Mapping the substrate scope of monoamine oxidase (MAO-N) as a synthetic tool for the enantioselective synthesis of chiral amines

Article abstract of DOI:10.1016/j.bmc.2017.07.023

A library of 132 racemic chiral amines (α-substituted methylbenzylamines, benzhydrylamines, 1,2,3,4-tetrahydronaphthylamines (THNs), indanylamines, allylic and homoallylic amines, propargyl amines) was screened against the most versatile monoamine oxidase (MAO-N) variants D5, D9 and D11. MAO-N D9 exhibited the highest activity for most substrates and was applied to the deracemisation of a comprehensive set of selected primary amines. In all cases, excellent enantioselectivity was achieved (e.e. >99%) with moderate to good yields (55–80%). Conditions for the deracemisation of primary amines using a MAO-N/borane system were further optimised using THN as a template addressing substrate load, nature of the enzyme preparation, buffer systems, borane sources, and organic co-solvents.

Full text of DOI:10.1016/j.bmc.2017.07.023

Accepted Manuscript
Mapping the Substrate Scope of Monoamine Oxidase (MAO-N) as a Synthetic
Tool for the Enantioselective Synthesis of Chiral Amines
Susanne Herter, Florian Medina, Simon Wagschal, Cyril Benhaïm, Friedemann
Leipold, Nicholas J. Turner
PII:
S0968-0896(17)30391-7
DOI:
http://dx.doi.org/10.1016/j.bmc.2017.07.023
BMC 13860
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
24 February 2017
23 June 2017
12 July 2017
Please cite this article as: Herter, S., Medina, F., Wagschal, S., Benhaïm, C., Leipold, F., Turner, N.J., Mapping the
Substrate Scope of Monoamine Oxidase (MAO-N) as a Synthetic Tool for the Enantioselective Synthesis of Chiral
Amines, Bioorganic & Medicinal Chemistry (2017), doi: http://dx.doi.org/10.1016/j.bmc.2017.07.023
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Mapping the Substrate Scope of
Monoamine Oxidase (MAO-N) as a
Synthetic Tool for the Enantioselective
Synthesis of Chiral Amines.
Susanne Herter,†,• Florian Medina,‡,• Simon Wagschal,*,‡ Cyril Benhaïm,‡ Friedemann
Leipold† and Nicholas J. Turner*,†
• These authors contributed equally to the work.
†School of Chemistry, University of Manchester, Manchester Institute of Biotechnology, 131
Princess Street, M1 7DN, Manchester, United Kingdom.
‡Pharmaceutical Development and Manufacturing Sciences, Janssen Pharmaceutica,
Turnhoutseweg 30, B-2340 Beerse, Belgium.
KEYWORDS: Monoamine oxidase (MAO-N), enantiomerically pure chiral amines, borane,
deracemisation, enzyme catalysis, chirality.
1

Abstract
A
library of 132 racemic chiral amines ( -substituted methylbenzylamines,
benzhydrylamines, 1,2,3,4-tetrahydronaphthylamines (THNs), indanylamines, allylic and
homoallylic amines, propargyl amines) was screened against the most versatile monoamine
oxidase (MAO-N) variants D5, D9 and D11. MAO-N D9 exhibited the highest activity for
most substrates and was applied to the deracemisation of a comprehensive set of selected
primary amines. In all cases, excellent enantioselectivity was achieved (e.e. > 99%) with
moderate to good yields (55-80%). Conditions for the deracemisation of primary amines
using a MAO-N/borane system were further optimised using THN as a template addressing
substrate load, nature of the enzyme preparation, buffer systems, borane sources, and organic
co-solvents.
2

Table of contents
Introduction
The synthesis of enantiomerically pure compounds employing biocatalysts in place of
conventional chemocatalysts has recently gained momentum.1 The use of biocatalysis offers a
sustainable alternative technology for the preparation of a broad range of highly
functionalised intermediates and building blocks. In this context, the development of efficient
and low waste biocatalytic methodologies is of high interest in both academia and industry.2
Enantiomerically pure amines constitute an important class of chiral building blocks with
40% of Active Pharmaceutical Ingredients (APIs) containing a chiral amine moiety (Figure
3

1).1 For instance, 1 was found to be a potent anti-cancer drug inhibiting apoptosis proteins3
and rasagiline 2 has been commercialised for the treatment of Parkinson’s disease.4
Furthermore, dilevalol 3 is administered against hypertension5 and compound 4 has found
application in pain treatments.6 Hence, synthetic methods for controlling stereochemistry in
organic synthesis are of great importance in the development of new drugs. In this context,
chirality is an essential dimension in pharmacology thus guiding industries to focus on the
synthesis of single enantiomeric drug entities.
Figure 1. Representative examples of biologically active compounds featuring
enantiomerically pure amine scaffolds.
A variety of strategies such as diastereomeric resolution, homogeneous catalysis or
biocatalytic approaches have been reported for the preparation of chiral amines.7 The use of
biocatalysis has grown remarkably in the past decades as exemplified by the use of -
transaminases,8 phenylalanine ammonia lyases,9 amine dehydrogenases10 and imine
reductases.11 An attractive alternative to the kinetic resolution of amines involves the use of
monoamine oxidase, an enzyme class which only requires aerial oxygen with no need for
additional or sacrificial co-factors.12 We have previously developed variants of the
4

