Amine-boranes: Effective reducing agents for the deracemisation of DL-amino acids using L-amino acid oxidase from Proteus myxofaciens

Article abstract of DOI:10.1016/S0040-4039(01)02233-X

The deracemisation of DL-α-amino acids using L-amino acid oxidase from Proteus myxofaciens and amine-boranes as chemical reducing agents has been investigated. Amine-boranes were found to be of particular interest in terms of reactivity and chemoselectivity compared to sodium borohydride and cyanoborohydride. Starting from the racemate, a range of D-amino acids were obtained in yields of up to 90% and e.e. >99%.

Full text of DOI:10.1016/S0040-4039(01)02233-X

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 707–710
Amine–boranes: effective reducing agents for the deracemisation
of DL-amino acids using -amino acid oxidase from
L
Proteus myxofaciens
a
b
b
a
Fran c¸ ois-Ren e´ Alexandre, David P. Pantaleone, Paul P. Taylor, Ian G. Fotheringham,
c
a,
David J. Ager and Nicholas J. Turner *
a
Department of Chemistry, Centre for Protein Technology, University of Edinburgh, King’s Buildings, West Mains Road,
Edinburgh EH9 3JJ, UK
b
Great Lakes Fine Chemicals, 601 E. Kensington Rd., Mt. Prospect, IL 60056, USA
c
MediChem Research, Inc., 2501 Davey Road, Woodridge, Il 60517, USA
Received 16 August 2001; revised 12 November 2001; accepted 22 November 2001
Abstract—The deracemisation of DL-a-amino acids using
L-amino acid oxidase from Proteus myxofaciens and amine-boranes as
chemical reducing agents has been investigated. Amine-boranes were found to be of particular interest in terms of reactivity and
chemoselectivity compared to sodium borohydride and cyanoborohydride. Starting from the racemate, a range of
D-amino acids
were obtained in yields of up to 90% and e.e. >99%. © 2002 Elsevier Science Ltd. All rights reserved.
The kinetic resolution of racemates is a highly success-
ful strategy for the synthesis of enantiomerically pure
chiral compounds and has found widespread usage in
ple of the latter process is the cyclic sequence involving
an enantioselective oxidation (e.g. enzymatic) leading to
an achiral intermediate, followed by a nonselective
reduction (Scheme 1). The repetition of this sequence
rapidly leads to a single enantiomer in a theoretical
100% yield and 100% e.e.
1
industry. However, the theoretical yield of each enan-
tiomer can never exceed 50% and thus to avoid the
wastage of the unwanted enantiomer, reactions that
2
give >50% yield have received increasing interest. Such
2
6a
reactions can be divided into three types, namely3
Based on the original reports by Hafner et al. and
6b,c
7
desymmetrisation of meso or prochiral compounds,
Soda et al.,
racemisation of acyclic and alicyclic a-amino acids using
a combination of porcine kidney -amino acid oxidase
we have recently investigated the de-
4
dynamic kinetic resolutions, and less exploited,
5
deracemisation of an enantiomeric mixture. An exam-
D
Scheme 1. Deracemisation of DL-amino acids using L-amino acid oxidase (L-AAO).
Keywords: amine–borane;
D
-amino acid;
L-amino acid oxidase; deracemisation.
*
Corresponding author. Tel./fax: (44) 131 650 4717; e-mail: n.j.turner@ed.ac.uk
0
040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)02233-X

