Enantioselective approach to securinega alkaloids. Total synthesis of securinine and (-)-norsecurinine

Article abstract of DOI:10.1021/jo901059n

(Chemical Equation Presented) The most representative securinega alkaloids have been synthesized through a new strategy involving the palladium-catalyzed enantioselective allylation of a cyclic imide, a vinylogous Mannich reaction, and a ring-closing meta

Full text of DOI:10.1021/jo901059n

pubs.acs.org/joc
Enantioselective Approach to Securinega Alkaloids. Total Synthesis
of Securinine and (-)-Norsecurinine
ꢀ
ꢀ
ꢀ
David Gonzalez-Galvez, Elena Garcıa-Garcıa, Ramon Alibes,
ꢀ
Pau Bayon,* Pedro de March, Marta Figueredo,* and
Josep Font
ꢁ
´
Universitat Autonoma de Barcelona, Departament de Quımica, 08193 Bellaterra,
Spain
marta.figueredo@uab.es
Received May 25, 2009
The most representative securinega alkaloids have been synthesized through a new strategy
involving the palladium-catalyzed enantioselective allylation of a cyclic imide, a vinylogous
Mannich reaction, and a ring-closing metathesis process, as the key steps. The diastereoselectivity
of the vinylogous Mannich reaction was in partial agreement with DFT theoretical calculations
performed in a model system. The synthesis of (-)-norsecurine has been accomplished in nine steps
from succinimide and 14% overall yield and that of securinine in 10 steps from glutarimide and
20% overall yield. Both syntheses compare favorably with those previously described. The three
key transformations have been performed in a synthetically useful scale (more than 500 mg).
Moreover, since the enantioselectivity was originated by a chiral phosphine ligand, the antipode of
which is readily available, the same route is expected to give access to (þ)-norsecurinine and
virosecurinine.
Introduction
diuretics2 and antipyretics,3 in the treatment of hepa-
tic disorders,4 and against skin eruptions.5 Securinine
(1, Chart 1), the most abundant of these alkaloids, was first
isolated in 1956, and its structure was fully established in the
1960s.6 Since then, around 30 related alkaloids have
been isolated and characterized.7 Typically, the skeleton of
The securinega alkaloids1 comprise a group of compounds
initially isolated from some plants of the Securinega (also
named Flueggea) and Phyllanthus species, belonging to the
Euphorbiaceae family, and later also found in other
Euphorbiaceae species, such as Margaritaria and Breynia,
and in Zygogynum pauciflorum (Winteraceae). Due to their
medicinal properties, some of these plants have been widely
used for years in traditional folk medicine in China and
Amazonia. In particular, they have found application as
(4) Nadkarni, A. K. Indiana Materia Medica; Popular Prakashan Prt.
Ltd.: Bombay, 1976; Vol. 1, p 948.
(5) Chopra, R. N.; Nayar, S. L.; Chopra, I. C. Gloss. Indian Med. Plants
1956191; Chem. Abstr. 1957, 51, 31637.
(6) Isolation: (a) Murav'eva, V. I.; Ban'kovskii, A. I. Dokl. Akad. Nauk
SSSR 1956, 110, 998-1000;
Chem. Abstr. 1957, 51, 43396. Structural
(1) Reviews: (a) Snieckus, V. In The Alkaloids; Manske, R. H. F., Ed.;
Academic Press: New York, 1973; Vol. 14, pp 425-503. (b) Beutler, J. A.;
Brubaker, A. N. Drugs Future 1987, 12, 957–976. (d) Bayon, P.; Busque, F.;
Figueredo, M. Targets Heterocycl. Syst. 2005, 9, 281–310. (c) Weinreb, S. M.
Nat. Prod. Rep. 2009, 26, 758–775.
assignment : (b) Satoda, I.; Murayama, M.; Tsuji, J.; Yoshii, E. Tetrahedron Lett.
1962, 3, 1199–1206. (c) Horii, Z.; Tanaka, T.; Tamura, Y.; Saito, S.; Matsumura,
C.; Sugimoto, N. J. Pharm. Soc. Jpn. 1963, 83, 602-605; Chem. Abstr. 1963,
59, 9087c. (d) Saito, S.; Kotera, K.; Shigematsu, N.; Ide, A.; Sugimoto, N.; Horii,
Z.; Hanaoka, M.; Yamawaki, Y.; Tamura, Y. Tetrahedron 1963, 19, 2085–2099.
(e) Horii, Z.; Ikeda, M.; Yamawaki, Y.; Tamura, Y.; Saito, S.; Kodera, K.
Tetrahedron 1963, 19, 2101–2110. Absolute configuration: (f) Imado, S.; Shiro,
M.; Horii, Z. Chem. Ind. (London) 1964, 40, 1691. (g) Imado, S.; Shiro, M.; Horii,
Z. Chem. Pharm. Bull. 1965, 13, 643–651.
ꢀ
ꢀ
(2) van der Woerd, L. A. Geneeskd. Tijdschr. Ned.-Indie 1941, 81, 1963-
1980; Chem. Abstr. 1942, 36, 1025.
(3) Serra, R. M. An. Univ. Santo Domingo 1944, 8, 295-297; Chem.
Abstr. 1947, 41:8643.
DOI: 10.1021/jo901059n
r
Published on Web 07/24/2009
J. Org. Chem. 2009, 74, 6199–6211 6199
2009 American Chemical Society

ꢀ
Gonzalez-Galvez et al.
ꢀ
JOCArticle
CHART 1. Representative Examples of Securinega Alkaloids
The securinega alkaloids have been associated with a
number of biological activities, some of which are well
documented. Securinine is an ACE inhibitor9 and a stimu-
lant of the central nervous system10 and has shown anti-
malarial11 and antibacterial activities.12 It has been described
that several securinega alkaloids are specific GABAA recep-
tor antagonists,13 and some of them act also as antitumor
agents.14 (þ)-Norsecurinine inhibits spore germination of
some plant pathogenic fungi.15
Despite their attractive potential as pharmacological
agents, published synthetic investigations related to the
securinega alkaloids are rather limited. Racemic securinine
was synthesized by Horii et al. a few years after its isolation,16
while in 1987 Heathcock and von Geldern disclosed the first
synthesis of racemic norsecurinine.17 Two additional total
syntheses18 and one formal synthesis19 of (()-securinine
have since been reported, as well as one total synthesis of
(()-norsecurinine.20 In this last work, a small quantity of the
oxygen-bridged alkaloid (()-nirurine was also obtained. The
first nonracemic total synthesis of a securinega alkaloid was
reported in 1989 by Jacobi and co-workers, who described
the enantioselective syntheses of (þ)- and (-)-norsecurinine,
starting from L- and D-proline, respectively.21 In 2000, the
Weinreb group described the second successful approach
to (-)-norsecurinine,22 together with the first preparation
of (-)-dihydronorsecurinine, both alkaloids derived from
a common synthetic intermediate. Simultaneously, and
through an analogous sequence, these authors also reported
the preparation of phyllantine, the first nonracemic synthesis
among the securinine subgroup, and more recently, they
disclosed the first synthesis of secu’amamine A.23 Two rather
securinega alkaloids encloses a 6-azabicyclo[3.2.1]octane
system (rings B and C) fused to a 2-furanone (ring D) and
either a piperidine or pyrrolidine (ring A), with the size of this
last ring characterizing the securinine and norsecurinine type
subgroups, respectively. The members within each one of
these subgroups differ in the configuration of their stereo-
genic centers and/or they present slight functionality altera-
tions. The enantiomer of securinine, virosecurinine, and their
epimers at C2, allosecurinine (2) and viroallosecurinine, have
all been isolated from natural sources. (-)-Norsecurinine
[(-)-3], the first isolated and most representative member of
its subgroup,8 and its antipode (þ)-norsecurinine are both
naturally occurring too. There are also several securinega
alkaloids that show significant modifications of the typical
structural framework, as is the case of phyllantidine, which
presents rings A and C connected through an oxygen bridge;
nirurine, where the benzofuranone moiety (rings C and D) is
connected to the nitrogen atom of the pyrrolidine subunit
through a different position, or secu’amamine A, wherein
rings B and C conform an azabicyclo[4.2.1]octane system.
(9) Friess, S. L.; Durant, R. C.; Whitcomb, E. R.; Reber, L. J.; Thomme-
sen, W. C. Toxicol. Appl. Pharmacol. 1961, 3, 347-357; Chem. Abstr., 1961,
55, 132861.
(10) Russian Pharmacopeia, 10th ed.; Ministry of Health of the USSR:
Moscow, 1968; pp 612-614.
(11) Weenen, H.; Nkunya, M. H.; Bray, D. H.; Mwasumbi, L. B.;
Kinabo, L. S.; Kilimali, V. A.; Wijnberg, J. B. Planta Med. 1990, 56, 371–373.
(12) (a) Mensah, J. L.; Lagarde, I.; Ceschin, C.; Michel, G.; Gleye, J.;
Fouraste, I. J. Ethnopharmacol. 1990, 28, 129–133. (b) Bakshu, L. M.; Jeevan
Ram, A.; Venkata Raju, R. Fitoterapia 2001, 72, 930–933.
(13) (a) Beutler, J. A.; Karbon, E. W.; Brubaker, A. N.; Malik, R.; Curtis,
D. R.; Enna, S. J. Brain Res. 1985, 330, 135–140. (b) Xu, L.; Zhang, J. Neurol.
Res. 2004, 26, 792–796.
(14) (a) Tatematsu, H.; Mori, M.; Yang, T. H.; Chang, J. J.; Lee, T. T. Y.;
Lee, K. H. J. Pharm. Sci. 1991, 80, 325–327. (b) Liu, W.; Gu, Z.; Zhou, W.; Guo,
C. Zhongguo Yaolixue Tongbao 1997, 13, 529-532; Chem. Abstr. 1998, 129,
211309. (c) Dong, N. Z.; Gu, Z. L.; Chou, W. H.; Kwok, C. Y. Acta Pharmacol.
Sin. 1999, 20, 267-270; Chem. Abstr. 1999, 130, 346993n.
(15) Goel, M.; Maurya, S.; Pandey, V. B.; Singh, V. P.; Singh, A. K.;
Singh, U. P. Mycobiology 2002, 30, 225–227.
(16) Horii, Z.; Hanaoka, M.; Yamawaki, Y.; Tamura, Y.; Saito, S.;
Shigematsu, N.; Kotera, K.; Yoshikawa, H.; Sato, Y.; Nakai, H.; Sugimoto,
N. Tetrahedron 1967, 23, 1165–1174.
(7) (a) Mulchandani, N. B.; Hassarajani, S. A. Planta Med. 1984, 50, 104-
105; Chem. Abstr. 1984, 101, 107371. (b) Chen, M.; Hou, L. Zhiwu Xuebao
1985, 27, 625-629; Chem. Abstr. 1986, 104, 183282. (c) Joshi, B. S.; Gawad, D.
H.; Pelletier, S. W.; Kartha, G.; Bhandary, K. J. Nat. Prod. 1986, 49, 614–620.
(d) Petchnaree, P.; Bunyapraphatsara, N.; Cordell, G. A.; Cowe, H. J.; Cox, P. J.;
Howie, R. A.; Patt, S. L. J. Chem. Soc., Perkin Trans. 1 1986, 1551–1556.
(e) Negi, R. S.; Fakhir, T. M. Phytochemistry 1988, 27, 3027–3028. (f) Ahond,
A.; Guilhem, J.; Hamon, J.; Hurtado, J.; Poupat, C.; Pusset, J.; Pusset, M.; Sevenet,
T.; Potier, P. J. Nat. Prod. 1990, 53, 875–881. (g) Hassarajani, S. A.; Mulchandani,
N. B. Indian J. Chem. 1990, 29B, 801-803; Chem. Abstr. 1991, 114; 20973; (h)
Arbain, D.; Birkbeck, A. A.; Byrne, L. T.; Sargent, M. V.; Skelton, B. W.; White,
A.H.J.Chem. Soc.,Perkin Trans.11991, 1863–1869. (i)Lajis, N.H.;Guan, O.B.;
Sargent, M. V.;Skelton, B. W.;White, A. H. Aust. J.Chem. 1992, 45, 1893–1897. (j)
Luger, P.; Weber, M.; Dung, N. X.; Ky, P. T.; Long, P. K. Acta Crystallogr. 1995,
C51, 127–129. (k) Houghton, P. J.; Woldemariam, T. Z.; O'Shea, S.; Thyagarajan,
S. P. Phytochemistry 1996, 43, 715–717. (l) Babady-Bila; Gedris, T. E.; Herz, W.
Phytochemistry 1996, 41, 1441–1443. (m) Dehmlow, E. V.; Guntenhoner, M.;
Van Ree, T. Phytochemistry 1999, 52, 1715–1716. (n) Ohsaki, A.; Ishiyama, H.;
Yoneda, K.; Kobayashi, J. Tetrahedron Lett. 2003, 44, 3097–3099.
(17) Heathcock, C. H.; von Geldern, T. W. Heterocycles 1987, 25, 75–78.
(18) (a) Xi, F. D.; Liang, X. T. Acta Pharm. Sin. 1992, 27, 349–352. (b)
Liras, S.; Davoren, J. E.; Bordner, J. Org. Lett. 2001, 3, 703–706.
(19) (a) Honda, T.; Namiki, H.; Kudoh, M.; Watanabe, N.; Nagase, H.;
Mizutani, H. Tetrahedron Lett. 2000, 41, 5927–5930. (b) Honda, T.; Namiki,
H.; Kudoh, M.; Nagase, H.; Mizutani, H. Heterocycles 2003, 59, 169–187.
ꢀ
(20) (a) Magnus, P.; Rodrıguez-Lopez, J.; Mulholland, K.; Matthews, I.
J. Am. Chem. Soc. 1992, 114, 382–383. (b) Magnus, P.; Rodrıguez-Lopez, J.;
ꢀ
Mulholland, K.; Matthews, I. Tetrahedron 1993, 49, 8059–8072.
(21) (a) Jacobi, P. A.; Blum, C. A.; DeSimone, R. W.; Udodong, U. E. S.
Tetrahedron Lett. 1989, 30, 7173–7176. (b) Jacobi, P. A.; Blum, C. A.;
DeSimone, R. W.; Udodong, U. E. S. J. Am. Chem. Soc. 1991, 113, 5384–
5392.
(22) (a) Han, G.; LaPorte, M. G.; Folmer, J. J.; Werner, K. M.; Weinreb, S.
M. Angew. Chem., Int. Ed. 2000, 39, 237–240. (b) Han, G.; LaPorte, M. G.;
Folmer, J. J.; Werner, K. M.; Weinreb, S. M. J. Org. Chem. 2000, 65, 6293–6306.
(23) Liu, P.; Hong, S.; Weinreb, S. M. J. Am. Chem. Soc. 2008, 130, 7562–
7563.
(8) Isolationand structuralassignment of (-)-norsecurinine: (a) Iketubosin,
G. O.; Mathieson, D. W. J. Pharm. Pharmacol. 1963, 15, 810–815. Absolute
configuration: (b) Saito, S.; Tanaka, T.; Kotera, K.; Nakai, H.; Sugimoto, N.;
Horii, Z.; Ikeda, M.; Tamura, Y. Chem. Pharm. Bull. 1965, 13, 786–796.
Isolation of (þ)-norsecurinine: (c) Rouffiac, R.; Parello, J. Plant. Med.
Phytother. 1969, 3, 220-223; Chem. Abstr. 1970, 72, 32094.
6200 J. Org. Chem. Vol. 74, No. 16, 2009

