A new synthesis of pyrrolidines by way of an enantioselective Mannich/diastereoselective hydroamination reaction sequence

Article abstract of DOI:10.1021/ol201566u

A new two-step synthesis of highly substituted pyrrolidines has been developed. Chiral silane Lewis acid promoted enantioselective Mannich reactions of silyl ketene imines with acylhydrazones may be used to access bishomoallylic benzoic hydrazides that in

Full text of DOI:10.1021/ol201566u

ORGANIC
LETTERS
2011
Vol. 13, No. 15
4056–4059
A New Synthesis of Pyrrolidines by Way of
an Enantioselective Mannich/
Diastereoselective Hydroamination
Reaction Sequence
Jenny M. Baxter Vu and James L. Leighton*
Department of Chemistry, Columbia University, New York, New York 10027,
United States
leighton@chem.columbia.edu
Received June 10, 2011
ABSTRACT
A new two-step synthesis of highly substituted pyrrolidines has been developed. Chiral silane Lewis acid promoted enantioselective Mannich
reactions of silyl ketene imines with acylhydrazones may be used to access bishomoallylic benzoic hydrazides that in turn may be cyclized to
pyrrolidines by way of the thermal hydroamination reaction reported recently by Beauchemin. Importantly, excellent diastereoselectivity may be
realized in the hydroamination reactions.
The development of experimentally simple, efficient,
and highly enantioselective methods to synthesize chiral
pyrrolidines is an important goal for chemical synthesis.
While great strides have recently been made in the devel-
opment of enantioselective azomethine ylide cycloaddition
reactions,1 this approach has significant limitations in
scope, and there is a continuing need for new methods to
access a greater range of pyrrolidine structures, especially
those with a high degree of substitution. Beauchemin and
co-workers recently reported that bishomoallylic benzoic
hydrazides may be converted to pyrrolidines by way of a
thermal hydroamination reaction,2 and because our acyl-
hydrazone-silane Lewis acid platform provides access to a
structurally diverse array of benzoic hydrazides,3 the cou-
pling of these two processes seemed a worthwhile pursuit
(Scheme 1A). Realization of this idea would require (1) an
enantioselective hydrazone addition reaction that would
lend itself to the synthesis of highly substituted bisho-
moallylic benzoic hydrazides and (2) the development of
diastereoselective variants of the hydroamination reaction,
ꢀ
(1) (a) Najera, C.; Sansano, J. M. Angew. Chem., Int. Ed. 2005, 44,
6272. (b) Pandey, G.; Banerjee, P.; Gadre, S. R. Chem. Rev. 2006, 106,
ꢀ
4484. (c) Pellissier, H. Tetrahedron 2007, 63, 3235. (d) Najera, C.;
Sansano, J. M. Chem. Rev. 2007, 107, 4584. (e) Stanley, L. M.; Sibi,
ꢀ
~
M. P. Chem. Rev. 2008, 108, 2887. (f) Alvarez-Corral, M.; Munoz-
Dorado, M.; Rodrıguez-Garcıa, I. Chem. Rev. 2008, 108, 3174. (g) Patil,
N. T.; Yamamoto, Y. Chem. Rev. 2008, 108, 3395.
(4) For early studies on the generation of silyl ketene imines and their
use in reactions with electrophiles, see: (a) Gornowicz, G. A.; West, R. J.
Am. Chem. Soc. 1971, 93, 1714. (b) Watt, D. S. Synth. Commun. 1974, 4,
127. (c) Cazeau, P.; Llonch, J.-P.; Simonin-Dabescat, F.; Frainnet, E.
J. Organomet. Chem. 1976, 105, 145. (d) Cazeau, P.; Llonch, J.-P.;
Simonin-Dabescat, F.; Frainnet, E. J. Organomet. Chem. 1976, 105,
(2) Roveda, J.-G.; Clavette, C.; Hunt, A. D.; Gorelsky, S. I.; Whipp,
C. J.; Beauchemin, A. M. J. Am. Chem. Soc. 2009, 131, 8740.
(3) (a) Berger, R.; Duff, K.; Leighton, J. L. J. Am. Chem. Soc. 2004,
126, 5686. (b) Shirakawa, S.; Berger, R.; Leighton, J. L. J. Am. Chem.
Soc. 2005, 127, 2858. (c) Shirakawa, S.; Lombardi, P. J.; Leighton, J. L.
J. Am. Chem. Soc. 2005, 127, 9974. (d) Notte, G. T.; Leighton, J. L. J.
Am. Chem. Soc. 2008, 130, 6676. (e) Valdez, S. C.; Leighton, J. L. J. Am.
Chem. Soc. 2009, 131, 14638. (f) Lee, S. K.; Tambar, U. K.; Perl, N. R.;
Leighton, J. L. Tetrahedron 2010, 66, 4769. (g) Notte, G. T.; Baxter Vu,
J. M.; Leighton, J. L. Org. Lett. 2011, 13, 816.
€
157. (e) Meier, S.; Wurthwein, E.-U. Chem. Ber 1990, 123, 2339.
(5) For two recent examples of the use of silyl ketene imines in
asymmetric reactions, see: (a) Mermerian, A. H.; Fu, G. C. Angew.
Chem., Int. Ed. 2005, 44, 949. (b) Denmark, S. E.; Wilson, T. W.; Burk,
M. T.; Heemstra, J. R., Jr. J. Am. Chem. Soc. 2007, 129, 14864.
r
10.1021/ol201566u
Published on Web 07/12/2011
2011 American Chemical Society