7
08
F.-R. Alexandre et al. / Tetrahedron Letters 43 (2002) 707–710
and sodium borohydride or sodium cyanoborohydride.
For acyclic substrates we have shown that the interme-
diate imino acid 3 is sufficiently stable to be reduced by
sodium borohydride although some hydrolysis to the
keto acid 4 is observed. However, both of these hydride
reducing agents have certain limitations in terms of
practical application. Sodium borohydride is sensitive
to water, is easily decomposed at low and neutral pH
requiring large excesses to be added which often leads
boranes) at neutral or basic pH but they are also
soluble and unreactive toward a wide range of protic
and aprotic solvents. Their reducing abilities are greatly
dependent on the complexing amine, thus a range of
reduction capabilities are obtainable providing high
versatility and selectivity. Furthermore, most aliphatic
amine–boranes are white crystalline solids and stable at
13
room temperature.
8
to rapid deactivation of oxidase enzymes. Sodium9
The deracemisation of DL-proline with
D-amino acid
cyanoborohydride is more stable at low to neutral pH,
however its industrial application is limited because of
its toxicity and its reduced reactivity compared with
sodium borohydride. For these reasons we decided to
investigate alternative reducing agents as well as using a
oxidase from porcine kidney was initially investigated
in order to compare the different reducing systems
(Scheme 2). The reactions were performed according
7
,14a
to the previously described procedure.
HPLC
15
monitoring showed that the reduction step is slower
10
new
to develop an efficient method for the synthesis of
-amino acids.
L-amino acid oxidase from Proteus myxofaciens,
with CTH (HCOONH /Pd/C) than with the amine–
4
boranes. Ammonia–borane and pyridine–borane gave
the best results (see Table 1). Only 6–10 equiv. of
amine–boranes, compared to the 500 necessary equiva-
D
6
The only alternative reducing system that has been
lents for NaBH4, were required to achieve complete
1
1
reported for deracemisation reactions is Catalytic
deracemisation. Because of the associated toxicity of
pyridine–borane, ammonia–borane and tert-butyl-
amine–borane were chosen for further studies.
1
2
Transfer Hydrogenation (CTH). The intermediate
imino acid 3 is reduced by hydrogen which is released
from a metal or metal complex and a hydrogen source
as the reaction proceeds. We now report utilisation of
amine–boranes as reducing agents in the deracemisa-
tion system. These compounds have received relatively
little attention compared to their chemical cousins
We next examined the ability of the
oxidase -AAO) from Proteus myxofaciens
catalyse the deracemisation of DL-mixtures to produce
-amino acids. The lad gene encoding -amino acid
L-amino acid
10
(
L
to
D
L
borane–THF,
borane–Me S,
borohydride
and
2
oxidase has been cloned and overexpressed in
Escherichia coli K12. The enzyme has a broad substrate
specificity catalysing the oxidation of both natural and
1
3
cyanoborohydride. Nevertheless, amine–boranes pos-
sess interesting properties that could overcome the dis-
advantages of using NaBH41 and NaBH CN.
3
unnatural
L-amino acids. The reaction has an absolute
3
Amine–boranes are stable in water (except arylamine–
requirement for O , releases NH but does not produce
2
3
detectable levels of H O and hence the enzyme has also
been termed a deaminase. The optimum pH is 7.5. We
2
2
1
0
Table 1.
first studied the deracemisation of DL-leucine using
Entry
Reducing agent
Yield (%)a
e.e. (%)a
10,14
disrupted cells
of L-AAO. The reactions were
14b
performed at pH 6.7 in 50 mM ammonium formate
solution. The variation of the quantity of reducing
agent revealed that 40 equiv. of ammonia–borane were
a
b
NH –BH3 (9 equiv.)
97
82
\99
\99
3
1 M HCOONH /Pd/C
4
(10%)
needed to afford a 98% yield of optically pure D-leucine
c
d
BH –Pyridine (6 equiv.)
99
\99
\99
3
t-BuNH –BH (10 equiv.) 92
(Scheme 3). This result suggests that the concentration
of reducing agent must be high enough to ensure
immediate reduction of the intermediate a-imino acid
2
3
a
Yield and e.e. based on chiral HPLC analysis.15
Scheme 2. Deracemisation of DL-proline using
D-amino acid oxidase and reducing agents.
Scheme 3. Deracemisation of DL-leucine with L-amino acid oxidase and amine–borane.

