A chemoenzymatic approach to enantiomerically pure amines using dynamic kinetic resolution: application to the synthesis of norsertraline

Article abstract of DOI:10.1002/chem.200802303

Racemization catalyst 5c and the enzyme Candida antarctica lipase B were combined in a one-pot dynamic kinetic resolution (DKR) of primary amines in which a wide range of amines were transformed to their corresponding amides in up to 95% isolated yield and > 99% ee. The DKR protocol was applicable with either isopropyl acetate or dibenzyl carbonate as the acyl donor. In the latter case, release of the free amine from the carbamate products was carried out under very mild conditions. The racemization of (S)-1-phenylethylamine with several different Ru catalysts was also evaluated. Catalyst 5c, of the Shvo type, was able to selectively racemize amines and was also compatible with the reaction conditions used for DKR. A racemization study of three different amines with varying electronic properties was also performed. Competitive racemization of a 1:1 mixture of the deuterated and non-deuterated amine was carried out with 5c and a primary kinetic isotope effect was observed for all three amines, providing support that the rate-determining step is β-hydride elimination. The chemoenzymatic DKR protocol was applied to the synthesis of norsertraline (16) by using a novel route starting from readily available 1,2,3,4-tetrahydro- 1-naphthy lamine (1o).

Full text of DOI:10.1002/chem.200802303

FULL PAPER
DOI: 10.1002/chem.200802303
A Chemoenzymatic Approach to Enantiomerically Pure Amines Using
Dynamic Kinetic Resolution: Application to the Synthesis of Norsertraline
Lisa K. Thalꢀn†, Dongbo Zhao, Jean-Baptiste Sortais, Jens Paetzold,
Christine Hoben, and Jan-E. Bꢁckvall*[a]
Abstract: Racemization catalyst 5c and
the enzyme Candida antarctica lipase B
were combined in a one-pot dynamic
kinetic resolution (DKR) of primary
of (S)-1-phenylethylamine with several
different Ru catalysts was also evaluat-
ed. Catalyst 5c, of the Shvo type, was
able to selectively racemize amines and
was also compatible with the reaction
conditions used for DKR. A racemiza-
tion study of three different amines
with varying electronic properties was
also performed. Competitive racemiza-
tion of a 1:1 mixture of the deuterated
and non-deuterated amine was carried
out with 5c and a primary kinetic iso-
tope effect was observed for all three
amines, providing support that the
rate-determining step is b-hydride
amines in which
a wide range of
amines were transformed to their cor-
responding amides in up to 95% isolat-
ed yield and >99% ee. The DKR pro-
tocol was applicable with either isopro-
pyl acetate or dibenzyl carbonate as
the acyl donor. In the latter case, re-
lease of the free amine from the carba-
mate products was carried out under
very mild conditions. The racemization
elimination.
The
chemoenzymatic
DKR protocol was applied to the syn-
thesis of norsertraline (16) by using a
novel route starting from readily avail-
able
1,2,3,4-tetrahydro-1-naphthyla-
Keywords: amines
catalysis dynamic
resolution · enzymes · ruthenium
·
asymmetric
mine (1o).
·
kinetic
Introduction
problems can be overcome by dynamic kinetic resolution
(DKR).[8] This is a process in which the unreacted starting
material is racemized during the KR allowing for a 100%
Chiral amines are of particular interest due to their poten-
tial usage in the pharmaceutical and agrochemical industries.
Several methods have been developed for the production of
chiral amines, such as hydrogenation of imines and enam-
ines,[1,2] alkylation of imines,[3] aminohydroxylation,[4] and re-
ductive amination.[5] Due to its simplicity, kinetic resolution
(KR) using enzymes continues to be one of the easiest
methods to produce chiral amines and entails the transfor-
mation (e.g., acylation) of one enantiomer in a racemic mix-
ture quicker than its counterpart into product.[6,7] However,
KR has two major drawbacks; a maximum of only half of
the starting material can be utilized, and the product has to
be separated from the remaining starting material. These
Scheme 1. Dynamic kinetic resolution.
[a] L. K. Thalꢀn, Dr. D. Zhao, Dr. J.-B. Sortais, Dr. J. Paetzold,
Dr. C. Hoben, Prof. Dr. J.-E. Bꢁckvall
theoretical yield as opposed to the mere 50% theoretical
yield with KR alone (Scheme 1).
Department of Organic Chemistry, Stockholm University
Arrhenius Laboratory, 10691 Stockholm (Sweden)
Fax : (+46)8-15-4908
DKR of secondary alcohols has been thoroughly investi-
gated by our group[9–13] and others,[14–21] whereas DKR of
amines has, until recently, been a less-explored area. The
first example of DKR of benzylic amines using a transition-
metal complex was reported by Reetz,[22] however, the reac-
E-mail: jeb@organ.su.se
[†] nꢀe Kanupp
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200802303.
