A fully synthetic route the neurotrophic illicinones: Syntheses of tricycloillicinone and bicycloillicinone aldehyde

Article abstract of DOI:10.1021/ja000521m

Tricycloillicinone (1) and bicycloillicinone asaronacetal (2) were both isolated from extracts of the wood of Illicium tashiroi. These compounds were found to enhance the action of choline acetyltransferase, which catalyzes the synthesis of acetylcholine from its precursor. Since one of the characteristic symptoms of Alzheimer's disease involves degeneration of cholinergic neurons, resulting in markedly reduced levels of acetylcholine, compounds with the properties of 1 or 2 could well serve as agents in the treatment of such disorders. In this paper, we report the total synthesis of I and the construction of the core structure 3 of 2. The tricycloillicinone synthesis employed a novel strategy to control the regiochemistry of two ortho Claisen rearrangements. A sulfonyl group introduced at C₂ of an allyl group effectively suppressed its unwanted rearrangement to the para position (23 → 24). Subsequently, ortho Claisen rearrangement was conducted using a reverse O- prenylated derivative 31 to furnish the desired 32, selectively. 32 was used as a common precursor for the syntheses of both 1 and 3. The application of Corey-Snider oxidative cyclization and the Barton-McCombie deoxygenation provided a direct route to 1. For bicycloillicinone aldehyde 3, a new tandem reaction using the Et₂AlCN to construct the cage structure (39 → 41) was employed. Flexible syntheses of the polycyclic illicinones should provide access to analogous structures for future biological and SAR studies.

Full text of DOI:10.1021/ja000521m

6160
J. Am. Chem. Soc. 2000, 122, 6160-6168
A Fully Synthetic Route to the Neurotrophic Illicinones: Syntheses of
Tricycloillicinone and Bicycloillicinone Aldehyde
Thomas R. R. Pettus,†,‡ Masayuki Inoue,§,| Xiao-Tao Chen,†, and
Samuel J. Danishefsky*,†,§
Contribution from the Laboratory for Bioorganic Chemistry, Sloan-Kettering Institute for Cancer
Research, 1275 York AVenue, New York, New York 10021, and the Department of Chemistry,
Columbia UniVersity, HaVemeyer Hall, New York, New York 10027
ReceiVed February 11, 2000
Abstract: Tricycloillicinone (1) and bicycloillicinone asaronacetal (2) were both isolated from extracts of the
wood of Illicium tashiroi. These compounds were found to enhance the action of choline acetyltransferase,
which catalyzes the synthesis of acetylcholine from its precursor. Since one of the characteristic symptoms of
Alzheimer’s disease involves degeneration of cholinergic neurons, resulting in markedly reduced levels of
acetylcholine, compounds with the properties of 1 or 2 could well serve as agents in the treatment of such
disorders. In this paper, we report the total synthesis of 1 and the construction of the core structure 3 of 2. The
tricycloillicinone synthesis employed a novel strategy to control the regiochemistry of two ortho Claisen
rearrangements. A sulfonyl group introduced at C2 of an allyl group effectively suppressed its unwanted
rearrangement to the para position (23 f 24). Subsequently, ortho Claisen rearrangement was conducted using
a reverse O-prenylated derivative 31 to furnish the desired 32, selectively. 32 was used as a common precursor
for the syntheses of both 1 and 3. The application of Corey-Snider oxidative cyclization and the Barton-
McCombie deoxygenation provided a direct route to 1. For bicycloillicinone aldehyde 3, a new tandem reaction
using the Et2AlCN to construct the cage structure (39 f 41) was employed. Flexible syntheses of the polycyclic
illicinones should provide access to analogous structures for future biological and SAR studies.
