Alkaloid synthesis using chiral δ-amino β-ketoesters: A stereoselective synthesis of (-)-lasubine II

Article abstract of DOI:10.1021/ol0061438

(equation presented) A highly stereoselective asymmetric synthesis of the quinolizidine alkaloid (-)-lasubine II from a δ-amino β-hydroxy ketone, a new polyfunctionalized chiral building block, is described.

Full text of DOI:10.1021/ol0061438

ORGANIC
LETTERS
2000
Vol. 2, No. 17
2623-2625
Alkaloid Synthesis Using Chiral
δ-Amino â-Ketoesters: A
Stereoselective Synthesis of
(−)-Lasubine II
Franklin A. Davis* and Bin Chao
Department of Chemistry, Temple UniVersity, Philadelphia, PennsylVania 19122
fdaVis@astro.ocis.temple.edu
Received June 1, 2000
ABSTRACT
A highly stereoselective asymmetric synthesis of the quinolizidine alkaloid (−)-lasubine II from a δ-amino â-hydroxy ketone, a new
polyfunctionalized chiral building block, is described.
The Lythraceaes are a large family of naturally occurring
alkaloids, a number of which contain the 4-arylquinolizidine
substructures.1 Many biologically important and structurally
interesting alkaloids have the quinolizidine skeleton.2 For
example, (-)-lasubine I (1) and (-)-lasubine II (2), which
syntheses of 1 and 2 have been described, only four
asymmetric syntheses have been reported.4 The key step in
Comins’ synthesis of 1 was a diastereoselective (86% de)
addition of a Grignard reagent to a chiral 1-acylpyridinium
salt.5 This alkaloid was also recently prepared via an aza-
Diels-Alder reaction employing a resolved chiral arylalde-
hyde tricarbonylchromium complex.6 A chiral 2-isoxazoline
was employed in the synthesis of 2,7 and both 1 and 2 were
prepared via the intramolecular cyclization of an N-acylimin-
ium ion derived from a chiral amino ester.4 However, in this
synthesis the yield (16%) and diastereoselectivity in favor
of 2 was poor.
As part of a program aimed at exploring the utility of
δ-amino â-ketoester chiral building blocks for alkaloid
synthesis, we describe a highly stereoselective asymmetric
synthesis of (-)-lasubine II (2).8
have the cis- and trans-quinolizidine skeletons, respectively,
were isolated by Fuji and co-workers from the leaves of
Lagerstroemia subcostata Koehne.3 While many racemic
(4) For references to the synthesis of racemic lasubine alkaloids, see:
Chalard, P.; Remuson, R.; Gelas-Mialhe, Y.; Gramain, J.-C. Tetrahedron:
Asymmetry 1998, 9, 4361.
(5) Comins, D. L.; LaMunyon, D. H. J. Org. Chem. 1992, 57, 5807.
(6) Ratni, H.; Ku¨ndig, E. P. Org. Lett. 1999, 1, 1997.
(7) Ukaji, Y.; Ima, M.; Yamada, T.; Inomata, K. Heterocycles 2000, 52,
563.
(1) (a) Golebiewski, W. M.; Wrobel, J. T. In The Alkaloids; Academic
Press: New York, 1981; Vol. 18, Chapter 4, p 263. (b) Fuji, K. In The
Alkaloids; Academic Press: New York, 1989; Vol. 35, Chapter 3, p 155.
(2) Michael, J. P. Nat. Prod. Rep. 1997, 14, 21.
(3) Fuji, K.; Yamada, T.; Fujita, E.; Murata, H. Chem. Pharm. Bull. 1978,
26, 2515.
10.1021/ol0061438 CCC: $19.00 © 2000 American Chemical Society
Published on Web 08/03/2000