as this question was largely unaddressed in the Beauchemin
report. With regard to the first point, we recently
reported a few examples of the addition of silyl ketene
imines4,5 (SKI) to chiral silane Lewis acid activated
acylhydrazones,3g,6 and one of those reactions involved
treatment of the complex derived from hydrazone 1 and
phenylsilane 2 with SKI 3 to give bishomoallylic benzoic
hydrazide 4 in good yield and moderately high enantios-
electivity (Scheme 1B).7 The further development of this
reaction seemed an excellent starting point for the present
study, and we detail herein our efforts to do exactly that
and the results of an investigation into the behavior of such
hydrazidesinBeauchemin-type hydroamination reactions.
studies eventually revealed that when the reaction of 5 was
conducted in CH2Cl2 at À20 °C,8 the product 6 was
obtained in 79% yield (5:1 dr) and 87% ee. Under other-
wise identical conditions substituted SKI 8 gave 9 in 71%
yield (10:1dr) and90%ee. Diallyl SKI10was also effective
under these optimized conditions leading to 11 in 73%
yield and 86% ee.
Scheme 2
Scheme 1
Unfortunately, these conditions were significantly less
effective for aliphatic aldehyde-derived hydrazone 1. Re-
optimization of the reaction between hydrazone 1 and SKI
3 promoted by 7 eventually led to the discovery that this
reaction was best conducted in trifluorotoluene at À20 °C,
conditions which resulted in the isolation of 4 in 59% yield
(9:1 dr) and 85% ee (Scheme 3). While these results are
essentially identical to those described in Scheme 1B, these
optimized conditions proved to be far more general with
respect to the SKI. Thus, reaction of the complex derived
from 1 and 7 with substituted SKIs 12À15 gave products
16À19, respectively, in the yields and stereoselectivities
shown. Hydrazones 20 and 21 were also treated with SKI 3
under the same conditions leading to products 22 and 23 in
improved yields (relative to the reactions of hydrazone 1)
albeit with reduced enantioselectivity. The mechanistic
origin of the consistently lower yields observed with hy-
drazone 1 remains unexplained, but the reactions of 20 and
21 (as well as the reactions in Scheme 2) seem to indicate
that 1 is an outlier in this regard. There does seem to be a
At the outset we were optimistic that the result described
in Scheme 1B would prove general for a range of both
hydrazones and substituted allyl-bearing SKIs. Unfortu-
nately, this was not the case as, for example, benzaldehyde-
derived hydrazone 5 performed poorly under otherwise
identical reaction conditions (Scheme 2). Prior experience
alerted us to the fact that these hydrazone addition reac-
tions can often be optimized simply by varying the specta-
tor group (Ph in the case of 2) on the silane Lewis acid.3d,f,g
Indeed, when the reaction of hydrazone 5 with SKI 3 was
repeated with previously reported tert-butoxysilane 7,3d
6
was produced in 94% ee, albeit with only moderate
diastereoselectivity and efficiency. Extensive optimization
(6) For other examples of Mannich-type reactions with acylhydra-
zones, see: Sugiura, M.; Kobayashi, S. Angew. Chem., Int. Ed. 2005, 44,
5176.
(8) As is the case with all of the pseudoephedrine-derived silanes, 7 is
prepared, isolated, and employed as a mixture (in this case 2.5:1) of
diastereomers. We have provided evidence that this is of no consequence
as the mixture converges on a single complex when reacted with
acylhydrazones (see refs 3a,3d). In the case of complexation reactions
of 7 with acylhydrazones carried out in CH2Cl2, we have found that the
complexation and convergence onto a single complex takes considerably
longer (details are provided in the Supporting Information), and this is
the reason for the significant “ageing” time in the first part of these
reactions.
(7) The effective use of disubstituted enolates or enolate equivalents
in enantioselective Mannich reactions for the generation of all-carbon
quaternary stereocenters is extremely rare (and, with hydrazones as the
electrophile, nonexistent). See: (a) Poulsen, T. B.; Alemparte, C.; Saaby,
S.; Bella, M.; Jørgensen, K. A. Angew. Chem., Int. Ed. 2005, 44, 2896. (b)
Tian, X.; Jiang, K.; Peng, J.; Du, W.; Chen, Y.-C. Org. Lett. 2008, 10,
3583. (c) Cheng, L.; Liu, L.; Jia, H.; Wang, D.; Chen, Y.-J. J. Org. Chem.
2009, 74, 4650. (d) Yin, L.; Kanai, M.; Shibasaki, M. J. Am. Chem. Soc.
2009, 131, 9610.
Org. Lett., Vol. 13, No. 15, 2011
4057

