A facile method for the solution and solid-phase synthesis of substituted [3.3.1] bicycles

Article abstract of DOI:10.1021/ol990791d

Formula presented Interest in bicyclic natural products from the Guttiferae classification has led to the development of an improved method for the selenium-mediated cyclization of alkenyl-substituted β-dicarbonyls (I) to form a variety of bicyclo[3.3.1]n

Full text of DOI:10.1021/ol990791d

ORGANIC
LETTERS
1999
Vol. 1, No. 5
807-810
A Facile Method for the Solution and
Solid-Phase Synthesis of Substituted
[3.3.1] Bicycles
K. C. Nicolaou,* Jeffrey A. Pfefferkorn, Guo-Qiang Cao, Sanghee Kim, and
Jilali Kessabi
Department of Chemistry and The Skaggs Institute for Chemical Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road,
La Jolla, California 92037, and Department of Chemistry and Biochemistry,
UniVersity of California, San Diego, 9500 Gilman DriVe, La Jolla, California 92093
kcn@scripps.edu
Received July 7, 1999
ABSTRACT
Interest in bicyclic natural products from the Guttiferae classification has led to the development of an improved method for the selenium-
mediated cyclization of alkenyl-substituted â-dicarbonyls (I) to form a variety of bicyclo[3.3.1]nonan-9-ones (II) both in solution and on solid
support.
Garsubellin A (1, Figure 1) is a polyisoprenylated phloro-
glucinol derivative from Garcinia subelliptica found on the
Okinawa Islands of Japan.1 It is characterized by a highly
oxygenated and densely substituted bicyclo[3.3.1]nonane-
1,3,5-trione core fused to a tetrahydrofuran ring and appended
with lipophilic side chains. Interest in garsubellin A (1) stems
from its in vitro neurotrophic properties making it a potential
candidate for the treatment of neurodegenerative diseases.2
The potential importance of garsubellin A (1) to biology and
medicine prompted us to pursue a total synthesis,3 the success
Figure 1. Structure of garsubellin A (1) and nemorosone II (2).
of which was contingent upon developing a reliable meth-
odology for constructing its unusual bicyclic skeleton.
Moreover, since the core of garsubellin A (1) is conserved
among a number of other natural products of the Guttiferae
classification,4 including nemorosone II5 (2), a versatile
construction of this framework might find application in
future synthetic investigations of this biologically diverse
family of compounds. Herein we describe an improved
method for the construction of the 2,2-dimethylbicyclo[3.3.1]-
nonan-9-one carbon framework of these molecules.
From a synthetic vantage point, one of the challenging
features of this skeleton is the highly congested C-1 through
C-3 region (Figure 1) with a quaternary gem-dimethyl group
at C-2 flanked by a quaternary bridgehead carbon (C-1) and
a tertiary prenylated carbon (C-3). This congestion, com-
pounded by the fact that the remaining bridgehead position
(C-5) is also quaternary, essentially precludes construction
(1) Fukuyama, Y.; Kuwayama, A.; Minami, H. Chem. Pharm. Bull. 1997,
45, 947-949.
(2) Hefti, F. J. Annu. ReV. Pharmcol. Toxicol. 1997, 37, 239-267. Hefti,
F. J. Neurobiol. 1994, 1418-1435.
(3) Nicolaou, K. C.; Pfefferkorn, J. A.; Kim, S.; Wei, H. J. Am. Chem.
Soc. 1999, 121, 4724-4725.
10.1021/ol990791d CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/06/1999