Introduction
In 1995, Fukuyama and colleagues isolated tricycloillicinone
from the toxic plant Illicium tachiroi, collected on Ishigaki island
in Japan.1,2 They determined its chemical structure to be that
shown as structure 1, mainly by extensive NMR analysis (Figure
1). Two years later, they reported the isolation of bicycloilli-
cinone asaronacetal (2) from the same plant. The entire structure
of 2 was elucidated through heavy reliance on spectroscopic
analysis and chemical degradation. Not surprisingly, acid
hydrolysis of 2 led to aldehyde 3, in addition to the catechol
portion. This finding confirmed the presence of the acetal
moiety.
The highly compact arrangement of the polycyclic structures
of 1-3 and the presence of the methylenedioxy function in a
nonaromatic setting served to pose interesting challenges to the
science of chemical synthesis. In the case of 2, the complexity
of the problem rises further, since provision must be arranged
for an additional formyl carbon at C3 in the form of a rare
catechol-derived acetal. Perhaps an equally compelling factor
in attracting our attention to the construction of these targets is
that they were found to be inducers of choline acetyltransferase
Figure 1. Structures of tricycloillicinone (1) and bicycloillicinone
asaronacetal (2).
(ChAT). ChAT is responsible for the biosynthesis of the
neurotransmitter acetylcholine, which has been identified as a
bioregulator for cholinergic neuron function. Compounds 1 and
2 were reported to increase ChAT activity at 30 µM.1-3 Since
the dementia associated with neurodegenerative diseases has
been partially attributed to atrophy of cholinergic neurons and
corresponding deficiencies in acetylcholine levels, a search for
chemical substances that could boost Ach levels at the locus of
cholinergic neurons has continued for more than a decade.4,5
† Columbia University.
‡ Current address: Department of Chemistry, University of California,
Santa Barbara, Santa Barbara, CA 93106.
(3) Hatanaka, H.; Tsukui, H.; Nihonmatsu, I. DeV. Brain Res. 1988, 39,
85.
(4) (a) Hefti, F. J. Neurobiol. 1994, 1418. (b) Hefti, F. Annu. ReV.
Pharmacol. Toxicol. 1997, 37, 239.
§ Sloan-Kettering Institute for Cancer Research.
| Current address: Department of Chemistry, Graduate School of Science,
Tohoku University, Sendai 980-8578, Japan.
Current address: DuPont Pharmaceuticals Co., Experimental Station,
P.O. Box 80500, Wilmington, DE 19880-0500.
(1) Fukuyama, Y.; Shida, N.; Kodama, M.; Chaki, H.; Yugami, T. Chem.
Pharm. Bull. 1995, 43, 2270.
(5) The cholinesterase inhibitors such as denepezil (Aricept) and tacrine
(Cognex), which are capable of increasing Ach levels, are now in use as
Alzheimer’s disease therapeutics. For a recent review, see: Giacobini, E.
Neurochem. Int. 1998, 32, 413. We thank the reviewers for calling these
agents to our attention.
(2) Fukuyama, Y.; Hata, Y.; Kodama, M. Planta Med. 1997, 63, 275.
10.1021/ja000521m CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/13/2000

Synthetic Route to the Neurotrophic Illicinones
J. Am. Chem. Soc., Vol. 122, No. 26, 2000 6161
Several proteinoid factors such as nerve growth factor (NGF)
have been found, and their potential roles in the development
of the mature nervous system and in preventing neuronal
degeneration have been suggested.6 However, such proteins have
expectedly poor pharmacokinetic characteristics and they fail
to cross through the blood-brain barrier. This situation prevents
their systemic application for treatment of central neurodegen-
erative diseases.7 Accordingly, non-peptide small molecules,
which are capable of increasing neurotransmitter biosynthesis
and possibly supporting the survival of cholinergic neurons,
attract considerable interest. Not surprisingly, the few non-
peptide small molecules that manifest apparent activity in this
challenge draw considerable notice.8
Scheme 1.a Photochemical Reaction of the Illicinone Series
Furthermore, two FDA-approved Alzheimer therapeutics
(Aricept and Coganex) apparently increase acetylcholine levels
in the brain.8e
a Reagents and conditions: (a) Et2O, 400-W mercury lamp, 20-30
°C, 1 h; (b) Et2O, 400-W mercury lamp, 20-30 °C, 30 min.