Chem. Eur. J. 2009, 15, 3403 – 3410
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3403

tion proceeded with moderate efficiency and gave 64% of
the enantiomerically pure amide after eight days. Recently,
an efficient process for the DKR of benzylic and aliphatic
amines was developed by our group.[23] The process employs
a combined ruthenium and enzyme catalysis to provide the
Scheme 2. Dynamic kinetic resolution of primary amines.
corresponding amides in high yields and high enantiomeric
excess from a racemic mixture (Scheme 2). Independent
work by the Jacobs–de Vos group showed that a palladium/
enzyme catalyst combination can be used for practical DKR
of benzylic amines.[24–26] Subsequent work has provided addi-
tional procedures for chemoenzymatic DKR of amines in
which various racemization methods are combined with en-
zymatic resolution.[27–33]
Figure 1. Racemization catalysts.
As mentioned above, a highly efficient process for the
DKR of amines using a Ru catalyst and a lipase to provide
the corresponding amides in high yields and high enantio-
meric excess has been developed within our group. Here we
report on the development of this process. An investigation
of several Ru catalysts for the racemization of (S)-1-phenyl-
ethylamine was performed. It was found that catalyst 5c, of
the Shvo type, was able to selectively racemize amines and
could be effectively combined with a lipase for DKR. Inves-
tigations of the racemization mechanism, the substrate
scope for DKR, and the use of various acyl donors were car-
ried out. The protocol was found to be general and applica-
ble with the acyl donors isopropyl acetate and dibenzyl car-
bonate. We have also developed a new route to norsertra-
line (16) starting from readily available 1,2,3,4-tetrahydro-1-
naphthylamine (1o) utilizing DKR as one of the key steps.
Preliminary reports on some aspects of this work were pre-
sented recently.[23,31]
Table 1. Racemization of (S)-1-phenylethylamine (1a).[a]
Entry
Catalyst
T [8C]
ee [%][b]
Selectivity for 1a [%][c]
1[d]
2[d]
3
3
4
100
100
100
100
100
100
90
90
80
90
100
90
85
46
53
36
55
25
55
55
78
34
0
>98
>98
82
25
95
5a
5b
5c
5c
5c
5c
5c
6
4
5
6[e]
7[e]
8[f]
90
95
>98
>98
>98
>98
>98
9[f]
10[g]
11[g,h]
12[g,h]
6
7
94
[a] Conditions: 0.50 mmol (S)-phenylethylamine, 0.02 mmol Ru source,
2 mL of toluene, 30 mL of pentadecane as internal standard, 24 h. [b] De-
termined as corresponding acetamides using chiral GC. [c] Determined
by GC. [d] Activated with 0.02 mmol KOtBu. [e] With 0.04 mmol 5c.
[f] With 0.04 mmol 5c, 8 mL of toluene. [g] Conditions: 1 mmol of (S)-
phenylethylamine, 0.01 mmol of catalyst, 0.02 mmol of tBuOK, 60 mL of
pentadecane as internal standard, 1 mL of toluene. [h] 19 h.
Results and Discussion
Racemization study: Several Ru catalysts (3–7, Figure 1)
were tested for their ability to racemize (S)-1-phenylethyla-
mine (1a, 99% ee) and the loss of optical purity over time
was monitored by GC. Ru catalysts 3, 4, 6, and 7 are base
activated and were chosen because catalysts 3 and 4 have
given high racemization rates for secondary alcohols,[34,35]
and catalysts 6 and 7 are commonly used in hydrogen-trans-
fer reactions.[36,37] Catalysts 3 and 7 showed very little activi-
ty for the racemization of (S)-1a, resulting in enantiomeric
excess values of 85 and 94%, respectively (Table 1, entries 1
and 12). Catalyst 4 showed modest activity resulting in an
enantiomeric excess of 46% after 24 h (Table 1, entry 2). In-
terestingly, catalyst 6, which is structurally similar to catalyst
7, gave full racemization (0% ee) after only 19 h at 1008C
(Table 1, entry 11). Shvoꢃs catalyst, 5a, and the electron-rich
p-methoxy substituted derivative 5c showed modest activity,
giving reasonable enantiomeric excess values of 53 and
55%, respectively, after 24 h (Table 1, entries 3 and 5). A
lower enantiomeric excess of 36% was achieved with the
electron-deficient p-fluoro substituted derivative 5b
(Table 1, entry 4). The molecular structure of 5c was con-
firmed by X-ray diffraction analysis (Figure 2).
Notably, catalysts 3, 4, 6, and 7 showed high selectivity
(>98%, Table 1), whereas catalysts 5a–c showed lower se-
3404
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3403 – 3410

Dynamic Kinetic Resolution
FULL PAPER
faster racemization, and that
the re-addition step was slowed,
extending the lifetime of the
free imine resulting in an in-
crease in side-product (8 and 9)
formation. Relative to 5b, the
electron-rich catalyst, 5c, gave
a slower yet much more selec-
tive racemization. We consider
that, although the dehydrogena-
tion step is slowed, the rate of
hydrogen re-addition is acceler-
ated, thus reducing the lifetime
of the free imine.
