Dynamic thermodynamic and dynamic kinetic resolution of 2-lithiopyrrolidines

Article abstract of DOI:10.1021/ja061963m

Dynamic resolution has been studied as a method for the asymmetric synthesis of 2-substituted pyrrolidines. Highly enantioselective electrophilic substitutions of racemic 2-lithiopyrrolidines in the presence of a chiral ligand have been achieved. The organolithium compounds were prepared by tin-lithium exchange from the corresponding tributylstannanes and n-butyllithium or by deprotonation of N-(tert-butyloxycarbonyl)-pyrrolidine with sec-butyllithium. A range of N-substituents and chiral ligands were investigated for the dynamic resolution. Electrophilic quench of the resolved diastereomeric 2-lithiopyrrolidine-chiral ligand complexes provided the enantiomerically enriched 2-substituted pyrrolidines. With N-alkyl derivatives, the resolution occurs conveniently at (or just below) room temperature and either enantiomer of the product can be formed by appropriate choice of the chiral ligand. The asymmetric induction occurs as a result of a thermodynamic preference for one of the diastereomeric complexes. The minor complex was found to have a faster rate of reaction with the electrophile. The use of N-allylic derivatives provides a means to prepare the N-unsubstituted pyrrolidine products. Best results were obtained with the N-2,3-dimethylbut-2-enyl derivative, and this N-substituent could be cleaved using 1-chloroethyl chloroformate. With N-Boc-2- lithiopyrrolidine, the enantioselectivity arises by a kinetic resolution and high levels of asymmetric induction in the presence of excess n-butyllithium can be obtained. Dynamic kinetic resolution of the N-Boc derivative is limited in the scope of electrophile that can be used.

Full text of DOI:10.1021/ja061963m

Published on Web 07/27/2006
Dynamic Thermodynamic and Dynamic Kinetic Resolution of
2-Lithiopyrrolidines
Iain Coldham,* Samuel Dufour, Thomas F. N. Haxell, Jignesh J. Patel, and
Graciela Sanchez-Jimenez
Contribution from the Department of Chemistry, UniVersity of Sheffield, Sheffield S3 7HF, U.K.
Received March 22, 2006; E-mail: i.coldham@sheffield.ac.uk
Abstract: Dynamic resolution has been studied as a method for the asymmetric synthesis of 2-substituted
pyrrolidines. Highly enantioselective electrophilic substitutions of racemic 2-lithiopyrrolidines in the presence
of a chiral ligand have been achieved. The organolithium compounds were prepared by tin-lithium exchange
from the corresponding tributylstannanes and n-butyllithium or by deprotonation of N-(tert-butyloxycarbonyl)-
pyrrolidine with sec-butyllithium. A range of N-substituents and chiral ligands were investigated for the
dynamic resolution. Electrophilic quench of the resolved diastereomeric 2-lithiopyrrolidine-chiral ligand
complexes provided the enantiomerically enriched 2-substituted pyrrolidines. With N-alkyl derivatives, the
resolution occurs conveniently at (or just below) room temperature and either enantiomer of the product
can be formed by appropriate choice of the chiral ligand. The asymmetric induction occurs as a result of
a thermodynamic preference for one of the diastereomeric complexes. The minor complex was found to
have a faster rate of reaction with the electrophile. The use of N-allylic derivatives provides a means to
prepare the N-unsubstituted pyrrolidine products. Best results were obtained with the N-2,3-dimethylbut-
2-enyl derivative, and this N-substituent could be cleaved using 1-chloroethyl chloroformate. With N-Boc-
2-lithiopyrrolidine, the enantioselectivity arises by a kinetic resolution and high levels of asymmetric induction
in the presence of excess n-butyllithium can be obtained. Dynamic kinetic resolution of the N-Boc derivative
is limited in the scope of electrophile that can be used.
Scheme 1
Introduction
Organolithiums are arguably the most widely used organo-
metallic reagents in the whole of synthetic chemistry. With the
steady increase in importance of stereoselective synthesis over
the past few decades, it is not surprising that there is a significant
and growing interest in chiral organolithium species (sp3-
hybridized carbanion center).1 There are two methods by which
asymmetric induction with such species can be promoted. One
approach is to prepare the organolithium species in enantioen-
riched form, for example by an asymmetric deprotonation or a
stereospecific transmetalation (particularly tin-lithium ex-
change) under conditions that maintain the configurational
stability of the organolithium species. Stereoselective electro-
philic quench of the chiral organolithium species then provides
the enantioenriched product. Alternatively, an asymmetric
substitution of the organolithium in the presence of some
stereochemical bias, such as a chiral ligand (L*) or chiral
auxiliary, can be effective. This latter method can make use of
a chiral organolithium that is racemic (or enantioenriched) and
which ideally is configurationally unstable, such that dynamic
resolution occurs (Scheme 1; k1, k2 * 0). Such dynamic
resolution and asymmetric induction on quench with an elec-
trophile (E+) can result either from a thermodynamic preference
for one of the two diastereomeric organolithium-chiral ligand
complexes (referred to as dynamic thermodynamic resolution),2
which is followed by stereoselective quench with the electrophile
(with retention or inversion of configuration at the carbanion
center), or from a kinetic preference (k3 * k4 or k3′ * k4′) for
reaction of one of these complexes with the electrophile (referred
to as dynamic kinetic resolution). In dynamic thermodynamic
resolution, electrophilic quench is faster than interconversion
of the diastereomeric complexes (k3, k4 > k1, k2), although there
can still be a kinetic preference for reaction of either of the two
complexes. In dynamic kinetic resolution, electrophilic quench
is slower than interconversion of the diastereomeric complexes
(1) (a) Basu, A.; Thayumanavan, S. Angew. Chem., Int. Ed. 2002, 41, 716. (b)
Clayden, J. Organolithiums: SelectiVity for Synthesis; Pergamon: Am-
sterdam, 2002. (c) Organolithiums in EnantioselectiVe Synthesis; Hodgson,
D. M., Ed.; Springer-Verlag: Heidelberg, 2003. (d) Gawley, R. E.;
Coldham, I. In The Chemistry of Organolithium Compounds; Rappoport,
Z., Marek, I., Eds.; Wiley: Chichester, 2004; p 997. (e) Hoppe, D.;
Christoph, G. In The Chemistry of Organolithium Compounds; Rappoport,
Z., Marek, I., Eds.; Wiley: Chichester, 2004; p 1055.
(2) Beak, P.; Anderson, D. R.; Curtis, M. D.; Laumer, J. M.; Pippel, D. J.;
Weisenburger, G. A. Acc. Chem. Res. 2000, 34, 715.
9
10.1021/ja061963m CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 10943-10951
10943

A R T I C L E S
Coldham et al.
