Deracemisation and stereoinversion of alpha-amino acids using D-amino acid oxidase and hydride reducing agents.

Article abstract of DOI:10.1039/b107580m

The deracemisation and stereoinversion of both cyclic and acyclic DL-alpha-amino acids, using porcine kidney D-amino acid oxidase (DAAO) and a hydride reducing agent (NaCNBH3-NaBH4), has been investigated.

Full text of DOI:10.1039/b107580m

Deracemisation and stereoinversion of a-amino acids using -amino acid
D
oxidase and hydride reducing agents
Timothy M. Beard and Nicholas J. Turner*
Department of Chemistry, Centre for Protein Technology, The University of Edinburgh, King’s Buildings,
West Mains Road, Edinburgh, UK EH9 3JJ. E-mail: n.j.turner@ed.ac.uk
Received (in Cambridge, UK) 21st August 2001, Accepted 11th December 2001
First published as an Advance Article on the web 11th January 2002
The deracemisation and stereoinversion of both cyclic and
2H-
-amino acids and (iii) can be used to deracemise a wide
range of acyclic amino acids in high yield and optical purity
provided that NaBH4 is used.
L
acyclic DL-a-amino acids, using porcine kidney
D-amino
acid oxidase (DAAO) and a hydride reducing agent
(NaCNBH3–NaBH4), has been investigated.
We selected proline as the substrate for initial optimisation
experiments. One problem associated with the original proce-
dure7,8 was the requirement for large excesses of NaBH4 (up to
500 equiv.) owing to its high reactivity with water. In addition,
the use of NaBH4 results in an operating pH of ~ 10 which is
close to the pH at which the enzyme begins to undergo
irreversible denaturation.10 NaCNBH3 is a milder, water stable,
hydride reducing agent11 and was able to deracemise proline
( > 99% yield; > 99% ee) with as little as 3 molar equivalents.
We were able to apply this improved methodology12 to the
deracemisation of DL-piperazine-2-carboxylic acid 3, a compo-
nent of the HIV-protease inhibitor Crixivan,13 in 86% yield and
99% ee (Scheme 2).
Enzyme-catalysed processes have proven to be extremely
useful for the synthesis of enantiopure amino-acids1 and indeed
a number of commercially important amino acids are now
manufactured using enzymatic methods.2 The most efficient
reactions are those that result in > 50% yield and in general this
outcome can be achieved by either the asymmetric transforma-
tion of a prochiral substrate3 or by a dynamic kinetic
resolution.4 A third approach involves the deracemisation of a
mixture of enantiomers in which a racemic mixture is converted
to the single enantiomer form.5 Deracemisation has been
achieved in two ways, either (i) by using two or more enzymes
(e.g. dehydrogenases) with complementary enantioselectivity6
or (ii) by combining an enantioselective enzyme-catalysed
oxidation reaction with a non-selective, non-enzymatic reduc-
tion reaction. The latter approach is the subject of this
communication.
In 1971, Hafner and Wellner reported the generation of
alanine 1a and -leucine 1b from the corresponding
enantiomers by the use of porcine kidney
L
-
L
D-
D
-amino acid oxidase
Scheme 2 Deracemisation of DL-piperazine-2-carboxylic acid 3.
(DAAO) (EC 1.4.3.3.) and NaBH4 (Scheme 1). Although the
yield was low the principle of stereoinversion was established
Interestingly, replacement of NaBH4 by NaCNBH3 caused a
change in the rate limiting step of the reaction. With NaBH4 as
by enzyme catalysed oxidation of the
D
-enantiomer to the
corresponding achiral imino acid 2 followed by reduction in situ
reducing agent, the total concentration of
D
and
L-proline
by the borohydride generating a mixture of
acids.7
D
- and -amino
L
remained almost constant throughout the reaction indicating
that the imino acid intermediate did not accumulate to any
significant degree. However, with NaCNBH3, the increase of
Subsequently Soda et al. extended this method to the
deracemisation of DL-proline8a and DL-pipecolic acid,8b again
the
L
enantiomer did not follow the decrease in
D
enantiomer
using a combination of -amino acid oxidase and NaBH4. In
D
suggesting that the reduction of the imine was slower than with
NaBH4 (Fig. 1), although by the end of the reaction the yield of
both cases the yields and optical purities were high ( > 98%).
The high yields obtained with cyclic substrates are a con-
sequence of the increased stability of the respective imino acid
intermediates allowing for a sufficient number of oxidation–
reduction cycles to occur leading to complete deracemisa-
tion.9
Herein we now report some significant developments of this
deracemisation process, in particular that the reaction (i) can be
made more efficient for cyclic substrates by the use of
NaCNBH3 rather than NaBH4, (ii) can be applied to the
stereoinversion of cyclic substrates and also used to prepare a-
the -isomer was > 98%. The NaCNBH3 concentration was
L
varied (3–100 equivalents) but did not significantly affect the
overall rate of reduction.
NaCNBH3 was found to be equally effective for the
stereoinversion of
of DAAO and NaCNBH3,
in > 99% yield and > 99% ee. This transformation could be
D- to
L
-a-amino acids. Thus in the presence
D
-proline was converted to -proline
L
particularly useful in cases where the -isomer is more readily
D
available than the racemate as a starting material.
Scheme 1 Stereoinversion of
D- to
L-amino acids.
Fig. 1 Comparison of NaCNBH3 with NaBH4.
246
CHEM. COMMUN., 2002, 246–247
This journal is © The Royal Society of Chemistry 2002

