Total synthesis of (±)-aspidospermidine

Article abstract of DOI:10.1021/jo991599s

(±)-Aspidospermidine (1) has been synthesized from readily available methyl 3-ethyl-2-oxocylopentanecarboxylate (17) in 5.9% yield over 13 steps. The key step of the synthesis is an intramolecular cascade reaction that simultaneously forms the B, C, and D rings of 1. A high-yielding method of closing the remaining E ring is also described.

Full text of DOI:10.1021/jo991599s

2642
J . Org. Chem. 2000, 65, 2642-2645
Tota l Syn th esis of (()-Asp id osp er m id in e
Matthew A. Toczko and Clayton H. Heathcock*
Department of Chemistry, University of California, Berkeley, California 94720
Received October 13, 1999
(()-Aspidospermidine (1) has been synthesized from readily available methyl 3-ethyl-2-oxocylo-
pentanecarboxylate (17) in 5.9% yield over 13 steps. The key step of the synthesis is an
intramolecular cascade reaction that simultaneously forms the B, C, and D rings of 1. A high-
yielding method of closing the remaining E ring is also described.
Sch em e 1
A considerable amount research has been devoted to
the synthesis of aspidospermidine1,2 (1), the parent
compound of the Aspidosperma alkaloids,3 which com-
prise a large family of diverse structures. Current inter-
est in the synthesis of these alkaloids partially results
from the pharmacological activity exhibited by a few of
its members.4 Although 1 is not of pharmacological
interest, its lack of functionality makes it an attractive
target for the development of new synthetic pathways
to these compounds.
The core structure of the Aspidosperma alkaloids is
typified by the [6.5.6.6.5] ABCDE ring system. Herein
we report a novel intramolecular cascade reaction in
which monocyclic precursor 4 undergoes a tricyclization
to form the B, C, and D rings. Furthermore, a facile
method of closing the E ring in high yield, which has thus
far been a major hindrance in the efficient realization of
this pentacyclic ring system, is discussed.5
Our retrosynthesis of 1, shown in Scheme 1, begins
with the displacement of an acyl halide by the three
position of the indole (2), closing the E ring, followed by
reduction of the amide and indolenine. The cornerstone
of the new synthesis was to be a cascade reaction wherein
compound 4 would undergo two cyclization reactions,
resulting in an acylimmonium ion and an indole (3). The
indole should undergo a Mannich-like ring closure,
forming 2. Cascade precursor 4 could be derived from the
hypothetical precursor 5 by simultaneous reduction of the
azide and nitrile functions, followed by hydrolysis of the
acetals. We hoped to prepare 5 by coupling a suitable
organometallic derivative of o-azidotoluene with acyl
halide 6, which could arise from two Michael reactions
of butanal with methyl acrylate and acrylonitrile.
this compound, attempts were made to prepare 7 from
both 87 and 98 by Michael addition of the corresponding
pyrrolidine enamines9 with acrylonitrile or methyl acry-
late, respectively. However, refluxing either enamine
with excess Michael acceptor in a variety of solvents for
extended periods of time failed to produce 7, typically
resulting in a return of starting material with a 30-40%
loss to decomposition. Since enamine alkylations of
similarly substituted enamines with more electrophilic
reagents such as allyl bromide are known,10 we rational-
ized that in the case of 9, addition of a Lewis acid might
sufficiently increase the electrophilicity of methyl acry-
late to overcome the steric hindrance imposed by the
trisubstituted enamine. Indeed, the incorporation of BF3‚
Et2O into the reaction mixture provided 7 in 44% yield
from 9. Because of problems encountered later in this
synthetic route, further optimization of this reaction was
not explored.
The synthesis of 7 has actually been reported, albeit
in low yield (ca. 14%).6 To improve the accessibility of
(1) For leading references, see: (a) d′Angelo, J .; Desmale D. J . Org.
Chem. 1994, 59, 2292. (b) Urrutia, A.; Rodr´ıguez, J . G. Tetrahedron
1999, 55, 11095. (c) Forns, P.; Diez, A.; Rubiralta, M. J . Org. Chem.
1996, 61, 7882.
(2) The numbering scheme is that of Biekmann, K.; Spiteller-
Friedmann, M.; Spiteller, G. J . Am. Chem. Soc. 1963, 85, 631.
(3) Cordell, G. A. In The Alkaloids; Manske, R. H. F., Rodrigo, R.
G. A., Eds.; Academic Press: New York, 1979; Vol. 17, pp 199-384.
Saxton, J . E. Nat. Prod. Rep. 1993, 10, 349-395; 1994, 11, 493.
(4) Antitumor Bisindole Alkaloids from Carentheus roseus. In The
Alkaloids; Brossi, A., Suffness, M., Eds.; Academic Press: San Diego,
1990; Vol. 37.
(7) Norman M. H.; Heathcock, C. H. J . Org. Chem. 1988, 53, 3370.
(8) Ziegler, F. E.; Zoretic, P. A. Tetrhedron Lett. 1968, 22, 2639.
(9) Stork, G.; Brizzolara, A.; Landesman, H.; Szmuszkovicz, J .;
Terrell, R. J . Am. Chem. Soc. 1963, 85, 207.
(5) For leading references, see: Padwa, A.; Price, A. T. J . Org. Chem.
1998, 63, 556.
(6) Fleming, I.; Harley-Mason, J . J . Chem. Soc. 1964, 6, 2165.
(10) Opitz, G.; Mildenberger, H. Angew. Chem. 1960, 72, 169.
10.1021/jo991599s CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/13/2000