(k3, k4 < k1, k2). There can, of course, be a combination of both
thermodynamic and kinetic factors that contribute to the
enantiomer ratio (er) of the products, depending on the relative
rates of these processes.
Significant advances in the use of chiral organolithium species
were made with the discovery that certain dipole-stabilized
R-alkoxy- and R-amino-organolithium species could be gener-
ated enantioselectively by asymmetric deprotonation with bu-
tyllithium and (-)-sparteine as the chiral ligand.3 These
organolithium species are configurationally stable at low tem-
perature (-78 °C is normally used) and react stereoselectively
with a range of electrophiles.4,5 In contrast, R-thio- and R-seleno-
organolithium species are configurationally labile at low tem-
perature, and this has allowed asymmetric substitution by
dynamic thermodynamic or dynamic kinetic resolution, with
high selectivities being achieved particularly using chiral bis-
(oxazoline) ligands.6 Benzylic and allylic organolithium species
are also prone to racemize at low temperature, and these
organolithiums can be suitable substrates for asymmetric
substitution.2,7
Figure 1. Barrier, rate constant, and half-life of enantiomerization of
2-lithiopyrrolidines in hexanes-Et2O.8
compares with the rate of reaction with the electrophile. The
majority of reactions that make use of chiral organolithium
species are carried out at low temperature, and this has perhaps
fuelled the misconception that only allylic, benzylic, or R-thio-/
R-seleno-organolithium species can readily undergo dynamic
resolution. In reality, any chiral organolithium species can
potentially undergo equilibration, particularly on warming, and
successful dynamic resolution requires a rate of enantiomer-
ization that is faster than its rate of decomposition or other side
reactions.
Hence, although chiral R-amino-organolithium species are
configurationally stable at low temperature,5 at room temperature
enantiomerization takes place readily (as found initially during
a study of the enantioselectivity on anionic cyclization to give
indolizidines).5b To gain further understanding of the factors
affecting the enantiomerization, the barrier to inversion of
several chiral 2-lithiopyrrolidines was determined in the solvent
hexanes-Et2O (4:1) (Figure 1).8 It is clear from these data that
N-alkyl-2-lithiopyrrolidines have a slightly higher barrier to
inversion (22 kcal/mol) than the corresponding N-(tert-buty-
loxycarbonyl) (N-Boc) derivative (20 kcal/mol), and conse-
quently, a higher temperature is required to allow dynamic
equilibration of N-alkyl-2-lithiopyrrolidines. In the presence of
a chiral ligand, it should be possible to effect dynamic resolution
of such organolithiums, and we have verified that this is, indeed,
the case.9-11 Extremely high selectivities can be obtained in
To promote successful asymmetric induction with a chiral
organolithium species, it is important to have some understand-
ing of its rate of enantiomerization and, ideally, how this rate
(3) (a) Hoppe, D.; Hintze, F.; Tebben, P. Angew. Chem., Int. Ed. Engl. 1990,
29, 1422. (b) Kerrick, S. T.; Beak, P. J. Am. Chem. Soc. 1991, 113, 9708.
(c) Hoppe, D.; Hense, T. Angew. Chem., Int. Ed. Engl. 1997, 36, 2282. (d)
Beak, P.; Basu, A.; Gallagher, D. J.; Park, Y. S.; Thayumanavan, S. Acc.
Chem. Res. 1996, 29, 552.
(4) For some recent examples with R-alkoxy-organolithium species, see: (a)
Deiters, A.; Mu¨ck-Lichtenfeld, C.; Fro¨hlich, R.; Hoppe, D. Chem. Eur. J.
2002, 1833. (b) Arrasate, S.; Lete, E.; Sotomayor, N. Tetrahedron:
Asymmetry 2002, 13, 311. (c) Simov, B. P.; Rohn, A.; Brecker, L.; Giester,
G.; Hammerschmidt, F. Synthesis 2004, 2704. (d) Mart´ınez, M. M.; Hoppe,
D. Eur. J. Org. Chem. 2005, 1427. (e) Zeng, W.; Fro¨hlich, R.; Hoppe, D.
Tetrahedron 2005, 61, 3281. (f) Wu¨rthwein, E.-U.; Hoppe, D. J. Org. Chem.
2005, 70, 4443. (g) Safir, I.; Candela Lena, J. I.; Finet, L.; Birlirakis, N.;
Arseniyadis, S. Tetrahedron: Asymmetry 2005, 16, 3436. (h) U¨ naldi, S.;
Fro¨hlich, R.; Hoppe, D. Synthesis 2005, 2507.
(5) For some recent examples with R-amino-organolithium species, see: (a)
Gawley, R. E.; Low, E.; Zhang, Q.; Harris, R. J. Am. Chem. Soc. 2000,
122, 3344. (b) Ashweek, N. J.; Coldham, I.; Snowden, D. J.; Vennall, G.
P. Chem. Eur. J. 2002, 8, 195. (c) Bragg, R. A.; Clayden, J.; Menet, C. J.
Tetrahedron Lett. 2002, 43, 1955. (d) Clayden, J.; Menet, C. J.; Tcha-
banenko, K. Tetrahedron 2002, 58, 4727. (e) Ncube, A.; Park, S. B.; Chong,
J. M. J. Org. Chem. 2002, 67, 3625. (f) Gibson, S. E.; Smith, M. H. Org.
Biomol. Chem. 2003, 1, 676. (g) Tomoyasu, T.; Tomooka, K.; Nakai, T.
Tetrahedron Lett. 2003, 44, 1239. (h) Hermet, J.-P. R.; Porter, D. W.;
Dearden, M. J.; Harrison, J. R.; Koplin, T.; O’Brien, P.; Parmene, J.; Tyurin,
V.; Whitwood, A. C.; Gilday, J. Smith, N. M. Org. Biomol. Chem. 2003,
1, 3977. (i) Kells, K. W.; Ncube, A.; Chong, J. M. Tetrahedron 2004, 60,
2247. (j) Clayden, J.; Knowles, F. E.; Baldwin, I. R. J. Am. Chem. Soc.
2005, 127, 2412. (k) Deng, X.; Mani, N. S. Tetrahedron: Asymmetry 2005,
16, 661. (l) Kocienski, P. J.; Christopher, J. A.; Bell, R.; Otto, B. Synthesis
2005, 75. (m) Kim, D. D.; Lee, S. J.; Beak, P. J. Org. Chem. 2005, 70,
5376. (n) Morita, Y.; Tokuyama, H.; Fukuyama, T. Org. Lett. 2005, 7,
4337. (o) McGrath, M. J.; O’Brien, P. J. Am. Chem. Soc. 2005, 127, 16378.