The requisite δ-amino â-ketoester, (RS,S)-(-)-methyl-3-
oxo-N-(p-toluenesulfinyl)-5-amino-5-(3,4-dimethoxy-phen-
yl)pentanoate (5), was prepared in one pot by treating (R)-
(+)-(3,4-dimethoxybenzylidene)-p-toluenesulfinamide (4) with
4.0 equiv of the sodium enolate of methyl acetate at -78
°C (Scheme 1). Flash chromatography gave (-)-5 in 93%
skeleton. For efficiency it was important to accomplish this
with a minimum of manipulation and/or protecting group
chemistry. However, attempts failed to generate the N-
trimethylsilyl lactam11 or the imidoyl chloride12 of 7, for
coupling with an appropriate organolithium reagent. The 1H
NMR suggests that O-silylation occurred rather than N-
silylation. Similarly, reaction of the N-Cbz derivative of 8
with LiHMDS followed by treatment of the resulting enolate
with N-(5-chloro-2-pyridyl)triflimide resulted in decomposi-
tion. Comins has reported that enol triflates, prepared from
N-acyllactams, readily react with organometallic reagents to
give enecarbamates.13 These failures may be a consequence
of steric congestion at the amide nitrogen and/or instability
of the intermediate enolate.
Scheme 1
Since the synthesis of ketones from carboxylic acids and
organolithium reagents is well-known,14,15 we next explored
the conversion of the carbomethoxy group in â-amino alcohol
6 to ketone 9 (Scheme 2). The amino alcohol 6 was treated
Scheme 2
yield and >95% de. Condensation of commercially available
(R)-(-)-p-toluenesulfinamide (3) with 3,4-dimethoxybenz-
aldehyde in the presence of Ti(OEt)4 afforded sulfinimine
(R)-(+)-4 in 95% yield.9 To produce the trans arrangement
of the 1-aryl and 3-hydroxy groups in 2, it was necessary to
reduce the 3-oxo group in 5 with syn selectivity. This was
readily accomplished using metal chelation control and Zn-
(BH4)2.10 Thus, treatment of (-)-5 with 3.2 equiv of Zn-
(BH4)2 at -78 °C for 1 h gave a 50:6 syn:anti ratio of
â-amino alcohol (-)-6 in 87% yield. In the syn isomer the
C(2) protons appear at δ 2.38 ppm, whereas in the anti isomer
one of these protons is shifted downfield to 2.49 ppm. These
assignments were confirmed by cyclization of 6 to hydroxy
piperidone 7 with TFA/NaHCO3 in quantitative yield, which
was converted into 2 (see below). In 7 the C(4) J3,4 proton
coupling constant is less than 2 Hz.
with 1.0 equiv of LiOH, to hydrolyze the ester to the acid,
and the solvent was removed to dryness. Azeotropic treat-
ment with benzene was used to remove the last traces of
water. When the lithium salt of 6 was reacted with 4 equiv
of n-BuLi, ketone 9 (R ) n-BuLi) was isolated in 60% yield.
However, with lithium 4-benzyloxybutane, prepared from
4-benzyloxybutyl phenyl sulfide and lithium naphthalene,16
ketone 9 was obtained in less than 20% yield.
To proceed with our synthesis of (-)-2, it was necessary
to replace the 2-oxo group in 7 with a substituent
N-Methoxy-N-methylamides, Weinreb amides, are impor-
tant carbonyl equivalents and have been extensively explored
-
[BnO(CH2)4 ] that could be cyclized to give the quinolizidine
(11) Hua, D. H.; Miao, S. W.; Bharathi, S. N.; Katsuhira, T.; Bravo, A.
A. J. Org. Chem. 1990, 55, 3682.
(12) Tamao, K.; Kodama, S.; Nakajima, I.; Kumada, M.; Minato, A.;
Suzuki, K. Tetrahedron 1982, 38, 3347.
(13) Foti, C. J.; Comins, D. L. J. Org. Chem. 1995, 60, 2656.
(14) Rubottom, G. M.; Kim, C.-w. J. Org. Chem. 1983, 48, 1550.
(15) For a review on the reation of carboxylic acids with organolithium
reagents, see: Jorgenson, M. J. Org. React. 1970, 18, 1.
(16) Screttas, C. G.; Micha-Screttas, M. J. Org. Chem. 1978, 43, 1064.
(8) Davis, F. A.; Chao, B.; Fang, T.; Szewczyk, J. M. Org. Lett. 2000,
2, 1041.
(9) (a) Davis, F. A.; Zhang, Y. Andemichael, Y.; Fang, T.; Fanelli, D.
L.; Zhang, H. J. Org. Chem. 1999, 64, 1403. (b) Fanelli, D. L.; Szewczyk,
J. M.; Zhang, Y.; Reddy, G. V.; Burns, D. M.; Davis, F. A. Org. Synth.
1999, 77, 50.
(10) Kathawala, F. G.; Prager, B.; Prasad, K.; Repic, O.; Shapiro, M. J.;
Stabler, R. S.; Widler, L. HelV. Chim. Acta 1986, 69, 803.
2624
Org. Lett., Vol. 2, No. 17, 2000

for the synthesis of ketones.17,18 Reaction of 6 with 5 equiv
of lithium N,O-dimethylhydroxylamine gave amide 10 in
92% isolated yield (Scheme 2). Remarkably, the N-sulfinyl
group remained intact and is a likely consequence of
formation of the lithium N-sulfinyl amide which protects it
from further reaction. With 10.0 equiv of 4-(benzyloxy)-1-
butane magnesium bromide, 10 gave the desired ketone 9
in 60% yield. The sulfinyl group was removed with 4 N
HCl, and neutralization with concentrated NH4OH gave the
cyclic imine 11 (Scheme 3). The crude imine 11 was not
isolated but was reduced with LAH/MeONa to give, exclu-
sively, the cis hydroxy piperidine 12 in 72% overall yield
for the conversion of 9 to 12. Deprotection of the benzyl
group with H2/Pd gave a 90% yield of alcohol 13. The
quinolizidine ring was formed from 13 on reaction with
pyridine and TsCl, affording (-)-lasubine II (2) in 71% yield
with properties in agreement with literature values.
Scheme 3
In summary, a highly stereoselective asymmetric synthesis
of (-)-lasubine II (2) from δ-amino â-ketoester (RSS)-
(-)-5 is presented. Noteworthy features of the synthesis
include the preparation of δ-amino â-hydroxy ketone (-)-9
from Weinreb amide 10, its ready conversion to the quino-
lizidine skeleton, and the absence of protecting group
chemistry in these transformations. This new methodology
further illustrates the utility of polyfunctionalized chiral
building blocks such as 5 and 9 for alkaloid synthesis.
Acknowledgment. This work was supported by the
National Institutes of Health (GM51982)
Supporting Information Available: Experimental pro-
cedures and spectroscopic data for compounds 2-13. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL0061438
(17) For a review on Weinreb amides see: Sibi, M. P. Org. Prep. Proc.
Int. 1993, 25, 15.
(18) Davis, F. A.; Kasu, P. V. N. Tetrahedron Lett. 1998, 39, 6135.
Org. Lett., Vol. 2, No. 17, 2000
2625