Preparation of methyl (1R,2S,5S)- and (1S,2R,5R)-2-amino-5-terf-butyl- cyclopentane-1-carboxylates by parallel kinetic resolution of methyl (RS)-5-tert-butyl-cyclopentene-1-carboxylate

Article abstract of DOI:10.1039/b308855c

Comparison of the kinetic and parallel kinetic resolutions of methyl (RS)-5-tert-butyl-cyclopentene-1-carboxylate allows for the efficient synthesis of both (1R,2S,5S)- and (1S,2R,5R)-enantiomers of methyl 2-amino-5-tert-butyl- cyclopentane-1-carboxylate.

Full text of DOI:10.1039/b308855c

Preparation of methyl (1R,2S,5S)- and
(1S,2R,5R)-2-amino-5-tert-butyl-cyclopentane-1-carboxylates by parallel
kinetic resolution of methyl (RS)-5-tert-butyl-cyclopentene-1-carboxylate
Stephen G. Davies,*a David Díez,b Mohamed M. El Hammouni,b A. Christopher Garner,a Narciso M.
Garrido,b Marcus J. C. Long,a Rachel M. Morrison,a Andrew D. Smith,a Miles J. Sweeta and Jonathan
M. Witheya
a The Dyson Perrins Laboratory, University of Oxford, South Parks Road, Oxford, UK OX1 3QY.
E-mail: steve.davies@chemistry.oxford.ac.uk
b Departmento de Química Orgánica, Universidad de Salamanca, Plaza de los Caídos 1-5, 37008
Salamanca, Spain
Received (in Cambridge, UK) 29th July 2003, Accepted 14th August 2003
First published as an Advance Article on the web 26th August 2003
Comparison of the kinetic and parallel kinetic resolutions of
methyl (RS)-5-tert-butyl-cyclopentene-1-carboxylate allows
for the efficient synthesis of both (1R,2S,5S)- and
(1S,2R,5R)-enantiomers of methyl 2-amino-5-tert-butyl-cy-
clopentane-1-carboxylate.
Cispentacin (2-aminocyclopentane carboxylic acid) has been
the subject of much commercial and synthetic interest,1 with the
asymmetric synthesis of derivatives required for both structural
and bioactivity studies.2 We have previously shown that the
conjugate addition of homochiral lithium amides to a,b-
Scheme 1 Reagents and conditions: (i) lithium (RS)-N-benzyl-N-a-
methylbenzylamide 2, THF, 278 °C then NH4Cl (aq.).
unsaturated esters represents an efficient and versatile method-
ology for the asymmetric synthesis of b-amino acid deriva-
tives.3 This methodology has recently been employed for the
preparation of (1R,2S,3R)-3-methyl-cispentacin by the kinetic
resolution of tert-butyl (RS)-3-methyl cyclopentene-1-carbox-
ylate.4 Although kinetic resolution methodology leads to the
generation of only one stereoisomeric reaction product, it is
practically limited by the need to compromise either yield or
selectivity in the reaction, due to the negative effects of mass
action as the reaction progresses.5 It was envisaged that the
development of a parallel kinetic resolution protocol would
maximise the e.e. of the product and facilitate the isolation of
both product enantiomers from a single reaction. The study
demonstrated herein describes the efficient parallel kinetic
resolution of methyl (RS)-5-tert-butyl-cyclopentene-1-carbox-
ylate 1 with a pseudoracemic mixture of chiral lithium
amides.