Aspergillus niger monoamine oxidase N (MAO-N) such as D5 which was, unlike the wild-
type enzyme, the first MAO-N generated in our laboratories to accept cyclic secondary and
tertiary amines.13 For example, (R)-1-phenylethylamine, (R)-coniine and (R)-1-methyl-2-
phenylpyrrolidine were obtained in an enantioselective fashion with MAO-N D5
demonstrating a high (S)-enantiopreference. In further rounds of rational protein design
targeting the generation of MAO-N mutants with increased volume of the active-site pocket,
variant D9 was generated and found to display significantly enhanced activity towards more
bulky amines.14 Several natural products such as crispine A, leptaflorin and eleagnine were
synthesised with MAO-N D9 in an enantiomerically pure fashion while retaining the parental
enantioselectivity of MAO-N D5. Furthermore, MAO-N D11 has been established as the
‘high-end’ biocatalyst in the MAO-N family tree with the largest active-site and entrance
channel. This variant was successfully applied to the deracemisation of primary, secondary
and tertiary hindered amines such as substituted benzhydrylamines, 1-phenyl-
tetrahydroisoquinoline and tetrahydrocarbolines.14,15
With this versatile set of MAO-N biocatalysts in hand, we set out to investigate the
applicability of variants D5, D9 and D11 in the enantioselective synthesis of amines
frequently found in APIs (Figure 1). To this end, we firstly turned our attention towards the
screening of 5 amine libraries comprising (i) -substituted benzylamines, (ii) -substituted
allyl-, homoallyl- and propargylamines, (iii) -substituted aliphatic amines and (iv) 1,2,3,4-
tetrahydronaphthylamine (THN) and aminoindane derivatives via a previously established
high-throughput assay16 (Figure 2).
5

Figure 2: General structure of the different amine libraries screened in this study.
Results and Discussion
In order to gain deeper insight into a broader variety of previously untested amines that
could be accessed by MAO-N-catalysed oxidations, we firstly conducted a structure-activity
relationship study starting from simple mono- and disubstituted -substituted benzylamines (
1-phenylethan-1-amine related compounds, MBAs) (Tables 1 and 2).
The prototype substrate MBA 5 was oxidised with high activity by MAO-N D9 and D11
which showed a 3- to 5-fold increased specific activity when compared to variant D5 (Table
1, entry 1). The specific activity displayed by MAO-N variant D9 towards MBA 5 was set to
100% (relative activity) and henceforth used as the benchmark for both the activities shown
by MAO-N D5 and D11 and all subsequent substrates tested. Fluorine substitution at C-2
gave the highest activities for the three enzymes but D9 was the only variant to produce good
activity with the three fluorinated substrates 6-8 (Table 1, entry 2). The ortho-chlorinated
and brominated compounds 9 and 12 were only moderately oxidised by MAO-N D9 and low
to moderate activity was displayed in all cases with the meta-substituted compounds 10 and
13. In contrast, both MAO-N D9 and D11 displayed excellent activities towards the para-
6

chloro and para-bromo substituted amines 11 and 14, respectively (Table 1, entries 3 and 4).
Iodine substitution was only accepted in the para-position with a good activity obtained with
both MAO-N D9 and D11 confirming the higher tolerance of those enzymes for bulky
substrates (Table 1, entry 5). In general, methyl substituents were not tolerated at all by
MAO-N D5 but high activities were obtained for MAO-N D9 and D11 with the para-methyl-
substituted compound 19. As previously described for the chloro- and bromo-substituted
compounds, ortho- and meta-methylated substrates (17 and 18) produced moderate to low
activities with both MAO-N D9 and D11 (Table 1, entry 6). Similarly, replacing the methyl
by a trifluoromethyl group at the 4-position gave excellent activity whereas a CF substituent
3
at C-2 yielded no activity (Table 1, entry 7). MBAs bearing a methoxy group at the 2- or the
4-position were efficiently oxidised by MAO-N D9 while no activity was displayed towards
3-methoxy MBA with any of the variants (Table 1, entry 8). Furthermore, hydroxy groups
on the aromatic ring did not lead to any oxidation of the substrates by the three variants
(Table 1, entry 9). Regarding the influence of a substituent on the nitrogen atom, N-methyl
substituted MBAs were included into the screening with compound 27 revealing lower
activities than model substrate 5 with the three variants and no activity in case of the tertiary
amine 28 (Table 1, entry 10), stressing the importance of the nitrogen atom for an efficient
interaction with the flavin co-factor.17
7

Table 1. Influence of substituents on the aromatic ring of α-methylbenzylamines MBAs
5-28 on the relative activities of MAO-N D5, D9 and D11.
Relative activity [%]
Entry Amine
R
H
D5
20
48
13
3
D9
100
175
30
123
19
2
D11
65
131
31
6
1
5
6
2-F
2
7
3-F
8
4-F
9
2-Cl
8
4
3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
3-Cl
2
1
4-Cl
1
246
3
145
1
2-Br
3-Br
4-Br
2-I
1
4
5
6
7
8
0
1
0
37
6
147
3
99
0
4-I
0
39
23
3
60
14
5
2-CH3
3-CH3
4-CH3
2-CF3
4-CF3
2-OCH3
3-OCH3
4-OCH3
3-OH
4-OH
H
1
0
0
211
1
149
0
0
6
125
59
0
80
29
0
13
0
0
111
0
58
0
0
9
0
3
2
3
14
0
0
10
CH3
0
0
Enzyme activities were measured as described in the Supporting Information in a total
volume of 200 µL and with 0.025 mg/mL of enzyme. The specific activity determined for
MAO-N D9 towards MBA 5 (0.91 U mg-1) was set to 100 % relative activity and used as
benchmark for MAO-N D5 and D11 and further compounds tested. Relative activity scale: •1
= low activity; 2-10 = moderate; 11-100 = good; >100 = excellent.
8