ꢀ
Gonzalez-Galvez et al.
ꢀ
JOCArticle
SCHEME 1. Retrosynthetic Analysis for Securinine- and Norsecurinine-Type Alkaloids
SCHEME 2. Enantioselective Allylation of Imides 8 and 9a
similar diastereoselective syntheses of securinine have been
developed by the group of Honda24 and ours,25 both of them
starting from (R)-pipecolinic acid. As part of the same
investigations, the Japanese group has also accomplished
the synthesis of (þ)-viroallosecurinine,26 and we have
obtained a previously unknown epimer at C2 of (-)-norse-
curinine, named (-)-allonorsecurinine.25 Starting from
4-hydroxypyroglutamic acid, and through a closely related
strategy, a third synthesis of securinine has been disclosed
very recently.27 Also in recent publications,28 Kerr and co-
workers described the syntheses of (þ)-phyllantidine and
(-)-allosecurinine from enantiopure cyclopropane precur-
sors, which were prepared through multistep sequences from
commercially available materials. Making use of a chiral
auxiliary approach, we have also accomplished the asym-
metric syntheses of allosecurinine and viroallosecurinine.29
In parallel to these studies, we were trying to develop a
general and enantioselective strategy for the synthesis of
both types of securinega alkaloids. Our goal was originating
enantioselectivity in a catalytic process rather than deriving
it from a chiral pool material, in order to make either
antipode of the target alkaloid equally available. The
efficiency of our enantioselective approach has been illu-
strated by concluding the total syntheses of (þ)-securinine
and (-)-norsecurinine30 in a small number of steps and
excellent overall yields. A full account of this work is
reported herein.
aKey (a) [η3-C3H5PdCl]2, L*, NaHCO3, solvent, room temperature; (b)
TBDPSCl, imidazole, CH2Cl2, 0 °C f room temperature.
containing rings B and C. The stereogenic center of the
iminium intermediate 7 would furnish the configuration at
C7 in the alkaloid, while the diastereoselectivity of its addi-
tion to 6 would determine the relative configuration at C2.
The additional stereocenter simultaneously generated at C9
will be lost in the formation of the enolate, which is required
for the final intramolecular alkylation leading to the
tetracyclic skeleton of the alkaloid. Therefore, the main
challenges for the stereochemical control of the synthesis
were an enantioselective preparation of a suitable precursor
of 7 and getting a good diastereoselectivity in the vinyl-
ogous Mannich reaction. The final cyclization step paralle-
lizes that used by Jacobi and co-workers in their synthesis of
norsecurinine, where, except for the nitrogen protec-
tion, for n = 1 and R = Ms the last intermediate is iden-
tical to 4.
Results and Discussion
Synthetic Plan. Scheme 1 shows the retrosynthetic analysis
of our enantioselective approach, where key steps are a
vinylogous Mannich reaction between an acyliminium ca-
tion 7 and silyloxyfuran 6, providing, respectively, rings A
and D of the alkaloid, and a ring-closing metathesis (RCM)
reaction, which would furnish the seven-membered cycle
The Enantioselective Imide Allylation. According to the
above synthetic plan, our first task was the enantiodirected
introduction of the alkyl substituent at the N-position of the
starting imides. Recently, Trost et al. described the conver-
sion of racemic butadiene monoepoxide (10) into a single
enantiomeric product through a palladium-catalyzed asym-
metric allylic alkylation of phthalimide in the presence
of a chiral diphosphine ligand.31 With this work in mind,
we studied the reaction of succinimide (8) and glutarimide (9)
with epoxide 10 (Scheme 2) as the starting point for the
(24) Honda, T.; Namiki, H.; Kaneda, K.; Mizutani, H. Org. Lett. 2004, 6,
87–89.
ꢀ
ꢀ
ꢀ
(25) Alibes, R.; Ballbe, M.; Busque, F.; de March, P.; Elias, L.; Figueredo,
M.; Font, J. Org. Lett. 2004, 6, 1813–1816.
(26) Honda, T.; Namiki, H.; Watanabe, M.; Mizutani, H. Tetrahedron
Lett. 2004, 45, 5211–5213.
(27) Dhudshia, B.; Cooper, B. F. T.; Macdonald, C. L. B.; Thadani, A. N.
Chem. Commun. 2009, 463–465.
(28) (a) Carson, C. A.; Kerr, M. A. Angew. Chem., Int. Ed. 2006, 45, 6560–
6563. (b) Leduc, A. B.; Kerr, M. A. Angew. Chem., Int. Ed. 2008, 47, 7945–
7948.
ꢀ
ꢀ
ꢀ
ꢀ
(29) Bardajı, G. G.; Canto, M.; Alibes, R.; Bayon, P.; Busque, F.; de
March, P.; Figueredo, M.; Font, J. J. Org. Chem. 2008, 73, 7657–7662.
ꢀ
(30) A preliminary account of this work has been published: Alibes, R.;
Bayon, P.; de March, P.; Figueredo, M.; Font, J.; Garcıa-Garcıa, E.;
ꢀ
(31) Trost, B. M.; Bunt, R. C.; Lemoine, R. C.; Calkins, T. L. J. Am.
Chem. Soc. 2000, 122, 5968–5976.
ꢀ
Gonzalez-Galvez, D. Org. Lett. 2005, 7, 5107–5109.
ꢀ
J. Org. Chem. Vol. 74, No. 16, 2009 6201