Initial investigations concentrated upon the preparation of
methyl (RS)-1, which was synthesised by the method of
Villiéras et al., via the copper catalysed conjugate addition of
tert-butyl magnesium chloride to methyl (RS)-5-O-acetyl-
cyclopentene-1-carboxylate.6 The maximum level of enantior-
ecognition between (RS)-1 and (RS)-lithium N-benzyl-N-a-
methylbenzylamide 2 was evaluated by the application of a
mutual kinetic resolution protocol in order to determine the
suitability of the reacting components for resolution.4,7 The
advantage of this protocol, in which both reaction components
are racemates, is that the stereoselectivity factor (E) (equal to
the ratio of product diastereoisomers) can be determined,
independent of the conversion. Thus, treatment of (RS)-1 with
(RS)-lithium amide 2 furnished (1RS,2SR,5SR,aSR)-3 in 81%
isolated yield and in 99 ± 1% d.e., with E > 99. The
configuration at C(2) within b-amino ester 3 relative to the N-a-
methylbenzyl stereocentre was assigned by analogy with
previous models developed to explain the diastereoselective
addition of lithium amide 2 to a,b-unsaturated acceptors,8 with
the relative (1RS,2SR,5SR,aSR) configuration within 3 con-
firmed by analysis of 1H NOE difference data (Scheme 1).
Having ascertained that a high degree of mutual enantio-
recognition is exhibited between the (RS)-1 and the (RS)-lithium
amide 2, the efficiency of the kinetic resolution of (RS)-1 with
the homochiral9 lithium amide (S)-2 was investigated. Addition
of 0.8 equivalents of lithium amide (S)-2 to (RS)-1 gave, at
approximately 53% conversion, b-amino ester (1R,2S,5S,aS)-3
in 88% d.e. and 42% isolated yield, with the major diastereo-
1
isomer showing identical H NMR spectroscopic properties to
b-amino ester 3 resulting from the mutual kinetic resolution
protocol. The resolved acceptor (S)-1 was isolated in 41% yield
and in 98 ± 1% e.e.,10 consistent with E > 65.11 Like all kinetic
resolutions, the selectivity observed upon reaction of the
homochiral lithium amide (S)-2 with (RS)-1 is lower than that
observed with the (RS)-lithium amide and (RS)-acceptor, due to
deleterious mass action effects operating in the former process.
b-Amino ester (1R,2S,5S,aS)-3 was next subjected to N-
debenzylation via hydrogenolysis to give the b-amino ester
(1R,2S,5S)-4 {[a]2D4 27.0 (c 1.8, CHCl3)} in 95% yield, 98%
d.e. and 88% e.e.12,13 Ester hydrolysis and ion exchange
chromatography afforded the free b-amino acid (1R,2S,5S)-5 in
68% yield and 98% d.e. (Scheme 2).
Scheme 2 Reagents and conditions: (i) (S)-2 (0.8 eq.), THF, 278 °C then
NH4Cl (aq.); (ii) Pd(OH)2 on C, MeOH, H2 (5 atm.), rt; (iii) 20% HCl (aq.),
D then Dowex 50W-X8.
Although high levels of selectivity are observed upon kinetic
resolution of (RS)-1 with lithium amide (S)-2, the use of two
2410
CHEM. COMMUN., 2003, 2410–2411
This journal is © The Royal Society of Chemistry 2003

pseudoenantiomeric lithium amides in a parallel kinetic resolu-
tion14 is expected to eliminate the deleterious effect of mass
action during the resolution process. A further advantage of this
protocol is elimination of the inherent sensitivity of the e.e. of
product and substrate to conversion, since each kinetic
resolution automatically stops at 50% conversion, giving the
products with improved e.e. (S)-2 (98% e.e.)15 and (R)-N-
3,4-dimethoxybenzyl-N-a-methylbenzylamide 6 (98% e.e.)15
were chosen as the components of the pseudoracemic mixture
required for this protocol, as they show complementary
stereoselectivity upon addition to achiral a,b-unsaturated
acceptors.16 Furthermore, it was predicted that the polar nature
of the N-3,4-dimethoxybenzyl protecting group would facilitate
the separation of the b-amino ester products arising from
conjugate addition.17 Thus, (RS)-1 (1 eq.) in THF was added to
a 50 : 50 mixture of amides (S)-2 (1.5 eq.) : (R)-6 (1.5 eq.) in
THF and stirred for two hours prior to the addition of NH4Cl
(aq.). Examination of the 1H NMR spectra of the crude reaction
mixture showed a 50 : 50 mixture of b-amino esters
(1R,2S,5S,aS)-3 (identical by 1H NMR to that arising from both
the mutual and kinetic resolution protocols) and (1S,2R,5R,aR)-
718 in 98 ± 1% d.e. in each case, consistent with E > 65 for each
process. Chromatographic purification enabled the isolation of
the two readily separable b-amino esters 3 and 7 in 39 and 35%
yield respectively, and in 98 ± 1% d.e. in each case. Subsequent
N-debenzylation of (1R,2S,5S,aS)-3 by hydrogenolysis gave b-
amino ester (1R,2S,5S)-4 in 75% yield, 98% d.e. and 96 ± 1%
other cases (R = iPr, Ph, Mes), and the extension of this
methodology for the synthesis of a range of 3-, 4- and 5-alkyl
substituted cispentacin analogues is currently under investiga-
tion within our laboratory.
