Dynamic resolution of N-Boc-2-lithiopiperidine

Article abstract of DOI:10.1039/b810988e

Dynamic thermodynamic resolution of N-Boc-2-lithiopiperidine is possible using a chiral ligand; the two enantiomers of this organolithium can be resolved with selectivities of up to 85: 15 from a selection of 26 chiral diamino-alkoxide ligands screened. The Royal Society of Chemistry.

Full text of DOI:10.1039/b810988e

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Dynamic resolution of N-Boc-2-lithiopiperidinew
Iain Coldham,*a Sophie Raimbault,a Praful T. Chovatia,a Jignesh J. Patel,a
Daniele Leonori,a Nadeem S. Sheikha and David T. E. Whittakerb
Received (in Cambridge, UK) 27th June 2008, Accepted 29th July 2008
First published as an Advance Article on the web 8th August 2008
DOI: 10.1039/b810988e
Dynamic thermodynamic resolution of N-Boc-2-lithiopiperidine
is possible using a chiral ligand; the two enantiomers of this
organolithium can be resolved with selectivities of up to 85 : 15
from a selection of 26 chiral diamino-alkoxide ligands screened.
2 under these conditions and report here some of the results of
this study.
It appeared that diamino-alkoxide ligands such as 5 were
most promising to promote, on complexation, a preference for
one of the chiral organolithiums 2. We therefore concentrated
our efforts on such chiral ligands. A selection of 24 ligands is
shown in Fig. 1. These were prepared either by coupling of an
N-protected amino-acid with an amino-ester, followed by
LiAlH4 reduction,7 or by addition of proline methyl ester to
an intermediate aziridinium ion8 followed by reduction (see
Supplementary Informationw). As a general procedure for the
dynamic resolution, N-Boc-piperidine 1 and 1.1–1.2 equiv. of
TMEDA in Et2O were treated with 1.1–1.2 equiv. of sec-BuLi
at ꢀ78 1C for 3 h,9 followed by the addition of the chiral
ligand (1.2–1.5 equiv.). The ligand was pre-treated with an
equivalent of n- or sec-BuLi in Et2O to form the lithium
alkoxide. The temperature was then raised to allow equilibra-
tion (between ꢀ40 and ꢀ10 1C).10 After 1 or 1.5 h at this
temperature the mixture was cooled to ꢀ78 1C and 3 equiv. of
the electrophile TMSCl were added. Sufficient time was
allowed for electrophilic quench before methanol was added.
Purification of the silane was followed by enantiomer ratio
determination by chiral GC (b-cyclodextrin column).5
The results of this study are summarised in Table 1. The
simplest chiral ligand 6, with one stereocentre, is clearly poor
at resolving the S- and R-organolithiums 2. With two stereo-
centres some good selectivities can be achieved. Comparison
of the ligands 4 and 5 with ligands 7 and 8 (derived from
valine) shows that having a branched (iPr) substituent alpha to
the nitrogen atom enhances the selectivity. This is confirmed
with the ligands 9 and 10 (derived from isoleucine) and the
poorer results using ligands 11 and 12 (derived from alanine).
Interestingly, the ligand 13 was good, giving the product
3 with er 85 : 15 (S : R), whereas its diastereomer 14 was
A popular and valuable method for asymmetric synthesis
involves the formation and reaction of chiral organolithiums.1
Of particular importance is the asymmetric deprotonation of
substrates such as O-alkyl-carbamates and N-(tert-butoxy-
carbonyl)pyrrolidine (N-Boc-pyrrolidine) using sec-butyllithium
and (ꢀ)-sparteine as the chiral base.1,2 However, a problem
with this chemistry is that it is limited to certain substrates. For
example, extension of asymmetric deprotonation to the im-
portant piperidine ring system (using N-Boc-piperidine) occurs
with low yields and/or low selectivities.3 Recently, building on
work on the analogous pyrrolidines,4 we have found that
highly enantioselective substitution of N-Boc-2-lithiopiperidine
is possible.5 In this case, the asymmetry arises from the faster
reaction of one of the diastereomeric complexes between the
chiral organolithium and the chiral ligand. Hence, using the
ligand L* = 4, the two organolithiums S-2 and R-2 can
interconvert at ꢀ20 1C, and organolithium S-2 reacts much
faster than R-2 to give the product 3 with high enantioselectivity
(er 93 : 7 S : R in Et2O and er 95 : 5 in THF) (Scheme 1).5
However, high selectivities were obtained only with trimethylsi-
lyl chloride (TMSCl) as the electrophile. To provide a general
approach to 2-substituted piperidines we needed either to opti-
mize this process for different electrophiles or find a ligand that
is suitable for dynamic resolution under thermodynamic con-
trol.6 Under thermodynamic conditions (by cooling to
ꢀ78 1C to prevent further interconversion) the ligand 4 gives
the product 3 with poor selectivity (er 42 : 58 S : R in Et2O,
representing a small preference for the organolithium (R)-2).