¯
Figure 2. Molecular structure of compound 5c. X-ray crystal data for Ru2C70O14H57: triclinic, P1 (no. 2), a=
14.848(1) b=15.3882(9) c=16.2590(7) ꢄ, a=101.508(5) b=106.972(5) g=116.302(6)8, V=2941.8(3) ꢄ3, Z=2,
T=100(2) K, 1calcd =1.495 gcmÀ3, R
ACTHNUTRGNE(UNG int)=0.045, R(F)=0.0426 for 788 parameters and 8904 observed reflec-
The amount of catalyst, 5c,
was doubled in an attempt to
make the racemization process
tions (I>2s(I)). All non-hydrogen atoms were refined with anisotropic displacement parameters. All hydrogen
atoms, except the hydride between the two Ru atoms, were geometrically positioned and refined in riding posi-
tion. The H atom between the Ru atoms was refined isotropically. CCDC 718665 contains the supplementary
crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallo-
graphic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
more
efficient
(Table 1,
entry 6). The enantiomeric-
excess value dropped as expect-
ed, however, the degree of side-
product (8 and 9) formation
lectivity (Table 1, entries 3–5). Condensation products 8 and
9 were observed in the less-selective racemizations and are
probably due to a slower re-addition rate of hydrogen to the
intermediate imine. The racemization mechanism consists of
a combination of dehydrogenation of the chiral amine and
re-addition of hydrogen to the imine (Scheme 3).[38–43] The
electron-deficient catalyst 5b gave both the fastest and the
least-selective racemization. This indicates that the b-hy-
dride elimination step was accelerated, giving an overall
also increased. This problem was solved by reducing the re-
action temperature and by dilution (Table 1, entries 7 and
8). In principle, the racemization is also applicable at 808C;
however, the reaction rate is too slow to be practical for
DKR.
Kinetic isotope effect: Three different amines (1a, 1 f, and
1i) with varying electronic properties from electron-deficient
to electron-rich were compared.
The racemization of (S)-[H]-1/
(S)-[D]-1 with 5c was carried
out in a competitive manner
and the ratio of the initial rate
of formation of (R)-[H]-1 and
(R)-[D]-1 was determined by a
GC-MS method in which
a
chiral GC was coupled to a
mass spectrometer (Scheme 4).
In this way, the isotope effect
could be determined.
The results show that the
electron-deficient amine (S)-1 f
has the largest isotope effect,
kH/kD =3.53. This was also the
slowest-reacting amine. Amines
(S)-1a and (S)-1i showed much
lower isotope effects with kH/kD
values of 1.70 and 1.73, respec-
tively. These amines also react-
ed faster than the electron-defi-
cient substrate (S)-1 f, with (S)-
1i being the fastest. The order
of reaction rate (S)-1 f<(S)-
1a<(S)-1i is consistent with a
Scheme 3. Catalytic cycle of catalysts 5a–c.
rate-determining
b-hydride
Chem. Eur. J. 2009, 15, 3403 – 3410
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3405

J.-E. Bꢁckvall et al.
Scheme 4. Racemization of (S)-[H]-1/(S)-[D]-1 with 5c.
elimination because hydrogen abstraction from the b posi-
tion to give an imine should be slower for an electron-defi-
cient substrate than for an electron-rich one.
The isotope effect observed for the electron-deficient
amine (S)-1 f (kH/kD =3.53, Table 2) confirms a rate-deter-
mining b-hydride elimination for this amine. Also, for the
~
&
Figure 3. Racemization of (S)-[H]-1a ( ) and (S)-[D]-1a ( ) by 5c car-
ried out as separate reactions in a non-competitive manner. Conditions:
0.50 mmol of (S)-amine, 0.02 mmol 5c, 5 mL of toluene, 30 mL of penta-
decane as internal standard. The reaction mixtures were analyzed by GC-
MS (chiral GC).
as a small isotope effect (Ru-H/Ru-D) was observed for
ruthenium hydride addition to imines with 10a.[38,41,43]
Therefore, the large isotope effect observed herein is in ac-
cordance with the rate-determining step being b-hydride
elimination for (S)-1a.
Table 2. Kinetic isotope effects for the racemization of amines by 5c.[a]
[c]
Entry
Compound
R
t[b] [h]
kH/kD
1
2
3
(S)-1 f
(S)-1a
(S)-1i
CF3
H
OMe
44
20
7
3.53Æ0.62
1.70Æ0.12
1.73Æ0.30
Dynamic kinetic resolution: After optimization of the race-
mization protocol, application of DKR using Candida ant-
arctica lipase B (CALB, Novozym 435) as the enzyme and
isopropyl acetate as the acylating agent was attempted. The
more-selective racemization catalysts 3, 4, and 6, which are
activated by base, proved ineffective in the DKR of amines.
We then tried the combination of CALB with racemiza-
tion catalyst 5c in the presence of isopropyl acetate at 908C
for three days, however, moderate yields only slightly above
those of a KR were observed. Surprisingly, after sodium car-
bonate was added to the reaction mixture, an efficient DKR
process took place (Table 3, entry 2). The effect of sodium
carbonate is not clear, but could be due to the neutralization
of traces of acids that originate from the polyacrylate sup-
port, from enzymatic hydrolysis of the acyl donor, or from
the enzyme itself.
[a] Conditions: 0.50 mmol of a 1:1 mixture of the deuterated and nondeu-
terated (S)-amine, 0.02 mmol 5c, 5 mL of toluene, 30 mL of pentadecane
as internal standard. The reaction mixtures were analyzed by GC-MS
(chiral GC). [b] Time to reach 7% formation of (R)-1 (i.e., ee=86%).