(p) Ferna´ndez, I.; Go´mez, G. R.; Alfonso, I.; Iglesias, M. J.; Ortiz, F. L.
Chem. Commun. 2005, 5408. (q) Patwardhan, A. P.; Pulgam, V. R.; Zhang,
Y.; Wulff, W. D. Angew. Chem., Int. Ed. 2005, 44, 6169. (r) Campos, K.
R.; Klapars, A.; Waldman, J. H.; Dormer, P. G.; Chen, C. J. Am. Chem.
Soc. 2006, 128, 3538. See also: (s) Dieter, R. K.; Chen, N.; Watson, R. T.
Tetrahedron 2005, 61, 3221. (t) Dieter, R. K.; Oba, G.; Chandupatla, K.
R.; Topping, C. M.; Lu, K.; Watson, R. T. J. Org. Chem. 2004, 69, 3076.
(u) Watson, R. T.; Gore, V. K.; Chandupatla, K. R.; Dieter, R. K.; Snyder,
J. P. J. Org. Chem. 2004, 69, 6105.
(6) (a) Kaiser, B.; Hoppe, D. Angew. Chem., Int. Ed. Engl. 1995, 34, 323. (b)
Hoffmann, R. W.; Klute, W.; Dress, R. K.; Wenzel, A. J. Chem. Soc.,
Perkin Trans. 2 1995, 1721. (c) Hoffmann, R. W.; Klute, W. Chem. Eur.
J. 1996, 2, 694. (d) Nakamura, S.; Nakagawa, R.; Watanabe, Y.; Toru, T.
J. Am. Chem. Soc. 2000, 122, 11340. (e) Nakamura, S.; Ito, Y.; Wang, L.;
Toru, T. J. Org. Chem. 2004, 69, 1581. (f) Nakamura, S.; Aoki, T.; Ogura,
T.; Wang, L.; Toru, T. J. Org. Chem. 2004, 69, 8916. See also: (g) Gaul,
C.; Scha¨rer, K.; Seebach, D. J. Org. Chem. 2001, 66, 3059. (h) Gaul, C.;
Arvidsson, P. I.; Bauer, W.; Gawley, R. E.; Seebach, D. Chem. Eur. J.
2001, 7, 4117. (i) Wang, L.; Nakamura, S.; Toru, T. Org. Biomol. Chem.
2004, 2, 2168. (j) Wang, L.; Nakamura, S.; Ito, Y.; Toru, T. Tetrahedron:
Asymmetry 2004, 15, 3059. (k) Otte, R.; Fro¨hlich, R.; Wibbeling, B.; Hoppe,
D. Angew. Chem., Int. Ed. 2005, 44, 5492.
(7) (a) Tomooka, K.; Wang, L.-F.; Okazaki, F.; Nakai, T. Tetrahedron Lett.
2000, 41, 6121. (b) Pippel, D. J.; Weisenburger, G. A.; Faibish, N. C.;
Beak, P. J. Am. Chem. Soc. 2001, 123, 4919. (c) Johnson, T. A.; Jang, D.
O.; Slafer, B. W.; Curtis, M. D.; Beak, P. J. Am. Chem. Soc. 2002, 124,
11689. (d) Whisler, M. C.; Beak, P. J. Org. Chem. 2003, 68, 1207. (e)
Stratmann, O.; Kaiser, B.; Fro¨hlich, R.; Meyer, O.; Hoppe, D. Chem. Eur.
J. 2001, 7, 423. (f) Marr, F.; Fro¨hlich, R.; Wibbeling, B.; Diedrich, C.;
Hoppe, D. Eur. J. Org. Chem. 2002, 2970. (g) Garc´ıa-Ruano, J. L.; Aranda,
M. T.; Aguirre, J. M. Tetrahedron 2004, 60, 5383. (h) Clayden, J.; Johnson,
P.; Pink, J. H.; Helliwell, M. J. Org. Chem. 2000, 65, 7033. (i) Derdau,
V.; Snieckus, V. J. Org. Chem. 2001, 66, 1992. (j) Kimachi, T.; Takemoto,
Y. J. Org. Chem. 2001, 66, 2700. (k) Enders, D.; Vignola, N.; Berner, O.
M.; Harnying, W. Tetrahedron 2005, 61, 3231. (l) Adam, S.; Pannecoucke,
X.; Combret, J.-C.; Quirion, J.-C. J. Org. Chem. 2001, 66, 8744. (m)
Clayden, J.; Helliwell, M.; Pink, J. H.; Westlund, N. J. Am. Chem. Soc.
2001, 123, 12449. (n) Arrasate, S.; Lete, E.; Sotomayor, N. Tetrahedron:
Asymmetry 2002, 13, 311. (o) Capo´, M.; Saa´, J. M. J. Am. Chem. Soc.
2004, 126, 16738. (p) Azzena, U. J. Chem. Soc., Perkin Trans. 1 2002,
360. (q) Pe´rard-Viret, J.; Prange´, T.; Tomas, A.; Royer, J. Tetrahedron
2002, 58, 5103. (r) Laumer, J. M.; Kim, D. D.; Beak, P. J. Org. Chem.
2002, 67, 6797. (s) Marr, F.; Fro¨hlich, R.; Hoppe, D. Tetrahedron:
Asymmetry 2002, 13, 2587. (t) Kuo, S.-C.; Chen, F.; Hou, D.; Kim-Meade,
A.; Bernard, C.; Liu, J.; Levy, S.; Wu, G. G. J. Org. Chem. 2003, 68,
4984. (u) Wilkinson, J. A.; Rossington, S. B.; Leonard, J.; Hussain, N.
Tetrahedron Lett. 2004, 45, 5481. (v) Castaldi, M. P.; Gibson, S. E.; Rudd,
M.; White, A. J. P. Angew. Chem., Int. Ed. 2005, 44, 3432. (w) Kim, D.
D.; Lee, S. J.; Beak, P. J. Org. Chem. 2005, 70, 5376.
(8) Ashweek, N. J.; Brandt, P.; Coldham, I.; Dufour, S.; Gawley, R. E.;
Haeffner, F.; Klein, R.; Sanchez-Jimenez, G. J. Am. Chem. Soc. 2005, 127,
449.
(9) Coldham, I.; Dufour, S.; Haxell, T. F. N.; Howard, S.; Vennall, G. P. Angew.
Chem., Int. Ed. 2002, 41, 3887.
(10) Coldham, I.; Dufour, S.; Haxell, T. F. N.; Vennall, G. P. Tetrahedron 2005,
61, 3205.
(11) Coldham, I.; Patel, J. J.; Sanchez-Jimenez, G. Chem. Commun. 2005, 3083.