ꢀ
Gonzalez-Galvez et al.
ꢀ
JOCArticle
TABLE 1. Enantioselective Allylation of Imides 8 and 9a
entry
imide
ligand
solvent
conversion (%)
C:Tb
ee of (þ)-11 or 12 (%)
1
2
3
4
5
6
7
8c
8
8
8
8
8
8
9
9
(þ)-L*-1
(þ)-L*-1
(þ)-L*-1
(þ)-L*-1
(þ)-L*-1
(þ)-L*-2
(þ)-L*-2
(þ)-L*-2
CH2Cl2
toluene
THF
CH3CN
CH3CN/H2O 3/1
CH2Cl2
92
100
96
45
15
99
64
100
10:1
16:1
64
29
46
81
90
87
90:1
10:1
CH2Cl2
CH2Cl2
>99
a Standard conditions: 0.4 mol % of [(C3H5)PdCl]2; 1.2 mol % of ligand*; 5.0 mol % of Na2CO3; room temperature. bRatio between central and
terminal attack product (11:13 or 12:14). cIn this experiment, double concentration of Pd complex, ligand, and base were used.
syntheses of norsecurinine and securinine, respectively. The
results are summarized in Table 1. The initial experiments
were performed using the most readily available ligand
L*-1.32 From the reaction between 8 and 10, under the
conditions previously described for phthalimide (entry 1)
(0.4 mol % of the π-allylpalladium chloride dimer, 1.2 mol %
of the chiral diphosphine ligand, and 5.0 mol % of Na2CO3,
in dichloromethane), after 14 h, the expected alkylation
product (þ)-11 was isolated in 83% yield, along with a minor
quantity (9%) of its regioisomer 13, resulting from the
terminal attack of the imide to the intermediate palladium
complex. In this run, the enantiomeric excess of (þ)-11 was
only moderate (64%). In an effort to improve this result, the
same reaction was performed in different solvents (entries
2-5). It was observed that an increase in polarity was
beneficial for the enantioselectivity, reaching 90% ee in
acetonitrile/water 3/1, but unfortunately, in polar solvents
the reaction became too slow. Consequently, we went back
to dichloromethane and changed the ligand to the bulkier
diphosphine L*-2, which was prepared as described.31
From this last experiment (entry 6), the desired regioisomer
(þ)-11 could be isolated in 91% yield with 87% ee. Surpris-
ingly, when the same conditions were applied to glutari-
mide (entry 7), we observed that both the regio- and
enantioselectivity decreased disappointingly. After some
experimentation, it was found that excellent yield, regios-
electivity, and ee can be obtained by using double catalyst
loading while keeping the Pd/ligand/base stoichiometry.
In this way, the allylated glutarimide (þ)-12 was obtained
in almost quantitave yield with >99% ee, and no traces of
its regioisomer 14 were detected (entry 8). The same
modification was then applied to the allylation of succini-
mide, but intriguingly, no improvement in the regio- or
enantioselectivity of the reaction was observed for this
substrate.
2-propanol furnished enantiomerically pure material in
overall 81% yield from succinimide, while enantiopure
silylether (-)-16 was obtained in quantitative yield from
glutarimide.
The Vinylogous Mannich Reaction. According to the
plan, we next performed the regioselective reduction of the
N-alkylated imides (-)-15 and (-)-16 to the corresponding
hemiaminals, required to generate the acyliminium Mannich
acceptors. For this transformation, LiEt3BH in THF worked
more efficiently than other attempted reducing agents
(NaBH4 or DIBALH), furnishing the mixtures of epimeric
hemiaminals 17 and 18 in 87% and 90% yield, respectively
(Scheme 3). Triisopropylsilyloxyfuran 6 was prepared in
97% yield from 4-vinyl-2(5H)-furanone, which was in turn
synthesized from β-tetronic acid in two steps and 68%
overall yield following a previously reported procedure.33
In the last years, the vinylogous Mannich reaction has
attracted the attention of several researchers in the field of
natural product synthesis.34 In relation to our work, parti-
cularly interesting were those precedents using a pyrrolinium
cation and a 2-silyloxyfuran as the reaction partners. Studies
on the stereochemical induction produced by a stereogenic
center enclosed in the pyrrolinium ring34e or by a chiral
auxiliary group attached to a carbamate nitrogen protec-
tion34c have been reported, but to the best of our knowledge,
an example with the stabilizing acyl group inside the ring and
the stereogenic motif outside has not been described yet.
A priori, the vinylogous Mannich reaction between silyloxy-
furan 6 and the iminium ions derived from 17 or 18 may
produce up to four diastereomers of 20 or 21, respectively,
among which the (2R,7R,9R) and (2R,7R,9S) isomers corre-
late with (-)-norsecurinine (n=1) or securinine (n=2), while
the (2S,7R,9S) and (2S,7R,9R) isomers of 21 correlate with
allosecurinine. To synthesize the corresponding, and also
natural, enantiomeric alkaloids, (þ)-norsecurinine, virose-
curinine, and viroallosecurinine, a starting hemiaminal with
the opposite configuration at C7 would be required. We
anticipated that the configurational assignment of adducts
20/21 would be quite complex, and hence, we decided to
acquire some theoretical knowledge on the stereochemical
course of the reaction.
The ee of the glutarimide derivative (þ)-12 was determined
by CHPLC. Surprisingly, in the case of the succinimide
derivative (þ)-11 the peak resolution was not possible and
the ee was established by NMR through preparation of the
diastereomeric Mosher’s esters. The absolute configuration
of the major enantiomers (þ)-11 and (þ)-12 was tentatively
assigned as R by analogy to the phthalimide derivative
obtained by Trost in the presence of the same ligand. The
enantiomerically enriched alcohols (þ)-11 and (þ)-12 ob-
tained under the conditions of entries 6 and 8 were converted
into the corresponding tert-butyldiphenylsilyl (TBDPS)
ethers (-)-15 and (-)-16. Crystallization of (-)-15 from
(33) Lattmann, E.; Hoffmann, H. M. R. Synthesis 1996, 155–163.
(34) (a) Martin, S. F.; Barr, K. J.; Smith, D. W.; Bur, S. K. J. Am. Chem.
Soc. 1999, 121, 6990–6997. (b) Bur, S. K.; Martin, S. F. Org. Lett. 2000, 2,
3445–3447. (c) D’Oca, M. G. M.; Pilli, A. R.; Vencato, I. Tetrahedron Lett.
2000, 41, 9709–9712. (d) Liras, S.; Lynch, C. L.; Fryer, A. M.; Vu, B. T.;
Martin, S. F. J. Am. Chem. Soc. 2001, 123, 5918–5924. (e) Reichelt, A.; Burr,
S. K.; Martin, S. F. Tetrahedron 2002, 58, 6323–6328. (f) Barnes, D. M.;
Bhagavatula, L.; DeMattei, J.; Gupta, A.; Hill, D. R.; Manna, S.;
MacLaughlin, M. A.; Nichols, P.; Premchandran, M.; Rasmussen, M. W.;
Tian, Z.; Wittenberger, S. J. Tetrahedron: Asymmetry 2003, 14, 3541–3551.
(32) Trost, B. M.; Van Vranken, D. L. Angew. Chem., Int. Ed. 1992, 31,
228–230.
6202 J. Org. Chem. Vol. 74, No. 16, 2009

ꢀ
Gonzalez-Galvez et al.
ꢀ
JOCArticle
SCHEME 3. Vinylogous Mannich Reaction Leading to Trienes 20 and 21
SCHEME 4. Models for the Density Functional Theory Calculations on the Vinylogous Mannich Reaction
Some years ago, Martin and Bur reported the results of ab
initio calculations concerning the vinylogous Mannich reac-
tion between 2-methoxyfuran and an achiral pyrrolinium
ion. These authors envisaged two kinds of transition states
(TS), “Diels-Alder-like TS” and “open-like TS”, and ratio-
nalized the empirical preference of the process to result in
threo over erythro products.34a,34d Taking this work as a
reference, we analyzed the reaction between the acyliminium
ion 22 and the furanyloxysilane 23 (Scheme 4), which we
judged appropriate models for our reagents, as a way to
evaluate the influence that the stereogenic center external to
the pyrrolinium ring may have on the stereoselectivity of the
process. Density functional theory calculations were carried
out to identify the structures of the eight competitive transi-
tion states A-H considered, using the program package
Gaussian 03 (rev E.01),35 the B3LYP36 combination of
functionals, and the 6-31G(d) as the basis. The relative
energies calculated for the eight saddle points suggested that
orientation A had the lowest energy, followed by C (0.3 kcal/
mol higher). This result predicted that threo adducts would
predominate, with the stereogenic center of the iminium
cation favoring (2R,7R,9R) over (2S,7R,9S) relative config-
uration in a ratio of 2:1 at 0 °C. The preferred adduct,
(2R,7R,9R)-24, had a suitable relative configuration to be
used as a precursor in the synthesis of norsecurinine.
Orientations H and E, coming next in stability (2.7 and 3.0
kcal/mol higher than A, respectively), would lead to erythro
diastereomers, which, consequently, would be less expected
among the reaction products. Finally, orientations D, B, F,
and G gave energies between 4.4 and 6.0 kcal/mol higher
than A. These values are too far from the minimum to have
an important effect on the diastereomeric ratio.
(35) Calculations using B3LYP/6-31G* as implemented by Gaussian 03,
Revision E.01 : Frisch, M. J. et al. Gaussian, Inc., Wallingford, CT, 2004.
(36) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785–789. (b)
Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652. (c) Stephens, P. J.; Devlin,
F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98, 11623–
11627.
To prove that the optimized geometries of the eight
calculated TS match the hybridization change on going from
the reactants to the product, the dihedral angles N-C2-C3-
C4 were observed. All of them were between 6° (TS C) and
J. Org. Chem. Vol. 74, No. 16, 2009 6203