[c] At less than 7% formation of (R)-1.
other amines, the rate-determining step is most likely b-hy-
dride elimination and the lower isotope effect (kH/kD ꢀ1.7)
may be explained by the competitive experiment, which
leads to an enrichment of the coordinated deuterated amine
due to slower b hydrogen-bond cleavage (see below).
The racemization of (S)-[H]-1a and (S)-[D]-1a was, there-
fore, performed in a non-competitive manner, in which the
individual racemizations were carried out in separate reac-
tion vessels. An initiation period was observed for (S)-[D]-
1a, which is most likely due to activation of the dimer into
two monomers and coordination of the amine to Ru mono-
mer 11.[44] The initial rate of each reaction was determined
and the kH/kD was found to be 4.2Æ1.0 (Figure 3). This is
much larger than the kH/kD of 1.7 observed for (S)-1a
(Table 2, entry 1) when the racemization was carried out in
a competitive manner.
Substrate variability: The scope of the reaction was studied
by applying the DKR conditions (Ru catalyst 5c, CALB,
isopropyl acetate, and sodium carbonate in toluene) to a
wide array of substrates (1a–r). The conditions were general
and amides 2a–r were obtained in high yields and high en-
antiomeric excess (Scheme 2, Table 3).
The difference in kH/kD in the competitive and non-com-
petitive methods indicates that the amine is coordinated to
the catalyst and that dissociation of the amine from the cata-
lyst is slower than b-hydride elimination. This would lead to
Benzylic amines could be converted to the corresponding
optically pure amides in yields of 69–95% with enantiomeric
excess values of 95–99%. Several functional groups, such as
fluorine, bromine, ether, nitrile, nitrate, and trifluoromethyl,
were tolerated (entries 6, 8, 10, 12, and, 15–17). Aliphatic
primary amines, which are not accessible through asymmet-
ric reduction of the corresponding imine,[45–48] were also ob-
tained in high yields with high enantiomeric excess values of
93–97% (entries 23–25). The heteroatomic thiophene deriv-
atives were also tested and gave high yields and an enantio-
meric excess of 99% (entries 4 and 14). Slightly more steri-
a
slower conversion of (S)-[H]-1a in the competitive
method than in the non-competitive method; the deuterated
amine (S)-[D]-1a will stay coordinated to the catalyst longer
than (S)-[H]-1a due to slow b-hydride elimination and thus
prevent (S)-[H]-1a from coordinating and reacting. Previ-
À
ously, a large isotope effect was observed for C H bond
cleavage (b-hydride elimination) of amines with 11a, where-
3406
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3403 – 3410

Dynamic Kinetic Resolution
FULL PAPER
Table 3. Dynamic kinetic resolution with isopropyl acetate as the acyl-group donor.[a]
cally hindered substrates were
also compatible with the
enzyme (entries 4, 6, 19–22, and
24) and gave high yields and
high enantiomeric excesses. In
some cases, the reaction time
was extended due to the slow
rate of acylation (entries 20 and
21).
Entry
Catalyst
Substrate
Product
Yield [%]
ee [%]
1
2
5a
5c
1a
1a
2a
2a
85
90
98
98
3
4
5a
5c
1b
1b
2b
2b
87
88
99
>99
5
6
5a
5c
1c
1c
2c
2c
78
99
>99
84[b]
Even the less-selective cata-
lysts 5a and 5b could be used
for DKR, giving (R)-amides
2a–2 f in 70–87% yield and 98–
99% ee. Catalyst 5a was tested
on a series of substrates and in
general gave a slightly lower
yield than when 5c was used
(Table 3, entries 1, 3, 5, 7, 9,
and 13). The difference was
more pronounced for electron-
rich substrates, which have a
slower re-addition step. It was
possible to use catalyst 5b as
the racemization catalyst with
the electron-deficient substrate
1 f giving the amide in 77%
yield and 99% ee after three
days (Table 3, entry 11).
7
8
5a
5c
1d
1d
2d
2d
73[b]
73[b]
99[c]
99[c]
9
10
5a
5c
1e
1e
2e
2e
70
69
>99
>99
11
12
5b
5c
1 f
1 f
2 f
2 f
77
91
99
99[d,e]
13
14
5a
5c
1g
1g
2g
2g
72
85
99
99
15
16
17
5c
5c
5c
1h
1i
2h
2i
78
95
84
99
99[d]
98
Hydrolysis of the in situ-
formed imine was found to be a
minor problem. The only by-
product detected was the corre-
1j
2j
sponding
N-isopropylamine.
18
19
20
21
5c
5c
5c
5c
1k
1l
2k
2l
69
92
85
80
99[d]
95[d]
However, under dry conditions,
the formation of this side prod-
uct was kept below 5%.
Alternate acyl donor: Isopropyl
acetate was used as the acyl
donor in all of the DKR reac-
tions discussed so far (Table 3).
The conditions for cleaving off
the acetyl group from the
amide products are quite harsh,
thus it is desirable to develop a
system with an acyl donor that
leads to products from which
the acyl group can be cleaved
off under mild conditions and
possibly orthogonally to other
protecting groups.