9
10944 J. AM. CHEM. SOC. VOL. 128, NO. 33, 2006

Dynamic Resolution of 2-Lithiopyrrolidines
A R T I C L E S
Table 1. Resolution of N-Alkyl-2-lithiopyrrolidines and Quench
Scheme 2
with Trimethylsilyl Chloride
yield
19
Chart 1
chiral
er
entry
R
ligand
conditions
(%)
(R:S)
1
2
3
4
5
6
7
8
9
Me
4
4
4
5
6
7
8
9
10
11
12
13
14
15
16
16
17
hexane, 30 min
hexane, 1 h
hexane, 1 h
81 73:27
74 74:26
78 85:15
CH2CH2CH3
CHMe2
Me
Me
CHMe2
Me
hexanes-Et2O (1:7), 30 min 48 70:30
hexanes-Et2O (4:1), 30 min 50 53:47
hexanes-Et2O (4:1), 1 h
45 46:54
hexanes-Et2O (4:1), 30 min 50 50:50
CHMe2
CHMe2
hexanes-Et2O (4:1), 1 h
hexanes-Et2O (4:1), 1 h
hexanes-Et2O (4:1), 1 h
hexanes-Et2O (4:1), 1 h
hexanes-Et2O (4:1), 1 h
hexanes-Et2O (4:1), 1 h
hexanes-Et2O (4:1), 1 h
hexane, 30 min
0
60 50:50
41 50:50
41 43:57
45 39:61
38 44:56
76 58:42
10 22:78
75 3:97
42 50:50
10 CHMe2
11 CHMe2
12 CHMe2
13 CHMe2
14 CHMe2
15 Me
16 CHMe2
17 CHMe2
Et2O, 1 h
Et2O, 1 h
the asymmetric substitution of both N-alkyl- and N-Boc-2-
lithiopyrrolidines. Remarkably, reactions occur by either a
dynamic thermodynamic resolution (using the N-alkyl deriva-
tives) or a dynamic kinetic resolution (using the N-Boc
derivative). A full account of these dynamic resolutions is given
here, including results not previously reported on the influence
of a wide selection of different ligands and the extension from
N-alkyl to N-allylic derivatives that can be cleaved and thus
provide a method to access the N-unsubstituted pyrrolidine
products.
did not promote any asymmetric induction, and on addition of
the bisoxazoline 9 (entry 8), a deep yellow color was obtained
and none of the desired product 19 was formed. However, the
diamino-alcohol 16 (entry 16) was found to give excellent levels
of asymmetric induction. This reaction was best carried out in
Et2O as solvent, since the lithium alkoxide generated from
addition of n-BuLi to the ligand 16 was not very soluble in the
less polar solvent hexane. The ligand 17 (entry 17), which is
the methyl ether of ligand 16, gave no asymmetric induction.
From these studies, it was clear that the best ligand was the
diamino-alcohol 16. This ligand is commercially available but
can also be prepared readily in two steps from proline methyl
ester.14 In regard to the 2-lithiopyrrolidine substrate, on increas-
ing the steric bulk in the substituent on the nitrogen atom, slight
improvements to the level of asymmetric induction were
obtained (entries 1-3). Thus, compound 18, R ) CHMe2,
together with ligand 16 were selected to study the scope of the
reaction with a variety of electrophiles (Table 2). In each case,
the stannane 3 was treated with n-BuLi in Et2O at room
temperature for 1 h prior to addition of the ligand 16. The ligand
16 was pretreated with n-BuLi (1.2 equiv) in Et2O at -30 °C
for 10 min and then allowed to warm to room temperature. The
electrophilic quench was best carried out after the mixture was
cooled.
Results
Dynamic Resolution of N-Alkyl-2-lithiopyrrolidines. A
selection of N-alkyl-2-tributylstannylpyrrolidines 3 was prepared
from N-Boc-2-tributylstannylpyrrolidine (1).12 Removal of the
N-Boc group using B-bromocatecholborane, followed directly
by acylation, gave the amides 2, which were reduced with
LiAlH4 to give the desired stannanes 3 (Scheme 2). Different
chiral ligands (Chart 1) were probed for the subsequent
asymmetric substitution reaction, and the results of this study
are given in Table 1. Chiral ligands with an OH or NH group
were treated with 1.2 equiv of n-BuLi prior to addition to the
organolithium derived from the stannane 3 and n-BuLi.
Asymmetric substitution with the ligand (-)-sparteine (4,
entries 1-3 in Table 1) gave low to moderate levels of
selectivity.10 On increasing the time or the temperature, similar
or lower yields and selectivities were obtained. The major
enantiomer of the product 19 had the R-absolute configuration,
which contrasts with the major enantiomer obtained using (-)-
sparteine as a chiral ligand for asymmetric deprotonation of
N-Boc-pyrrolidine.13 Other diamine ligands (5-7, entries 4-6)
or â-amino-alcohol-derived ligands (10-15, entries 9-14) also
gave low levels of selectivity. The ligand BINOL (8, entry 7)
The products 19-24 were obtained with excellent levels of
enantioselectivity. The absolute configuration of the major
enantiomer was determined by comparison of the specific
rotation with that of an authentic sample, prepared from the
enantioenriched stannane 1 (see Supporting Information). These
asymmetric substitution reactions were successful with a variety
of electrophiles, including acetone (Table 2, entry 3) that bears
enolizable protons. Only with benzaldehyde (entry 5) was the
er reduced, possibly due to partial single-electron transfer.5a The
(12) Beak, P.; Lee, W. K. J. Org. Chem. 1993, 58, 1109.
(13) Beak, P.; Kerrick, S. T.; Wu, S.; Chu, J. J. Am. Chem. Soc. 1994, 116,
3231.
(14) Mukaiyama, T.; Soai, K.; Sato, T.; Shimizu, H.; Suzuki, K. J. Am. Chem.
Soc. 1979, 101, 1455.
9
J. AM. CHEM. SOC. VOL. 128, NO. 33, 2006 10945

A R T I C L E S
Coldham et al.
Table 2. Resolution of N-Isobutyl-2-lithiopyrrolidine with Ligand 16
a Cool to -70 °C prior to electrophilic quench.