Considering the biological activity and the intriguing struc-
tures of the neurotrophic illicinones, we undertook a program
directed to their total synthesis. In this paper, we report (i) the
total synthesis of tricycloillicinone (1) and (ii) the synthesis of
the aldehyde version 3 of bicycloillicinone asaronacetal (2). In
the course of this synthetic journey, we solved a problem that
arose in the context of conducting sequential aromatic Claisen
rearrangements. The method employed is apt to be pertinent to
other instances of bond reorganization reactions.
access to this unusual skeleton. Such a systematic synthesis
might provide a sound strategy to reach more complex neu-
rotrophic illicinones such as 2. The most challenging synthetic
problem in synthesizing this molecular class appeared to be that
of construction of the cagelike ring system. The results of
Furukawa10 suggested that the tricyclic structure of 1 could
perhaps be constructed via the formation of a radical at C2 of
4. Given that irradiation-induced cyclization was inefficient for
a selective synthesis, an alternative approach to generate the
radical might prove to be more effective. Accordingly, we
decided to apply an oxidative radical reaction of 1,3-dicarbonyl
compounds using manganese(III). Such chemistry had been
introduced by Corey and subsequently expanded by Snider for
the construction of polycyclic systems (Scheme 2: 10 f 9).11
In principle, the initial loss of the C2 hydrogen by oxidation
would give a radical intermediate 11. Sequential cyclizations
could lead to 12. Oxidation of this radical, associated with
elimination of a hydrogen atom from C8, would afford 3-oxy-
tricycloillicinone (9). It was anticipated that this radical method
could selectively construct the desired tricyclic system without
undue complications of the other possible radical pathways.
Four-electron reductive deoxygenation of 9 from C3 would
afford the target 1. Importantly, the oxygen functionality of 9
at C3 could serve as a conduit to synthesize analogues for further
biological study.
Synthetic Plan for Tricycloillicinone
Although 1 was isolated in 1995 from I. tachiroi, the
compound had actually been generated earlier, through inadvert-
ence, by Furukawa in 1984. In the course of searching for
biologically active natural products, Japanese workers deter-
mined, in 1980, the structures of four novel phytoquinoids,9
illicinones-A-D, from I. tashiroi.
The information, which was of the greatest value for our
project, came from subsequent chemical investigations of
illicinone-A (4, Scheme 1).10 Irradiation of 4 with a mercury
lamp for 1 h was claimed to afford compounds 1, 6, and 7 in
24%, 21%, and 5% yield, respectively. Additionally, reaction
of illicinole (5) under the same conditions for 30 min furnished
1, 4, 6, 7, and 8 in low yields. The structure of 1, isolated from
these reactions, was unambiguously established by a single-
crystal X-ray analysis. Thus, in 1995, 1 was reintroduced to
the chemical community by Fukuyama et al. as tricycloillici-
none, in the context of his natural product isolation exercise.1
The earlier studies of Furukawa suggest 5 or 4 to be plausible
intermediates in the biosynthesis of tricycloillicinone.
Compound 10 would be prepared by ortho Claisen rearrange-
ment of the reverse O-prenylated phenol 13. Although we could
not predict, a priori, the selectivity of the rearrangement, one
literature example encouraged us to examine the possibilities
inherent in such a dearomatization reaction.12 Moreover, we
expected that the sterically demanding nature of the dimethyl
group on C12 might prevent the prenyl group in 10 from
undergoing para rearrangement to C6. The allyl group in 13
itself could be constructed by ortho Claisen rearrangement of
On basis of the above-described disclosure, we could in
principle reach tricycloillicinone through preparation of illicinole
(5) or illicinone-A. However, we were interested in a more
logical, flexible, and chemically satisfying solution to gain
(6) Seiger, A.; Nordberg, A.; vonHolst, H.; Backman, L.; Ebendal, L.;
Alafuzoff, I.; Ambria, K.; Herlitz, A.; Lilja, A. BehaV. Brain Res. 1993,
57, 255.