trend in the data presented in Schemes 2 and 3 that more
hindered SKIs (8 and 12À14) consistently lead to higher
levels of enantioselectivity (90À96% ee) than do SKIs 3,
10, and 15 (82À87% ee), while no trends at all are
discernible that might explain the fairly dramatic varia-
tions in diastereoselectivity. Nevertheless, the use of tert-
butoxysilane 7 under the optimized conditions as outlined
in Schemes 2 and 3 does allow for the synthesis of a
reasonably diverse array of bishomoallylic benzoic
hydrazides.
pure pyrrolidine product 24 was thereby obviated. That
this interesting resolution was due to the minor diaster-
eomer simply not reacting (as opposed to decomposing by
an alternate pathway) was corroborated by the hydroami-
nation (benzene, 60 °C) of 22, which was produced,
isolated, and employed as a 4:1 mixture of diastereomers
(see Scheme 3). A single pyrrolidine product (25) was
isolated in 51% yield, along with the unreacted minor
diastereomer (22a) of the starting material in 19% yield.
Hydroamination of 23 (9:1 dr) in toluene at 70 °C led to the
isolation of 26 as the sole pyrrolidine product both provid-
ing another example of this resolution and establishing
that complete regioselectivity for two alkenes can be
realized based solely on the gem-disubstituent effect.9,10
Scheme 3
Scheme 4
With access to a range of bishomoallylic benzoic hydra-
zides secured, weturnedour attention toanexamination of
the Beauchemin hydroamination reaction. Indeed, even
before we formally initiated this study we unwittingly
performed our first hydroamination when a solution of 4
in CDCl3 was allowed to stand at ambient temperature
overnight. Reacquisition of the 1H NMR spectrum of this
solution revealed the presence of a new compound, and
after 9 days the conversion to this new species was com-
plete and quite clean. This new compound was subse-
quently identified as 24, and it was found that if a
solution of 4 (produced, isolated, and employed as a 9:1
mixture ofdiastereomers) inCHCl3 was heated at60°C for
14 h, 24 could be isolated in 74% yield as a single
diastereomer (Scheme 4). Among the notable features of
this transformation are the especially mild conditions
necessary to effect it, and the diastereoselectivity: not only
did the major diastereomer of 4 undergo hydroamination
with excellent diastereoselectivity, but the minor diaster-
eomer of 4 apparently undergoes hydroamination much
more slowly, if at all. A potentially difficult separation of
the diastereomersof4 inorder toobtain diastereomerically
The impact of alkene substitution on the reactivity and
diastereoselectivity was examined next, and although
(E)-alkene 19 underwent hydroamination to give 27 in 68%
yield (isolated yield of pure major diastereomer), more
forcing conditions (xylenes, 120 °C, 24 h) were required
(Beauchemin and co-workers made similar observations)
which in turn caused a drop in the diastereoselectivity (7:1
dr) of the hydroamination event (Scheme 5). Remarkably,
even at this elevated temperature the resolution was op-
erative as the minor diastereomer present in the starting
material did not react. Despite the more forcing conditions
and reduced diastereoselectivity, this result nevertheless
establishes that groups other than methyl (i.e., CH2R) can
be installed at C(5) of the pyrrolidine with useful levels of
efficiency and diastereoselectivity. We turned next to an
examination of the impact of substitution on the internal
carbon of the alkene using substrates 16À18. In every case
(9) (a) Beesley, R. M.; Ingold, C. K.; Thorpe, J. F. J. Chem. Soc. 1915,
107, 1080. (b) Ingold, C. K. J. Chem. Soc. 1921, 119, 305. (c) Ingold,
C. K.; Sako, S.; Thorpe, J. F. J. Chem. Soc. 1922, 120, 1117.
(10) Jung, M. E.; Piizzi, G. Chem. Rev. 2005, 105, 1735.
4058
Org. Lett., Vol. 13, No. 15, 2011