Total Synthesis of (()-Aspidospermidine
J . Org. Chem., Vol. 65, No. 9, 2000 2643
Sch em e 2
As shown in Scheme 2, protection of the aldehyde as
the dioxolane and subsequent ester hydrolysis provided
acid 10 in quantitative yield. Acid 10 was converted into
the corresponding acyl chloride, which was treated with
o-azidobenzylzinc bromide in the presence of palladium
according to the method of Negishi et al.11 However, all
attempts to link these two segments met with failure.
In hope of remedying this problem, a brief model study
was performed. Acryloyl chloride was treated with the
bromozinc derivatives, derived from reaction of benzyl
bromide (12a ) and o-azidobenzyl bromide (12b) with zinc,
in the presence of tetrakis(triphenylphosphine)palladium.
In the case of benzyl bromide, the coupling proceeds
smoothly at -78 °C to provide the desired enone 14a in
88% yield. Unfortunately, the introduction of nitrogen-
containing functionalities at the ortho position inhibits
the coupling reaction. In the case of o-azidobenzyl
bromide (12b) the only product obtained is o-azidotoluene
(13b).12 This indicates that the organozinc species formed
but was not reactive in the acylation reaction. An attempt
to prepare the organozinc derivative from o-nitrobenzyl
bromide gave only decomposition products, whereas
o-phthalamidobenzyl bromide13 was unreactive when
treated with zinc. Because of this failure, we looked to
alternate methods for generating 4.
Our second synthetic route (Scheme 3) takes advantage
of the strategic placement of carbonyl functionality, which
allows access to 4 via oxidative cleavage of a properly
substituted cyclopentene (15). This cyclopentene could be
generated by reduction of 16, followed by chloroacylation.
We envisioned 16 arising from alkylation and cyano-
ethylation of the readily available â-ketoester 17.14
Alkylation of 17 with o-azidobenzyl bromide proceeds
in superb yield if one uses Cs2CO3 in lieu of K2CO3 as
the base. Initial attempts at cyanoethylation of 18
utilizing standard Michael addition conditions gave only
moderate yields (40-55%) of the desired Michael adduct
19, often returning starting material as well as numerous
side products. Optimization of this reaction revealed that
treatment of 18 with Cs2CO3 in t-BuOH proceeds so
cleanly that it can be taken on to the hydrolysis/
decarboxylation step without purification to provide 20
in 77% yield for the two steps (Scheme 4).
Sch em e 3
phosphate, and the vinyl triflate. Circumvention of this
problem necessitated reduction of the ketone and sub-
sequent elimination of the alcohol. Ketone 20 reduced
cleanly in the presence of CeCl3, but the subsequent
dehydration proved to be difficult. Treatment with either
thionyl chloride or phosphorus oxychloride in the pres-
ence of various bases resulted in poor yields (25-30%),
with most of the material being lost to decomposition.
Additionally, sulfonation reactions were slow and poor
yielding, often giving many products, and although the
trifluoroacetate could be formed cleanly, all elimination
attempts resulted in either hydrolysis or no reaction.
Ultimately, the use of PCl5 was found to provide a 4:1
mixture of inseparable olefin isomers 16 and 21 in 63%
yield.
With 16 in hand, simultaneous hydride reduction of
the azide and cyano groups provided diamine 22, and
chloroacylation of the primary amine proceeded readily
to give 15. Although attempts to access 4 directly from
15 resulted solely in decomposition, after protection of
the aniline nitrogen as the tert-butyl carbamate (23),
ozonolysis successfully cleaved the alkene to provide 24
in modest yield. Efforts were made to accomplish oxida-
tive cleavage by dihydroxylation and oxidative cleavage
of the syn diols. However, subjecting 23 to various
dihydroxylation conditions resulted in either return of
starting material or slow decomposition. These results
were not unexpected because of the steric hindrance
about the alkene. Alternatively, it was thought that
epoxidation and formation of the trans diol could give
access to 24 upon treatment with Pb(OAc)4. Although
epoxidation proceeded readily, the compound suffered
rearrangement to what appeared to be a quinoline upon
acidification.
Preliminary attempts were made to convert 20 directly
to the alkene. However, steric hindrance about the ketone
prevented the formation of the tosylhydrazone, the vinyl
(11) (a) Negishi, E.; Bagheri, V.; Chatterjee, S.; Luo, F. T.; Miller,
J . A.; Stoll, T. A. J . Am. Chem. Soc. 1983, 24, 5181. (b) Sato, T.; Naruse,
K.; Enokiya, M.; Fujisawa, T. Chem. Lett. 1981, 1135.
(12) Mornet, R.; Leonard, N. J .; Theiler, J . B.; Doree, M. J . Chem.
Soc., Perkin Trans. 1 1984, 879.
(13) Chakravarty, A. K.; Dastidas, P. P. G.; Pakreshi, S. C. Tetra-
hedron 1982, 38, 1797.
As a result of the difficulty in purifying 24 it was
typically carried on to the cascade reaction crude. Treat-
ment of 24 with a 1:1 CH2Cl2/TFA solution provides 2 in
(14) Saquer, D.; Smadju, S. Recl. Trav. Chim. Pays.-Bas. 1976, 95,
172.