(7) Resenburg, S. In Annual Reports in Medicinal Chemistry; Academic
Press: New York; 1992; Vol. 27, p 41.
(8) cf. Isodunnianin: (a) Fukuyama, Y. Shida, N.; Kodama, M. Planta
Med. 1993, 59, 181. Garsubellin A: (b) Fukuyama, Y.; Kuwayama, A.;
Minami, H. Chem. Pharm Bull. 1997, 45, 947. Synthetic studies on
garsubellin A: (c) Nicolaou K. C.; Pfefferkorn, J. A.; Kim, S.; Wei, H. X.
J. Am. Chem. Soc. 1999, 121, 4724. K-252a derivatives: (d) Kaneko, M.;
Saito, Y.; Saito, H.; Matsumoto, T.; Matsuda, Y.; Vaught, J. L.; Dionne,
C. A.; Angeles, T. S.; Glicksman, M. A.; Neff, N. T.; Rotella, D. P.; Kauer,
J. C.; Mallamo, J. P.; Hudkins, R. L. Murakata, C. J. Med. Chem. 1997,
40, 1863.
(11) (a) Corey, E. J.; Kang, M.-C. J. Am. Chem. Soc. 1984, 106, 5384.
(b) Snider, B. B. Chem. ReV. 1996, 96, 339. (c) Kates, S. A.; Dombroski,
M. A.; Snider, B. B. J. Org. Chem. 1990, 55, 2427. (d) Snider, B. B.; Merritt,
J. E.; Dombroski, M. A.; Buckman, B. O. J. Org. Chem. 1991, 56, 5544.
(e) Snider, B. B.; Mohan, R.; Kates, S. A. J. Org. Chem. 1985, 50, 3659.
For a review, see: (f) Snider, B. B. Chem. ReV. 1996, 96, 339.
(12) Murray et al. showed that coumarin derivative i underwent facile
ortho Claisen rearrangement in the course of their extensive studies on
coumarin-type natural products: Murray, R. D. H.; Sutcliffe, M.; M.
Hasegawa. Tetrahedron 1975, 31, 2966.
(9) Yakushijin, K.; Sekikawa, J.; Suzuki, R.; Morishita, T.; Furukawa,
H.; Murata, H. Chem. Pharm. Bull. 1980, 28, 1951.
(10) Yakushijin, K.; Furukawa, H.; McPhail, A. T. Chem. Pharm. Bull.
1984, 32, 23.

6162 J. Am. Chem. Soc., Vol. 122, No. 26, 2000
Pettus et al.
Scheme 2. Synthetic Strategy en Route to Tricycloillicinone
(1)
Scheme 3. First Attempts for Ortho Claisen Rearrangement
Scheme 4.a Regioselective Ortho Claisen Rearrangement
a Reagents and conditions: (a) 22, K2CO3, acetone, room temper-
ature, 85%; (b) toluene, 165 °C, 12 h; (c) K2CO3, acetone, reflux, 94%
(two steps); (d) 10% Na-Hg, MeOH-EtOAc, -20 °C, 87%.
which could, in turn, could give rise either to 17 or to 18 (via
subsequent Cope rearrangement followed by tautomerization).
An independent experiment excluded the possibility of subse-
quent conversion of 17 to 18 under reaction conditions and
suggested that the para/ortho ratio was under kinetic control.
Apparently, the tautomerization of 16 is slower than the second
sigmatropic rearrangement. Of course, in offering this analysis,
we have not considered the possibility of a “direct” para Claisen
rearrangement, which would not involve ortho Claisen inter-
mediates.
We also investigated the case of substrate 15 with an electron-
donating group at the C1 position (see 15). The reaction course
did not change, though a slight yield improvement was observed.