F.-R. Alexandre et al. / Tetrahedron Letters 43 (2002) 707–710
709
and avoid hydrolysis to the keto acid. In order to
increase the concentration of the a-imino acid versus
keto acid, ammonium formate was added to the reac-
tion (Scheme 1). It appears that the concentration of
the a-imino acid must be reasonably high, since an
attempt to start from 4-methyl-2-oxo-pentanoic acid to
aromatic, ethylenic, alkyl and thiol containing groups
(Table 3). It should be noted that the double bond of
allyglycine (entry e) remained unaffected in the de-
racemisation reaction, showing that ammonia–borane is
1
3
a poor hydroborating system under these conditions.
produce
D-leucine under the same conditions was
In conclusion, we have shown that amine–boranes are
efficient reducing agents for deracemisation reactions.
Further studies will be focused on their chemoselective
reductive abilities in choosing the appropriate amine
substituent according to the reductive functionalities in
the substrate.
unsuccessful. Yields obtained with tert-butylamine-
borane are also lower than with ammonia–borane as
shown in compared reaction profiles (Fig. 1) pre-
sumably as a result of the lower reactivity and solubility
of tert-butylamine-borane in water. Furthermore,
ammonia–borane releases ammonia as the reaction pro-
1
6
14c
ceeds. Experiments using whole cells
acid oxidase were also performed and deracemisation
of DL-leucine provided -leucine in 90% yield and e.e.
99% (Table 2).
of L-amino
References
D
>
1
. Liese, A.; Seelbach, K.; Wnadrey, C. Industrial Biotrans-
formations—A Comprehensive Handbook; Wiley-VCH:
Weinheim, 2000.
Several racemic mixtures of natural and unnatural
amino acids were then subjected to the deracemisation
reaction using ammonia–borane and
2. (a) Strauss, U. T.; Felder, U.; Faber, K. Tetrahedron:
Asymmetry 1999, 10, 107–117.
L
-amino acid
1
4
oxidase. The a-functionality was varied to include
3
. For amino acids see: (a) Enzyme Catalysis in Organic
Synthesis: A Comprehensive Handbook; Drauz, K.; Wald-
mann, H., Eds.; VCH, 1995, pp. 633–641; (b) Kelly, N.
M.; Reid, R. G.; Willis, C. L.; Winton, P. L. Tetrahedron
Lett. 1996, 37, 1517–1520.
4
. For amino acids, see for example: Turner, N. J.; Winter-
man, J. R.; McCague, R.; Paratt, J. S.; Taylor, S. J. C.
Tetrahedron Lett. 1995, 36, 1113–1116.
5. For mathematical treament, see: Kroutil, W.; Faber, K.
Tetrahedron: Asymmetry 1998, 9, 2901–2913.
6
. (a) Hafner, E. W.; Wellner, D. Proc. Natl. Acad. Sci.
USA 1971, 68, 987–991; (b) Huh, J. W.; Yokoigawa, K.;
Esaki, N.; Soda, K. J. Ferment. Bioeng. 1992, 74, 189–
190; (c) Huh, J. H.; Yokoigawa, K.; Esaki, N.; Soda, K.
Biosci. Biotech. Biochem. 1992, 56, 2081–2082.
. Beard, T. M.; Turner, N. J. Chem. Commun. 2001, in
press.
7
Figure 1. Deracemisation of DL-leucine with L-AAO and 20
equiv. of NH –BH (red. 1) or t-BuNH –BH (red. 2).
3
3
2
3
8
. Quay, S.; Massey, V. Biochemistry 1977, 16, 3348–3354.
. Borch, R. F.; Bernstein, M. D.; Durst, H. D. J. Am.
Chem. Soc. 1971, 93, 2897–2904.
9
Table 2.
Entry
Reducing agent
Yield (%)a
98
e.e. (%)a
10. (a) Pantaleone, D. P.; Geller, A. M.; Taylor, P. P. J. Mol.
Cat. B: Enzym. 2001, 11, 795–803; (b) Cells were dis-
rupted using a french press, centrifuged, and the superna-
tant was used for experiments.
1. Presented in part at Chiral USA organised by Scientific
Update, Marriott Longwharf, Boston, USA; May 3–4,
1999.
a
b
c
NH –BH3 (40 equiv.)
\99
\99
\99
3
t-BuNH –BH (20 equiv.) 90
2
3
1 M HCOONH /Pd/C
65
4
1
(10%)
a
Yield and e.e. based on chiral HPLC analysis.15
Table 3.
Entry
Substrate
Yield (%)a
e.e. (%)a
Entry
Substrate
Yield (%)a
e.e. (%)a
a
b
c
d
e
f
Norvaline
81
86
82
82
79
90
\99
\99
\99
\99
\99
\99
g
h
i
j
k
l
O-Benzyl serine
a-Aminobutyric acid
Valine
Histidine
Cyclopentylglycine
Tyrosine
87
64
64
67
87
79
93
96
28
65
\99
\99
Norleucine
Tryptophan
Phenylalanine
Allylglycine
Methionine
b
b
a
Yield and e.e. based on chiral HPLC analysis,15 reaction time 5 h.
Lower e.e. values are due to incomplete deracemisation.
b

7
10
F.-R. Alexandre et al. / Tetrahedron Letters 43 (2002) 707–710
ing agent. (c) using whole cells of L-AAO: 1.5 mL of 2.5
1
1
2. Johnstone, R. A. W.; Wilby, A. H.; Entwistle, I. D.
Chem. Rev. 1985, 85, 129–170.
mM DL-leucine in HCOONH4 50 mM, 20 mL of a
whole cells solution (10 mg in 100 mL buffer) and
ammonia–borane (40 eq.).
3. (a) Hutchins, R. O.; Hutchins, M. K. In Comprehensive
Organic Synthesis; Trost, B. M.; Fleming, I., Eds.; Perg-
amon, 1991; Vol. 8, p. 25; (b) Hutchins, R. O.; Learn,
K.; Nazer, B.; Pytlewski, D. Org. Prep. Proced. Int.
984, 16, 335–372 and references cited therein.
4. Reaction procedures: (a) using –AAO of porcine kid-
ney (pk): 100 mL, 25 mM DL–amino acid in HCOONH4
50 mM or 1 M (CTH)], 100 mL, 0.8 mM FAD solution
in HCOONH , 800 mL HCOONH , then 1 mg catalase,
15. Chiral HPLC analyses were performed with Chirex 3126
column (supplied by Phenomenex), using CuSO4 1.5
mM solution/MeOH as the mobile phase. Yields were
determined by integration of the peaks against internal
standards.
1
1
D
[
16. (a) White, S. S., Jr.; Kelly, H. C. J. Am. Chem. Soc.
1968, 90, 2009–2011; (b) White, S. S., Jr.; Kelly, H. C.
J. Am. Chem. Soc. 1970, 92, 4203–4209; (c) Wolfe, T.
C.; Kelly, H. C. J. Chem. Soc., Perkin Trans. 2 1973,
1948–1950.
4
4
2
mg
rupted cells of
acids in HCOONH 50 mM, 20 mL
D
-AAO (pk) and reducing agent. (b) using dis-
-AAO: 1.5 mL of 2.5 mM DL-amino
-AAO and reduc-
L
L
4