In general, di-halogenated substrates produced better activity in the case of fluorine compared
to chlorine with MAO-N D9 revealing the highest activities in most cases (Table 2, entry 1).
Given the high activity of MAO-N D9 towards fluorinated substrates, we wondered if the
presence of an additional fluorine substituent on the aromatic ring would increase the activity
of the enzymes for a given MBA (Table 2, entries 2-5). With MAO-N D9, a fluorine atom in
the para-position led in most cases to higher activities. To our delight, a 5- to 9-fold increase
in activity was observed for substrates 34 and 35 bearing halogens at C-2 (Table 2, entries 2
and 3). Fluorinated substrates 36 and 37 bearing methyl substituents respectively gave a 20%
decrease and a 25-fold increase in activity with MAO-N D9 (Table 2, entries 4 and 5).
Table 2. Influence of di-substitution of the aromatic ring of α-methylbenzylamines on
the relative activities of MAO-N D5, D9 and D11.
Relative activity [%]
Entry Amine
R1
R2
H
D5
48
10
4
D9
175
148
8
D11
131
25
8
6
2-F
2-F
3-F
2-F
2-Cl
2-Cl
29
4-F
5-F
6-F
4-Cl
5-Cl
H
30
1
31
18
0
48
1
1
32
33
7
0
2
0
9
8
19
96
3
4
2
2-Cl
2-Br
34
4-F
H
5
10
1
12
1
3
35
4-F
H
0
28
23
19
3
0
17
1
14
0
4
2-CH3
3-CH3
36
4-F
H
8
18
0
5
5
37
4-F
0
74
2
9

Enzyme activities were measured as described in the Supporting Information in a total
volume of 200 µL and with 0.025 mg/mL of enzyme. The specific activity determined for
MAO-N D9 towards MBA 5 (0.91 U mg-1) was set to 100 % relative activity and used as
benchmark for MAO-N D5 and D11 and further compounds tested. Relative activity scale: •1
= low activity; 2-10 = moderate; 11-100 = good; >100 = excellent.
We next turned our attention towards heteroaromatic rings and naphthyl derivatives and
found that MAO-N D5 and D9 were able to oxidise the 4-pyridyl substrate 38 as well as the
furan 41 and the thiophene 42 to a moderate extent (Table 3, entries 1 and 2). Methoxy-
substituted 2-naphthyl and 1,2,3,4-tetrahydro-2-naphthyl substrates 44 and 45 gave moderate
to good activities with MAO-N D9 and D11 but were unreactive with the D5 variant (Table
3, entry 3).
Table 3. Influence of the nature of substituents on the aromatic ring of α-
methylbenzylamines on the relative activities of MAO-N D5, D9 and D11.
Relative activity [%]
Entry Amine
D5
D9
D11
38
12
0
11
0
11
0
1
2
3
39
40
41
42
43
44
45
1
2
0
9
9
3
14
0
10
0
0
0
0
4
3
0
18
9
Enzyme activities were measured as described in the Supporting Information in a total
volume of 200 µL and with 0.025 mg/mL of enzyme.. The specific activity determined for
MAO-N D9 towards MBA 5 (0.91 U mg-1) was set to 100 % relative activity and used as
benchmark for MAO-N D5 and D11 and further compounds tested. Relative activity scale: •1
= low activity; 2-10 = moderate; 11-100 = good; >100 = excellent.
We then evaluated the influence of the α-substituent and tested benzylamine derivatives
46-49 (Table 4, entry 1). Ethyl and cyclopropyl groups, respectively, led to good and high
10

activity with MAO-N D9 but no oxidation could be observed with a benzylic moiety. Also,
α-trifluoromethyl benzylamine 48 was presumably too electron-deficient to be a suitable
substrate. The bulkier benzhydrylamine 50 was accepted by the three enzymes and MAO-N
D11 was the only one accepting 4-substituted benzhydrylamines 51-53 (Table 4, entry 2).14
However, the less hindered substrate 55 containing a furan ring was only moderately oxidised
by MAO-N D9. We also tested substrates 56-64 (structures detailed in supporting
information) that were synthesised in a single step via the Petasis multicomponent reaction
(Table 4, entry 3).18 Regardless of the substitution pattern, no activity could be measured in
all cases, probably due to the substrates’ hindered benzylic position and/or the phenolic
moiety.
Table 4. Influence of the nature of the α-substituent of α-methylbenzylamines on the
relative activities of MAO-N D5, D9 and D11.
Relative activity [%]
Entry Amine
D5
D9
D11
46
0
0
34
0
0
0
1
47/48
49
0
184
9
42
10
2
50
11
0
51
0
52
0
0
2
2
3
53
54
0
0
7
0
0
0
55
0
3
1
56/57-
64
0
0
0
Enzyme activities were measured as described in the Supporting Information in a total
volume of 200 µL and with 0.025 mg/mL of enzyme.. The specific activity determined for
MAO-N D9 towards MBA 5 (0.91 mU mg-1) was set to 100 % relative activity and used as
benchmark for MAO-N D5 and D11 and further compounds tested. Relative activity scale: •1
= low activity; 2-10 = moderate; 11-100 = good; >100 = excellent.
11

To further assess the substrate scope of the MAO-N variants, we then tested amines with
synthetic potential such as α-substituted allyl-19 and homoallylamines20 as well as
propargylamines21 (Table 5). 2-Phenyl allylamine 65 generally gave only moderate activities
with MAO-N D5 and D9 (Table 5, entry 1). para-Substituents were not tolerated by the
three variants and only a weak activity was measured for amine 68 using D5. Similarly,
homoallylamine 69 gave moderate activities with the three variants and a chlorine atom on
the aromatic ring was tolerated only when using MAO-N D9 and D11 (Table 5, entries 2
and 3).22 Regardless of the substitution pattern, low to moderate activities could be measured
for homoallylamines 73-76. However, we were pleased to find moderate to good activities of
MAO-N D9 and D11 for 4-pyridyl-, 2-naphthyl- and cinnamyl-substituted homoallylamines
77-79, further stressing the high substrate tolerance of these variants (Table 5, entry 5).
Propargylamines are a known class of MAO-inhibitors that, once oxidised, form a covalent
bond with the flavin co-factor via Michael addition (Figure 1, Rasagiline 2).23 We wondered
if the presence of a substituent at the - position or at the terminal carbon of the triple bond
would prevent the enzyme inhibition while allowing the oxidation of the propargylic position.
However, no oxidation could be detected for any of the five substrates tested (Table 5, entry
6) as additionally confirmed by inhibition studies (see Supporting information for details
and structures of 80-84).
12