Scheme 3
lectivity arise from the dynamic resolution of N-isobutyl-2-
lithiopyrrolidine, a drawback with these compounds is that the
N-isobutyl group cannot be cleaved from the 2-substituted
products. We therefore sought to find an alternative N-
substituent that allowed efficient dynamic resolution and
subsequent cleavage from the product. An initial investigation
using N-benzyl-2-lithiopyrrolidine provided a mixture of prod-
ucts containing only a small amount of the desired product,
together with the destannylated compound N-benzylpyrrolidine
and some 1,2-rearrangement product 2-benzylpyrrolidine. We
therefore turned to N-allylic derivatives. These were prepared
from N-Boc-2-tributylstannylpyrrolidine (1) by removal of the
N-Boc group (using B-bromocatecholborane) and acylation,
followed by reduction in the same way as shown above (Scheme
2, except using reduction with AlH3 formed from LiAlH4 and
AlCl3), or by removal of the N-Boc group and allylation with
the required bromide.16 Initially we prepared the stannane 26,
which was treated with n-BuLi to effect tin-lithium exchange
(Table 3). The transmetalation was carried out at 0 °C to avoid
any potential sigmatropic rearrangement at elevated tempera-
ture.16 Using the ligand 16 (deprotonated with n-BuLi prior to
addition), good levels of asymmetric induction were obtained
(Table 3, entry 1). This was a marked improvement over the
use of the ligand (-)-sparteine (4), which gave (entry 2, in
hexane rather than Et2O and after quench with Me3SiCl) the
product N-prenyl-2-trimethylsilylpyrrolidine in 69% yield and
with er 25:75 (opposite absolute configuration of the major
product with this ligand). We decided to screen some other chiral
ligands, as shown in Chart 2. High levels of selectivity were
obtained using ligand 28 (Table 3, entry 3), which is related to
16 but lacks the additional stereogenic center. However, the
ligand 29 was poorly selective (entry 4), thus demonstrating
the requirement for both nitrogen atoms in the ligand. Enhancing
the steric bulk of the ligand (using ligand 30, entry 5) did not
improve the results in comparison with those obtained with
ligand 16 (the yields of these reactions are unoptimized). Finally,
an attempt with the ligand 31 failed to give any product 27.
chiral ligand could be recovered after column chromatography
and distillation, without any loss of enantiopurity.
If the chiral ligand 16 was added to the organolithium species
after cooling to -78 °C, then electrophilic quench gave the
desired products with no enantioselectivity (er 50:50). However,
if only a small amount of the electrophile Me3SiCl was added
to the organolithium‚16 complex at -78 °C, then the product
19 (R ) CHMe2) was formed with low selectivity in favor of
the opposite major enantiomer (er 39:61, S:R). These results
suggest that the asymmetric induction arises from a dynamic
thermodynamic resolution. Hence, the ratio of enantiomers given
in Tables 1 and 2 arises from the ratio of the diastereomeric
complexes 18.
The ligand 16 contains two stereogenic centers, and it was
important to probe the extent of asymmetric induction with its
diastereoisomer. Ligand 25 was therefore prepared from D-
proline methyl ester according to the published procedure.15
Efficient dynamic resolution was achieved with this ligand,
quenching with phenyl isocyanate to give the product 20 with
high enantioselectivity (Scheme 3). The chiral ligands 16 and
25 promote the formation of different stereoisomeric products
(compare Table 2, entry 1, with Scheme 3), and the asymmetry
therefore arises from the absolute configuration of the (lithiated)
hydroxymethyl group of the ligand.
Dynamic Resolution of 2-Lithiopyrrolidines Bearing N-
Allylic Substituents. Although excellent levels of enantiose-
(15) Gallagher, D. J.; Wu, S.; Nikolic, N. A.; Beak, P. J. Org. Chem. 1995, 60,
(16) Gawley, R. E.; Zhang, Q.; Campagna, S. J. Am. Chem. Soc. 1995, 117,
8148.
11817.
9
10946 J. AM. CHEM. SOC. VOL. 128, NO. 33, 2006

Dynamic Resolution of 2-Lithiopyrrolidines
A R T I C L E S
Table 3. Resolution of N-Prenyl-2-lithiopyrrolidine
Scheme 4
Scheme 5
chiral
liganda
yield 27
er
entry
(%)
(R:S)
1
2
3
4
5
6
16
4
28
29
30
31
72
69b
40
35
20
0
87:13
25:75
83:17
47:53
83:17
a The lithium alkoxides of ligands 16 and 28-31 were used. b Quench
with Me3SiCl.
Chart 2
products of decomposition of BuLi enhance the rate of trans-
metalation. The dynamic resolution was effective for a selection
of electrophiles, as shown in Table 4. Using PhNCO gave the
product (R)-33, E ) CONHPh, and the enantiomer ratio was
determined by chiral HPLC (see Supporting Information). The
enantiomer ratio of the product 33, E ) SiMe3, from quench
with Me3SiCl was determined by removal of the N-allyl group
(see below), protection with Boc2O, and chiral GC. The
electrophile Me2SO4 gave the product 33, E ) Me (together
with some protodestannylated product), and the electrophile MeI
gave none of the desired product. The enantiomer ratio of the
product 33, E ) Me, was determined by chiral GC analysis.
The electrophile cyclohexanone (Table 4, entry 5) gave the
product 33, E ) C(OH)C5H10, and the enantiomer ratio was
determined by chiral GC analysis. Use of the electrophile Bu3-
SnCl was successful (yield of 33, E ) SnBu3 ) 32 was 60%),
but a reduced enantiomer ratio (er R:S 23:77 as judged by
specific rotation) was obtained due to the presence of residual
starting material 32 that had resisted transmetalation; by allowing
a longer time (3 h) for transmetalation at higher temperature
(-5 °C), some improvement in the enantiomer ratio was
obtained (er R:S 14:86, yield 56%).
Table 4. Resolution of N-Trimethylallyl-2-lithiopyrrolidine
yield 33
er
entry
E+
E
(%)
(R:S)
1
PhNCO
PhNCO
Me3SiCl
Me2SO4
cC6H10O
CONHPh
CONHPh
SiMe3
Me
C(OH)C5H10
51
70
62
59
48
96:4
96:4
5:95
6:94
95:5
2a
3
4
5
a No TMEDA, but 2.7 equiv of aged BuLi and 1.5 equiv of 16 were
used.
The levels of asymmetric induction obtained using the
N-prenyl compound 26, although good, were not as high as that
obtained using the N-isobutyl compound 3 (R ) CHMe2). We
therefore prepared the branched compound 32, which is more
structurally related to compound 3 (R ) CHMe2). The stannane
32 was reluctant to undergo transmetalation, but this could be
induced in the presence of the additive TMEDA (1.2 equiv, 1
h). This concerned us somewhat, as the ligand TMEDA could
potentially compete with the chiral ligand for complexation to
the organolithium (18). However, we were pleased to find that
excellent levels of asymmetric induction were obtained using
the chiral ligand 16 (Table 4; ligand 16 was deprotonated with
n-BuLi in Et2O prior to addition). The best results were obtained
when the mixture was allowed sufficient time (>1 h) before
quenching with the electrophile. This should allow the formation
of the thermodynamic ratio of diastereomeric complexes.