The diastereomeric ligand 5, although poor in the dynamic
kinetic resolution, was better in this regard, providing the
product 3 with reasonable selectivity (er 77 : 23 S : R in Et2O)
[the ligand (ꢀ)-sparteine gave poor selectivity (55 : 45 S : R in
Et2O)].5 Due to the importance of the piperidine ring system in
natural products and medicinal drugs, we have screened a
number of ligands to effect dynamic resolution of organolithium
a Department of Chemistry, University of Sheffield, Sheffield, UK S3
7HF. E-mail: i.coldham@sheffield.ac.uk; Fax: (+44) 114 222 9346;
Tel: (+44) 114 222 9428
b AstraZeneca, Mereside, Alderley Park, Macclesfield, Cheshire, UK
SK10 4TG
w Electronic supplementary information (ESI) available: Experimental
procedures, spectroscopic data and GC traces. See DOI: 10.1039/
b810988e
Scheme 1 Dynamic kinetic resolution.5 (a) sec-BuLi (1.2 equiv.),
THF, TMEDA (1.2 equiv.), ꢀ78 1C, 3 h, then 4 (1.5 equiv.) (pre-
treated with n-BuLi in Et2O), then ꢀ20 1C, 2 min then 4 equiv. TMSCl
added over 50 min at ꢀ20 1C, 60%, er 95 : 5, S : R.
ꢁc
This journal is The Royal Society of Chemistry 2008
4174 | Chem. Commun., 2008, 4174–4176

View Article Online
Scheme
2
Quenching with different electrophiles. (a) sec-BuLi
(1.1 equiv.), Et2O, TMEDA (1.1 equiv.), ꢀ78 1C, 3 h, then 7
(1.2 equiv.) (pre-treated with sec-BuLi in Et2O), then ꢀ40 1C, 90 min
then ꢀ78 1C, 3 equiv. electrophile Me3SiCl, Bu3SnCl or DMF;
(b) as (a) but, prior to addition of the electrophile allyl bromide, was
added ZnCl2 (1.3 equiv.) in THF then CuCNꢂ2LiCl (1.2 equiv.) in THF.
organolithium from N-Boc-2-tributylstannylpiperidine and
minimising the amount of TMEDA present (the use of n-BuLi
and 10 mol% TMEDA allows efficient tin–lithium exchange).
This is consistent with the diamino-alkoxides being better
ligands than TMEDA.4 The chiral ligands can be recovered
after the reaction by column chromatography and distillation.
The yields in these dynamic resolutions were typically in the
30–60% region, with about 30% recovered starting material 1.
Interestingly, the yields were best using aged bottles of sec-
BuLi. Aged sec-BuLi contains a significant amount of sec-
BuOLi (verified by 1H NMR spectroscopy)11 and it was
possible to obtain satisfactory yields (450%, together with
some recovered starting material) using either aged sec-BuLi
or freshly acquired sec-BuLi in the presence of a small amount
Fig. 1 Structures of ligands 6–29.
Table 1 Enantiomer ratio of 3 by dynamic thermodynamic resolu-
tion using ligands 4–29
Ligand
3, er (S : R)a
Ligand
3, er (S : R)a
i
4
5
6
7
8
9
10
11
12
13
14
15
16
42 : 58
77 : 23
41 : 59
80 : 20
15 : 85
82 : 18
22 : 78
51 : 49
51 : 49
85 : 15
49 : 51
65 : 35
23 : 77
17
18
19
20
21
22
23
24
25
26
27
28
29
56 : 44
20 : 80
55 : 45
72 : 28
46 : 54
55 : 45
50 : 50
52 : 48
51 : 49
57 : 43
51 : 49
81 : 19
30 : 70
(5 mol%) of PrOLi.