ꢀ
Gonzalez-Galvez et al.
ꢀ
JOCArticle
FIGURE 1. Stick drawings of the saddle point structures of the most representative TS. Relative energies (rel ΔH) are given in kcal/mol.
TABLE 2. Vinylogous Mannich Reaction between Silyloxyfuran 6 and Iminium Precursors 17-19a
entry
iminium precursor
Lewis acid
TIPSOTfc
TIPSOTfc
T (°C)
conversion (%)
products (ratio)
20 or 21RRR:RRS:SRS:SRRb
1
2
3
4
5
17
17
17
17
17
17
17
17
17
17
17
18
18
18
18
18
18
18
18
18
19
19
19
19
19
19
-78
0
75
85
0
20
20
43:29:14:14
c
Cu(OTf)2
Sn(OTf)2
0
0
0
c
0
BF3 OEt2
3
>99
92
85
90
95
20
20
20
20
20
20
20
57: 29:5:9
62: 25:4:9
6
BF3 OEt2
3
-10
-20
0
7
BF3 OEt2
3
8d
9e
BF3 OEt2
3
BF3 OEt2
3
0
0
0
0
0
0
0
0
f
10
11
12
13h
14
15
16i
17
18
19
20j
21j
22j
23j
24i
25
26
BF3 OEt2
BF3 OEt2
60
85
3
g
3
BF3 OEt2
BF3 OEt2
>99
>99
>99
>99
>99
>99
20
<2
>99
>99
>99
>99
>99
25
21/26 (1:1.5)
26
26
48:4:24:24
3
3
TIPSOTf
Sn(OTf)2
BF3 OEt2
26
21/26 (1.5:1)
21/26 (2.3:1)
21/26 (3:1)
48:4:24:24
48:4:24:24
48:4:24:24
3
c
c
BF3 OEt2
-20
-40
-78
-20
-20
-20
-20
-20
-40
-78
3
BF3 OEt2
BF3 OEt2
3
3
BF3 OEt2
BF3 OEt2
21/26 (4:1)
21/26 (>20:1)
21/26 (>20:1)
26
21/26 (1:2.3)
21/26 (4:1)
48:4:24:24
50:-:35:15
59:-:24:17
3
3
n-Bu2BOTf
(R,R)-IpcBCl k
(R,R)-IpcBOTfk
n-Bu2BOTf
60:-:22:18
53:-:33:14
n-Bu2BOTf
<2
aStandard conditions: 1 equiv of aminal (17, 18, or 19); 1.2 equiv of 6 (prepared from 1.2 equiv of 4-vinyl-2(5H)-furanone, 1.6 equiv of Et3N, and 1.3
equiv of TIPSOTf); 2.5 equiv of Lewis acid; diethyl ether; 30 min. bDiastereomeric ratio was measured by 1H NMR analysis of the crude product of the
vinylogous Mannich reaction. cReaction time 120 min. d1 equiv of 6 was used. e1.1 equiv of 6 was used. f1 equiv of Lewis acid was used. g1.75 equiv of
Lewis acid were used. hA 3-fold excess of Et3N was used in the preparation of 6. iFuran 6 was prepared in situ. jThe reaction was run in CH3CN. kIpc =
isopinocamphenyl.
20° (TS E), indicating that the five-membered ring adopts a
shallow envelop conformation. The formation-bond dis-
(TIPSOTf) as the Lewis acid promoter, in diethyl ether at
-78 °C (entry 1). Under these conditions, after 2 h of
reaction, the conversion of the hemiaminal was incomplete
(around 75%), and a complex mixture of products was
formed. Increasing the temperature to 0 °C (entry 2) im-
proved the conversion, but still some unreacted hemiaminal
remained. Next, alternative Lewis acid promoters were tried
˚
tances were also analyzed, with the shortest being 2.25 A
˚
and the longest 2.43 A, but no relation between this para-
meter and the relative stability of the TS was observed.
Figure 1 shows the saddle point structures for the most
representative TS. The calculated data show that a stereo-
genic center attached to the five-membered ring by the
nitrogen atom could affect the diastereoselectivity of the
reaction, despite its distance to the reactant carbon atom and
even neither of the faces of the pyrrolinium cation being
clearly blocked.
Table 2 summarizes the experimental results of the vinyl-
ogous Mannich reactions leading to trienes 20 and 21. The
first experiment was performed with hemiaminal 17, using a
20% excess of furan 6 and 2.5 equiv of triisopropyl triflate
(entries 3-5), and it was found that BF3 Et2O furnished the
3
best conversion, with a moderate enhancement of the dia-
stereoselectivity compared to the reactions promoted by
TIPSOTf. Attempts to enhance it further by lowering the
temperature (entries 6 and 7) led to lower conversions of 17,
probably due to the competitive hydrolysis of the silylox-
yfuran. Some experiments were also performed by decreas-
ing the relative amounts of furan 6 (entries 8 and 9) or Lewis
acid (entries 10 and 11), but none of them improved the
6204 J. Org. Chem. Vol. 74, No. 16, 2009

ꢀ
Gonzalez-Galvez et al.
ꢀ
JOCArticle
SCHEME 5. Pathways for the Formation of Enamide 26
this intermediate. Thus, the change to acetonitrile favored
the formation of adducts 21 over enamide 26 (entry 20).
Since hemiaminal analogues lacking the free hydroxyl
group are often used as precursors of acyliminium ions,38
we performed the acetylation of hemiaminal 18 under stan-
dard conditions,39 and the reaction was attempted starting
from its acetyl derivative 19 (entry 21). Rewardingly, with
this new substrate, only trace amounts of enamide 26 were
detected, and the diastereoselectivity of the addition process
was also improved. Encouraged by this good result, we
decided to study the effect of the boron substituents over
the reaction (entries 22-24). The best result was obtained
with nBu2BOTf, which gave good conversion, negligible
enamide formation, and the higher proportion of the major
isomer of 21. A pair of experiments at lower temperatures
were also carried on (entries 25-26), but further improve-
ment was not accomplished.
For synthetic purposes, the conditions of entry 22 were
found to be the most convenient. As in the case of their five-
membered-ring analogues 20, the vinylogous Mannich
adducts 21 are not stable to long contact with silica gel or
alumina, but a fast flash chromatography on silica gel
allowed the separation of the mixture in two fractions with
two diastereomers in each, RRR- and RRS-21 (50% yield)
and SRS- and SRR-21 (35% yield). We attribute the forma-
tion of small quantities of adduct RRS-21, which was not
present in the reaction crude material, to partial epimeriza-
tion of C-9 in RRR-21 during the chromatographic purifica-
tion. Fortunately, both isomers correlate with securinine and
the synthesis could therefore be continued with the mixture.
As above, the stereochemical assignment of the isomeric
adducts was established after the subsequent RCM step,
when the more rigid tricyclic structures facilitated their
elucidation.
Ring-Closing Metathesis Process. In order to determine
the relative configuration of the vinylogous Mannich ad-
ducts 20 and 21, RCM reactions40 starting from mixtures
containing diverse proportions of two, three, or four isomers
of 20 and 21 were performed (Scheme 6). These reactions
were run in dichloromethane at room temperature, with 15%
molar of the second-generation Grubbs’ catalyst. Starting
from a mixture of the four diastereomers of 20, we obtained
27, also as a mixture of four diastereomers, matching the
relative proportion of the starting trienes, in ca. 70% yield.
The tricyclic olefins 27 could be chromatographically
separated in three fractions (RRR, SRRþSRS, and RRS)
and their configuration determined with the help of NOESY
experiments (Figure 2). For the major isomer of 27, a clear
interaction was observed between the three protons at the
stereogenic centers, H2, H7, and H9, indicating that they are
sharing the same face of the tricycle, and hence, it was
assigned as the threo isomer RRR-27. The NOE interactions
shown by the other isomer isolated as a single compound,
which correlates with the second most abundant product of
the vinylogous Mannich reaction, were only consistent with
the erythro isomer RRS-27. Hence, both isolated isomers
results attained under the conditions established in entry 5.
All attempts to separate the isomers of 20 by chromatogra-
phy on silica gel or alumina or by fractional crystallization
met with failure, due to the low stability of these compounds,
but a purified mixture of all the isomers could be obtained by
quick filtration of the crude product through silica gel. On
the other hand, on standing at room temperature overnight,
the major adduct crystallized from a reaction mixture with-
out any solvent, and it could be separated by filtration in
51% yield. The relative configuration of the isomers of 20
was established by performing a RCM experiment with a
mixture containing all the isomers (vide infra). By correla-
tion, the isolated vinylogous Mannich adduct was identified
as RRR-20 and the second most abundant one as its erythro
epimer RRS-20. This stereochemical assignment is in partial
agreement with the theoretical calculations, which antici-
pated the preferential formation of the major adduct RRR-
20 but predicted a stronger predominance of threo over
erythro isomers. Hence, it has to be concluded that the
DFT calculations underestimate the influence of the stereo-
genic center external to the ring on the threo/erythro selec-
tivity of the process.
The first trials with the six-membered cyclic hemiaminal
18 were performed under the best conditions found for its
five-membered analogue (entry 12). Disappointingly, we
found that, instead of the expected vinylogous Mannich
adducts 21, the major product of the reaction was enamide
26 (Scheme 5), resulting from dehydration of 18 through
direct β-elimination or with the intermediacy of the iminium
cation 25. A control experiment under the same conditions,
except for the absence of furan 6, rendered enamide 26
quantitatively after only 5 min. To preclude silyloxyfuran
hydrolysis as the cause of the undesired pathway, furan 6 was
prepared in the presence of a large excess of triethylamine,37
but the formation of the enamide was even more favorable
(entry 13). The same occurred when TIPSOTf or Sn(OTf)2
were used as the reaction promoters (entries 14-15). Certain
improvement was accomplished when furan 6 was prepared
in situ (entry 16), and a decrease in the formation of enamide
was also observed performing the reaction at -20 °C, but
lowering further the temperature stopped the conversion
(entries 17-19). As it was obvious that the stability of the
iminium cation 25 played a crucial role, we decided to assay
the reaction in a more polar solvent, which could stabilize
(38) Speckamp, W. N.; Moolenaar, M. J. Tetrahedron 2000, 56, 3817–
3856.
_
ꢀ
(39) Kazuza, Z.; Mostowicz, D.; Doze-ga, G.; Mroczko, K.; Wojcik, R.
Tetrahedron 2006, 62, 943–953.
€
(37) The isolation of furan 6 is problematic because it hydrolyzes very fast
in the presence of any trace of acid.
(40) Reviews: (a) Furstner, A. Angew. Chem., Int. Ed. 2000, 39, 3012–
3043. (b) Grubbs, R. H. Tetrahedron 2004, 60, 7117–7140.
J. Org. Chem. Vol. 74, No. 16, 2009 6205