Transmetalation is slow in the absence of TMEDA but improves
when using n-BuLi that has aged. This result suggests that the
The same transformation using the diastereomeric ligand 25,
quenching with phenyl isocyanate, gave the product 33, E )
CONHPh, with high enantioselectivity (Scheme 4).
Deallylation of the products 33 was investigated under a
selection of different conditions. Although the N-prenyl group
could be cleaved using Pd(dba)2, dppb, and thiosalicylic acid
in THF, these conditions were unsuccessful for the more
hindered compound 33. However, the desired N-unsubstituted
products 34 could be obtained by heating with R-chloroethyl
chloroformate in PhMe, followed by heating in MeOH (Scheme
5). Both (R)- and (S)-34 could be obtained in this way. The
absolute configurations of these compounds were verified from
their specific rotations, which matched those in the literature.17
9
J. AM. CHEM. SOC. VOL. 128, NO. 33, 2006 10947

A R T I C L E S
Scheme 6
Coldham et al.
Scheme 8
Scheme 9
Scheme 7
SiCl. However, very good er’s could be obtained when a
substoichiometric amount of Me3SiCl was added or when Me3-
SiCl was added slowly at -20 °C. This was particularly so when
excess n-BuLi was present. For example, formation of the
organolithium in the presence of 1.5 equiv of the chiral ligand
16 and an overall excess of 10 equiv of n-BuLi gave, after
quenching with 5 equiv of Me3SiCl, the product 35 in 42% yield
and with er 98:2 (S:R). The use of 10 equiv of excess n-BuLi
gave better selectivities than 3.25 or 6.25 equiv of excess
n-BuLi. The high er must be a reflection of the improved relative
rate of reaction between the diastereomeric complexes. To
improve the yield in this kinetic resolution, we needed to carry
out the electrophilic quench at a temperature that allows dynamic
equilibration. Using the stannane 1 and the chiral ligand 16 as
above, with 10 equiv of excess n-BuLi, followed by warming
to -20 °C and slow addition of Me3SiCl (16 equiv) over 90
min at this temperature, gave the product (S)-35 with er 96:4
(Scheme 8). Likewise, the use of the diastereomeric ligand 25
gave the product (R)-35 with er 92:8. In the absence of excess
n-BuLi, poor er’s were obtained.11
The same procedure was applied to the silane 33, E ) SiMe3,
to give, after protection with Boc2O, the product 35 (Scheme
6).
Dynamic Resolution of N-Boc-2-lithiopyrrolidine. The
kinetic data shown in Figure 1 indicate that it should be possible
to effect dynamic resolution of the organolithium species derived
from the N-Boc compound 1. As the stannanes used in our
dynamic resolutions described above are prepared from com-
pound 1, such chemistry may avoid the need to convert
compound 1 to the N-alkyl or N-allyl derivatives. Formation of
N-Boc-2-lithiopyrrolidine from compound 1 (or from deproto-
nation of N-Boc-pyrrolidine) is well known, and although this
organolithium maintains its configurational stability at -78 °C
for several hours, it is known to lose optical purity at -40 °C.13
Beak and co-workers reported that no asymmetric substitution
of this organolithium occurs in the presence of (-)-sparteine
(4) as a chiral ligand.13 This experiment is reported to have been
conducted at low temperature with excess Me3SiCl to quench
the organolithium, and under such conditions no resolution
would be observed. We therefore probed this reaction and found
that the product N-Boc-2-trimethylsilylpyrrolidine (35) is formed
as a racemic mixture by using higher temperatures (e.g. -20
or 0 °C that should allow the dynamic equilibration) or
substoichiometric amounts of Me3SiCl as the electrophile. The
chiral ligand (-)-sparteine (4) is therefore not suitable to
promote dynamic resolution of N-Boc-2-lithiopyrrolidine. De-
spite this, we thought that it would be appropriate to test whether
the ligand 16 could be used in this chemistry.
First, we tested the chiral ligand 16 (pretreated with n-BuLi)
at -20 °C for 20 min, followed by cooling to -78 °C and
quenching with Me3SiCl. The product 35 was isolated as a
mixture of enantiomers in a ratio 42:58, with the (R)-enantiomer
as the major absolute configuration (determined by comparison
with an authentic sample) (Scheme 7).13 Longer reaction times
or elevated temperature gave similar results. However, quench-
ing with 0.4 equiv of Me3SiCl gave the product 35 with er 81:
19 in favor of the (S)-enantiomer. It appears, therefore, that the
chiral ligand complex with the (2S)-N-Boc-2-lithiopyrrolidine
reacts faster than that with the (2R)-enantiomer.
As an alternative method that avoids tin-lithium exchange,
N-Boc-pyrrolidine 36 was treated with sec-BuLi (2.6 equiv) and
the chiral ligand 16 (1.5 equiv) in Et2O at -78 °C. After 6 h,
10 equiv of n-BuLi was added. Slow addition of Me3SiCl (16
equiv) at -20 °C over 30 min gave the product (S)-35 with er
95:5 (Scheme 9). Likewise, the use of the diastereomeric ligand
25 gave the product (R)-35 (er 9:91). In this case, it was
preferable to complete the electrophilic quench over about 30
rather than 90 min to minimize the formation of the byproduct
Bu2CO, presumably formed from attack of the excess n-BuLi
with the Boc group.
The transformation given in Scheme 9 is actually an asym-
metric deprotonation at -78 °C, followed by equilibration at
-20 °C. Indeed, Beak and co-workers have described this
asymmetric deprotonation with the ligand 16 and quench at low
temperature to give product 35 (er 87:13 S:R).15 To verify that
our enhanced enantioselectivity arises from a kinetic resolution,
we carried out the following three experiments, all using
enantioenriched stannane 1 (er 97:3 S:R) and the chiral ligand
25, representing the “mismatched” arrangement (Scheme 10).
Allowing a short period of time at -20 °C before cooling and
quenching with Me3SiCl gave the product 35, predominantly
as the (S)-enantiomer (er 84:16), indicating that there was
insufficient time for equilibration. However, after 60 min at -20
°C, the enantiomer ratio had changed to 21:79 in favor of the
(R)-enantiomer. Crucially, allowing just 20 min at -20 °C
Attempts were made to optimize this er by altering the
solvent, concentration, amount of n-BuLi, temperature, and rate
of addition of Me3SiCl. In all cases, the er was poor when the
mixture was cooled to -78 °C and quenched with excess Me3-
(17) Mukaiyama, T. Tetrahedron 1981, 37, 4111.