Table 5. Relative activities of MAO-N D5, D9 and D11 towards allylamine,
homoallylamine and propargylamine derivatives.
Relative activity [%]
Entry Amine
D5
3
0
2
2
0
0
0
1
1
0
0
0
3
0
0
D9
4
D11
0
65
1
2
3
66/67
68
0
0
0
0
69
5
2
70
2
1
71
4
3
72
20
9
9
73
6
74
4
3
4
75
0
0
76
16
28
22
3
11
14
11
2
77
5
6
78
79
80-84
0
0
Enzyme activities were measured as described in the Supporting Information in a total
volume of 200 µL and with 0.025 mg/mL of enzyme.. The specific activity determined for
MAO-N D9 towards MBA 5 (0.91 U mg-1) was set to 100 % relative activity and used as
benchmark for MAO-N D5 and D11 and further compounds tested. Relative activity scale: •1
= low activity; 2-10 = moderate; 11-100 = good; >100 = excellent.
Next we investigated simple aliphatic amines to study whether the oxidation would take
place in the absence of an aromatic ring (Table 6). Gratifyingly, MAO-N D9 gave by far the
best activities for 2-amino linear alkyl chains (85-86) but also for bulky amines such 87, 88
and 89. tert-Butylamine 90 was only weakly oxidised by MAO-N D9 and α-trifluoromethyl
ethylamine 91 was incompatible with the three variants (Table 6, entry 1). Homoallylamine
92 gave good activity but its analogue 93 was less reactive possibly due to the longer alkyl
chain (Table 6, entry 2). We next tested substrates featuring an aromatic on the alkyl chain
and found that the homobenzylic substrate 94 was only weakly oxidised by the three variants.
A longer alkyl chain was well tolerated by MAO-N D9 but an additional fluorine substituent
13

at the ortho-position produced weaker activity (Table 6, entry 3). Finally, we evaluated
amino alcohols derivatives 97-100 and found only weak activities in all cases (Table 6, entry
4).
Table 6. Relative activities of MAO-N D5, D9 and D11 towards aliphatic amines.
Relative activity [%]
Entry Amine
D5
3
D9
67
87
69
202
50
1
D11
2
NH2
NH2
R
85
86
87
Alk
n
3
2
16
15
2
17
54
1
85
86
87
88
89
90
91
(Alk = Et)
94
95
96
(n = 1, R = 4-F)
(n = 2, R = H)
(n = 2, R = 2-F)
(Alk = Pent)
(Alk = i-Bu)
(Alk = Cy)
(Alk = adamantyl)
(Alk = t-Bu)
(Alk = CF3)
1
88
89
90
0
0
NH2
91
0
0
0
RO
n
92
2
27
3
19
2
2
3
93
0
97
98
99
(n = 1, R = H)
(n = 2, R = H)
(n = 1, R = Me)
NH2
94
0
0
0
Alk
95
3
22
16
1
3
100
(n = 1, R = Bn)
96
0
1
92
93
(Alk = i-Bu)
(Alk = hydrocinnamyl)
97/98
99
0
0
4
0
0
0
100
0
2
0
Enzyme activities were measured as described in the Supporting Information in a total
volume of 200 µL and with 0.025 mg/mL of enzyme.. The specific activity determined for
MAO-N D9 towards MBA 5 (0.91 U mg-1) was set to 100 % relative activity and used as
benchmark for MAO-N D5 and D11 and further compounds tested. Relative activity scale: •1
= low activity; 2-10 = moderate; 11-100 = good; >100 = excellent.
We finally investigated bicyclic compounds 101-104 bearing an amino group at the
benzylic position to gain more insights on the influence of a more rigid backbone on the
enzymatic activity (Table 7). The 4-membered ring 101 was preferentially accepted by
MAO-N D5 and D9 but not at all by the D11 variant. 1-Aminoindane 102 turned out to be an
excellent substrate for both MAO-N D9 and D11 together with THN 103 producing very
good activity with MAO-N D9. On the contrary, only moderate activities could be detected
14

for bigger ring-size (Table 7, entry 1). In the next step, a comprehensive set of 1-
aminoindanes and THNs as well as related heterocycles were screened against the three
MAO-N variants (Tables 7 and 8).
Similar to the MBA series (cf. Tables 1 and 2), 1-aminoindanes harboring a fluorine
substitution at C-4, C-5 and C-6 were particularly reactive with MAO-N D9 (Table 7, entry
2). A chlorine or a bromine atom at the 4-position gave moderate activity for MAO-N D9
whereas 6-chloro-1-aminoindane 109 turned out to be a good substrate for MAO-N D9
(Table 7, entries 3 and 4). A comparison between these results and the ones obtained for
meta-chloro MBA 10 (Table 1) clearly showed the positive influence of a rigid backbone on
the enzyme activity. A methyl substituent at the 5-position gave good activity with MAO-N
D9, whereas 6-methyl-1-aminoindane was only oxidized by MAO-N D11 (Table 7, entry 5).
The dimethoxy substituted compound 113 was only weakly oxidised by the three variants
(Table 7, entry 6).
In the THN series, a similar trend was observed with fluorinated compounds being
excellent substrates for MAO-N D9 (Table 7, entry 7). The presence of a methyl group at the
7-position yielded good activities but an additional methyl group at the 5-position generally
suppressed the activity of the three variants (Table 7, entry 8). 6-methoxy- and
6-cyano-THN 118 and 121 respectively gave good and excellent activity for both MAO-N
D9 and D11 (Table 7, entries 9 and 10). Noteworthy, the bicyclic substrates 101-121 turned
out to be highly compatible towards MAO-N D9 or D11 with 14 substrates giving good to
excellent activities with either of the two variants.
15