The authors wish to thank Pfizer for an industrial CASE
award (J. M. W.), the Rhodes Trust (M. J. S.), New College,
Oxford for a Junior Research Fellowship (A. D. S.) and the
CICYT, Junta de Castilla y Leon and FSE for financial support
(SA 036/01).
Notes and references
1 For a review on the chemistry of this class of b-amino acid see F. Fülöp,
Chem. Rev., 2001, 2181; for selected syntheses see S. G. Davies, O.
Ichihara, I. Lenoir and I. A. S. Walters, J. Chem. Soc., Perkin Trans. 1,
1994, 1411; V. K. Aggarwal, S. J. Roseblade, J. K. Barrell and R.
Alexander, Org. Lett., 2002, 4, 1227.
2 For biological activity of the parent compound see M. Konishi, M.
Nishio, K. Saitoh, T. Miyaki, T. Oki and H. Kawaguchi, J. Antibiot.,
1989, 42, 1749; T. Oki, M. Hirano, K. Tomatsu, K. Numata and H.
Kamei, J. Antibiot., 1989, 42, 1756. For the secondary structure of
oligomers of cispentacin see T. A. Martinek, G. K. Táth, E. Vass, M.
Hollósi and F. Fülöp, Angew. Chem., Int. Ed., 2002, 41, 1718.
3 S. G. Davies, N. M. Garrido, O. Ichihara and I. A. S. Walters, Chem.
Commun., 1993, 1153; S. G. Davies, O. Ichihara and I. A. S. Walters,
Synlett, 1993, 461; S. G. Davies, O. Ichihara, I. Lenoir and I. A. S.
Walters, J. Chem. Soc., Perkin Trans. 1, 1994, 1411.
4 S. Bailey, S. G. Davies, A. D. Smith and J. M. Withey, Chem. Commun.,
2002, 2910; M. Bunnage, A. M. Chippindale, S. G. Davies, R. M.
Parkin, A. D. Smith and J. M. Withey, Org. Biomol. Chem., 2003, 1,
DOI: 10.1039/b306935b.
5 For excellent reviews on the subject of kinetic resolution see H. B.
Kagan and J. C. Fiaud, Top. Stereochem., 1988, 18, 249; H. B. Kagan,
Tetrahedron, 2001, 57, 2449.
e.e.12 {[a]D 27.5 (c 1.0, CHCl3)}, while N-deprotection of
24
(1S,2R,5R,aR)-7 via oxidative deprotection with DDQ and
subsequent hydrogenolysis gave b-amino ester (1S,2R,5R)-4 in
43% yield (over two steps), 98% d.e. and 97 ± 1% e.e.12 {[a]D24
+7.6 (c 1.2, CHCl3)} (Scheme 3).
6 V. Dambrin, M. Villiéras, P. Janvier, L. Toupet, H. Amri, J. Lebreton
and J. Villiéras, Tetrahedron, 2001, 57, 2155.
7 The concept that mutual kinetic resolution allows the true E value for a
process to be evaluated was first proposed by A. C. R. Horeau,
Tetrahedron, 1975, 31, 1307.
8 J. F. Costello, S. G. Davies and O. Ichihara, Tetrahedron: Asymmetry,
1994, 5, 3919.
