained if only the (R)-IRED was used with the
[4a]
corresponding imine.
Entries 1, 2, 7, 9, and 10 of Table 2
highlight examples for which the ee of the resulting amine is
enhanced with the 6-HDNO/IRED system relative to that ob-
tained with the use of the IRED alone with the imine. In con-
trast, entry 4 (Table 2) shows a lower ee because of the low ac-
tivity of this substrate with 6-HDNO.
The use of an IRED in place of ammonia borane should
enable more rapid deracemization owing to recycling of the
imine to the desired enantiomer. Figure 2 shows that deracem-
ization of 1c by using the 6-HDNO/IRED system is complete
To enable comparison of the IRED-based system with that
using ammonia borane, and to gain access to both (R) and (S)
products, we employed the enantiocomplementary mono-
amine oxidase (MAO-N) and 6-HDNO enzymes. Deracemization
of a range of racemic N-heterocyclic amines (Figure 1) was per-
formed by using either MAO-N or 6-HDNO variants in combi-
nation with NH BH . Racemic piperidines were deracemized
3
3
with high ee values in 16–96 h (Table 3). For many substrates,
the products were isolated from preparative-scale reactions
(
100–500 mg). MAO-N was active on 16 out of 19 substrates
and 6-HDNO was active on 12 out of 19 substrates. Although
-HDNO was active on piperidines and thiomorpholines, it was
6
not active with morpholines and piperazines, unlike MAO-N,
which was active on all four types of substrate. This system,
therefore, provides a method for the asymmetric synthesis of
2
-substituted morpholines, thiomorpholines, and piperazines,
[
8]
examples of which are rare in the literature. For the majority
of substrates, the desired enantiomer can be obtained simply
by appropriate selection of either MAO-N or 6-HDNO. The ad-
vantage of this approach is that because deracemization does
not depend on the complementary substrate scope for two
enzymes, a larger panel of amines can be explored. However,
typically, longer reaction times are observed than for the IRED/
6-HDNO system.
Figure 2. Comparison of the deracemization of racemic 2-(4-methoxyphe-
nyl)piperidine (1c) to the (S) enantiomer by using 6-HDNO combined with
(
R)-IRED (black) or ammonia borane (gray).
(
99% conversion, 99%ee) after 2–4 h, whereas the use of am-
monia borane requires at least 6 h, after which the ee is only
7%.
Using the 6-HDNO/IRED com-
bination, (S)-2-phenylpiperidine
1a) was produced on a 0.25 g
9
Table 3. Deracemization parameters for substrates 1 and 2.
(
scale starting from the racemic
amine (99%ee, 82% yield) and
also starting from the imine
Substrate
MAO-N
6-HDNO E350L/E352D
[
a]
[a]
Variant
ee [%]
Time [h]
Yield [%]
ee [%]
Time [h]
Yield [%]
1
a
D5
D5
D9
D9
D5/D9
D5
D5/D9
D5/D9
D5
D9
D5
D5
D9
D9
D5
D5
D9
99 (R)
99 (R)
99 (R)
99 (R)
inactive
99 (R)
inactive
inactive
99 (R)
16
48
96
96
–
85
89
53
–
–
89
–
–
–
–
89
84
–
63
–
68
88
47
45
99 (S)
99 (S)
99 (S)
99 (S)
99 (S)
99 (S)
90 (S)
99 (S)
99 (S)
16
16
16
16
–
78
–
–
–
–
—
–
–
–
–
–
–
–
–
62
–
–
–
(
99%ee, 80% yield), which there-
1b
1c
by demonstrates the practical
application of this methodology.
This approach of combining
an AO with an IRED is currently
partially limited by the substrate
scope of the imine reductases.
However, new IREDs are rapidly
being discovered, for which the
substrate scope will likely differ
1
1
1
e
f
g
[
b]
b]
48
16
48
72
48
96
–
16
–
[
1i
1j
–
1
1
2
k
l
a
72
96
16
72
60
24
96
48
48
72
90
62 (n.a.)
99 (S)
82 (S)
inactive
inactive
inactive
inactive
inactive
99 (R)
99 (R)
inactive
inactive
2b
2c
2d
99 (S)
–
–
-
–
72
48
–
99 (n.a.)