Table 7. Relative activities of MAO-N D5, D9 and D11 towards 1-aminoindane and
THN derivatives.
Relative activity [%]
Entry
1
Amine
D5
20
5
D9
12
199
92
6
D11
0
101 (n = 1)
NH2
102 (n = 2)
165
10
4
103 (n = 3)
1
n
104 (n = 4)
0
105 (R = 4-F)
106 (R = 5-F)
107 (R = 6-F)
108 (R = 4-Cl)
109 (R = 6-Cl)
110 (R = 4-Br)
111 (R = 5-Me)
112 (R = 6-Me)
113 (R = 5,6-OMe)
114 (R = 6-F)
115 (R = 7-F)
116 (R = 7-Me)
117 (R = 5,7-Me)
118 (R = 6-OMe)
119 (R = 7-OMe)
120 (R = 6,7-OMe)
121 (R = 6-CN)
0
101
242
143
3
1
2
1
75
39
0
0
0
3
4
5
6
7
0
103
4
13
0
0
23
0
61
0
20
20
1
1
2
0
155
90
39
0
107
25
26
0
18
20
0
8
0
49
2
11
1
9
0
0
1
0
10
0
103
32
Enzyme activities were measured as described in the Supporting Information in a total
volume of 200 µL and with 0.025 mg/mL of enzyme. The specific activity determined for
MAO-N D9 towards MBA 5 (0.91 U mg-1) was set to 100 % relative activity and used as
benchmark for MAO-N D5 and D11 and further compounds tested. Relative activity scale: •1
= low activity; 2-10 = moderate; 11-100 = good; >100 = excellent.
Next, the influence of a substituent on the saturated alkyl chain was studied. Weak to
moderate activities were measured for aminoindanes 122-124 and a moderate activity for this
group of substrates was found for MAO-N D9 towards 125 (Table 8, entry 1). Oxygenated
heterocycle 126 was oxidised by the D9 variant but low activity was found for diamine 127.
We next attempted to replace the aromatic ring with heterocycles such as thiophene 128 and
16

furan 129 for comparison with MBA derivatives 41 and 42 (Table 3, entry 2). To our
delight, a dramatic improvement in activity was measured (up to 25-fold) with the more rigid
backbones 128 and 129. However, the oxidation of non-benzylic positions in the bulky
compounds 131-132 remains a limitation of the current MAO-N variants (Table 8, entry 4).
Table 8. Relative activities of MAO-N D5, D9 and D11 towards 1-aminoindane and
THN derivatives.
Relative activity [%]
Entry Amine
D5
0
D9
1
D11
1
122
123
0
4
4
1
124
0
5
0
125
8
14
22
2
8
126
0
0
2
127
1
0
128
1
33
214
12
0
1
3
4
129
130
131
132
6
10
0
0
11
0
0
1
0
Enzyme activities were measured as described in the Supporting Information in a total
volume of 200 µL and with 0.025 mg/mL of enzyme.. The specific activity determined for
MAO-N D9 towards MBA 5 (0.91 U mg-1) was set to 100 % relative activity and used as
benchmark for MAO-N D5 and D11 and further compounds tested. Relative activity scale: •1
= low activity; 2-10 = moderate; 11-100 = good; >100 = excellent.
In summary, this extensive and wide-ranging investigation of the substrate scope of
MAO-N has highlighted the versatility of this biocatalyst for the synthesis of
enantiomerically enriched chiral amines. Although the variants D5, D9 and D11 were initially
selected for activity towards simple model chiral amines, screening against a much broader
range of pharmaceutically relevant amines has identified new applications beyond those
originally envisaged.
17

Optimisation of biotransformations with 1-aminoindane and THN
Based on the results from the activity measurements, 1-aminoindanes (102, 105-113) and
THNs (103, 114-121) were identified as new classes of substrates compatible with MAO-N
catalysed oxidations. The efficient biocatalytic amination of 1-indanone and 1-tetralone still
remains a challenge with transaminases24 which makes these particular substrates of interest
for enantioselective MAO-N oxidations. Accordingly, kinetic constants for 1-aminoindane
102 and THN 103 were measured with mutants D9 and D11 (Table 9).
Both MAO-N D9 and D11 were found to possess higher affinities (K ) towards 103 with
m
D9 possessing a 11-fold elevated Km as compared to D11. As MAO-N D9 displayed the
highest affinities (Km) and catalytic efficiencies (kcat/Km) towards both amine substrates when
compared to D11, we opted for this variant for subsequent experiments.
Table 9. Kinetic constants determined for MAO-N variants D9 and D11 towards 102
and 103.
Vmax
[U mg-1]
kcat
[s-1]
kcat/Km
[s-1mM-1]
Amine
Variant
Km [mM]
1.47
7.06
0.51
5.85
3.11
1.31
2.882
1.213
0.491
0.342
1.958
0.172
0.963
0.058
D9
D11
D9
0.53
0.369
D11
Conditions: 4-AAP/TBHBA/horseradish peroxidase-coupled assay (200 µL) in KPi buffer
(100 mM, pH 7.8) with substrates 102 and 103 dissolved in DMSO (co-solvent concentration
thoroughly adjusted to 6%, v/v), =510 nm, 30 °C (Figure S1-S4).
18

Using the MAO-N variants developed in our laboratories in presence of a non-selective
chemical reductant was previously found to constitute an efficient deracemisation process for
the preparation of enantiomerically pure primary, secondary and tertiary amines. In order to
find improved conditions broadly applicable for MAO-N catalysed deracemisation reactions,
we next revisited the influence of the reaction milieu. Primary amines are particularly
challenging substrates due to their propensity for competing hydrolysis of the intermediate
imine to the corresponding ketone, thereby reducing the enrichment of the desired amine
enantiomer. Therefore, we considered the use of more basic buffers in place of the commonly
used potassium phosphate buffer (KPi, 100 mM, pH 7.8) to produce a higher pH of the
reaction milieu shifting the equilibrium to the imine side. Ammonia buffer (250 mM) was
found to be not suitable in combination with MAO-N D9 as enzyme inhibition, evidenced by
drastically reduced conversions of 103, was observed (see Supporting Information). We
next focused on the use of Tris/HCl buffer (pH 9.0) and were pleased to find a remarkable
enrichment of (R)-103 (Scheme 1). Generally, increasing ammonia borane concentrations
were well tolerated by the MAO-N enzyme and gradually led to excellent and almost
quantitative concentrations of the enantiomerically pure amine 103.25 The excess borane
reduced the amount of ketone 133 to alcohol 134, both being present in the reaction mixture
at the end of the reaction.
19