9 (S)-N-Benzyl-N-a-methylbenzylamine used in the kinetic resolution
protocol was prepared by reductive amination of commercially
available a-methylbenzylamine and subsequent recrystallisation of its
HCl salt; an e.e. of > 99% was established by 1H NMR chiral shift
experiments with (S)-O-acetyl mandelic acid and comparison with an
authentic racemic standard.
10 As determined by chiral shift 1H NMR experiments using Eu(hfc)3 and
comparison with an authentic racemic sample.
11 E was calculated using the kinetic resolution calculation developed by
the Goodman group in Cambridge (see http://www.ch.cam.ac.uk/
magnus/); for further information see J. M. Goodman, A.-M. Köhler and
S. C. M. Alderton, Tetrahedron Lett., 1999, 40, 8715.
12 As indicated by derivatisation with homochiral and racemic Mosher’s
1
acid chlorides and comparison of the H and 19F NMR spectra of the
Scheme 3 Reagents and conditions: (i) (S)-2 (1.5 eq.), (R)-6 (1.5 eq.), THF,
278 °C then NH4Cl (aq.); (ii) Pd(OH)2 on C, MeOH, H2 (5atm), rt; (iii)
DDQ (2.1 eq.), DCM–H2O (3 : 1), rt.
resulting amides.
13 The e.e. of (1R,2S,5S)-4 (88% e.e.) equals the d.e. of (1R,2S,5S,aS)-5
(88% d.e.), indicating that the minor diastereoisomer from kinetic
resolution has the (1S,2R,5R,aS) configuration.
In conclusion, the kinetic resolution of (RS)-1 has been
achieved with lithium amide (S)-2, while the parallel kinetic
resolution of (RS)-1 has been achieved using a pseudoenantio-
meric mixture of lithium amides (S)-2 and (R)-6. High levels of
selectivity are observed in these resolution reactions, which
simultaneously generate the C(1) and C(2) stereogenic centres
of the b-amino ester addition product stereoselectively while
kinetically resolving the C(5)-stereogenic centre. The parallel
kinetic resolution protocol allows for the efficient synthesis of
each enantiomer of (1R,2S,5S)- and (1S,2R,5R)-5-tert-butyl-
cispentacin in a single reaction, and in higher e.e. than that
afforded by standard kinetic resolution. The use of parallel
kinetic resolution in this manner (E > 65 in each case) is
equivalent to a simple kinetic resolution with an apparent E of
up to 1000 (99% e.e. at 50% conversion), and effectively
introduces an automatic barrier for each resolution at the
optimum 50% conversion. The generality of this highly
selective protocol has been demonstrated further for a range of
14 For our initial work on parallel kinetic resolution see S. C. Preston, D.
Phil Thesis, University of Oxford, 1989: for other examples of parallel
kinetic resolution see; E. Vedejs and X. Chen, J. Am. Chem. Soc., 1997,
119, 2584; T. M. Pedersen, J. F. Jensen, R. E. Humble, T. Rein, D.
Tanner, K. Bodmann and O. Reiser, Org. Lett., 2000, 2, 535; F.
Bertozzi, P. Crotti, F. Macchia, M. Pineschi and B. L. Feringa, Angew.
Chem., Int. Ed., 2001, 40, 930; A. G. Al-Sehemi, R. S. Atkinson and C.
K. Meades, Chem. Commun., 2001, 2684.
15 In this parallel kinetic resolution reaction the e.e. of (S)-N-benzyl-N-a-
methylbenzylamine and (R)-N-3,4-dimethoxybenzyl-N-a-methylben-
zylamine was established as 98% by 1H NMR chiral shift experiments
with (R)-O-acetyl mandelic acid and comparison with an authentic
racemic standard.
16 S. G. Davies and O. Ichihara, Tetrahedron: Asymmetry, 1991, 2, 183.
17 The mutual kinetic resolution of (RS)-1 with (RS)-N-3,4-dimethox-
ybenzyl-N-a-methylbenzylamide gave (1RS,2SR,5SR,aSR)-7 in 99 ±
1% d.e. (E > 99).
18 The relative configuration within b-amino ester (1S,2R,5R,aR)-7 was
determined by 1H NOE difference analysis.
CHEM. COMMUN., 2003, 2410–2411
2411