9
10948 J. AM. CHEM. SOC. VOL. 128, NO. 33, 2006

Dynamic Resolution of 2-Lithiopyrrolidines
A R T I C L E S
Scheme 10
before quenching the organolithium slowly over 30 min at -20
°C, the product 35 was formed with er 19:81 in favor of the
(R)-enantiomer. Hence, despite the fact that only a small amount
of the (R)-enantiomer (of the organolithium) is present under
these conditions, the (R)-enantiomer of the product is produced
predominantly.
Figure 2. Assumed energy diagram for dynamic thermodynamic resolution
of N-alkyl- and N-allyl-2-lithiopyrrolidines.
As the products from dynamic resolution and electrophilic
quench of the N-allylic substrate 32 are formed in good yield
and high enantiomeric excess, and the N-substituent can be
cleaved from the product, this methodology is favored over the
use of the N-alkyl substrates 3. The chemistry provides access
to either enantiomer of the 2-substituted pyrrolidine products,
depending on the choice of the chiral ligand.
The optimized method to allow dynamic kinetic resolution
is therefore to deprotonate N-Boc-pyrrolidine 36 with sec-BuLi
and the chiral ligand (16 or 25) and then add excess n-BuLi,
warm to -20 °C, and quench slowly. This is effective for
electrophilic quench with Me3SiCl but was unsuccessful with
other electrophiles, such as acetone and PhNCO. The electro-
phile MeI did give the desired product, but the yield and
enantioselectivity were moderate. The electrophile Bu3SnCl gave
racemic organostannane 1.
In an attempt to probe the structure and relative stability of
these organolithium species, we carried out ab initio calculations
on the diastereomeric complexes. All calculations were per-
formed using the Gaussian03 program package.18 Different basis
sets were used within the framework of density functional theory
using the B3LYP functional.19,20 However, all calculations of
the organolithium (with different N-alkyl or N-allyl substituents)
complexed with ligand 16 resulted in similar energies between
the diastereomeric complexes. Further calculations, including
solvent interactions, indicated that there is a significant effect
of binding Et2O to the vacant Li coordination site. Recalculating
the energies of our structures using second-order Møller-Plesset
perturbation theory within the counterpoise framework21 did give
relative energies that were in good agreement with experiment
in some cases (see Supporting Information for details). However,
in these cases the diastereomer predicted to be more stable was
the one that requires an inversion of configuration at the
carbanion center on going from the organolithium to the product.
This is less likely than a retentive electrophilic quench but cannot
be ruled out with certainty, as the absolute configuration of the
N-alkyl-2-lithiopyrrolidines complexed with the chiral ligand
is unknown. It is widely accepted that tin-lithium exchange
occurs with retention of configuration but that subsequent
electrophilic quench can occur with retention, inversion, or
racemization.1 As we can prepare the enantioenriched stannane
32, it is easy to confirm that tin-lithium exchange, followed
by electrophilic quench, occurs with overall retention of
configuration. For example, treatment of the enantioenriched
stannane (S)-32 [absolute configuration derived from (S)-1]13,22
(er 97:3) with BuLi/TMEDA at -10 °C for 90 min (to allow
transmetalation without complete racemization), followed by
cooling to -78 °C and quenching with PhNCO, gave the amide
Discussion
Dynamic Resolution of N-Alkyl- and N-Allyl-2-lithiopy-
rrolidines. Excellent levels of asymmetric induction are possible
from 2-lithiopyrrolidines bearing N-isobutyl- and N-2,3-dim-
ethylbut-2-enyl substituents (compounds 3, R ) CHMe2, and
32) using the chiral ligands 16 and 25 (Tables 2 and 4). The
asymmetry arises from a dynamic thermodynamic resolution,
as shown by the fact that similar er’s are obtained if the mixture
is cooled to -78 °C prior to quenching with the electrophile.
Also, if only a small amount of electrophile is added, then a
different er is obtained (showing that the asymmetry does not
arise from a kinetic resolution). Thus, the organolithium prepared
from the stannane 3, R ) CHMe2, and n-BuLi, on cooling to
-78 °C then addition of the chiral ligand 16 followed by
addition of only 0.3 equiv of Me3SiCl, gave the product 19 with
er 39:61 (S:R). Electrophilic quench of the organolithium species
is believed to occur with retention of configuration at the
carbanion center;10 hence, using the ligand 16, the enantiomer
(R)-2-lithiopyrrolidine reacts slightly faster than the (S)-enan-
tiomer (∆∆Gq must be about 0.7 kJ/mol). However, the (S)-
enantiomer predominates at equilibrium to the extent of at least
97:3 (S:R), and this ratio is reflected in the er of the products
given in Table 2. The reaction energy profile can be described
as shown in Figure 2. As the ratio of enantiomers of the products
reflects the ratio of diastereomeric complexes, the difference
in the energy of these complexes, ∆G°, must be about 8 kJ/
mol. The barrier to inversion of the diastereomeric complexes
will alter with the temperature; as there is a slight preference
for reaction of the minor (R)-diastereomer in these dynamic
thermodynamic resolutions, the electrophilic quenches are best
carried out at low temperature, where no interconversion takes
place. However, good er’s can be obtained by quenching at the
temperature required for equilibration.
(18) Frisch, M. J.; et al. Gaussian 03, Revision C.02; Gaussian, Inc.: Walling-
ford, CT, 2004.
(19) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980,
72, 650.
(20) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(21) Boys, S. B.; Bernardi, F. Mol. Phys. 1970, 19, 553.
(22) Gawley, R. E.; Narayan, S.; Vicic, D. A. J. Org. Chem. 2005, 70, 328.
9
J. AM. CHEM. SOC. VOL. 128, NO. 33, 2006 10949

A R T I C L E S
Coldham et al.
(R)-33 (80%, er 73:27 by chiral HPLC) (absolute configuration
by comparison of the specific rotation of the N-deprotected
compound proline-N-phenylamide with the literature17). This
result confirmed the overall retention of configuration from the
stannane to the amide (partial racemization of the organolithium
occurs under these conditions). Alongside this experiment, we
carried out the same reaction, but after cooling to -78 °C, the
ligand 16 (deprotonated with BuLi) was added. After 90 min
at -78 °C, the mixture was quenched with PhNCO. The amide
(R)-33 (65%, er 74:26) was formed. The fact that the same
enantiomer ratio and the same major enantiomer were produced
confirms that the electrophilic quench in the presence of the
chiral ligand 16 proceeds with the same absolute sense. These
results support a double retention on going from the stannane
to the product via the organolithium. Therefore, the ab initio
calculations appear to predict the wrong absolute configuration
of the organolithium. The reason for this is unclear, but it could
be that the monomeric organolithium (complexed with the chiral
ligand) that we have modeled does not match the solution
structure and there may be a dimeric or other oligomeric
complex present.