The stereoinversion of
be possible to prepare
with NaCNB2H3 although we recognised that the production of
-2H-proline early on in the reaction might slow down the
oxidation reaction. However no significant decrease in the rate
of consumption of the -enantiomer was observed suggesting
D
- to
L
-proline suggested that it should
reduced by addition of bovine serum albumin (BSA) to the
reaction which sacrificially reacted with the a-keto acid thereby
removing it from solution.
L
-2H-proline by replacing NaCNBH3
D
In conclusion we have developed a highly efficient chemo-
enzymatic method for the deracemisation and stereoinversion of
cyclic and acyclic a-amino acids. The reaction is attractive in
that only a single enzyme is required together with the readily
available enzyme DAAO and NaCNBH3 or NaBH4 We are
currently examining additional facets of this reaction namely
D
that DAAO only exhibits a small primary kinetic isotope effect
with proline as the substrate. The use of NaCNB2H3 did cause
a slight decrease in the rate of formation of the -enantiomer due
L
to slower reduction of the imine intermediate. This observation
is consistent with the fact that reduction of the imino acid is the
rate limiting step and suggests that cleavage of the B–2H bond
the identification of a wider range of
D
- and -amino acid
L
oxidases from microbial sources and exploring the use of
alternative chemical reducing agents.
is slower than cleavage of the B–H bond. The
L
-2H-proline
We are grateful to the BBSRC and Chirotech Ltd for financial
support (T. M. B.). We also wish to thank Dr Andy Kiener,
derived from this experiment was found to have a 96%
2
incorporation of H (82% yield, > 99% ee); ESI-MS: m/z (%)
Lonza AG, for providing samples of
D
-, - and DL-3.
L
117 (100) [M+ + 1].
To date there have been no reports of the deracemisation of
acyclic amino acids using DAAO. We have found that acyclic
substrates give good to excellent yields (75–90%) of high ee
( > 98%) products indicating that the corresponding acyclic
imino acids 2 are sufficiently stable to undergo reduction rather
than hydrolysis to the a-keto acid. However it was necessary to
use NaBH4 in these reactions rather than NaCNBH3 because the
Notes and references
1 A. Schmid, J. S. Dordick, B. Hauer, A. Kiener, M. Wubbolts and B.
Witholt, Nature, 2001, 409, 258; C. H. Wong and G. M. Whitesides,
Enzymes in Synthetic Organic Chemistry, Vol. 12, Pergamon, Oxford,
1994.
2 Chirality in Industry, ed. A. N. Collins, G. N. Sheldrake and J. Crosby,
Wiley, 1994, pp. 371–397.
3 Enzyme Catalysis in Organic Synthesis: A Comprehensive Handbook,
ed. K. Drauz and H. Waldmann, VCH, 1995, pp 633-641; J. R. Harding,
R. A. Hughes, N. M. Kelly, A. Sutherland and C. L. Willis, J. Chem.
Soc., Perkin Trans. 1, 2000, 3406.
4 N. J. Turner, J. R. Winterman, R. McCague, J. S. Parratt and S. J. C.
Taylor, Tetrahedron Lett., 1995, 36, 1113; H. Vanderdeen, A. D.
Cuiper, R. P. Hof, B. L. Feringa and R. M. Kellogg, J. Am. Chem. Soc.,
1996, 118, 3801; J. V. Allen and J. M. J. Williams, Tetrahedron Lett.,
1996, 37, 1859.
lower reactivity of the latter led to decreased yields of the -
L
amino acid. The a-functionality was varied (Table 1) to include