the reaction proceeded with reasonable efficiency to give
28À30 in good yields establishing that substitution at this
position is well-tolerated. Unfortunately, in the reactions
of 17 and 18, whose hydroamination creates a fully sub-
stitutedcarbon stereocenter, little tono diastereoselectivity
was observed.
Scheme 6
Scheme 5
and generality for this reaction. In addition, we have
documented that very high levels of diastereoselectivity
may be realized in the Beauchemin hydroamination reac-
tion. Coupled together, these reactions expand the pool of
complex pyrrolidine structures that may be easily accessed
by asymmetric chemical synthesis. Of course, the Mannich
chemistry described here is just one of many possible ways to
generate bishomoallylic benzoic hydrazides. We anticipate
that the broader principle established here;that the pairing
of our chiral silane-benzoylhydrazone methodology with
the diastereoselective Beauchemin hydroamination reaction
adds up to a powerful new pyrrolidine synthesis;will
facilitate access to a much broader range of valuable nitro-
gen-containing heterocycles.
Finally, the dramatic difference in rates at which the
diastereomers of the bishomoallylic hydrazides undergo
the hydroamination reaction suggested the intriguing pos-
sibility that the diastereotopic allyl groups of 11 might be
effectively differentiated in a reaction that would create the
C(5) stereocenter and a quaternary carbon stereocenter at
C(3) of the pyrrolidine product. Indeed, heating a CHCl3
solution of 11 at 60 °C for 16 h led to the selective production
of 31 with 20:2:1 diastereoselectivity (measured by 1H NMR
spectroscopic analysis of the unpurified reaction mixture).
Pyrrolidine 31 was isolated as a g10:1 mixture of diaster-
eomers in 79% yield (Scheme 6).
Acknowledgment. This work was supported by a grant
from the National Science Foundation (CHE-08-09659).
We thank Novartis for a Graduate Fellowship to J.M.B.
We thank Prof. Ged Parkin and Mr. Wesley Sattler
(Columbia University) for an X-ray structure analysis
(see Supporting Information), and the National Science
Foundation (CHE-0619638) is thanked for acquisition of
an X-ray diffractometer.
The present study has established a method for the
synthesis of bishomoallylic benzoic hydrazides by way of
an enantioselective Mannich-type reaction with silyl ke-
tene imines, and we have documented a reasonable scope
Supporting Information Available. Experimental pro-
cedures, characterization data, and stereochemical
proofs. This material is available free of charge via the
Internet at http://pubs.acs.org.
Org. Lett., Vol. 13, No. 15, 2011
4059