ꢀ
Gonzalez-Galvez et al.
ꢀ
JOCArticle
FIGURE 2. Significant NOEs observed in compounds 27 (CDCl3) and 28 (C6D6), R=TBDPS.
SCHEME 6. RCM Reaction of the Mixtures of Isomers 20 and 21
present the same C2/C7 relative configuration as in norse-
curinine.
NOEs observed between the protons at these positions. The
relative configuration of the other isomer accompanying the
former was more difficult to ascertain, but a NOE interac-
tion between H2 and H7 pointed to the 2R,7R,9S relative
configuration, meaning that both isomers of the mixture
would correlate with securinine. For the second major iso-
mer, which was the one isolated as a single product, positive
NOE’s between protons H2, H9, and H8 were consistent
with the relative 2S,7R,9S configuration as in allosecurinine.
Completion of the Synthesis of (-)-Norsecurinine. With the
tricyclic diene RRR-27 in hands, the two transformations
required to conclude the synthesis of (-)-norsecurinine were
reducing the lactam to amine and connecting the stereogenic
centers C7 and C9 through the one-carbon bridge to com-
plete the tetracyclic skeleton (Scheme 7). Conversion of
RRR-27 into the corresponding thiolactam 29 was accom-
plished in 82% yield by reaction with Lawesson’s reagent in
refluxing THF, but treatment of 29 with Raney nickel or
other reducing agents (including LiAlH4, AlH3, NaBH4, and
When the RCM reaction was applied to the crystallized
isomer RRR-20, the expected diene RRR-27, [R] = þ118
(c 0.61, CHCl3), was isolated in nearly quantitative yield.41
The RCM reaction of a mixture of the four diastereomers
of 21 furnished the corresponding mixture of 28. Unfortu-
nately, it was not possible to separate the diastereomers of 28
because prolonged contact with silica gel or alumina led to
partial decomposition, with the isolated yield dropping
under 20%. However, starting from the mixture of RRR-
and RRS-21, after quick filtration through silica gel, we
isolated a mixture of RRR- and RRS-28 in the same ratio as
the starting trienes and 97% total yield. Unexpectedly, the
complementary mixture of SRS- and SRR-21 reacted in the
presence of the second-generation Grubbs’ catalyst to give
the diastereomer SRS-28, [R]=þ87 (c 1.45, CHCl3), as the
only reaction product in 98% yield. In view of the high yield
isolated, this fact was attributed to epimerization of the
allylic stereocenter C9 in the presence of the RCM catalyst.
In fact, epimerization of vinylcyclopropanes in the presence
of the first-generation Grubbs’ catalyst or first-generation
Hoveyda-Grubbs’ catalyst has previously been described,42
and these catalysts have also been used for olefin isomeriza-
tion in substrates bearing an oxygen or nitrogen atom at the
allylic position,43 a process that has even been applied to
total syntheses.44 As above, the relative 2R,7R,9R config-
uration of the major isomer of 28 was deduced from the
BH3 THF under different reaction conditions) ended up
3
with decomposition products. Apparently, the lactone and/
or olefin functionalities were more reactive toward these
reducing agents than the thiolactam moiety. In view of these
failures, we decided to attempt the direct reduction of the
lactam to amine, which a priori had been judged more
problematic. Recently, it was reported the use of borane to
reduce a lactam without affecting neither the ester moiety
nor the olefins present in the substrate,45 but lactam RRR-27
displayed low reactivity and poor selectivity toward
BH3 THF. It is also known that AlH3 is able to reduce
3
(41) In a work scale below 100 mg, the yield was quantitative, although
for larger quantities of starting material (around 500 mg), the yield was
somewhat lower (around 80%).
(42) Zeng, X.; Wei, X.; Farina, V.; Napolitano, E.; Xu, Y; Zhang, L.;
Haddad, N.; Yee, N. K.; Grinberg, N.; Shen, S.; Senanayake, C. H. J. Org.
Chem. 2006, 71, 8864–8875.
(43) (a) Hu, Y.-J.; Dominique, R.; Das, S. K.; Roy, R. Can. J. Chem.
2000, 78, 838–845. (b) Nakashima, K.; Ito, R.; Sono, M.; Tori, M. Hetero-
cycles 2000, 53, 301–314. (c) Braddock, D. C.; Wildsmith, A. J. Tetrahedron
Lett. 2001, 42, 3239–3242. (d) Arisawa, M.; Terada, Y.; Nakagawa, M.;
Nishida, A. Angew. Chem., Int. Ed. 2002, 41, 4732–4734. (e) Sutton, A. E.;
Seigal, B. A.; Finnegan, D. F.; Snapper, M. L. J. Am. Chem. Soc. 2002, 124,
13390–13391.
lactams much faster than esters and that it does not react
with olefins, conjugated or not.46 The exclusive reduction
of an ester with this reagent in a substrate containing
an olefin was also described,47 but the selective reduction
of a lactam in a substrate bearing additional ester and olefin
functionalities has not been published. After extensive
ꢀ
ꢀ
(45) Dupont, C.; Guenard, D.; Tchertanov, L.; Thoret, S.; Gueritte, F.
Bioorg. Med. Chem. 1999, 7, 2961–2969.
(46) Brown, H. C.; Yoon, N. M. J. Am. Soc. Chem. 1966, 88, 1464–1472.
(47) Giraud, L.; Huber, V.; Jenny, T. Tetrahedron 1998, 58, 11899–11906.
(44) Wipf, P.; Rector, S. R.; Takahashi, H. J. Am. Chem. Soc. 2002, 124,
14848–14849.
6206 J. Org. Chem. Vol. 74, No. 16, 2009

ꢀ
Gonzalez-Galvez et al.
ꢀ
JOCArticle
SCHEME 7. Completion of the Synthesis of (-)-Norsecurinine
SCHEME 8. Completion of the Synthesis of Securinine
SCHEME 9. Attempted Preparation of SRS-35 en Route to
Allosecurinine, 2
experimentation, we found that treatment of lactam RRR-27
with freshly prepared 1 M AlH3 in THF at 0 °C for 5 min
produced the desired amine 30, [R]=þ89 (c 1.39, CHCl3), in
59% yield after chromatographic purification over silica gel,
with 10% of the starting lactam being recovered. The 12,13-
saturated amine 31 was identified as the main byproduct of
this reaction. Desilylation of 30 by different standard pro-
tocols (including treatment with TBAF in various solvents
and BF3 OEt2 in the presence of silica gel) met with failure,
3
but the free alcohol 32, [R]=þ123 (c 0.60, CH2Cl2), could be
finally obtained in 94% yield by reaction of 30 with an excess
of Et3N 3HF in THF at room temperature. Although
3
alcohol 32 is the penultimate intermediate in the synthesis
of (-)-norsecurinine published by Jacobi et al.,21 its optical
rotation value was not given. Hence, to assess the absolute
configuration of our synthetic material we completed the
synthesis of the alkaloid by mesylation and subsequent
treatment with potassium bis(trimethylsilyl)amide, accord-
ing to Jacobi’s protocol. These transformations were satis-
factorily reproduced in 78% yield for the two steps, and we
ended up with the levorotatory enantiomer of norsecurinine,
[R]=-270 (c 0.20, EtOH), lit.8b [R]=-270 (c 6.9, EtOH).
Overall, the synthesis of (-)-norsecurinine was completed in
nine steps from succinimide and 14% total yield.
AlH3 in THF, but in this case the reaction was much slower,
even at higher temperatures, and we were unable to obtain
the corresponding isolated amines in more than 40% yield.
Among many alternative reducing agents and conditions
tried, the best results were obtained when the mixture of
RRR- and RRS-28 was treated with a 7.5-fold molar amount
of BH3 THF at 0 °C for 45 min. This reaction furnished
3
amines RRR- and RRS-33 in 70% isolated yield, along with
another product identified as alcohol 34 (10% yield), and
recovered starting lactam (10%). Performing the reaction at
room temperature led to decomposition products, while at
lower temperatures (-20 °C) the conversion was very slow
and borane addition to the olefins was observed at prolonged
reaction times. It is worth mentioning that the ratio RRR-33:
RRS-33 of isolated amines was not the same as the ratio of
the starting lactams, and the yield of the reduction dimin-
ished if the initial mixture had a higher proportion of
Completion of the Synthesis of Securinine. To conclude the
synthesis of securinine, we sought to elaborate the mixture of
tricyclic dienes RRR- and RRS-28 (Scheme 8). Considering
the previous results within the norsecurinine sequence, the
reduction of the lactam was first attempted by treatment with
J. Org. Chem. Vol. 74, No. 16, 2009 6207

ꢀ
Gonzalez-Galvez et al.
ꢀ
JOCArticle
SCHEME 10. Complete Optimized Route to (-)-Norsecurinine and Securininea
aKey: (a) (()-10, [η3-C3H5PdCl]2, (þ)-L*-2, NaHCO3, CH2Cl2, room temperature; (b) TBDPSCl, imidazole, CH2Cl2, 0 °C f room temperature; (c) crystalli-
zation from 2-propanol; (d) LiHBEt3, THF, -78 °C; (e) 6, BF3 OEt2, Et2O, 0 °C; (f) Grubbs II, CH2Cl2, room temperature; (g) AlH3, THF, 0 °C; (h)
3
Et3N HF, THF, room temperature; (i) MsCl, Et3N, CH2Cl2, 0 °C, then KHMDS, CH2Cl2, -78 °C f room temperature; (j) Ac2O, DMAP, CH3CN, 0 °C,
3
then 6, n-Bu2OTf, CH3CN, -20 °C; (k) BH3 THF, THF, 0°C.
3
RRS-28. Intrigued by the formation of alcohol 34, we
investigated if this product was formed in the reaction
medium or during the methanolic workup by performing
1H NMR analysis of the reaction mixture before the quench-
ing, and we observed that alcohol 34 was already present in
the reaction medium. A reference experiment showed that,
under the reduction conditions, the lactam moiety of alcohol
34 remains unchanged. As a plausible explanation for the
formation of 34, we visualize a hydroboration-oxidation
process, in which residual traces of ruthenium species from
the former RCM step would act as the oxidizing agent.48
Cleavage of the silyl ether in RRR/RRS-33 by treatment
yield, but all attempts to desilylate SRS-33 with different
reagents and conditions failed. Although most standard
desilylation protocols were tried, the starting material re-
mained unchanged or it delivered complex mixtures of
products, among which the expected aminoalcohol SRS-35
was never detected. Trials to trap the alcohol by in situ
mesylation were also unsuccessful. A plausible explanation
to this pitfall could be the inherent instability of amino
alcohol SRS-35, which may be prone to undergo skeletal
rearrangements with the intermediacy of an aziridinium
cation such as 36.49 A related process has been postulated
for the biogenetic transformation of 3-β-hydroxyallosecur-
inine into secu’amamine A.50
with Et3N 3HF in THF at room temperature for 7 h
3
delivered the expected alcohols RRR/RRS-35 in 88% yield
after chromatographic purification. The alcohols were con-
verted in the corresponding mesylates, which showed low
stability and were therefore immediately treated with
KHMDS in dichloromethane, giving securinine as the un-
ique product in 79% yield from the alcohols 35. The NMR
spectra of the synthetic material matched that of the natural
alkaloid, and its specific rotation, [R] = -1061 (c 1.08,
CHCl3), was in agreement with the previously reported
values for securinine, [R]=-1082 (c 1.0, CHCl3).24 Overall,
the synthesis of securinine was completed in 10 steps from
glutarimide and 20% total yield.
Conclusion
In summary, we have developed a new synthetic strategy
for securinega alkaloids. The key transformations are an
enantioselective allylation of a cyclic imide, a vinylogous
Mannich reaction, and a ring-closing metathesis pro-
cess. The palladium-catalyzed enantioselective allylation
of phthalimide developed by Trost31 has been extended
to succinimide and glutarimide, providing a new and
(49) In SRS-35, the nitrogen lone pair and the “activated” hydroxyl
group are suitably oriented for an intramolecular SN2 displacement leading
to an aziridinium cation. This is not the case for RRR-35.
Attempted Synthesis of Allosecurinine. We then intended
to perform a parallel sequence starting from the tricyclic
diene SRS-28, in order to complete the synthesis of allose-
curinine (Scheme 9). Promisingly, treatment of SRS-28 with
BH3 THF in THF at 0 °C for 45 min furnished the corre-
3
sponding amine SRS-33, [R]=þ63 (c 0.73, CH2Cl2), in 97%
(48) For a related process, see: Yates, M. H. Tetrahedron Lett. 1997, 38,
2813–2816.
(50) Magnus, P.; Padilla, A. I. Org. Lett. 2006, 8, 3569–3571.
6208 J. Org. Chem. Vol. 74, No. 16, 2009