Due to the encountered complexities, we wondered about the
consequences of placing a group at C2 of the migrating allyl
moiety to improve the ortho/para ratio of the rearrangement. It
was hoped that such a group would suppress the migratory
aptitude of the allyl function with respect to the presumed Cope
phase of the tandem sequence discussed above. To apply this
concept in the context of the total synthesis, this group would
need to be introduced and removed both easily and selectively.
With this consideration in mind, we chose a phenylsulfonyl
group as a candidate substructure (Scheme 4). Alkylating agent
22 was prepared from allyl phenyl sulfide in two steps using a
known procedure.18 Base-induced combination of 2119 and 22
gave rise to the R,â-unsaturated sulfone 23 in 87% yield.
Presumably, under these conditions, HBr is eliminated from 22
either 14 or 15. It was envisioned that utilization of the same
hydroxyl group for the two consecutive Claisen rearrangements
would simplify the protecting group strategy required to build
the highly oxygenated precursor 9.13 However, it was recognized
that differentiation of phenolic hydroxyl functions in aromatic
compounds is not always a small matter.14 For instance, the
synthesis of 15 from methylenedioxyresorcinol could be com-
plicated by a lack of control in either the phenolic allylation or
silyl protection steps. Thus, following the elegant precedents
of Boger and Coleman,15 we decided to use a carbonyl group
as a surrogate for the hydroxy residue at C1. The methyl ester
in 14 could be converted to a tertiary alcohol at an appropriate
stage in the synthesis. Given such an alcohol, benzylic hydro-
peroxide rearrangement could be exploited to deliver the
required phenol at C1 (see 13).
Total Synthesis of Tricycloillicinone16
To test the feasibility of the first ortho Claisen rearrangement,
we prepared allyl ether 14 (Scheme 3). However, thermolysis
of 14 provided the undesired rearrangement product 18 as the
major product. The tendency for formation of para Claisen
products in such systems is well precedented.17 Mechanistically,
Claisen rearrangement of 14 would lead to intermediate 16
(13) The consecutive rearrangement strategy practiced here uses the same
aromatic hydroxyl to direct rearrangements to the two flanking ortho centers.
This is to be distinguished from the concept of tandem rearrangements;
see: Ziegler, F. E. Chem. ReV. 1988, 88, 1423.
(14) cf. (a) Sinhababu, A. K.; Borchardt, R. T. J. Org. Chem. 1983, 48,
1941. (b) Schneider, G. E.; Stevenson, R. J. Org. Chem. 1981, 46, 2969.
(c) McKittrick, B. A.; Stevenson, R. J. Chem. Soc., Perkin Trans 1 1984,
709.
(15) (a) Boger, D.; Coleman, R. S. J. Org. Chem. 1986, 51, 5436. For
recent examples of this concept, see: (b) Boger, D.; Coleman, R. S. J. Am.
Chem. Soc. 1987, 109, 2717. (c) Sa´nchez, A. J.; Konopelski, J. P. J. Org.
Chem. 1994, 59, 5445.
(16) For a preliminary account of this work, see: Pettus, T. R. R.; Chen,
X.-T.; Danishefsky, S. J. J. Am. Chem. Soc. 1998, 120, 12684.
(17) Rhoads, S. J.; Raulins, N. R. In Organic Reactions; Dauben, W.
G., Ed.; John Wiley & Sons: New York, London, Sydney, Toronto; 1975;
Vol. 22, p 1.
(18) For the preparation of 1,3-dibromo-2-phenylsulfonylpropane, see:
(a) Knochel, P.; Normant, J. F. Tetrahedron Lett. 1985, 26, 425. (b) Auvray,
P.; Knochel, P.; Normant, J. F. Tetrahedron Lett. 1985, 26, 4455.
(19) Compound 21 was prepared as previously reported in one step from
commercially available methyl gallate; see: (a) Keseru¨, G. M.; No´gra´di,
M.; Kajta´r-Peredy, M. Liebigs Ann. Chem. 1994, 361. (b) Kuo, G.-H.;
Eissenstat, M. A. Tetrahedron Lett. 1997, 38, 3343.