Using one of the best ligands in Table 1, namely ligand 7, we
prepared
a selection of 2-substituted N-Boc-piperidines
(Scheme 2).z Deprotonation of N-Boc-piperidine (1.2 g) with
sec-BuLi and TMEDA in Et2O followed by addition of the
deprotonated ligand 7, and warming to –40 1C for 1.5 h (to set
up the dynamic resolution), then cooling to –78 1C, followed
by addition of Bu3SnCl gave the 2-substituted product 30 in
65% yield. Similarly, quenching with DMF gave the aldehyde
31. As expected, the enantiomer ratios of the products 30–31
were similar (in comparison with the silane 3) on altering the
electrophile, which is consistent with a resolution in which (in
this case) the organolithium (S)-2 is preferred. The product 32
was prepared using allyl bromide as the electrophile after
transmetallation to the zinc-cuprate species.12 Reduction of
the alkene and hydrolysis of the carbamate provides an overall
three-step synthesis of the alkaloid coniine.13
a
The absolute configuration of the major enantiomer was determined
by comparison with an authentic sample prepared according to ref. 3.
poor. Further branching with the tert-butyl derivatives 15 and
16 did not enhance the selectivity and neither did the use of the
piperidine analogues 17 and 18. Incorporation of gem-
dimethyl groups in ligands 19 and 20 reduced the effectiveness
of the ligands. The prolinol motif seems to be important as
more flexible amino-alcohol portions (ligands 21–23) gave
essentially racemic product 3. We then turned to ligands
prepared from styrene epoxide (24–25) or N-methylephedrine
(26–27) or pseudoephedrine (28–29). Of these, only ligand
28 gave good selectivity in the dynamic resolution reaction.
The enantioselectivity was not improved by using longer
reaction times or higher temperatures prior to cooling and
quenching. Likewise, no improvement in the enantiomer ratio
was obtained by using other solvents or by forming the
The enantioselectivity could arise from a dynamic thermo-
dynamic resolution,6 or from a preference for the chiral ligand
to complex to one enantiomer of the interconverting organo-
lithiums. There was no crystallization and we have drawn the
organolithium as a monomer, although the structure has not
been determined. In contrast to previous studies,5 the selecti-
vity does not arise from a dynamic kinetic resolution. For
example, ligand 8 provides the silane product 3 with er
15 : 85 (S : R) (Table 1), whereas under conditions for dynamic
kinetic resolution (slow quench with TMSCl at –40 1C),5 the
same ligand 8 was found to give the product 3 with er 84 : 16
ꢁc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 4174–4176 | 4175

View Article Online
(S : R). This indicates that there is interconversion between the
diastereomeric complexes (either directly or via their uncom-
plexed forms) and that these complexes react at different rates,
with the minor complex reacting faster.
for Synthesis, Pergamon, Oxford, 2002; (c) Organolithiums in
Enantioselective Synthesis, ed. D. M. Hodgson, Springer-Verlag,
Heidelberg, 2003; (d) R. E. Gawley and I. Coldham, in
The Chemistry of Organolithium Compounds, ed. Z. Rappoport
and I. Marek, Wiley, Chichester, 2004, p. 997; (e) D. Hoppe and G.
Christoph, in The Chemistry of Organolithium Compounds, ed. Z.
Rappoport and I. Marek, Wiley, Chichester, 2004, p. 1055.
2 P. Beak, S. T. Kerrick, S. Wu and J. Chu, J. Am. Chem. Soc., 1994,
116, 3231.
3 (a) W. F. Bailey, P. Beak, S. T. Kerrick, S. Ma and K. B. Wiberg,
J. Am. Chem. Soc., 2002, 124, 1889; (b) M. J. McGrath, J. L. Bilke
and P. O’Brien, Chem. Commun., 2006, 2607; (c) I. Coldham, P.
O’Brien, J. J. Patel, S. Raimbault, A. J. Sanderson, D. Stead and
D. T. E. Whittaker, Tetrahedron: Asymmetry, 2007, 18, 2113.
4 (a) I. Coldham, S. Dufour, T. F. N. Haxell, S. Howard and G. P.
Vennall, Angew. Chem., Int. Ed., 2002, 41, 3887; (b) I. Coldham, J.
J. Patel and G. Sanchez-Jimenez, Chem. Commun., 2005, 3083;
(c) I. Coldham, S. Dufour, T. F. N. Haxell, J. J. Patel and G.
Sanchez-Jimenez, J. Am. Chem. Soc., 2006, 128, 10943.
5 I. Coldham, J. J. Patel, S. Raimbault and D. T. E. Whittaker,
Chem. Commun., 2007, 4534.