ꢀ
Gonzalez-Galvez et al.
ꢀ
JOCArticle
convenient access to enantiopure pyrrolidine and piperidine
derivatives. DFT calculations performed in a model system
predicted the predominance of the (2R,7R,9R) diastereos-
electivity experimentally observed in the vinylogous
Mannich reaction. Extremely concise syntheses of the most
representative members of each subtype of the securinega
alkaloids have been completed (Scheme 10). Thus,
(-)-norsecurine has been prepared from succinimide in nine
steps and 14% overall yield, while the synthesis of securinine
has been accomplished starting from glutarimide in 10 steps
and 20% overall yield. The absolute configuration of the
final alkaloids has demonstrated that the stereochemical
assignments of all the intermediates were correct. Both
syntheses can be performed in a synthetically useful scale
and compare favorably with the previously described routes,
which were more complicated and long and with signi-
ficantly lower overall yields. Moreover, all the previous
syntheses directed to a single enantiomer of the target
alkaloid made use of chiral pool precursors, while in our
approach, the enantioselectivity was originated by the
phosphine ligand (þ)-(S,S)-L*-2, the antipode of which is
equally available. Consequently, the same route is expected
to provide also access to (þ)-norsecurinine and virose-
curinine.
trough Celite. The filtrate was concentrated under vacuum, and
the resulting yellowish wax was purified by flash chromatogra-
phy on silica gel (gradient, hexanes/ethyl acetate 9:1 to 8:2) to
give (-)-16 (4.60 g, 10.9 mmol, 100%) as a white solid: Rf=0.58
(hexanes/ethyl acetate 1:1); mp = 88-89 °C (2-propanol);
[R]20D = -13 (c 1.20, CHCl3); IR (ATR) 2960, 2856, 1679,
1467, 1386, 1351, 1177, 1109; 1H NMR (250 MHz, CDCl3)
7.61 (m, 4H), 7.36 (m, 6H), 6.04 (ddd, J=17.5, 10.4, 7.3 Hz, 1H:
0
H2’), 5.52 (m, 1H:H1 ), 5.15 (dt, J=17.5, 1.4 Hz, 1H:H3 ), 5.11
0
0
(dt, J=10.4, 1.4 Hz, 1H:H3 ), 4.20 (t, J=9.7 Hz, 1H:H1 ), 3.81
00
00
(dd, J=9.7, 6.3 Hz, 1H:H1 ), 2.60 (t, J=6.6 Hz, 4H:2H3,2H5),
1.88 (qn, J = 6.6 Hz, 2H:2H4), 1.00 (s, 9H); 13C NMR (62.5
MHz, CDCl3) δ 172.3 (C2,C6), 135.3 (CPh), 135.2 (CPh), 133.4
(CPh), 133.3 (CPh), 133.1 (CPh), 129.5 (C2’), 129.4 (CPh), 127.6
0
00
(CPh), 118.4 (C3 ), 62.9 (C1 ), 56.1 (C1’), 33.4 (C3,C5), 26.7 (CMe),
19.1 (CtBu), 17.2 (C4); HRMS (ESIþ) calcd for C25H31NO3Si
444.1965 [M þ Na]þ, found 444.1957. Anal. Calcd for
C25H31NO3Si: C, 71.22; H, 7.41; N, 3.32. Found: C, 71.22; H,
7.41; N, 3.39.
(6RS)-1-[(1R)-1-(1-tert-Butyldiphenylsilyloxymethyl)-2-pro-
penyl]-6-hydroxyhexahydro-2-pyridinone (18). A solution of
LiBEt3H in THF (1M, 14 mL, 14 mmol) was added dropwise
to a solution of 16 (3.70 g, 8.78 mmol) in dry THF (34 mL) at
-78 °C, and the reaction mixture was stirred at the same
temperature for 1 h. Keeping the temperature at -78 °C,
saturated aqueous NaHCO3 (66 mL) and H2O2 (16.4 mL) were
added, and the mixture was allowed to warm slowly to room
temperature and then stirred for an additional 1 h. After
filtration through Celite, the solution was extracted with
CH2Cl2 (4 ꢀ 50 mL), and the combined organic extracts were
dried over anhydrous Na2SO4 and concentrated under va-
cuum. The oily residue was purified by flash chromatography
on silica gel (gradient, hexanes/ethyl acetate 9:1 to 7:3) to give a
mixture of isomers 18 (3.46 g, 8.17 mmol, 93%) as a colorless
Experimental Section
General experimental details are provided in the Supporting
Information. Securinine numbering was used for the signal assign-
ment of 1H and 13C NMR spectra of compounds 21, 28, 33, and 35.
1-[(1R)-1-(Hydroxymethyl)-2-propenyl]-2,6-piperidinedione
[(þ)-12]. A mixture of 33 mg (0.090 mmol) of π-allylpalladium
chloride dimer, 220 mg (0.278 mmol) of (þ)-L*-2, 122 mg
(1.15 mmol) of sodium carbonate, and 1.3 g (11.7 mmol) of
glutarimide, 9, was purged with nitrogen for 1 h. Dry CH2Cl2
(93 mL) was added, and the mixture was stirred at room
temperature for 10 min. Then, 950 μL (11.8 mmol) of buta-
diene monoepoxide, 10, was added, and the resulting mixture
was efficiently stirred under nitrogen for 14 h. The solvent was
removed under vacuum, and the oily residue was purified by
flash chromatography on silica gel (gradient hexanes/ethyl
acetate 1:1 to ethyl acetate) to give 2.14 g (11.7 mmol, 100%)
of (þ)-12 (99% ee) as a colorless oil: Rf=0.30 (ethyl acetate);
[R]20D = þ35 (c 0.95, CHCl3); IR (ATR) 3272, 2961, 1639,
1562, 1384, 1111; 1H NMR (250 MHz, CDCl3) δ 6.03 (dddd,
1
oil: IR (ATR) 3391, 3072, 2956, 2857, 1619, 1111; H NMR
(400 MHz, CDCl3) isomers A (major) and B (minor) δ 7.65
(m, 4HA þ 4HB: 4HPhA,4HPhB), 7.42 (m, 6HA þ 6HB:
0
6HPhA,6HPhB), 6.15 (ddd, J = 17.5, 10.5, 7.2 Hz, 1HB:H2 B),
0
5.76 (ddd, J = 17.3, 10.6, 5.9 Hz, 1HA:H2 A), 5.28-5.15
0
0
(complex, 3HA:H2 A,H1 A,OHA), 5.15-5.03 (complex,
0
0
2HAþ3HB:H2 A,H6A,2H2 B,H6B), 4.30 (dd, J = 10.3, 9.3 Hz,
00
1HB:H1 B), 4.26 (bs, 1HB:OHB), 3.94 (dd, J = 11.3, 3.4 Hz,
00
1HA:H1 A), 3.82 (dd, J=11.3, 6.0 Hz, 1HA:H1 A), 3.62 (dd, J=
00
00
10.3, 4.0 Hz, 1HB:H1 B), 2.62 (m, 1HA:H3A), 2.52 (m, 1HB),
2.41 (m, 1HA:H3A), 2.31 (m, 1HA:H3A), 2.17 (m, 1HB), 2.10 (m,
1HA:H4A), 2.08 (m, 1HB), 1.92 (m, 1HB), 1.82 (m, 1HB) 1.73 (m,
2HA:2H5A), 1.07 (s, 9HB:9HtBuB), 1.06 (s, 9HA:9HtBuA); 13C
NMR (100 MHz, CDCl3) isomer A δ 170.7 (C2), 135.7 (CPh),
0
J =17.6, 10.6, 7.0, 1.5 Hz, 1H:H2 ), 5.31 (bqd, J = 6.8, 1.2,
0
135.5 (CPh), 132.8 (C2 ), 131.9 (CPh), 130.2 (CPh), 127.9 (CPh),
0
1H:H1 ), 5.13 (bdd, J = 17.6, 1.2 Hz, 1H:H3 ), 5.12 (bdd,
0
0
00
127.9 (CPh), 118.7 (C3 ), 75.4 (C6), 64.9 (C1 ), 56.0 (C1’), 32.4
(C3), 29.5 (C5), 26.8 (CMe), 19.2 (CtBu), 15.4 (C4); HRMS
(ESIþ) calcd for C25H33NO3Si 446.2122 [M þ Na]þ, found
446.2109.
0
00
J = 10.6, 1.2 Hz, 1H:1H3 ), 3.93 (m, 1H:H1 ), 3.75 (dd, J =
00
11.4, 5.0 Hz, 1H:H1 ), 2.98 (bs, 1H:HOH), 2.60 (t,
C
J=6.4 Hz, 4H:2H3, 2H5), 1.88 (qn, J=6.3 Hz, 2H:2H4); 13
0
NMR (62.5 MHz, CDCl3) δ 173.2 (C2,C6), 133.1 (C2 ), 118.3
(6R)-1-[(1R)-1-(tert-Butyldiphenylsilyloxymethyl)-2-propenyl]-
6-[(2RS)-5-oxo-3-vinyl-2,5-dihydro-2-furyl]hexahydro-2-pyridi-
none (RRR/RRS-21). In a 250 mL Schlenk vessel, hemiaminal
18 (1.66 g, 3.92 mmol) was dissolved in dry acetonitrile (21.6
mL) under nitrogen. The solution was cooled to 0 °C, dimethyl-
aminopyridine (527 mg, 4.31 mmol) and acetic anhydride (1.5
mL, 15.68 mmol) were added, and the reaction mixture was
kept at the same temperature for 15 min. After this time, the
cooling bath was removed, and the mixture was kept at room
temperature for 1 h. It was then treated with a mixture of ice and
water (approx 10 mL), and the aqueous phase was extracted
with CH2Cl2 (3 ꢀ 15 mL). The combined organic extracts were
successively washed with saturated aqueous NaHCO3 (10 mL),
brine (10 mL), and cold water (10 mL). The organic phase was
dried over anhydrous Na2SO4 and concentrated under vacuum.
0
00
0
(C3 ), 62.6 (C1 ), 56.3 (C1 ), 33.2 (C3,C5), 16.9 (C4); HRMS
(ESIþ) calcd. for C9H13NO3 206.0793 [M þ Na]þ, found
206.0788.
For preparative purposes, the crude material can be used in
the next step without chromatographic purification, with the
overall yield being unaffected.
1-[(1R)-1-(tert-Butyldiphenylsilyloxymethyl)-2-propenyl]-2,6-
piperidinedione [(-)-16]. Imidazole (3.74 g, 54.9 mmol) and
TBDPSCl (5.7 mL, 21.9 mmol) were added to a solution of
(þ)-12 (2.00 g, 10.9 mmol) in dry CH2Cl2 (97 mL) under
nitrogen at 0 °C. Then, the cooling bath was removed and the
mixture stirred at room temperature, under nitrogen, for 14 h.
The solvent was evaporated under vacuum and replaced by ethyl
acetate (55 mL). The resulting mixture was stirred vigorously
and the insoluble fine white powder (imidazole HCl) filtered
3
J. Org. Chem. Vol. 74, No. 16, 2009 6209