1m
1n
2m
2n
99[d,e]
>99[d,e]
22
23
24
5c
5c
5c
1o
1p
1q
2o
2p
2q
70
99
85[b]
78
93
95[f]
25
5c
1r
2r
88
97
Previously, we found that
acyl donors, such as ethyl and
[a] Conditions: 0.50 mmol amine, 0.02 mmol 5c, 400 mL isopropyl acetate, 20 mg Novozyme-435, 20 mg isopropyl trifluoroacetate and
Na2CO3, 8 mL toluene, 908C, 3 days; isolated yields with >95% purity. [b] Conversion (%) determined by
GC. [c] 1008C. [d] 30 mg CALB, 30 mg Na2CO3. [e] 5 days. [f] 15 mg enzyme, 15 mg Na2CO3.
isopropyl chloroacetate, which
transfer more electron-deficient
acyl groups, lead to a significant
Chem. Eur. J. 2009, 15, 3403 – 3410
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3407

J.-E. Bꢁckvall et al.
Table 4. Dynamic kinetic resolution with dibenzyl carbonate as the acyl
donor.[a]
amount of non-catalyzed background reaction under the re-
action conditions required for DKR.[49] Thus, only a low
level of enantioselectivity could be obtained with these acyl
donors. We thus focused our attention on acyl donors that
lead to carbamates, as there is well-documented methodolo-
gy to release the free amine from these compounds under
mild conditions.[50] In particular, the protection of amines as
allyloxy, benzyloxy, or tert-butyloxy carbamates is popular
because deprotection requires either mild Pd catalysis or
weak acid catalysis. For the enzyme-catalyzed reaction of
amines to give carbamates we considered diallyl, dibenzyl,
and di-tert-butyl carbonates, 12–14, as appropriate acyl
donors because they gave no non-catalyzed background re-
action on stirring with 1-phenylethylamine (1a) in toluene
at 908C for 24 h (Scheme 5). Kinetic resolution of amine 1a
was carried out with CALB using carbonates 12–14 as acyl
donors. Diallyl carbonate, 12, gave only a moderate enantio-
selectivity (E=9),[51] whereas dibenzyl carbonate, 13, gave a
highly enantioselective reaction (E=156). The di-tert-butyl
carbonate, 14, was completely inactive in the enzymatic re-
action.
Entry Substrate
Product
Yield[b] [%] ee[c] [%]
1
2
1a
1h
15a 90
93
98
15b 95
3
4
1i
1j
15c 74[d]
97
99
15d 72[d]
5
6
7
1l
15e 64
15 f 89
15g 92
99
90
96
1p
1r
8
1s
15h 60[d]
99
[a] Conditions: 0.50 mmol amine, 0.02 mmol 5c, 1.25 mmol dibenzylcar-
bonate, 20 mg Novozyme-435, 20 mg Na2CO3, 8 mL toluene, 908C, 3 d.
[b] Isolated yields unless otherwise noted. [c] Enantiomeric excess deter-
mined by chiral HPLC (chiralcel ODH). [d] Determined by GC.
Scheme 5. CALB-catalyzed kinetic resolution of 1a with carbonates 12–
14.
Thus, dibenzyl carbonate, 13, was chosen as the acyl
donor for the DKR reactions. We were pleased to find that
carbonate 13 was compatible with the racemization catalyst
5c. Reaction of racemic 1a and carbonate 13 in the presence
of Ru catalyst 5c and CALB in toluene afforded the benzyl
carbamate, 15a, in 90% yield and 93% ee (Table 4, entry 1).
p-Substituted phenylethylamines were reacted under the
DKR conditions and gave the corresponding carbamates
15b–d in good to high yields with high enantiomeric excess
values (97–99% ee, entries 2–4). In addition, aliphatic
amines worked well in the chemoenzymatic DKR reaction
to give the corresponding carbamates in good to high enan-
tiomeric excess (entries 6–8).
The benzyloxycarbonyl group of products 15a–h can be
removed easily through hydrogenolytic cleavage.[50] Depro-
tection using 1 mol% palladium on charcoal (10%) under a
hydrogen atmosphere was demonstrated with carbamate
15a in which liberation of the amine was complete within
half an hour at room temperature. The amine was obtained
Scheme 6. Deprotection of carbamates obtained from DKR.
Synthesis of norsertraline: Sertraline (17) is marketed by
Pfizer as Zoloft for the treatment of depression as a selec-
tive serotonin reuptake inhibitor (SSRI). Norsertraline (16)
is structurally similar to 17 and is currently in clinical trials
at Sepracor for the treatment of central nervous system
(CNS) disorders. Researchers at Sepracor recently presented
a synthetic approach for 16.[52]
in
a quantitative yield and with full retention of ee
(Scheme 6). The H2 reaction was very clean giving only CO2
and toluene as side products. If the reaction was run for a
longer period of time or if a larger amount of catalyst was
employed some racemization of the product was observed.
3408
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3403 – 3410

Dynamic Kinetic Resolution
FULL PAPER
À18.5 ppm (s, 1H); 13C NMR (100 MHz, CDCl3): d=201.3, 158.9, 158.1,
We have developed a novel route to 16 starting from
readily available 1,2,3,4-tetrahydro-1-naphthylamine (1o)
utilizing DKR as one of the key steps. First, DKR was ap-
plied by using the standard procedure to obtain (R)-2o in
70% yield and 99% ee. Oxidation at the C-4 position was
affected with KMnO4 to provide (R)-18. The enolate was
then formed and trapped with N-phenyl-bis(trifluorometha-
nesulfonimide) to give (R)-19 that was subsequently reacted
with 3,4-dichlorophenylboronic acid in a Suzuki coupling to
provide (R)-20 in 96% yield and 99% ee. Hydrogenation
was carried out with trans selectivity using the Crabtree cat-
alyst to afford (1R,4S)-21 in 95% yield and with a trans/cis
ratio of >99:1. The acetamide was subsequently deprotected
under acidic conditions to provide norsertraline, (1R,4S)-16,
in 95% yield with full retention of ee and dr (99% ee, and
trans/cis >99:1, Scheme 7).