99 (+)
65 (S)
[6]
from that reported here. Addi-
tionally, crystal structures of
imine reductases are becoming
available, and this will enable en-
gineering of IREDs to expand
2
2
e
f
99 (S)
99 (S)
99 (S)
99 (+)
2g
2h
D9
D9
2
i
–
[
7]
the substrate range. Ideally, the
AO and IRED genes should be
coexpressed in the same cell to
[a] Absolute configuration of the product was determined by comparison with authentic standards or the spe-
cific rotation ([a] ) (see the Supporting Information). [b] Results published previously in Ref. [2f]; n.a.=not as-
D
signed.
ChemCatChem 2016, 8, 117 – 120
www.chemcatchem.org
119
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
In summary we have shown that an NADPH-dependent IRED
can be combined with an oxygen-dependent AO in a one-pot
process for the enantioselective synthesis of chiral amines by
deracemization. Significantly, an increased ee can be achieved
with the AO/IRED system relative to that obtained by using
the IRED alone for the reduction of the corresponding imine.
In principle, this approach could be further extended by ex-
changing 6-HDNO and (R)-IRED for MAO-N and/or (S)-IRED to
access the complementary enantiomeric amines. Equally, IREDs
could be combined with other amine oxidases such as the (S)-
[2] a) D. Ghislieri, A. P. Green, M. Pontini, S. C. Willies, I. Rowles, A. Frank, G.
Grogan, N. J. Turner, J. Am. Chem. Soc. 2013, 135, 10863–10869; b) M.
Alexeeva, A. Enright, M. J. Dawson, M. Mahmoudian, N. J. Turner, Angew.
Chem. Int. Ed. 2002, 41, 3177–3180; Angew. Chem. 2002, 114, 3309–
3312; c) D. Ghislieri, D. Houghton, A. P. Green, M. Pontini, S. C. Willies,
N. J. Turner, ACS Catal. 2013, 3, 2869–2872; d) I. Rowles, K. J. Malone,
L. L. Etchells, S. C. Willies, N. J. Turner, ChemCatChem 2012, 4, 1259–
1261; e) R. Carr, M. Alexeeva, M. J. Dawson, V. Gotor-Fernandez, C. E.
Humphrey, N. J. Turner, ChemBioChem 2005, 6, 637–639; f) R. S. Heath,
M. Pontini, B. Bechi, N. J. Turner, ChemCatChem 2014, 6, 996–1002.
[3] a) “Imine Reductases”: F. Leipold, S. Hussain, S. P. France, N. J. Turner, Sci-
ence of Synthesis Biocatalysis in Organic Synthesis 2 (Eds.: K. Faber, W.-D.
Fessner, N. J. Turner), Thieme, Stuttgart, 2015, Chap. 2.4.2, pp. 359–381;
b) J. H. Schrittwieser, S. Velikogne, W. Kroutil, Adv. Synth. Catal. 2015,
357, 1655–1685; c) F. Leipold, S. Hussain, D. Ghislieri, N. J. Turner, Chem-
CatChem 2013, 5, 3505–3508; d) K. Mitsukura, T. Kuramoto, T. Yoshida,
N. Kimoto, H. Yamamoto, T. Nagasawa, Appl. Microbiol. Biotechnol. 2013,
[9]
selective cyclohexylamine oxidase variants and the (R)-selec-
[
10]
tive engineered d-amino acid oxidase (pkDAO Y228L/R283),
which have substrate specificity complementary to that of the
MAO-N and 6-HDNO variants. NADPH regeneration is not a con-
cern for the AO/IRED system; the IRED is used as a whole-cell
biocatalyst, in which NADPH is regenerated by the cell with
only the addition of glucose, the cost of which is negligible, es-
pecially if compared to the cost of ammonia borane required
for the AO/NH BH system. Thus, with the future discovery and
9
7, 8079–8086.
[
4] a) S. Hussain, F. Leipold, H. Man, E. Wells, S. P. France, K. R. Mulholland,
G. Grogan, N. J. Turner, ChemCatChem 2015, 7, 579–583; b) K. Mitsu-
kura, M. Suzuki, S. Shinoda, T. Kuramoto, T. Yoshida, T. Nagasawa, Biosci.
Biotechnol. Biochem. 2011, 75, 1778–1782.
[5] a) F. Schwizer, V. Kçhler, M. Dürrenberger, L. Knçrr, T. R. Ward, ACS Catal.