ꢀ
Gonzalez-Galvez et al.
ꢀ
JOCArticle
The colorless oily residue that contained hemiaminal 19 (TLC
analysis) was immediately used in the next step to avoid
decomposition. In a 250 mL Schlenk vessel connected to a
nitrogen line, 4-vinyl-2(5H)-furanone (510 mg, 4.63 mmol) was
dissolved in dry diethyl ether (14 mL). The solution was cooled
to 0 °C, triethylamine (840 μL, 6.03 mmol) was added, and the
reaction mixture was kept at the same temperature for 30 min.
After this time, TIPSOTf (1.37 mL, 5.09 mmol) was slowly
added; the mixture was warmed to room temperature and left
until complete butenolide disappearence (basic alumina TLC
analysis). Next, the solution was cooled down to -20 °C, a
solution of hemiaminal 19 in acetonitrile (22 mL) and nBu2-
BOTf (10.1 mL, 1 M in CH2Cl2) were slowly added, and the
reaction mixture was kept at the same temperature for 30 min.
Then, it was treated with saturated aqueous NaHCO3 (30 mL),
the aqueous phase was extracted with CH2Cl2 (3 ꢀ 30 mL), and
the combined organic extracts were dried over anhydrous
Na2SO4 and concentrated under vacuum. The brownish oily
residue was purified by flash chromatography on silica gel
(gradient, hexanes/ethyl acetate 4:1 to 2:1) to give a 3.5:1
mixture of isomers RRR-21 and RRS-21 (983 mg, 1.96 mmol,
50%) as a colorless oil, a 6:1 mixture of isomers SRS-21 and
SRR-21 (687 mg, 1.37 mmol, 35%) as a colorless oil, and a
fraction with enamide 26 (65 mg, 0.16 mmol, 4%) as a colorless
oil. RRR-21 and RRS-21: Rf=0.4 (hexanes/ethyl acetate 1:1);
[R]20D=-25 (c 1.2, CHCl3); IR (ATR) 2929, 2856, 1757, 1639,
1259, 1104; 1H NMR (400 MHz, CD2Cl2) isomers RRR-21
(hexanes/ethyl acetate 1:1); IR (ATR) 2930, 2857, 1754, 1629,
1427, 1110; 1H NMR (400 MHz, C6D6) isomers RRR-28
(A, major) and RRS-28 (B, minor) δ 7.70 (m, 4HA þ 4HB:
4HPhA,4HPhB), 7.24 (m, 6HA þ 6HB:6HPhA,6HPhB), 6.40 (dd,
J=10.1, 4.1 Hz, 1HA:H15A), 5.92 (dd, J=11.2, 5.2 Hz, 1HB:
H
15B), 5.78 (d, J=11.2 Hz, 1HB:H14B), 5.75 (dd, J=10.1, 2.4 Hz,
1HA:H14A), 5.43 (bs, 1HA:H12A), 5.36 (bs, 1HB:H12B), 4.66 (bd,
J=7.9 Hz, 1HA:H9A), 4.52 (dd, J=10.0, 7.2 Hz, 1HA:H8A), 4.38
(bd, J=10.1 Hz, 1HB:H9B), 4.32 (dd, J=10.0, 6.0 Hz, 1HA:
H
8A), 4.19 (dd, J=8.6, 5.8 Hz, 1HB:H8B), 4.13 (m, 1HB:H7B),
4.08 (dd, J=8.6, 5.4 Hz, 1HB:H8B), 3.54 (ddd, J=10.9, 7.9, 5.5
Hz, 1HA:H2A), 3.42 (m, 1HA:H7A), 2.81 (dt, J=10.1, 5.0 Hz,
1HB:H2B), 2.06-1.91 (complex, 2HA þ 1HB:2H5A,H5B), 1.68-
1.46 (complex, 2HA þ 2HB:1H3A,1H4A,2H3B), 1.25 (m, 1HB:
H5B), 1.16 (s, 9HA:9HtBuA), 1.12 (s, 9HB:9HtBuB), 1.10-0.85
(complex, 1HA þ 2HB:H4A,2H4B), 0.60 (m, 1HA:H3A); 13C
NMR (100 MHz, C6D6) isomers RRR-28 (A, major) and
RRS-28 (B, minor) δ 171.8 (C11A), 171.0 (C11B), 170.0 (C6B),
169.0 (C6A), 164.2 (C13B), 161.8 (C13A), 146.3 (C15A), 139.5
(C15B), 136.0 (CPh), 134.0 (CPh), 133.8 (CPh), 133.6 (CPh), 130.3
(CPh), 128.3 (CPh), 128.2 (CPh), 122.7 (C14B), 121.8 (C14B), 117.3
(C12A), 117.2 (C12B), 82.1 (C9A), 81.2 (C9B), 65.4 (C8B), 65.0
(C8A), 62.8 (C2B), 60.8 (C7B), 59.7 (C7A), 59.5 (C2A), 33.3 (C5A),
32.2 (C5B), 27.2 (CMe), 27.1 (CMe), 26.4 (C3B), 22.9 (C3A), 19.5
(CtBu), 19.4 (CtBu), 18.1 (C4A), 17.5 (C4B); HRMS (ESIþ) calcd
for C29H33NO4Si: 510.2071 [M þ Na]þ, found 510.2062.
This reaction has been performed with quantities of RRR-21 þ
RRS-21 up to 700 mg. Yields of different runs range from 80%
to >98%.
(A, major) and RRS-21 (B, minor) δ 7.65 (m, 4HA
þ
4HB:4HPhA,4HPhB), 7.47 (m, 6HA þ 6HB:6HPhA,6HPhB), 6.58
(ddt, J=17.8, 11.2, 0.7 Hz, 1H B:H14B), 6.57 (ddt, J=18.5, 11.0,
0.7 Hz, 1HA:H14A), 6.22 (ddd, J=17.8, 10.4, 6.5 Hz, 1HB:H15B),
6.21 (ddd, J = 17.6, 10.5, 7.3 Hz, 1HA:H15A), 6.06 (m, 1HB:
(6R,11aR,11bRS)-6-tert-Butyldiphenylsilyloxymethyl-2,6,8,
9,10,11,11a,11b-octahydrofuro[2,3-c]pyrido[1,2-a]azepin-2-one
(RRR/RRS-33). In a 10 mL Schlenk vessel, a 3.5:1 mixture of
RRR-28 and RRS-28 (80 mg, 0.16 mmol) was dissolved in dry
THF (3.2 mL) under nitrogen. The solution was cooled to 0 °C,
H
1HA:H14 A), 5.69 (d, J=11.2 Hz, 1HA:H14 A), 5.67 (d, J=11.2
12B), 5.98 (bd, J=1.1 Hz, 1HA:H12A), 5.74 (d, J=17.8 Hz,
0
0
0
0
Hz, 1HB:H14 ’B), 5.65 (d, J=17.8 Hz, 1HB:H14 B), 5.48 (m, 1HB:
9B), 5.22 (dt, J=17.8, 1.4 Hz, 1HB:H15’B), 5.21 (dt, J=10.4, 1.4
a solution of BH3 THF (1 M in THF, 820 μL, 0.82 mmol) was
3
H
Hz, 1HB:H15 B), 5.18 (m, 1HA:H9A), 5.03 (dt, J=10.6, 1.2 Hz,
added, and the reaction mixture was kept at the same tem-
perature for 2 h. It was then treated with 1 M NaOH, the
aqueous phase was extracted with CH2Cl2 (3 ꢀ 5 mL), and the
combined organic extracts were dried over anhydrous Na2SO4
and concentrated under vacuum. The brownish oily residue
was purified by flash chromatography (gradient, hexanes/
ethyl acetate 49:1 to 1:1) to furnish a 9:1 mixture of RRR-
and RRS-33 (54 mg, 0.11 mmol, 70%) as a colorless oil,
alcohol 34 (8.0 mg, 0.016 mmol, 10%) as a yellow oil, and
recovered RRR/RRS-28 (8 mg, 0.016 mmol, 10%). An analy-
tical sample of RRR-33 could by obtained by repeated flash
chromatography. RRR-33 and RRS-33: Rf = 0.66 (hexanes/
ethyl acetate 1:1); IR (ATR) 2927, 1750, 1427, 1113; HRMS
(ESIþ) m/z calcd for C29H35NO3Si 474.2459 [M þ H]þ,
found:474.2457. RRR-33: 1H NMR (400 MHz, CD2Cl2) δ
7.68 (m, 4H:4HPh), 7.42 (m, 6H:6HPh), 6.58 (dd, J=12.8, 2.4
Hz, 1H:H14), 6.25 (dd, J=12.8, 4.1 Hz, 1H:H15), 5.77 (bs, 1H:
0
0
1HA:H15 A), 5.03 (dt, J=17.6, 1.3 Hz, 1HA:H15’A), 4.38 (dd, J=
10.2, 8.2 Hz, 1HB:H8B), 4.32 (dd, J=10.1, 9.1 Hz, 1HB:H8B),
4.12 (m, 1HA:H2A), 3.97 (m, 1HB:H2B), 3.88 (dd, J=10.2, 5.6
Hz, 1HB:H8B), 3.70 (m, 1HB:H7B), 3.62 (dd, J=10.3, 4.7 Hz,
1HA:H8A), 3.44 (m, 1HA:H7A), 2.31 (m, 2HA þ 2HB:2H5A
,
2H5B), 1.97 (m, 2HA þ 2HB:H3A,H4A,H3B,H4B), 1.85 (m, 1HA þ
1HB:H3A,H3B), 1.70 (m, 1HA:H4A), 1.60 (m, 1HB:H4B), 1.07 (s,
9HB:9HtBuB), 1.06 (s, 9HA:9HtBuA); 13C NMR (100 MHz,
CD2Cl2) isomers RRR-21 (A, major) and RRS-21 (B, minor)
δ 171.6 (C11B), 171.2 (C11A), 170.6 (C6A), 170.4 (C6B), 162.3
(C13A), 161.8 (C13B), 135.3 (CPh), 135.2 (CPh), 134.4 (C15A),
134.3 (C15B), 133.1 (CPh), 133.0 (CPh), 129.5 (CPh), 127.4 (CPh),
0
0
127.3 (C14A), 126.9 (C14B), 124.0 (C14 AþC14 B), 117.90 (C12B),
0
0
117.87 (C12A), 117.0 (C15 B), 116.4 (C15 A), 82.3 (C9A), 81.4
(C9B), 68.6 (C7A), 65.1 (C7B), 64.1 (C8A), 63.9 (C8B), 61.4
(C2A), 59.3 (C2B), 32.4 (C5B), 31.8 (C5A), 26.38 (C3A), 26.36
(CMeA), 26.31 (CMeB), 21.3 (C3B), 18.7 (CtBuB), 18.6 (CtBuA),
18.2 (C4B), 17.2 (C4A); HRMS (ESIþ) calcd for C31H37NO4Si:
538.2384 [M þ Na]þ, found 538.2397.
H
12), 5.39 (dd, J=11.5, 1.7 Hz, 1H:H9), 3.76 (dd, J=11.1, 8.6
Hz, 1H:H8), 3.73 (dd, J=11.0, 7.1 Hz, 1H:H8), 3.52 (m, 1H:
H7), 3.06 (dt, J=11.8, 4.1 Hz, 1H:H2), 2.80 (m, 1H:H6), 2.57
(m, 1H:H6), 1.95 (m, 1H:H3), 1.78 (m, 1H:H3), 1.67-1.47
(complex, 4H:2H4,2H5), 1.05 (s, 9H:9HtBu); 13C NMR (100
MHz, CD2Cl2) δ 172.9 (C11), 165.9 (C13), 140.6 (C15), 135.6
(CPh), 134.7 (CPh), 133.5 (CPh), 129.7 (CPh), 129.6 (CPh), 127.7
(CPh), 123.0 (C14), 115.7 (C12), 78.0 (C9), 65.8 (C8), 64.5 (C7),
63.4 (C2), 42.6 (C6), 29.7 (C3), 28.3 (C5), 26.6 (CMe), 19.1 (C4),
18.8 (CtBu). RRS-33: 13C NMR (100 MHz, CD2Cl2), observed
significant signals, δ 116.4 (C12), 66.5 (C8), 65.1 (C2), 65.0 (C7).
(6R,11aR,11bRS)-6-Hydroxymethyl-2,6,8,9,10,11,11a,11b-
octahydrofuro[2,3-c]pyrido[1,2-a]azepin-2-one (RRR/RRS-35). In
a 10 mL Schlenk vessel, a 9:1 mixture of RRR-33 and RRS-33
(40 mg, 0.085 mmol) was dissolved in THF (1.1 mL) under
(6R,11aR,11bRS)-6-tert-Butyldiphenylsilyloxymethyl-2,6,8,9,
10,11,11a,11b-octahydrofuro[2,3-c]pyrido[1,2-a]azepine-2,8-dione
(RRR/RRS-28). A solution of second-generation Grubbs’ cat-
alyst (46 mg, 0.054 mmol) in degassed anhydrous CH2Cl2 (14
mL) was slowly added to a solution of a 3.5:1 mixture of RRR-
21 and RRS-21 (170 mg, 0.34 mmol) in degassed CH2Cl2 (34
mL) under argon, and the reaction mixture was stirred for 14 h
at room temperature. Then, the solvent was evaporated under
vacuum, and the dark brown oily residue was purified by flash
chromatography (gradient, hexanes to hexanes/ethyl acetate
1:1) to furnish a 3.5:1 mixture of compounds RRR-28 and RRS-
28 (161 mg, 0.33 mmol, 97%) as a pale brownish oil: Rf=0.41
nitrogen, Et3N 3HF (83 μL, 0.51 mmol) was added, and the
3
6210 J. Org. Chem. Vol. 74, No. 16, 2009