153.7, 133.2, 132.3, 122.9, 122.6, 113.0, 112.9, 102.1, 87.5, 55.0, 54.9 ppm.
General procedure for racemization:
A flame-dried 20-mL reaction
vessel was charged with the ruthenium precursor (0.02 mmol Ru); the
vessel was closed, evacuated, and backfilled with argon three times. Tolu-
ene (2 mL), 1-(S)-phenylethyl amine ((S)-1a, 0.50 mmol, 65 mL), and pen-
tadecane (30 mL) were added subsequently by using a syringe. The mix-
ture was stirred at 1008C for 24 h. After the reaction, samples of 200 mL
were taken, transferred to a GC-vial, and diluted with toluene. The
amine 1a was transformed to the corresponding amide 2a by addition of
two drops of acetic anhydride and two drops of triethylamine to the GC
vial. Chiral GC analysis: injector 2508C program: 1258C/4 min/1408C/
38CminÀ1/2008C/408CminÀ1, 5 min, tS =10.39 min, tR =10.62 min.
General procedure for dynamic kinetic resolution with isopropyl acetate:
Synthesis of N-((R)-1-thiophen-2-yl-propyl)acetamide (2b):
A flame-
dried 20-mL reaction vessel was charged with Novozym 435 (20 mg),
Na2CO3 (0.20 mmol, 20 mg), and Ru complex 5c (26.5 mg, 0.02 mmol),
5a (21.7 mg, 0.02 mol), or 5b (24.6 mg, 0.02 mol), respectively. The vessel
was closed, evacuated, and backfilled with argon three times. Toluene
(8 mL), 1-thiophen-2-yl-propyl amine
(1b, 0.50 mmol, 70 mL), isopropyl ace-
tate (3.50 mmol, 400 mL), and pentade-
cane (30 mL) were added subsequently
by using syringe under argon and the
reaction mixture was stirred at 908C
for three days. The reaction was
cooled to RT, the solids were removed
by filtration and then subsequently
washed with dichloromethane (3ꢅ
20 mL). The solvent was removed in
vacuo and the crude product was puri-
fied by column chromatography
(SiO2, pentane/ethyl acetate 1:4)
yielding 2b as a white solid (81 mg,
0.44 mmol, 88%, 99% ee). The spec-
troscopic data of the product were in
agreement with those reported in the
literature.[53] Chiral GC-analysis: injec-
tor 2508C program: 1008C/5 min/
1558C/38CminÀ1
,
5 min/2008C/
Scheme 7. Synthetic route to norsertraline (COD=1,5-cyclooctadiene).
208CminÀ1, 5 min, tS =24.43 min, tR =
24.90 min;
CHCl3).
[a]2D7 =+188.9
(c=1.0,
General procedure for dynamic kinetic
resolution with dibenzyl carbonate: Synthesis of benzyl 1-phenylethylcar-
bamate (13a): A flame-dried 20-mL reaction vessel was charged with
complex 5c (26.5 mg, 0.02 mmol), Novozym 435 (20 mg), dibenzylcarbon-
ate (1.25 mmol, 303 mg), and Na2CO3 (0.20 mmol, 20 mg). The vessel was
closed, evacuated, and backfilled with argon three times. Dry toluene
(5 mL), 1-phenylethylamine (1a, 0.50 mmol, 65 mL), and pentadecane
(30 mL) were added subsequently by using a syringe under argon. The re-
action mixture was stirred at 908C. After three days the reaction was
cooled to RT and was filtered through a sintered-glass funnel. The sol-
vents were removed in vacuo and the crude product was purified by
column chromatography (SiO2, petroleum ether/ethyl acetate, varying
mixtures) yielding of 13a as a white solid (114.9 mg, 90%, 93% ee). The
spectroscopic data of the product were in agreement with those reported
in the literature.[54] Chiral GC-analysis: injector 2508C program: 1008C/
Conclusion
We have developed a highly efficient protocol for DKR of
amines that allows a variety of functionalized primary
amines to be transformed into one enantiomer in high yield
and with high enantioselectivity. The protocol is also practi-
cal by using dibenzyl carbonate as the acyl donor, which
allows release of the free amine from the carbamate prod-
ucts under very mild conditions. We have also developed a
new route to norsertraline (16) starting from readily avail-
able 1,2,3,4-tetrahydro-1-naphthylamine (1o) utilizing DKR
as one of the stereo-defining steps.
5 min/1558C/38CminÀ1
,
5 min/2008C/208CminÀ1
,
12 min,
tstandard =
20.8 min, tproduct =36.9 min; HPLC (Chiralcel ODH, isohexane/2-propanol
95:5): tR =25.8 min, tS =31.2 min.