of this framework via the R,R′-annulation strategies tradition-
ally employed in the construction of bicyclo[3.3.1]nonan-
9-ones.6 In fact, there has been very little work reported on
the synthesis of 2,2-dimethylbicyclo[3.3.1]nonan-9-ones with
both bridgehead positions substituted as required for con-
struction of the garsubellin A (1) skeleton.7 In light of this,
we considered a strategy as outlined in the simplified
retrosynthetic analysis of Scheme 1. Examination of the
C-1-C-2 bond with concomitant installation of a radical
precursor at C-3. Encouragingly, as shown in Scheme 1, Ley
et al. had previously accomplished the construction of bicycle
8 by treatment of an allyl-substituted â-keto ester 6 with a
variety of selenium donors and catalysts to initially afford
the kinetically favored O-cyclized furan 7, which was
subsequently equilibrated to the thermodynamically favored
carbocycle 8.8 Thus, we constructed the corresponding
prenylated â-keto ester model system 9 (Scheme 1) and using
similar conditions, obtained the O-cyclized furan 10 in 88%
yield. Unfortunately, furan 10 proved resistant toward
equilibration to the desired 2,2-dimethylbicyclo[3.3.1]nonan-
9-one 11 even upon prolonged exposure (24 h) to a variety
of Lewis acids (see Scheme 1) resulting only in decomposi-
tion. This failure forced us to consider a strategy for masking
the nucleophilicity of the enolic oxygen, thereby providing
direct access to the desired carbocyclic product. We postu-
lated that acetylation of the enolic oxygen would inhibit
O-cyclization9 and, as proposed in intermediate 13, might
lead to the desired C-cyclization through in situ, nucleophi-
cally assisted cleavage of the acetate during carbon-centered
attack on a pendent epi-selenium complex. To test this
hypothesis, we prepared enol acetate 12, and gratifyingly,
upon treatment with N-(phenylseleno)phthalimide (N-PSP)10
and SnCl4 at -78 °C for 5 min, it underwent exclusive
conversion to the desired carbocycle 11, in 94% yield.
Encouraged by the facility and selectivity of this cycliza-
tion at low temperature, we sought to determine its synthetic
utility focusing on three criteria: (a) tolerance toward
increasing substitution on the cyclohexane framework as
required for the construction of various natural and designed
products; (b) effect of olefin substitution on the regioselec-
Scheme 1. General Retrosynthetic Analysis and Model
Studies of Selenium-Mediated Cyclizationsa
(4) For examples, see: Gustafson, K. R.; Blunt, J. W.; Munro, M. H.
G.; Fuller, R. W.; McKee, T. C.; Cardelina, J. H.; McMahon, J.; Gragg, G.
M.; Boyd, M. R. Tetrahedron 1992, 48, 10093-10102. Iinuma, M.; Tosa,
H.; Tanaka, T.; Kanamaru, S.; Asai, F.; Kobayashi, Y.; Miyauchi, K.;
Shimano, R. Biol. Pharm. Bull. 1996, 18, 311-314. Alves, T. M .A.; Alves,
R. O.; Romanha, A. J.; Santos, M. H.; Nagem, T. J.; Zani, C. L. J. Nat.
Prod. 1999, 62, 369-371. Oliveira, C. M. A.; Porto, A. M.; Bittrich, V.;
Vencato, I.; Marsaioli, A. J. Tetrahedron Lett. 1996, 37, 6427-6430.
(5) Oliveira, C. M. A.; Porto, A. L. M.; Bittrich, V.; Marsaioli, A. J.
Phytochemistry 1999, 50, 1073-1079.
(6) For examples, see: Taeschler, C.; Sorenson, T. S. J. Org. Chem.
1998, 63, 5704-5705. Harding, K. E.; Clement, B. A.; Moreno, L.;
Katalinic, J. P. J. Org. Chem. 1981, 46, 940-948. Padwa, A.; Kline, D.
N.; Murphree, S. S.; Yeske, P. E. J. Org. Chem. 1992, 57, 298-306. Peters,
J. A. Synthesis 1979, 321-336. Lu, X.; Huang, Y. Tetrahedron Lett. 1986,
27, 1615-1616.
(7) To the best of our knowledge, no substituted 2,2-dimethylbicyclo-
[3.3.1]nonan-9-one wherein both bridgehead positions are quaternary has
been reported. For examples of the construction of substituted 2,2-
dimethylbicyclo[3.3.1]nonan-9-one with at least one quaternary bridgehead
position, see: Firrel, N. F.; Hickmott, P. W.; Hopkins, B. J. J. Chem Soc.
C 1970, 1477-1480. Dauben, W. G.; Bunce, R. A. J. Org. Chem. 1983,
48, 4642-4648. Gambacorta, A.; Fabrizi, G.; Bovicelli, P. Tetrahedron
1992, 48, 4459-4464. Evans, E. H.; Hewson, A. T.; March, L. A.; Nowell,
I. W.; Wadsworth, A. H. J. Chem. Soc., Perkin Trans. 1 1987, 137-149.
(8) Jackson, W. P.; Ley, S. V.; Morton, J. A. J. Chem. Soc., Chem.
Commun. 1980, 1028-1029. Jackson, W. P.; Ley, S. V.; Whittle, A. J. J.
Chem. Soc., Chem. Commun. 1980, 1173-1174. Jackson, W. P.; Ley, S.
V.; Morton, J. A. Tetrahedron Lett. 1981, 22, 2601-2604.
(9) For precedent, see: Edstrom, E. D.; Livinghouse, T. J. Org. Chem.
1987, 52, 951-953. Tamelen, E. E; Hwu, J. R.; Leiden, T. M. J. Chem.
Soc., Chem. Commun. 1983, 62-63. Evans, E. H.; Hewson, A. T.;
Wadsworth, A. H. Synth. Comm. 1985, 15, 243-247.
(10) Nicolaou, K. C.; Claremon, D. A.; Barnette, W. E.; Seitz, S. P. J.
Am. Chem. Soc. 1979, 101, 3704-3706. Nicolaou, K. C.; Petasis, N. A.;
Claremon, D. A. Tetrahedron 1985, 41, 4835-4841.
a Conditions: (a) 1.1 equiv of N-PSP, 1.0 equiv of SnCl4, CH2Cl2,
0 °C 10 min; (b) 1.1 equiv of (SnCl4, TiCl4, ZnCl2, AlCl3, or
BF3OEt2), CH2Cl2, 0 to 25 °C, 24 h; (c) Ac2O, 4-DMAP, 80 °C,
30 min; (d) 1.1 equiv of N-PSP, 1.0 equiv of SnCl4, CH2Cl2, -78
°C, 5 min. Abbreviations: 4-DMAP ) 4-(dimethylamino)pyridine.
N-PSP ) N-(phenylseleno)phthalimide.
general skeleton 3 led to the realization that the C-3
substituent might be disconnected via a radical addition
transform leading to intermediate 4 which revealed the retron
for a powerfully simplifying selenium-induced endocycliza-
tion of a pendent olefin onto a suitably positioned â-dicar-
bonyl moiety. Such a cyclization would establish, in one step,
the bicycle through construction of the sterically demanding
808
Org. Lett., Vol. 1, No. 5, 1999

tivity of the cyclization; and (c) applicability of solid-phase
selenium reagents,11 thereby allowing for construction of
resin-bound bicycles applicable to solid-phase natural product
and combinatorial synthesis.
Table 2. Construction of Aryl-Fused
Bicyclo[3.3.1]nonan-9-ones via Selenium-Mediated Cyclizations
We began by exploring the effects of increased substrate
functionalization with an emphasis on constructing suitable
frameworks for the total synthesis of garsubellin A (1),
nemorosone II (2), and related natural products. Of primary
interest were the lactone-functionalized systems of Table 1
Table 1. Construction of Frameworks for Bicyclic Natural
Products from the Guttiferae Classification
a Isolated yield after purification unless otherwise noted. b Yield deter-
mined H NMR of crude reaction mixture.
1
a Isolated yield after purification unless otherwise noted. b Yield at 40%
conversion by H NMR of crude reaction mixture.
1
is generally tolerant of even electron-rich aromatics, although
the product of substrate 22 was obtained only in low yield.
Notwithstanding the utility of these substituted 2,2-
dimethylbicyclo[3.3.1]nonan-9-ones, we hoped to expand the
usefulness of this cyclization by varying the substitution of
the olefin, thereby accommodating the construction of
differentially substituted bicycles. Preliminarily, we sought
to determine the effect (if any) that olefin substitution had
on the regioselectivity of the cyclization. As illustrated in
Table 3, we returned to the original model system, varying
the alkenyl substituents.
since the bicyclic products resulting from such substrates
possessed the minimum functionality required for elaboration
to these compounds (1 and 2), as we recently demonstrated
by converting the product of substrate 14 to the fully
substituted and oxygenated core of garsubellin A (1).3 As
illustrated in Table 1, the cyclization was not hampered by
the increased strain introduced by the fused lactone, and it
nicely accommodated a variety of substituents including
methoxy, methyl, isopropyl, phenyl, and even a conjugated
olefin. Unlike the previous model systems, these more
functionalized substrates required slightly higher tempera-
tures (-23 °C) for optimal conversion. The only limitation
noted was with substrate 16 as the increased steric demand
of the isopropyl substituent caused the reaction to proceed
sluggishly and never past 40% conversion. Attempts to
increase conversion using higher reaction temperatures or
longer reaction times resulted only in decomposition.
Quite unexpectedly, these substituent changes caused a
dramatic perturbation in cyclization selectivity. In fact, in
Table 3. Effects of Olefin Substitution on Cyclization
Regioselectivitya
Cognizant of the presence of [3.3.1] bicycles in a host of
other bioactive natural and designed structures, we sought
to extend our studies. As an example we considered
construction of a variety of fused aromatic bicycles such as
those shown in Table 2. Since several natural products (e.g.,
selagine12 and huperzine13) as well as several designed small
molecule analgesics14 and anticonvulsants15 possess related
skeletons, application of the current methodology would
allow for the construction of conformationally constrained
analogues.16 As shown in Table 2, the cyclization reaction
a Reaction conditions: 1.1 equiv of N-PSP, 1.0 equiv of SnCl4, CH2Cl2,
-23 °C, 15 min. b Isolated yield after purification unless otherwise noted.
c Yield as a mixture of stereoisomers.
(11) Nicolaou, K. C.; Pastor, J.; Barluenga, S.; Winssinger, N. Chem.
Commun. 1998, 1947-1948. We thank N. Winssinger for the generous
preparation of a sample of these resins.
Org. Lett., Vol. 1, No. 5, 1999
809

all cases, except for the original model 12 and the crotyl
substituted system 26, the valuable preference for carbon-
centered cyclization was lost. Although a complete explana-
tion of these results remains elusive, we have excluded the
possibility that the substrates may be suffering cleavage of
the enol acetate functionality prior to cyclization, as sub-
strates treated with SnCl4 under the same conditions but in
the absence of N-PSP were recovered in greater than 90%
yield in all cases. Hence, the results may be attributed to
the sterics governing approach of the epi-selenium complex
to the nucleophilic center.17
produced no detectable bicyclic products. With the success
of the former reagent, we attempted the cyclization of several
representative substrates as illustrated in Table 4. Solution
and solid-phase behavior were closely correlated in all cases.
Besides acting as a robust tether, the selenium linker can
undergo traceless or functionalizing cleavage as shown in
Scheme 2. The versatility of these cleavage options coupled
Scheme 2. Cleavage Protocols of Selenium Linkera
Last, with an improved understanding of the potential and
limits of this cyclization in hand, we turned our attention
toward solid-phase applications of the method. We previously
reported the construction of polystyrene-based selenium
bromide and selenium phthalimide resins;11 consequently,
we sought to determine whether such cyclization reactions
might be amenable to the construction of resin-bound
bicycles. Initially, we considered the simplified enol acetate
12 (see structure in Scheme 1) as a substrate for both resins.
By using slightly longer reaction times (Table 4) and a 3-fold
Table 4. Solid-Phase Cyclization Reactionsa
a Conditions: (a) 10.0 equiv of 30% H2O2, THF, 0 °C, 2 h; (b)
CCl4, 80 °C, 10 min; (c) 5.0 equiv of n-Bu3SnH, 0.1 equiv of AIBN,
C6H6, 80 °C, 2 h; (d) 5.0 equiv of n-Bu3SnCH2CHdCH2, 0.1 equiv
of AIBN, C6H6, 80 °C, 2 h.
with the economy of a single step loading/cyclization to
produce resin-bound polyfunctionalized bicycles opens the
possibility for the solid-phase synthesis of 1 and 2, and more
importantly, provides access to a novel class of rigid,
carbocyclic scaffolds amenable to further combinatorial
derivatization for the development of small molecule librar-
ies.18
a Reaction conditions: 1.0 equiv of resin, 3.0 equiv of substrate, 3.0
equiv of SnCl4, CH2Cl2, -23 °C. b See Scheme 1, Table 1, and Table 2 for
structure of substrates. c Loading is approximated by mass gain of resin
assuming a functionalization of 1.76 mmol/g. d Purity determined by 1H
NMR of crude cleavage product [via oxidation-elimination protocol
(Scheme 2)].
Acknowledgment. This work was supported by the
National Institutes of Health, Bethesda, MD, The Skaggs
Institute for Chemical Biology, and the Department of
Defense (fellowship to J.P.).
excess of substrate and SnCl4, 12 was independently reacted
with each resin in CH2Cl2. Subsequently, the resins were
treated with n-Bu3SnH and AIBN at 80 °C to affect cleavage
of any reaction products. Gratifyingly, the reaction with the
selenium bromide resin cleanly yielded the desired bicycle
in 89% yield on the basis of theoretical resin loading. In
contrast, the reaction with the selenium phthalimide resin
Supporting Information Available: Representative pro-
cedures and spectral data for the preparation and cyclization
of substrates 12, 14 (solid phase), 15, and 21. This material
is available free of charge via the Internet at http://pubs.acs.org.
OL990791D
(12) Shamma, M.; Jones, C. D.; Weiss, J. A. Tetrahedron 1969, 25, 4743.
(13) Review: Kozikowski, A. P.; Tuckmantel, W. Acc. Chem. Res. 1999,
32, to appear in the August issue on p 641 and available via ACS Articles
ASAP until print publication.
(14) Freed, M. E.; Potoski, J. R.; Freed, E. H.; Conklin, G. L. J. Med.
Chem. 1973, 16, 595-599.
(15) Gilbert, I. M.; Hewett, C. L.; Ree, D. R.; Redpath, J.; Savage, D.
S.; Sleigh, T. J. Chem. Soc., Perkin Trans. 1 1995, 133-139.
(16) For an example of the utility of the gem-dimethyl group in huperzine
analogs, see: Kozikowski, A. P.; Ding, Q.; Saxena, A.; Doctor, B. P. Bioorg.
Med. Chem. Lett. 1996, 6, 259-262.
(17) Molecular modeling studies are in progress.
(18) Gordon, E. M.; Barrett, R. W.; Dower, W. J.; Fodor, S. P. A.; Gallop,
M. A. J. Med. Chem. 1994, 37, 1385-1401. For a similar example, see:
Ley, S. V.; Mynett, D. M.; Koot, W. J. Synlett 1995, 1017-1020.
810
Org. Lett., Vol. 1, No. 5, 1999