Hence, this chemistry provides a general method to access
enantiomerically enriched 2-substituted piperidines.14 Either
enantiomer can be prepared by choice of the diastereomer (or
enantiomer) of the chiral ligand. This chemistry therefore
provides a solution to the poor yields and/or low levels of
asymmetric induction obtained using either asymmetric de-
protonation of N-Boc-piperidine or asymmetric substitution
with different electrophiles of N-Boc-2-lithiopiperidine by
dynamic kinetic resolution.
Notes and references
z General procedure: N-Boc-piperidine 1 (0.3 g, 1.6 mmol) and
TMEDA (0.27 mL, 1.8 mmol) in Et2O (3 mL) were treated with sec-
BuLi (1.4 mL, 1.8 mmol, 1.3 M in hexanes) at ꢀ78 1C. After 3 h, the
6 P. Beak, D. R. Anderson, M. D. Curtis, J. M. Laumer, D. J. Pippel
and G. A. Weisenburger, Acc. Chem. Res., 2000, 33, 715.
7 (a) T. Mukaiyama, Tetrahedron, 1981, 37, 4111; (b) D. J. Gallagher,
S. Wu, N. A. Nikolic and P. Beak, J. Org. Chem., 1995, 60,
8148.
deprotonated ligand
7 [prepared by adding sec-BuLi (1.6 mL,
2.0 mmol, 1.3 M in hexanes) to 7 (0.42 g, 1.9 mmol) in Et2O (3 mL)
at 0 1C] was added. The mixture was warmed to ꢀ40 1C. After 90 min
the mixture was cooled to ꢀ78 1C and the electrophile TMSCl (0.6 mL,
4.8 mmol) was added. The mixture was allowed to warm slowly (over
18 h) to room temperature and MeOH (2 mL) was added. The solvent
was evaporated and the residue was purified by column chromatogra-
phy on silica, eluting with light petroleum (bp 40–60 1C)–EtOAc (98 : 2)
to give the piperidine (S)-3 (213 mg, 51%), [a]D22 + 18.5 (0.5, CHCl3),
8 R. K. Dieter, N. Deo, B. Lagu and J. W. Dieter, J. Org. Chem.,
1992, 57, 1663.
9 For racemic proton abstraction using sec-BuLi, Et2O, TMEDA,
see P. Beak and W. K. Lee, J. Org. Chem., 1993, 58, 1109.
10 Kinetic studies are ongoing to determine the barrier for enantio-
merization, but equilibration takes place at and above ꢀ40 1C in
Et2O–TMEDA.
lit.5 for (S)-3, er 95 : 5, [a]D + 36.4 (1.95, CHCl3); other data as
24
reported;3 er 79 : 21 determined by GC [b-cyclodextrin-permethylated
120 fused silica capillary column 30 m ꢃ 0.25 mm i.d., 20% permethy-
lated b-cyclodextrin in SPB-35 poly(35% diphenyl/65% dimethyl)
siloxane, nitrogen carrier at 14 psi, retention times 31.5 min (major)
and 32.4 min (minor) (at 85 1C)]. The chiral ligand 7 can be recovered
by column chromatography, eluting with CH2Cl2–MeOH (7 : 3),
followed by evaporation of the solvent, acid/base wash (acidify with
2 M HCl, wash with CH2Cl2, basify with NaOH pellets and extract
with CH2Cl2) and distillation under reduced pressure.
11 T. R. Hoye, B. M. Eklov and M. Voloshin, Org. Lett., 2004, 6,
2567.
12 (a) R. K. Dieter, G. Oba, K. R. Chandupatla, C. M. Topping, K.
Lu and R. T. Watson, J. Org. Chem., 2004, 69, 3076; (b) J. P. N.
Papillon and R. J. K. Taylor, Org. Lett., 2002, 4, 119.
13 D. Passarella, A. Barilli, F. Belinghieri, P. Fassi, S. Riva, A.
Sacchetti, A. Silvani and B. Danieli, Tetrahedron: Asymmetry,
2005, 16, 2225.
14 For alternative procedures involving organolithiums, see (a) T. J.
Wilkinson, N. W. Stehle and P. Beak, Org. Lett., 2000, 2, 155;
(b) C. Serino, N. Stehle, Y. S. Park, S. Florio and P. Beak, J. Org.
Chem., 1999, 64, 1160.
1 For reviews, see (a) A. Basu and S. Thayumanavan, Angew. Chem.,
Int. Ed., 2002, 41, 716; (b) J. Clayden, Organolithiums: Selectivity
ꢁc
This journal is The Royal Society of Chemistry 2008
4176 | Chem. Commun., 2008, 4174–4176