3
3
2013, 3, 1752–1755; b) V. Kçhler, Y. M. Wilson, M. Dürrenberger, D. Ghi-
development of AOs and IREDs that have overlapping sub-
strate specificity, we anticipate that the AO/IRED deracemiza-
tion system will find even wider application for the deracemi-
zation of racemic amines.
slieri, E. Churakova, T. Quinto, L. Knçrr, D. Häussinger, F. Hollmann, N. J.
Turner, T. R. Ward, Nat. Chem. 2013, 5, 93–99; c) V. M. Robles, M. Dür-
renberger, T. Heinisch, A. Lledos, T. Schirmer, T. R. Ward, J. MarØchal, J.
Am. Chem. Soc. 2014, 136, 15676–15683.
6] a) P. N. Scheller, S. Fademrecht, S. Hofelzer, J. Pleiss, F. Leipold, N. J.
Turner, B. M. Nestl, B. Hauer, ChemBioChem 2014, 15, 2201–2204; b) D.
Gamenara, P. D. de María, Org. Biomol. Chem. 2014, 12, 2989–2992;
c) M. Ganda, H. Müllera, R. Wardengab, M. Hçhne, J. Mol. Catal. B 2014,
[
[
Acknowledgements
1
10, 126–132; d) D. Wetzl, M. Berrera, N. Sandon, D. Fishlock, M. Ebel-
ing, M. Müller, S. Hanlon, B. Wirz, H. Iding, ChemBioChem 2015, 16, DOI:
0.1002/cbic.201500218.
Research leading to these results received funding from the Euro-
pean Union’s Seventh Framework Programme FP7/2007–2013
under grant agreement no. 266025 (BIONEXGEN). We are grateful
for funding from the Marie Curie Initial Training Network (BIO-
TRAINS FP7-ITN-238531 to M.P.) and the Royal Society for a Wolf-
son Research Merit Award (N.J.T.).
1
7] a) M. Rodríguez-Mata, A. Frank, E. Wells, F. Leipold, N. J. Turner, S. Hart,
J. P. Turkenburg, G. Grogan, ChemBioChem 2013, 14, 1372–1379; b) T.
Huber, L. Schneider, A. Präg, S. Gerhardt, O. Einsle, M. Müller, Chem-
CatChem 2014, 6, 2248–2252; c) H. Man, E. Wells, S. Hussain, F. Leipold,
S. Hart, J. P. Turkenburg, N. J. Turner, G. Grogan, ChemBioChem 2015, 16,
1
052–1059.
8] C.-V. Vo, G. Mikutis, J. W. Bode, Angew. Chem. Int. Ed. 2013, 52, 1705–
708; Angew. Chem. 2013, 125, 1749–1752.
[
[
1
Keywords: amines
· deracemization · enantioselectivity ·
9] a) G. Li, J. Ren, P. Yao, Y. Duan, H. Zhang, Q. Wu, J. Feng, P. C. K. Lau, D.
Zhu, ACS Catal. 2014, 4, 903–908; b) G. Li, J. Ren, H. Iwaki, D. Zhang, Y.
Hasegawa, Q. Wu, J. Feng, P. C. Lau, D. Zhu, Appl. Microbiol. Biotechnol.
enzyme catalysis · nitrogen heterocycles
[
1] a) S. Yokoshima, K. Watanabe, F. Uejara, Y. Usui, H. Tanaka, Bioorg. Med.
Chem. Lett. 2014, 24, 5749–5751; b) M. R. Monaco, P. Renzi, D. M. S.
Schietroma, M. Bella, Org. Lett. 2011, 13, 4546–4549; c) M. Hussain, T. S.-
L. Banchelin, H. Andersson, R. Olsson, F. Almqvist, Org. Lett. 2013, 15,
2
014, 98, 1681–1689.
10] K. Yasukawa, S. Nakano, Y. Asano, Angew. Chem. Int. Ed. 2014, 53, 4428–
431; Angew. Chem. 2014, 126, 4517–4520.
[
4
5
4–57; d) L. V. Adriaenssens, C. A. Austin, M. Gibson, D. Smith, R. C.
Hartley, Eur. J. Org. Chem. 2006, 4998–5001; M. U. Luescher, C.-V. Vo,
J. W. Bode, Org. Lett. 2014, 16, 1236–1239.
Received: July 24, 2015
Published online on October 16, 2015
ChemCatChem 2016, 8, 117 – 120
www.chemcatchem.org
120
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim