Kinetic resolution of racemic pyrrolidine-2,5-diones using chiral oxazaborolidine catalysts

Article abstract of DOI:10.1039/b800510a

Kinetic resolution of racemic C-3 substituted pyrrolidine-2,5-diones has been achieved for the first time using highly efficient oxazaborolidine catalysts derived from cis-1-amino-indan-2-ol. The Royal Society of Chemistry.

Full text of DOI:10.1039/b800510a

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Kinetic resolution of racemic pyrrolidine-2,5-diones using chiral
oxazaborolidine catalystsw
Mike D. Barker,a Rachel A. Dixon,b Simon Jones*c and Barrie J. Marshc
Received (in Cambridge, UK) 11th January 2008, Accepted 15th February 2008
First published as an Advance Article on the web 11th March 2008
DOI: 10.1039/b800510a
Kinetic resolution of racemic C-3 substituted pyrrolidine-2,5-
diones has been achieved for the first time using highly efficient
oxazaborolidine catalysts derived from cis-1-amino-indan-2-ol.
carbonyl group. In preliminary experiments, oxazaborolidine
3 was employed as the catalyst with BH3ꢀTHF as the hydride
source, followed by reduction of the initially formed hydro-
xylactam product to the corresponding g-lactam for ease of
analysis (Table 1). All imides were prepared by existing
literature procedures.w
Asymmetric catalysis is perhaps one the most useful methods
available to the modern organic chemist, allowing the trans-
formation of an achiral material into an enantioenriched one.1
Chiral catalysts have been developed for use in various types
of organic transformations, with one of the most widely used
for the asymmetric reduction of prochiral ketones being chiral
oxazaborolidines, initially reported by Itsuno et al. then
championed by Corey and co-workers.2 Oxazaborolidines
have also been employed in the kinetic resolution of racemic
esters giving enantioenriched products.3
In all cases, full regiocontrol of the reaction was achieved,
reducing only the carbonyl at the C-5 position (Table 1, entries
1–7). The absence of any of the C-2 reduced species was
attributed to steric interactions between the catalyst and the
substituent at the C-3 position, making binding between the
C-2 carbonyl and the catalyst unfavourable. This selectivity
raised two questions: first, would the interaction between the
bound catalyst and substrate stereogenic centre be sufficient to
allow a stereoselective process to occur, and second, since the
steric environment around the C-5 carbonyl appeared to be
relatively uncongested, would the catalyst turn-over too
quickly with the usual loading of 10 mol%? The latter point
was addressed by careful optimisation of the reaction condi-
tions, using as little as 0.5 mol% of catalyst 2 or 3, BH3ꢀTHF
as reductant, at 0 1C for 120 min using catalyst 2 and for
60 min with catalyst 3.z Using these conditions, approximately
21–52% of the substrate was converted into product, depend-
ing upon the substituents present.
Our group has used various B-substituted oxazaborolidines
1 (Fig. 1), derived from cis-1-amino-indan-2-ol, as chiral
catalysts for the asymmetric reduction of prochiral ketones.4
Subsequently, various N-substituted oxazaborolidines 1–3
were probed as catalysts for the desymmetrisation of meso-
imides, demonstrating that unsubstituted oxazaborolidines 2
and 3 were the most efficient catalysts for this transformation.5
More recently, we have demonstrated the crucial choice of the
nitrogen substituent of the imide in obtaining high levels of
enantioselectivity in this transformation.6 Herein, we would
like to report the unprecedented use of B-substituted oxaza-
borolidines as catalysts for the kinetic resolution of racemic
C-3 substituted pyrrolidine-2,5-diones.
Applying these optimised conditions, the issue of enantio-
selectivity in the kinetic resolution of various C-3 substituted
pyrrolidine-2,5-diones was assessed. Results showed that by
increasing steric bulk at the C-3 position, the level of the
selectivity in the reaction was greatly enhanced. When a
methyl group was used selectivity was poor, giving selectivity
or s factors of between 1 and 2 (Table 2, entries 1 and 2). When
the substituent size was increased to phenyl or 3,5-dimethyl
phenyl, the level of selectivity increased to between 3.7 and 4.1
(Table 2, entries 3–6). Further increases were observed when
o-tolyl or 1-naphthyl substituents were employed, giving
Speckamp and co-workers showed that regioselectivity in
the reduction of imide species is controlled by the size and
electronic nature of the substituent at the C-3 position when
nucleophilic hydride sources are employed as the reducing
agent.7 The major isomer formed in many cases was the
hydroxylactam where reduction had occurred at the C-2
carbonyl, i.e. proximal to the bulky group. Since oxazaboro-
lidine catalysts function by pre-complexation followed by
intramolecular hydride delivery, it was initially unclear
whether reduction would take place at the proximal or distal
a GlaxoSmithKline Medicines Research Centre, Gunnels Wood Road,
Stevenage, Hertfordshire, UK SG1 2NY
b School of Natural Sciences (Chemistry), University of Newcastle
upon Tyne, Bedson Building, Newcastle upon Tyne, UK NE1 7RU
c Department of Chemistry, University of Sheffield, Dainton Building,
Brook Hill, Sheffield, UK S3 7HF. E-mail:
simon.jones@sheffield.ac.uk; Fax: +44 114 222 9346; Tel: +44 114
222 9483
w Electronic supplementary information (ESI) available: Copies of 1H
NMR spectra, full experimental details and characterisation. See DOI:
10.1039/b800510a
Fig. 1 cis-1-Amino-indan-2-ol oxazaborolidine catalysts.
ꢁc
This journal is The Royal Society of Chemistry 2008
2218 | Chem. Commun., 2008, 2218–2220

View Article Online
Table 1 Regioselective reduction of various C-3 substituted pyrroli-
dine-2,5-diones
Entry
Imide
R1
5 : 6a
Yield 5 (%)
Scheme 1 Kinetic resolution of imide 4h with oxazaborolidine 3.
1
2
3
4
5
6
7
a
4a
4b
4c
4d
4e
4f
Me
Ph
3,5-Me2C6H3
o-Tolyl
t-Bu
1-Napthyl
Mesityl
1
100 : 0
100 : 0
100 : 0
100 : 0
100 : 0
100 : 0
100 : 0
17
34
29
45
37
48
49
C–C bond, molecular modelling calculations predicting an
energy of rotation around the C–Ph bond of 17 kcal molꢂ1 8
.
This is significantly large enough to be a barrier to rotation,
effectively reducing the size of the mesityl group to that of a
methyl resulting in the low selectivities seen.
4g
Regioselectivity determined by H NMR spectroscopy.
The absolute stereochemistry in each case, except substrate
4e, was determined by comparison with literature data. In all
cases the (R) enantiomer was favoured, although in the case of
imide 4a the opposite sense of stereoselection was observed
due to the assignment governed by the Cahn–Ingold–Prelog
(CIP) rules.
excellent s factors of 12 and 18, respectively (Table 2, entries
7–10). The most pleasing result was seen for the t-Bu deriva-
tive, giving quite remarkable s factors of 20–23 for this
reaction (Table 2, entries 11 and 12). Surprisingly, the parti-
cularly large mesityl derivative gave very disappointing s
factors of between 1.3 and 2.5 (Table 2, entries 13 and 14).
This poor selectivity occurs since the methyl groups in the 2
and 6 positions of the phenyl ring restrict rotation around the
A working model for the selectivity observed in this reaction
has not yet been established, however, several key points
crucial to obtaining high levels of enantioselectivity have been
established. Firstly, the selectivity of the reaction increases
with the size of the C-3 substituent. In previous studies on the
desymmetrisation of meso-imides, a remarkable effect of the
imide protecting group on the stereoselectivity of the reaction
had been observed, with an N-phenyl group giving one
enantiomer of product.5 However, in the case of N-phenyl
imide 4h, the enantioselectivity dropped from an s factor of 4
to 1.2 (Scheme 1), clearly indicating that, unlike previous
examples, N-benzyl is the preferred protecting group. The role
of this group in determining the selectivity is unclear, however
it could be functioning as a chiral relay,9 controlling the
reactive conformation of the substrate.
Table 2 Kinetic resolution of various C-3 substituted pyrrolidine-2,5-
diones using oxazaborolidine catalysts 2 and 3
Additionally the original Corey–Bakshi–Shibata (CBS)
catalyst was not as efficient in promoting this reaction. When
used to resolve imide 4b, the selectivity obtained was particu-
larly poor, giving an s factor of 1.2 compared to 4 for the cis-
amino-indanol derived catalysts (Scheme 2). Not only does
this catalyst fail to give acceptable selectivity, but it also
Entry R1
Catalysta Conversion (%)b ee (%)c s factor
1
2
Me
Me
2
3
2
3
2
3
2
3
2
3
2
3
2
3
46
32
52
36
48
39
23
43
21
30
38
35
31d
34d
16
11
45
28
41
32
24
57
23
37
52
47
16
5
1.8
1.8
3.7
3.9
3.8
4.1
11.6
12.7
17.3
19.6
20.4
23.4
2.5
3
4
Ph
Ph
proceeds at
a much slower rate, providing only 10%
5
6
7
8
3,5-Me2C6H3
3,5-Me2C6H3
o-Tolyl
o-Tolyl
1-Naphthyl
1-Naphthyl
t-Bu
t-Bu
Mesityl
Mesityl
9
10
11
12
13
14
a
1.3
0.5 mol% catalyst was used in each case. Reactions performed with
catalyst 2 were carried out for 120 min, while those with catalyst 3 for
b
60 min. Conversion determined via 1H NMR spectroscopy using
c
maleimide as an external standard. ee determined via chiral phase
HPLC. Conversion determined via 1H NMR spectroscopy using
N-benzotriazole as an external standard.
d
Scheme 2 Kinetic resolution of imide 4b using CBS oxazaborolidine.
Chem. Commun., 2008, 2218–2220 | 2219
ꢁc
This journal is The Royal Society of Chemistry 2008