ꢀ
Gonzalez-Galvez et al.
ꢀ
JOCArticle
reaction mixture was kept at room temperature for 7 h. It was then
treated with saturated aqueous NaHCO3 (1 mL), and the aqueous
phase was extracted with CH2Cl2 (3 ꢀ 4 mL). The combined
organic extracts were dried over anhydrous Na2SO4 and concen-
trated under vacuum. The brownish oily residue was purified by
flash chromatography (gradient, CH2Cl2/MeOH 99:1 to 20:1) to
furnish a 9:1 mixture of RRR- and RRS-35 (17.6 mg, 0.075 mmol,
88%) as a brownish oil. Rf=0.46 (CH2Cl2/acetone 7:3); IR (ATR)
3423, 2927, 1749, 1629, 1102; 1H NMR (600 MHz, CDCl3) RRR-
35 δ 6.61 (ddd, J=11.6, 1.1, 0.4 Hz, 1H:H14), 5.83 (dd, J=0.6, 0.4
Hz, 1H:H12), 5.79 (dd, J=11.3, 3.9 Hz, 1H:H15), 5.46 (bdd, J=
10.6, 1.1 Hz, 1H:H9), 3.62-3.49 (complex, 3H:H7,2H8), 3.16 (ddd,
J = 10.7, 3.8, 3.0 Hz, 1H:H2), 2.82 (bs, 1H:OH), 2.78 (td,
J=11.8, 2.8 Hz, 1H:H6), 2.68 (dt, J=11.8, 3.9 Hz, 1H:H6), 2.14
(dddd, J=14.0, 6.6, 3.0, 1.4 Hz, 1H:H3), 1.91 (m, 1H:H3), 1.80 (m,
1H:H5), 1.73 (m, 2H:2H4), 1.66 (m, 1H:H5); RRS-35 observed
significant signals, 6.49 (d, J=12.4 Hz, 1H:H14), 6.07 (dd, J=12.4,
5.6 Hz, 1H:H15), 5.78 (m, 1H:H12), 5.50 (bd, J=10.5 Hz, 1H:H9);
13CNMR (100MHz, CD2Cl2) RRR-35, δ172.2 (C11), 164.9(C13),
137.3(C15), 124.3(C14), 116.5(C12), 77.2 (C9), 63.7 (C2), 63.2(C7),
59.8 (C8), 53.2 (C6), 28.1 (C3), 25.8 (C5), 18.9 (C4); HRMS (ESIþ)
m/z calcd for C13H17NO3 236.1287 [M þ H]þ, found 236.1300.
Securinine (1). In a 10 mL Schlenk vessel, a 9:1 mixture of
RRR-35 and RRS-35 (17 mg, 0.073 mmol) was dissolved in
anhydrous CH2Cl2 (2.6 mL) under nitrogen. The solution was
cooled at 0 °C, and Et3N (45 μL, 0.328 mmol) and MsCl (25 μL,
0.219 mmol) were added. The reaction mixture was kept at the
same temperature for 30 min. It was then treated with saturated
aqueous NaHCO3 (1 mL), the aqueous phase was extracted with
CH2Cl2 (3 ꢀ 5 mL), and the combined organic extracts were
dried over anhydrous Na2SO4 and concentrated under vacuum.
The brownish oily residue was purified by flash chromatogra-
phy (gradient, CH2Cl2/MeOH 99:1 to 20:1) to furnish the
mesylate as a brownish oil. Rf = 0.82 (CH2Cl2/acetone 7:3).
Immediately, this mesylate was dissolved in dry CH2Cl2 (13
mL), and a solution of KHMDS in toluene (0.5 M, 200 μL, 0.10
mmol) was added at -78 °C. After being stirred for 15 min, the
mixture was warmed to room temperature and stirred for an
additional 45 min. The reaction mixture was then cooled again
to -78 °C, and cold (0 °C) saturated aqueous NaHCO3 (13 mL)
was added. After the mixture was warmed to room temperature,
the organic phase was decanted and the aqueous layer was
extracted with CH2Cl2 (3 ꢀ 15 mL). The combined organic
extracts were dried over anhydrous Na2SO4, the solvent was
removed under vacuum, and the remaining oily material was
purified by flash chromatography on silica gel (gradient, hexanes/
AcOEt 1:1 to AcOEt) to furnish 1 (12.6 mg, 0.058 mmol, 79%) as
a yellow solid: Rf=0.18 (ethyl acetate); [R]20D=-1061 (c 1.08,
CHCl3); 1H NMR (600 MHz, CD2Cl2) δ 6.59 (d, J=9.1, Hz, 1H:
H14), 6.39 (dd, J=9.1, 5.3 Hz, 1H:H15), 5.49 (s, 1H:H12), 3.80
(t, J=4.6 Hz, 1H:H7), 2.95 (dt, J=10.5, 3.7 Hz, 1H:H6), 2.44 (dd,
J=9.2, 4.1 Hz, 1H:H8), 2.39 (ddd, J=10.5, 6.5, 4.9 Hz, 1H:H6),
2.07 (dd, J=11.3, 2.3 Hz, 1H:H2), 1.85 (m, 1H:H4), 1.75 (dd,
J=9.2, 0.6 Hz, 1H:H8), 1.62-1.54 (complex, 2H:H3,H5), 1.54
(dd, J=7.6, 3.8 Hz, 1H:H5), 1.50 (qd, J=12.1, 3.8 Hz, 1H:H3),
1.22 (m, 1H:H4); 13C NMR (100 MHz, CD2Cl2) δ 173.1 (C11),
169.9 (C13), 139.9 (C15), 120.9 (C14), 104.5 (C12), 89.1 (C9),
62.7 (C2), 58.5 (C7), 48.4 (C6), 42.1 (C8), 27.1 (C3), 25.8 (C5),
24.4 (C4).
Acknowledgment. This paper is dedicated to Professor
Luis Castedo on the occasion of his 70th birthday. We
acknowledge financial support from MEC (Grant Nos.
CTQ2004-2539 and CTQ2007-60613) and fellowships from
MEC (to D.G.-G.) and from DGR (to E.G.-G.). We are
grateful to Dr. Gregori Ujaque for his helpful assistance on
the theoretical calculations.
Supporting Information Available: Experimental procedure
for the preparation of SRS-28 and SRS-33 and listing of
their physical data; physical data of SRS/SRR-21, 26, and 34;
1
significant H and 13C NMR spectra of new compounds and
synthetic securinine, CHPLC analysis of (þ)-12, Cartesian
coordinates, and imaginary frequencies of 24. This material is
available free of charge via the Internet at http://pubs.acs.org.
J. Org. Chem. Vol. 74, No. 16, 2009 6211