Experimental Section
General methods and additional procedures are provided in the Support-
ing Information.
Spectroscopic data for 5c (corrected data): 1H NMR (400.1 MHz,
CDCl3): d=7.04 (d, J=8.0 Hz, 8H), 6.92 (d, J=8.0 Hz, 8H), 6.56 (d, J=
8.0 Hz, 8H), 6.53 (d, J=8.0 Hz, 8H), 3.78 (s, 12H), 3.69 (s, 12H),
Acknowledgements
Lars Eriksson is gratefully acknowledged for his X-ray crystal-structure
contribution. The Swedish Research Council, the Swedish Foundation for
Chem. Eur. J. 2009, 15, 3403 – 3410
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3409

J.-E. Bꢁckvall et al.
Strategic Research, the K&A Wallenberg Foundation, Lavoisier grant of
France (J.B.S.), and the German Research Foundation (DFG: J.P. and
C.E.H.) are gratefully acknowledged for financial support.
[26] A. Parvulescu, D. De Vos, P. Jacobs, Chem. Commun. 2005, 5307–
5309.
[27] S. Gastaldi, S. Escoubet, N. Vanthuyne, G. Gil, M. P. Bertrand, Org.
Lett. 2007, 9, 837–839.
[28] M. J. Kim, W. H. Kim, K. Han, Y. K. Choi, J. Park, Org. Lett. 2007,
9, 1157–1159.
[29] K. Han, J. Park, M.-J. Kim, J. Org. Chem. 2008, 73, 4302–4304.
[30] M. Stirling, J. Blacker, M. I. Page, Tetrahedron Lett. 2007, 48, 1247–
1250.
[31] C. E. Hoben, L. Kanupp, J.-E. Bꢁckvall, Tetrahedron Lett. 2008, 49,
977–979.
[32] C. J. Dunsmore, R. Carr, T. Fleming, N. J. Turner, J. Am. Chem. Soc.
2006, 128, 2224–2225.
[33] M. A. J. Veld, K. Hult, A. R. A. Palmans, E. W. Meijer, Eur. J. Org.
Chem. 2007, 5416–5421.
[34] O. Pꢊmies, J.-E. Bꢁckvall, Chem. Rev. 2003, 103, 3247–3261.
[35] M.-J. Kim, Y. Ahn, J. Park, Curr. Opin. Biotechnol. 2002, 13, 578–
587.
[36] W. Baratta, M. Bosco, G. Chelucci, A. Del Zotto, K. Siega, M. To-
niutti, E. Zangrando, P. Rigo, Organometallics 2006, 25, 4611–4620.
[37] W. Baratta, G. Chelucci, S. Gladiali, K. Siega, M. Toniutti, M. Za-
nette, E. Zangrando, P. Rigo, Angew. Chem. 2005, 117, 6370–6375;
Angew. Chem. Int. Ed. 2005, 44, 6214–6219.
[1] H. U. Blaser, F. Spindler in Hydrogenation of Imino Groups, Vol. 1
(Eds.: E. N. Jacobsen, A. Pfaltz, H. Yamamoto), Springer, Berlin,
1999, pp. 247–265.
[2] B. Singaram, C. T. Goralski in The Reduction of Imines and Enam-
ines with Transition Metal Hydrides, Vol. 2 (Eds.: M. Beller, C.
Bolm), Wiley-VCH, Weinheim, 1998, pp. 147–154.
[3] S. E. Denmark, O. J. C. Nicaise in Alkylation of Imino Groups,
Vol. 1 (Eds.: E. N. Jacobsen, A. Pfaltz, H. Yamamoto), Springer,
Berlin, 1999, pp. 923–961.
[4] C. Bolm, J. P. Hildebrand, K. Muniz in Recent Advances in Asym-
metric Dihydroxylation and Aminohydroxylation (Ed.: I. Ojima),
Wiley-VCH, Weinheim, 2000, pp. 399–428.
[5] V. I. Tararov, A. Bçrner, Synlett 2005, 203–211.
[6] M. Breuer, K. Ditrich, T. Habicher, B. Hauer, M. Keßeler, R.
Stꢆrmer, T. Zelinski, Angew. Chem. 2004, 116, 806–843; Angew.
Chem. Int. Ed. 2004, 43, 788–824.
[7] K. Drauz, H. Waldmann Enzyme Catalysis in Organic Synthesis: A
Comprehensive Hanbook, Wiley-VCH, Weinheim, 1995.
[8] J. Steinreiber, K. Faber, H. Griengl, Chem. Eur. J. 2008, 14, 8060–
8072.
[38] A. H. ꢋll, J. B. Johnson, J.-E. Bꢁckvall, Chem. Commun. 2003,
1652–1653.
[39] O. Pamies, A. H. Ell, J. S. M. Samec, N. Hermanns, J.-E. Bꢁckvall,
Tetrahedron Lett. 2002, 43, 4699–4702.
[9] B. Martꢇn-Matute, M. Edin, K. Bogꢈr, F. B. Kaynak, J.-E. Bꢁckvall,
J. Am. Chem. Soc. 2005, 127, 8817–8825.
[40] J. S. M. Samec, J.-E. Bꢁckvall, P. G. Andersson, P. Brandt, Chem.
Soc. Rev. 2006, 35, 237–248.