Figure 3. Assumed energy diagram for dynamic kinetic resolution of
N-Boc-2-lithiopyrrolidine in the presence of excess BuLi.
The energy profile for a dynamic kinetic resolution typically
has a relatively high barrier to reaction with the electrophile in
comparison to interconversion of the complexes. In our case,
reactions are best carried out at -20 °C, where interconversion
takes place (albeit slowly), and high enantiomer ratios with
yields greater than 50% are possible if the electrophile is added
slowly. Hence, the barrier to inversion of the diastereomeric
complexes may be higher than the barriers to reaction with the
electrophile, but slow addition of the electrophile allows re-
equilibration so that there is always present some of the faster-
reacting enantiomer (in this case, the (S)-enantiomer using ligand
16).
Dynamic Resolution of N-Boc-2-lithiopyrrolidine. Excellent
levels of asymmetric induction are possible from N-Boc-2-
lithiopyrrolidines using the chiral ligands 16 and 25 (Schemes
8 and 9). The asymmetry arises from a dynamic kinetic
resolution, as shown by the fact that a warm-cool protocol gives
lower selectivities and that different er’s are obtained under
different reaction conditions and using different rates of addition
of the electrophile. The organolithium N-Boc-2-lithiopyrrolidine
is believed to react with Me3SiCl with retention of stereochem-
istry at the carbanion center;13 hence, using the chiral ligand
16 and excess n-BuLi, the (S)-2-lithiopyrrolidine enantiomer
reacts faster than the (R)-enantiomer (∆∆Gq must be about 8
kJ/mol). However, in the absence of excess n-BuLi, there is
less difference in the rate of reaction. The ratio of the
diastereomeric complexes is close to 50:50 in the absence of
excess n-BuLi, although this ratio appears to be of the order of
80:20 in the presence of excess n-BuLi. This thermodynamic
selectivity matches the kinetic preference for reaction. Thus,
using ligand 16, the (S)-enantiomer of N-Boc-2-lithiopyrrolidine
is formed by an asymmetric deprotonation (er ∼87:13), and this
ratio is maintained for extended periods in the presence of excess
n-BuLi at -20 °C, as judged by cooling the complexes to -78
°C after different time periods, quenching the reaction mixture
using Me3SiCl, and measuring the er (by chiral GC). Hence,
using the optimized conditions (Scheme 9) with the ligand 16,
an er of approximately 87:13 (S:R) of N-Boc-2-lithiopyrrolidine
(obtained initially by asymmetric deprotonation) converts on
warming to -20 °C to a ratio of approximately 80:20 (S:R) (by
dynamic equilibration to the thermodynamic ratio of complexes),
and then reaction with Me3SiCl at -20 °C occurs faster with
the (S)-enantiomer of N-Boc-2-lithiopyrrolidine to give the
product 35 (er 95:5 S:R). However, even when the mismatched
arrangement is present, such that the enantiomer that reacts faster
is present as the minor component, high er’s of the product 35
can be obtained as a result of faster reaction of the minor
diastereomeric complex in an equilibrating mixture (Scheme
10). These facets imply that the energy profile of the reaction
can be described as shown in Figure 3 (using the ligand 16).
The addition of electrophiles other than Me3SiCl gave poor
results. This indicates that Me3SiCl reacts with the 2-lithiopy-
rrolidine, whereas other electrophiles react competitively with
BuLi, thereby giving mixtures of products and reduced enan-
tioselectivity. For example, addition of Bu3SnCl gave racemic
organostannane 1. This is likely due to repeated tin-lithium
exchange and final quench of the organolithium in the absence
of excess BuLi, although a contributory factor could come from
a variation in the relative rate of reaction of the diastereomeric
organolithium-chiral ligand complexes, depending on the nature
of the electrophile.
The rate of enantiomerization for N-Boc-2-lithiopyrrolidine
has been determined to be of the order of 1.5 × 10-5 s-1 at
-22 °C in hexanes-Et2O (4:1) (Figure 1). In Et2O as the
solvent, the rate of enantiomerization is about 3.6 × 10-4 s-1
at -17 °C and 2.8 × 10-5 s-1 at -26 °C.8 These values would
equate to a half-life for enantiomerization of about 30 min at
-17 °C and about 7 h at -26 °C, and they appear to be
somewhat slow for efficient dynamic resolution at the temper-
ature that we use (-20 °C). It is likely, however, that the chiral
ligand exerts an influence on the rate of enantiomerization. To
probe this issue, the enantioenriched organolithium formed from
asymmetric deprotonation of N-Boc-pyrrolidine 36 using the
ligand 16 and no excess BuLi (compare with Scheme 10) was
allowed to age over different time periods at -20 °C, followed
by cooling to -78 °C and electrophilic quench with Me3SiCl.
The enantiomer ratios of the products 35 were determined by
chiral GC, and this showed a gradual loss of er over time. Only
a limited number of data points were collected, but these
followed an exponential decay, suggesting first-order kinetics.
Using these data, the rate of enantiomerization, k, was deter-
mined as approximately 1.5 × 10-3 s-1 at -20 °C, which
equates to a barrier to inversion, ∆Gq ) 18 kcal/mol and a
9
10950 J. AM. CHEM. SOC. VOL. 128, NO. 33, 2006

Dynamic Resolution of 2-Lithiopyrrolidines
A R T I C L E S
half-life of about 8 min. This represents a faster rate of
enantiomerization in the presence of the ligand.
Acknowledgment. We thank The Leverhulme Trust (F/
00118V) and the EPSRC (GR/N32532/01 and GR/S35202/01)
for support of this work. We are grateful to GlaxoSmithKline
for a CASE Award (to T.F.N.H.) and Steven Howard for helpful
discussions. We thank Anthony J. H. M. Meijer for carrying
out the ab initio calculations.
Conclusions
Dynamic resolution of 2-lithiopyrrolidines under either
thermodynamic or kinetic control can occur with excellent levels
of asymmetric induction. The best chiral ligand is the diproline
derivative 16; its diastereomer 25 can be used to provide the
opposite absolute configuration of the products. Resolution of
N-Boc-2-lithiopyrrolidine is complicated by the need for excess
n-BuLi for highly selective kinetic resolution. Resolution of
N-(2,3-dimethylbut-2-enyl)-2-lithiopyrrolidine occurs by a dy-
namic thermodynamic resolution, and the products can be
converted easily to their N-unsubstituted derivatives.
Supporting Information Available: Experimental and com-
putational details, including procedures and spectroscopic data
for compounds 19-24, 27, 29, 32-35, plus the complete ref
18. This material is available free of charge via the Internet at
http://pubs.acs.org.
JA061963M
9
J. AM. CHEM. SOC. VOL. 128, NO. 33, 2006 10951