View Article Online
the combined organic extracts dried over MgSO4. The solvent was
removed under vacuum to give a crude mixture. After analysis, the
crude mixture was purified by column chromatography, eluting with
25% EtOAc–petroleum ether.
conversion after 2 h compared to 52% for the cis-amino-
indanol analogue.
In conclusion we have shown oxazaborolidine catalysts 2
and 3 to be extremely effective catalysts at extremely low
loadings for the kinetic resolution of C-3 substituted pyrroli-
dine-2,5-dione species giving outstanding regio- and stereo-
control, complementing existing methods for the synthesis of
related targets.10 The origin of the enantioselectivity is still
unclear but relies on the bulk of the C-3 substituent, the nature
of the N-protecting group and the catalyst structure. Further
work aimed at expanding the substrate scope and further
elucidating the reaction mechanism is currently underway.
We thank the EPSRC (RAD & BJM) and GlaxoSmithKline
(BJM) for funding, and Shasun-Pharma Solutions for the
donation of (1R,2S)-cis-1-amino-indan-2-ol.
1 R. Noyori, Asymmetric Catalysis in Organic Synthesis, Wiley, UK,
1994.
2 A. Hirao, S. Itsuno, S. Nakahama and N. Yamazaki, J. Chem.
Soc., Chem. Commun., 1981, 315; E. J. Corey, K. R. Bakshi and S.
Shibata, J. Am. Chem. Soc., 1987, 109, 5551.
3 G. Bringmann, T. Pabst, P. Henschel, J. Kraus, K. Peters, E. M.
Peters, D. S. Rycroft and J. D. Connolly, J. Am. Chem. Soc., 2000,
122, 9127.
4 S. Jones and J. C. C. Atherton, Tetrahedron: Asymmetry, 2000, 11,
4543; S. Jones and N. J. Gilmore, Tetrahedron: Asymmetry, 2003,
14, 2115; S. Jones, N. J. Gilmore and M. P. Muldowney, Org.
Lett., 2004, 6, 2805.
5 S. Jones and R. A. Dixon, Tetrahedron: Asymmetry, 2002, 13,
1115.
6 S. Jones, R. A. Dixon, M. D. Barker and B. J. Marsh, Tetrahedron,
2006, 62, 11663.
Notes and references
7 J. B. P. A. Wijnberg, H. E. Schoemaker and W. N. Speckamp,
Tetrahedron, 1978, 34, 179; J. B. P. A. Wijnberg, H. E. Schoemaker
and W. N. Speckamp, Tetrahedron Lett., 1974, 46, 4073.
8 Calculations were performed using CaChe WorkSystem Pro Ver-
sion 6.1.12.33, r Fujitsu Ltd 2000–2004. The rotational energy
barrier was calculated from an energy map generated by minimisa-
tion of conformers from searching the dihedral angle in increments
of 151.
z Procedure for experiments performed as in Table 2: (1R,2S)-cis-1-
amino-indan-2-ol (149 mg, 1 mmol) was suspended in dry THF (5 cm3)
and treated with trimethylborate (88 ml, 1 mmol). After stirring at
room temperature for 30 min, THF (15 cm3) was added to give a
known concentration of catalyst 3 which was used immediately. A
solution of the imide 4a–4g (1 mmol) in THF (8 cm3) at 0 1C was
added to the oxazaborolidine (0.1 cm3, 0.5 mol%) and BH3ꢀTHF (1M
in THF, 1 mmol). The solution was allowed to stir at 0 1C for 60 min,
then quenched by addition of MeOH (5 cm3) and 1M HCl (5 cm3). The
product was then extracted with CH2Cl2 (1 ꢃ 10 cm3, 3 ꢃ 5 cm3) and
9 S. D. Bull, S. G. Davies, S. W. Epstein and J. V. A. Ouzman,
Chem. Commun., 1998, 659.
10 R. Shintani, W.-L. Duan, T. Nagano, A. Okada and T. Hayashi,
Angew. Chem., Int. Ed., 2005, 44, 4611.
ꢁc
This journal is The Royal Society of Chemistry 2008
2220 | Chem. Commun., 2008, 2218–2220