[10] B. Martꢇn-Matute, M. Edin, J.-E. Bꢁckvall, Chem. Eur. J. 2006, 12,
6053–6061.
[41] C. P. Casey, J. B. Johnson, J. Am. Chem. Soc. 2005, 127, 1883–1894.
[42] C. P. Casey, G. A. Bikzhanova, Q. Cui, I. A. Guzei, J. Am. Chem.
Soc. 2005, 127, 14062–14071.
[43] J. S. M. Samec, A. H. Ell, J. B. ꢄberg, T. Privalov, L. Eriksson, J.-E.
Bꢁckvall, J. Am. Chem. Soc. 2006, 128, 14293–14305.
[44] The initiation period is over after ca 1% conversion and is, there-
fore, not visible for (S)-[H]-1a.
[45] Amines 2p, 2q, and 2r can not be obtained by asymmetric hydroge-
nation or transfer hydrogenation because the reported examples re-
quire an aromatic substituent.
[46] J. Mao, D. C. Baker, Org. Lett. 1999, 1, 841–843.
[47] A. Trifonova, J. S. Diesen, C. J. Chapman, P. G. Andersson, Org.
Lett. 2004, 6, 3825–3827.
[11] L. Borꢀn, B. Martꢇn-Matute, Y. Xu, A. Cꢉrdova, J.-E. Bꢁckvall,
Chem. Eur. J. 2006, 12, 225–232.
[12] A. B. L. Fransson, Y. Xu, K. Leijondahl, J.-E. Bꢁckvall, J. Org.
Chem. 2006, 71, 6309–6316.
[13] D. Strꢆbing, P. Krumlinde, J. Piera, J.-E. Bꢁckvall, Adv. Synth. Catal.
2007, 349, 1577–1581.
[14] J. H. Choi, Y. K. Choi, Y. H. Kim, E. S. Park, E. J. Kim, M. J. Kim, J.
Park, J. Org. Chem. 2004, 69, 1972–1977.
[15] S. Akai, K. Tanimoto, Y. Kanao, M. Egi, T. Yamamoto, Y. Kita,
Angew. Chem. 2006, 118, 2654–2657; Angew. Chem. Int. Ed. 2006,
45, 2592–2595.
[16] A. Dijksman, J. M. Elzinga, Y.-X. Li, I. W. C. E. Arends, R. A. Shel-
don, Tetrahedron: Asymmetry 2002, 13, 879.
[48] N. Uematsu, A. Fujii, S. Hashiguchi, T. Ikariya, R. Noyori, J. Am.
Chem. Soc. 1996, 118, 4916–4917.
[17] M. J. Kim, Y. I. Chung, Y. K. Choi, H. K. Lee, D. Kim, J. Park, J.
Am. Chem. Soc. 2003, 125, 11494–11495.
[49] The racemization with catalyst 5c under the DKR conditions re-
quires a temperature of 908C.
[18] A. Berkessel, M. L. Sebastian-Ibarz, T. N. Mꢆller, Angew. Chem.
2006, 118, 6717–6720; Angew. Chem. Int. Ed. 2006, 45, 6567–6570.
[19] S. Akai, K. Tanimoto, Y. Kita, Angew. Chem. 2004, 116, 1431–1434;
Angew. Chem. Int. Ed. 2004, 43, 1407–1410.
[20] I. Hilker, G. Rabani, G. K. M. Verzijl, A. R. A. Palmans, A. Heise,
Angew. Chem. 2006, 118, 2184–2186; Angew. Chem. Int. Ed. 2006,
45, 2130–2132.
[21] B. A. C. van As, J. van Buijtenen, A. Heise, Q. B. Broxterman,
G. K. M. Verzijl, A. R. A. Palmans, E. W. Meijer, J. Am. Chem. Soc.
2005, 127, 9964–9965.
[22] M. T. Reetz, K. Schimmossek, Chimia 1996, 50, 668–669.
[23] J. Paetzold, J.-E. Bꢁckvall, J. Am. Chem. Soc. 2005, 127, 17620–
17621.
[50] T. W. Greene, P. G. M. Wuts, Protective Groups in Organic Synthesis,
Wiley Interscience, New York, 1999, pp. 503–550.
[51] It is not clear why the allyl carbonate gives only a moderate enantio-
selectivity as there was no background reaction in the control reac-
tion. A likely explanation is that the enzyme itself has some effect
on the non-catalyzed reaction, increasing the reaction pathway
through direct acylation.
[52] Z. Han, S. G. Koenig, H. Zhao, X. Su, S. P. Singh, R. P. Bakale, Org.
Process Res. Dev. 2007, 11, 726–730.
[53] M. J. Burk, Y. M. Wang, J. R. Lee, J. Am. Chem. Soc. 1996, 118,
5142–5143.
[54] C. Helgen, C. G. Bochet, J. Org. Chem. 2003, 68, 2483–2486.
[24] A. N. Parvulescu, E. van der Eycken, P. A. Jacobs, D. De Vos, J.
Catal. 2008, 255, 206–212.
[25] A. N. Parvulescu, P. A. Jacobs, D. E. De Vos, Chem. Eur. J. 2007, 13,
2034–2043.
Received: November 5, 2008
Published online: February 16, 2009
3410
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3403 – 3410