Scheme 1. Yields of (R)-103 in deracemisation reactions of ( )-103 in potassium
phosphate (KPi, 100 mM, pH 7.8) or Tris/HCl buffer (pH 9.0).
5
5
4,5
4.5
4
4
3,5
3.5
3
3
22.,55
2
2
0
100
200
300
c NH3:BH3 [mM]
Tris/HCl (pH 9.0)
Tris/HCl (pH 9.0)
KKppi i(p(pHH77.8.8))
Reaction conditions: 5 mM 103 (100 mM in DMSO), 20-300 mM ammonia borane
(NH3:BH3; 1 M stock solution made up in the respective reaction buffer), 0.5 mg/mL MAO-N
D9, KPi buffer (pH 7.8) or Tris/HCl buffer (pH 9.0), 1 mL final volume, 14 mL headspace,
37 °C, 250 rpm, 24 h. Error bars correspond to three replicates. Concentration determined by
HPLC.
With a view to the practicality of MAO-N catalysed deracemisation reactions for operative
requirements of pharmaceutical industries, we next studied further reaction parameters
20

including the nature of the biocatalyst employed, the use of a bi-phasic reaction system with
organic solvents, the substrate load and the use of alternative borane complexes. THN 103
was chosen as model compound based on the good kinetic constants obtained for variant D9.
Whole or lyophilized cells turned out to be not applicable in this instance, delivering only
moderate enrichments towards the desired enantiomer . The addition of co-solvent such as
toluene, 2-methylpentane, cyclohexane or 2-methyltetrahydrofuran to the Tris/HCl buffer in a
1:1 ratio (v/v) was attempted to prevent extensive hydrolysis of the imine intermediate in
reactions with whole cells. Although the enzyme operated without any evidence of inhibition
with all solvents except 2-methyltetrahydrofuran, again no clear enrichment of the desired
enantiomer was achieved. Additionally, the presence of organic solvents favoured the
formation of side-products. These results suggested that the imine hydrolysis is faster than its
reduction by the borane complex in presence of cells. Based on this we conclude that the use
of whole or lyophilised cells in deracemisation reactions of primary amines is not a feasible
option although the procedure has proven successfully with the more stable imine
intermediates of secondary and tertiary amines in previous studies.14
In the next set of experiments, we tested various borane sources at different concentrations
and found that borane pyridine and borane tert-butylamine complexes gave levels of (R)-103
up to 73% (100 mM borane pyridine) and 71% (200 mM borane tert-butylamine) (Figure 2).
On the contrary, dimethylamine and morpholine borane complexes turned out to be less
efficient reductants regardless of the borane concentration. Overall, ammonia borane
performed best at all concentrations as compared to the other boranes tested.
Figure 2. Yields of (R)-103 in deracemisation reactions with different borane complexes.
21

5
,
4.5
4
4
3.5
,
3
3
2.5
2,5
2
2
0
100
200
300
c Borane complex [mM]
Borane pyridine
Borane dimethylamine
Borane morpholine
Borane tert-butylamine
Ammonia borane
Reaction conditions: 5 mM 103 (100 mM in DMSO), 20-300 mM borane complexes (3 M
stock solutions made up in Tris/HCl buffer or DMSO), 0.5 mg/mL MAO-N D9, Tris/HCl
buffer (pH 9.0), 1 mL final volume, 14 mL headspace, 37 °C, 250 rpm, 24 h. Concentration
determined by HPLC.
With this new protocol in hand, we tested a comprehensive set of racemic amines including
MBAs, THNs, 1-aminoindanes and amino-substituted heterocycles previously identified to
produce the best activities with MAO-N D9 (relative activities ranging from 17 to 417%,
Tables 1, 7 and 8). MAO-N D9 showed excellent enantioselectivity in all oxidation and
deracemisation reactions (Table 10, e.e. > 99%). The reference compound 5 was obtained in
a good yield (67%) while moderate yields were achieved with MBAs bearing a fluorine
substituent, regardless of its position on the aromatic ring (Table 10, entries 1 and 2).
Several electron withdrawing and donating groups were tolerated at the 4-position, affording
the corresponding enantioenriched MBAs in a 55-62% yield (Table 10, entry 3). As already
presented in Table 7, MAO-N D9 exhibited only a moderate activity towards the bicyclic
compound 101 containing a 4-membered ring. To our delight, an encouraging yield of 55%
was reached, emphasising the broad range of applicability of this methodology (Table 10,
entry 4). The same yield was obtained for (R)-1-aminoindane 102 but one of the best results
22

was achieved in the case of THN 103, which could be deracemised towards the (R)-
enantiomer in 79% yield. As observed for the MBAs, fluorinated 1-aminoindanes were
efficiently deracemised by the MAO-N D9/borane system (Table 10, entry 5). A good
enrichment was also achieved in the case of 5-methyl- and 6-chloro-1-aminoindane (Table
10, entry 6). 6- and 7-fluoro substituted THNs could be deracemised in an excellent 80% and
72% yield, respectively (Table 10, entry 7). THNs bearing a methoxy group or a cyano
moiety on C-6 were also compatible with the deracemisation reaction conditions employed
(Table 10, entry 8). Finally, heterocycles 126 and 128 could be obtained in an enantiopure
fashion in up to 71% yield via this deracemisation approach (Table 10, entries 9 and 10).
Table 10. Deracemisation of selected amines with MAO-N D9/BH3-NH3 system.a
Relative
Activity [%]
100
Entry
Amine
Conv. [%]b
e.e.[%]c
1
5 (R = H)
6 (R = 2-F)
67
58
60
57
62
61
61
55
> 99 (R)
> 99
175
2
7 (R = 3-F)
30
> 99
8 (R = 4-F)
123
> 99
11 (R = 4-Cl)
14 (R = 4-Br)
19 (R = 4-CH3)
24 (R = 4-OCH3)
246
> 99 (R)
> 99
147
3
211
> 99
111
> 99
101 (n = 1)
102 (n = 2)
103 (n = 3)
12
199
92
55
55
79
> 99
4
5
> 99 (R)
> 99 (R)
105 (R = 4-F)
106 (R = 5-F)
107 (R = 6-F)
109 (R = 6-Cl)
111 (R = 5-CH3)
101
242
143
103
61
63
71
58
74
63
> 99
> 99
> 99
> 99
> 99
6
7
114 (R = 6-F)
115 (R = 7-F)
155
90
80
72
> 99
> 99
23