2644 J . Org. Chem., Vol. 65, No. 9, 2000
Toczko and Heathcock
Sch em e 4
Sch em e 6
showed that the formation of the C11-C12 bond late in
the synthesis can be difficult. Until recently, use of
intramolecular SN2 displacement to form the pyrollidine
ring resulted in low yields, especially in the case of
haloacetamides.16,17 This is, in part, due to steric imped-
ance about the pseudopentacoordinate transition state.
Magnus et al.17 had attempted the same ring closure with
both 2 and the bromide analogue under a variety of
conditions but were unsuccessful. Additionally they tried
to close the E ring by replacing the haloacetamide with
a 2-bromoethyl group but were again unsuccessful. To
circumvent this problem they resorted to closing the E
ring with a Pummerer reaction, which added several
steps to their overall synthesis. More recently Rubiralta
et al. have successfully closed the pyrollidine ring system
in high yield via base-mediated displacement using the
2-alkylsulfonylethyl derivative.18
Sch em e 5
We chose to pursue the closure of the E ring via the
haloacetamide, as we felt that it would require less
protecting-group manipulation and would better protect
the primary amine. Because the acyl iodide should be
significantly more reactive than its chloro and bromo
analogues, it was synthesized and treated with silver
trifluoroacetate (Scheme 6). Under these conditions,
indolenine 26 was isolated in 43% yield, along with a
nearly equal amount of an unstable side product in which
chloride appears to have been replaced by trifluoro-
acetate. Use of silver triflate eliminates this side product
and provides 26 in 86% yield, based on chloride 2.
Reduction of both the amide and the indolenine provides
aspidospermidine (1) in 82% yield.
In conclusion, the syntheses of aspidospermidine pro-
ceeds in 5.9% yield over 13 steps. This synthetic strategy
should prove useful for the construction of other members
of the Aspidosperma family because the piece-wise con-
struction of the cascade precursor allows for incorporation
of varied functionality. Furthermore, the discovery of a
facile method of closing the E ring may allow for the
implementation of synthetic routes that may have been
otherwise discounted.
53% yield from 23.15 To demonstrate the utility of this
cascade reaction, compound 24 was painstakingly puri-
fied and subjected to identical reaction conditions. These
conditions gave the deprotected tetracyclic indole deriva-
tive 2 in 81% yield. Under less acidic conditions, the Boc-
protected indole 25 is produced in 76% yield (Scheme 5).
Having achieved three out of the four cyclizations, we
next sought to close the E ring. Literature precedence
(16) Urrutia, A.; Rodriguez, J . G. Tetrahedron 1999, 55, 11095.
(17) Gallagher, T.; Magnus, P.; Huffman, J . C. J . Am. Chem. Soc.
1983, 105, 4750.
(18) (a) Forns, P.; Diez, A.; Rubiralta, M. J . Org. Chem. 1996, 61,
7882. (b) Rubiralta, M.; Diez, A.; Bosch, J .; Solans, X. J . Org. Chem.
1989, 54, 5591.
(15) The isolated yield of 25 is lower than the yield of the ozonolysis/
cascade because material must be sacrificed to obtain clean material.

Total Synthesis of (()-Aspidospermidine
J . Org. Chem., Vol. 65, No. 9, 2000 2645
material is available free of charge via the Internet at
http://pubs.acs.org.
Ack n ow led gm en t. This research was supported by
a research grant from the National Institutes of Health
(GM46057).
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures and full characterization for all compounds. This
J O991599S