side-chains containing aromatic 1c–1e, thiol 1f, alkyl (straight
chained 1g and branched 1b) and cyclic 1h groups. tert-Leucine
1i failed to act as a substrate, presumably due to the bulky nature
of the tert-butyl group. The reaction profile for the deracemisa-
tion of DL-leucine 1b (90% yield; > 99% ee) is given in Fig.
2.
The use of NaBH4 was found to result in partial inactivation
of the DAAO presumably due to reduction of the Schiff’s base
formed between some adventitious a-keto acid and an im-
portant amino group on the enzyme. This problem could be
5 For a recent review on the kinetics and stereochemistry of deracemisa-
tion reactions see: U. T. Strauss, U. Felfer and K. Faber, Tetrahedron
Asymmetry, 1999, 10, 107.
6 P. C. B. Page, A. J. Carnell and M. J. McKenzie, Synlett, 1998, 774; K.
Nakamura, Y. Inoue, T. Matsuda and A. Ohno, Tetrahedron Lett., 1995,
36, 6263; G. Fantin, M. Fogagnolo, P. P. Giovannini, A. Medici and P.
Pedrini, Tetrahedron Asymmetry, 1995, 6, 3047; N. Nakajima, N. Esaki
and K. Soda, J. Chem. Soc., Chem. Commun., 1990, 947.
7 E. W. Hafner and D. Wellner, Proc. Nat. Acad. Sci., 1971, 68, 987.
8 (a) J. W. Huh, K. Yokoigawa, N. Esaki and K. Soda, J. Ferment.
Bioeng., 1992, 74, 189; (b) J. W. Huh, K. Yokoigawa, N. Esaki and K.
Soda, Biosci. Biotech. Biochem., 1992, 56, 2081; ; See also: K. Soda, T.
Oikawa and K. Yokoigawa, J. Mol. Cat. B, 2001, 11, 149.
9 The final ee of the product from a deracemisation reaction is determined
Table 1 Deracemisation of RCH(NH2)CO2H
R
Yield (%)
Ee (%)
Ph
PhCH2
1c
1d
1e
1f
1g
1b
1h
1i
75
82
76
77
87
90
79
99
99
99
99
99
99
99
—
Indolyl-CH2
HSCH2
CH3CH2
(CH3)2CHCH2
Cyclopentyl
tBu
by the ratio of rate constants k1/k2 for oxidation of the
D and L-
enantiomers since the latter are in equilibrium via the imino acid. Hence
the ee is identical to that achieved by a standard kinetic resolution
process. The ee is also given by (E 2 1)/(E + 1) as for dynamic kinetic
resolutions; H. Stecher and K. Faber, Synthesis, 1997, 1.
Not a substrate
For conditions see reference 12.
10 S. Quay and V. Massey, Biochem. J., 1977, 16, 3348.
11 R. F. Borch, M. D. Bernstein and H. D. Durst, J. Am. Chem. Soc., 1971,
93, 2897.
12 General procedure: deracemisation of cyclic amino acids was carried
out in phosphate buffer (pH 8.0, 40 mM) containing amino acid (2.5
mM), FAD (80 mM), porcine kidney DAAO (5U, Sigma), catalase
(2000U) and NaCNBH3 (3 equivalents) and the solution incubated at 37
°C. Acyclic amino acids were deracemised using the same methodology
except that BSA (1 mg/1 ml) was added and the NaCNBH3 was replaced
by NaBH4. Samples were removed, filtered and analysed by chiral
HPLC (Chirex 3126 supplied by Phenomenex, using CuSO4–MeOH as
the mobile phase). For the deuteration experiments the NaCNBH3 was
replaced by NaCNBD3 and the DAAO was removed through a 10000
molecular weight cut off prior to analysis by ES-MS.
13 For an alternative amidase-based kinetic resolution of DL-3 see: E.
Eichhorn, J. P. Roduit, N. Shaw, K. Heinzmann and A. Kiener,
Tetrahedron Asymmetry, 1997, 8, 2533.
Fig. 2 Profile for deracemisation of DL-leucine 1b.
CHEM. COMMUN., 2002, 246–247
247