118 (R = 6-OCH3)
121 (R = 6-CN)
49
69
56
> 99
> 99
8
9
103
126
128
22
33
71
62
> 99
> 99
10
aReaction Conditions: 5 mM amine substrate (100 mM in DMSO, 5 % v/v co-solvent), 100
mM NH3:BH3 (1 M in Tris/HCl buffer, pH 9.0), 0.5 mg/mL MAO-N D9, Tris/HCl buffer (pH
b
9.0), 1 mL final volume, 14 mL headspace, 37 °C, 250 rpm, 24 h. Conversion of the
c
respective enantiomer in deracemisation reactions. (R)-enantiomer remained (S-selective
oxidation by MAO-N D9). Each experiment was accomplished in three replicates. Relative
activity scale: •1 = low activity; 2-10 = moderate; 11-100 = good; >100 = excellent.
We finally turned our attention to higher substrate loads (10-50 mM) in order to gain an
improved amine/reductant ratio and chose compound 103 as a model substrate (Figure 3). In
all cases, MAO-D9 oxidised the (S)-enantiomer quantitatively (e.e. > 99% for the R-
enantiomer) and the best conversion was achieved at 30 mM amine concentration with 6.6
equiv. of ammonia borane (84% conversion, 22.2 mM).
Figure 3. Yields of (R)-103 in oxidation and deracemisation reactions with increasing
concentrations of ( )-103.
24

35
30
25
20
15
10
5
0
10
20
30
40
50
c (±)-103 [mM]
Oxidation
Deracemisation
Reaction conditions: 10-50 mM 103 (1 M in DMSO), 200 mM ammonia borane (NH :BH ;
3
3
1 M stock solution made up in Tris/HCl buffer), 0.5 mg/mL MAO-N D9, Tris/HCl buffer (pH
9.0), 1 mL final volume, 14 mL headspace, 37 °C, 250 rpm, 24 h. Concentration determined
by HPLC.
Conclusion
In conclusion, screening extensively various types of amines with the most versatile variants
of MAO-N in a high-throughput fashion using a spectrophotometric assay has allowed us to
identify new classes of substrates compatible with these enzymes. The MAO-N D9 variant
stands out as the most active enzyme variant and a thorough optimisation of the reaction
conditions was carried out in the case of THN 103. A selection of primary amines was then
deracemised using the optimised MAO-N D9/ammonia borane system. The good results
observed in the case of the moderately active amine 126 suggest that at least 56 out of the 132
amines tested are substrates and 22 of them form enantioenriched amines (Table 10) in good
yield using MAO-N D9. In view of a potential application on scale, we showed that amine
103 can be obtained in an enantiomerically pure fashion using a six-fold increased substrate
load and 6.6 equiv. of reductant. Research towards an application of MAO-N on large scale is
currently on going in our laboratories and will be reported in due course.
25

ASSOCIATED CONTENT
Supporting Information. This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*swagscha@its.jnj.com
*nicholas.turner@manchester.ac.uk
Present Addresses
† Susanne Herter, Siemens Healthcare Diagnostics Products GmbH, Emil-von-Behring-
Strasse 76, D-35041 Marburg, Germany
Author Contributions
The manuscript was written through contributions of all authors. All authors have given
approval to the final version of the manuscript.
ǂThese authors contributed equally to the work.
Funding Sources
The research leading to these results received support from the Innovative Medicines
Initiative Joint Undertaking under the grant agreement no. 115360 (Chemical manufacturing
methods for the 21st century pharmaceutical industries, CHEM21), resources of which are
composed of financial contribution from the European Union’s Seventh Framework Program
26

(FP7/2007-2013) and EFPIA companies’ in-kind contributions (to SH, FM, SW, FL, CB &
NJT), and The Royal Society Wolfson Merit Award (to NJT).
(1) (a) Bornscheuer, U.T.; Huisman, G.; Kazlauskas, R.; Lutz, S.; Moore, J.; Robins, K.
Nature 2012, 485, 185-194. (b) Pollard, D. J.; Woodley, J. M. Trends Biotechnol. 2007, 25,
66 - 73. (c) Breuer, M.; Ditrich, K.; Habicher, T.; Hauer, B.; Kesseler, M.; Sturmer, R.;
Zelinski, T. Angew. Chem. Int. Ed. 2004, 43, 788 – 824; (d) Aleku, G.A.; France, S.P.;
Mangas-Sanchez, J.; Montgomery, S.L.; Leipold, F.; Hussain, S.; Man, H.; Sharma, M;
Grogan, G.; Turner, N.J. Nature Chem., 2017, 9, doi:10.1038/nchem.2782.
(2) (a) Billerbeck, S.; Harle, J.; Panke, S. Curr. Opin. Biotechnol. 2013, 24, 1037 - 1043.
(b) Constable, D. J. C.; Dunn, P. J.; Hayler, J. D.; Humphrey, G. R.; Leazer, J. L.; Linderman,
R. J.; Lorenz, K.; Manley, J.; Pearlman, B. A.; Wells, A.; Zaks, A.; Zhang, T. Y. Green
Chem. 2007, 9, 411 - 420. (c) Wasilke, J.-C.; Obrey, J. C.; Baker, R. T.; Bazan, G. C. Chem.
Rev. 2005, 105, 1001 – 1020; (d) Turner, N.J.; O’Reilly, E. Nature Chem. Biol., 2013,9, 285-
288; Hönig, M.; Sondermann, P.; Turner, N.J.; Carreira, E.M. Angew. Chem. Int. Ed., 2017,
DOI: 10.1002/anie.201612462;
(3) Kim, K. S.; Zhang, L.; Williams, D.; Perez, H .L.; Stang, E.; Borzilleri, R. M.; Posy, S.;
Lei, M.; Chaudhry, C.; Emanuel, S.; Talbott R. Bioorg. Med. Chem. Lett. 2014, 24, 5022 -
5029.
(4) Hoy, S.; Keating, G. Drugs 2012, 72, 643 - 669.
(5) Lalonde, R. L.; Tenero, D. M.; Kazierad, D. J. Pharmacotherapy 1990, 10, 15 - 28.
(6) Thiel, O. R.; Bernard, C.; King, T.; Dilmeghani-Seran, M.; Bostick, T.; Larsen, R. D;
Faul, M. M. J. Org. Chem. 2008, 73, 3508 - 3515.
27

(7) Chiral Amine Synthesis: Methods, Developments and Applications; Nugent, T. C., Ed.;
Wiley-VCH: Weinheim, 2010.
(8) For recent reviews, see: (a) Simon, R. C.; Richter, N.; Busto, E.; Kroutil, W. ACS
Catal. 2014, 4, 129 - 143. (b) Paul, C. E.; Rodríguez-Mata, M.; Busto, E.; Lavandera, I.;
Gotor-Fernández, V.; Gotor, V.; García-Cerrada, S.; Mendiola, J.; de Frutos, O.; Collado, I.
Org. Process Res. Dev. 2014, 18, 788 - 792. (c) Savile, C. K.; Janey, J. M.; Mundorff, E. C.;
Moore, J. C.; Tam, S.; Jarvis, W. R.; Colbeck, J. C.; Krebber, A.; Fleitz, F. J.; Brands, J.;
Devine, P. N.; Huisman, G. W.; Hughes, G. J. Science 2010, 329, 305 - 309.
(9) (a) N. J. Turner, Curr. Opin. Chem. Biol. 2011, 15, 234 - 240. (b) de Lange, B.; Hyett,
D. J.; Maas, P. J. D.; Mink, D.; van Assema, F. B. J.; Sereing, N.; de Vries, A. H. M.; de
Vries, J. G. ChemCatChem 2011, 3, 289 - 292.
(10) (a) F. G. Mutti, T. Knaus, N. S. Scrutton, M. Breuer and N. J. Turner, Science 2015,
349, 1525-1529. (b) Abrahamson, M. J.; Vazquez-Figueroa, E.; Woodall, N. B.; Bommarius,
A. S. Angew. Chem. Int. Ed. 2012, 51, 3969 - 3972.
(11) (a) Scheller, P. N.; Fademrecht, S.; Hofelzer, S.; Pleiss, J.; Leipold, F.; Turner, N. J.;
Nestl, B. M.; Hauer, B. ChemBioChem 2014, 15, 2201 - 2204. (b) Leipold, F.; Hussain, S.;
Ghislieri, D.; Turner, N. J. ChemCatChem 2013, 5, 3505 - 3508. (c) Mitsukura, K.; Suzuki,
M.; Tada, K.; Yoshida, T.; Nagasawa, T. Org. Biomol. Chem. 2010, 8, 4533 - 4535. (d)
Wetzl, D.; Berrera, M.; Sandon, N.; Fishlock, D.; Ebeling, M.; Mueller, M.; Hanlon, S.;
Wirz, B.; Iding, H. ChemBioChem 2015, 16, 1749-1756.
(12) (a) Turner, N. J. Chem. Rev. 2011, 111, 4073 - 4087. (b) Höhne, M.; Bornscheuer, U.
T. ChemCatChem 2009, 1, 42 - 51. (c) Silverman, R. B. Acc. Chem. Res. 1995, 28, 335 - 342.
28

(13) (a) Carr, R.; Alexeeva, M.; Enright, A.; Eve, T. S. C.; Dawson, M. J.; Turner, N.
Angew. Chem. Int. Ed. 2003, 42, 4807 - 4810. (b) Alexeeva, M.; Enright, A.; Dawson, M. J.;
Mahmoudian, M.; Turner, N. J. Angew. Chem. Int. Ed. 2002, 41, 3177 - 3180.
(14) Ghislieri, D.; Green, A. P.; Pontini, M.; Willies, S. C.; Rowles, I.; Frank, A.; Grogan,
G.; Turner, N. J. J. Am. Chem. Soc. 2013, 135, 10863 - 10869.
(15) Ghislieri, D.; Houghton, D.; Green, A. P.; Willies, S. C.; Turner, N. J. ACS Catal.
2013, 3, 2869 - 2872.
(16) M. Braun, M.; Kim, J. M.; Schmid, R. D. Appl. Biochem. Biotechnol. 1992, 37, 594 -
598.
(17) Dunn, R. V.; Munro, A. W.; Turner, N. J.; Rigby, S. E. J.; Scrutton, N. S.
ChemBioChem 2010, 11, 1228 - 1231.
(18) Candeias, N. R.; Veiros, L. F.; Afonso, C. A. M.; Gois, P.M. P. Eur. J. Org. Chem.
2009, 1859 - 1863. See supporting information for structure details.
(19) Lafrance, M.; Roggen, M.; Carreira, E. M. Angew. Chem. Int. Ed. 2012, 51, 3470 -
3473.
(20) Sugiura, M.; Hirano, K.; Kobayashi, S. J. Am. Chem. Soc. 2004, 126, 7182 - 7183.
(21) Bariwal, J. B.; Ermolat’ev, D. S.; Van der Eycken, E. V. Chem. Eur. J. 2010, 16, 3281
- 3284.
(22) Traces of pinacol were present in some of the homoallylamine substrates (up to 10%)
but no interaction with the enzymes could be found by inhibition studies (See Supporting info
for more details, Figure S11)
29