Synthetic Route to the Neurotrophic Illicinones
J. Am. Chem. Soc., Vol. 122, No. 26, 2000 6163
Scheme 5a
a Reagents and conditions: (a) MeLi, THF, -78 °C to room temperature, 12 h, 90%; (b) (CH3)2C(Cl)CtCH, KI, K2CO3, CH3COCH3, 60 °C,
48 h, 50-73%; (c) (CH3)2C(OH)CtCH, (CF3CO)2O, DBU, then 26, CuCl2‚H2O, DBU, room temperature, 12 h, 77% based on starting material
(89% conversion); (d) 50% H2O2-H2SO4 (5:4), CH2Cl2, 0 °C, 10 min, 85%; (e) TBDPSCl, imidazole, DMAP, CH2Cl2, room temperature, 97%;
(f) 10 mol % Pd-CaCO3 (Pb-poisoned), quinoline, EtOAc, room temperature, >95%; (g) toluene, 100 °C, 2 h, >95%.
to provide 3-bromo-2-phenylsulfonylpropene, which in turn
reacts with phenol 21. Upon heating the resultant phenyl ether
23 to 165 °C, a facile Claisen rearrangement occurred. Remark-
ably, only the desired regioisomer 24 was obtained in 94% yield.
We next studied means to cleave the phenylsulfonyl group
of 24 in a reductive fashion. Several preliminary attempts in
this regard were not successful. Presumably these failures
reflected side reactions open to the radicaloid or carbanionic
species formed upon desulfonylation of 24. Thus, we developed
the two-step procedure for removal of the phenyl sulfone.
Compound 24 was first converted to 25 via conjugate addition
of the phenolic oxygen to the R,â-unsaturated sulfone. Reductive
elimination of the sulfonyl group, by the action of sodium
amalgam, resulted in the desired 17 in 87% yield.20 Thus, the
reaction sequence had provided 17 from 21, without contamina-
tion from the para Claisen rearrangement product, in 69% yield
over four steps.
With the ortho Claisen rearrangement accomplished, focus
shifted to the introduction of the reverse prenyl group onto the
C3 phenol and to the conversion of the methoxycarbonyl linkage
to a C1 hydroxyl group (Scheme 5). In the event, excess MeLi
reacted with the ester function to give tertiary alcohol 26 in
90% yield. The phenolic hydroxyl group in this compound was
alkylated under standard conditions (K2CO3, KI, 3-chloro-3-
methylbut-1-yne).21 Unfortunately, the yields of this heteroge-
neous O-alkylation were variable, especially for large-scale
operations. After considerable experimentation, a copper-
catalyzed reaction proved to be more reliable on scale-up.22
Thus, by adding a homogeneous solution of DBU and propargyl
trifluoroacetate with a catalytic amount of CuCl2‚H2O at room
temperature to 26, the desired product 27 was isolated in 68%
yield on multigram scale.
period of time.15 Under these conditions, the hydroperoxide 28
was apparently generated. Rearrangement of this compound
afforded the desired 29 in 85% yield. Protection of its free
phenolic hydroxyl as a silyl ether gave 30 in 97% yield.
In anticipation of a second ortho Claisen rearrangement step,
the triple bond in 30 was reduced to olefin 31 with Lindlar’s
catalyst using quinoline as an additive.12,23 Happily, warming
of 31 to 100 °C, led to its rearrangement to the desired
cyclohexadienone 32 in almost quantitative yield. Noteworthy
was the finding that the reaction occurred, albeit slowly, even
at room temperature.
With 32 in hand, attention was directed to construction of
the tricyclic ring system (Scheme 6). To test the feasibility of
the light-induced reaction of illicinone-A reported by Fu-
rukawa,10 we first prepared 4 from 32 using standard synthetic
manipulations. Treatment of 32 with L-Selectride, followed by
acetylation and â-elimination provided illicinone-A in 82% yield
for two steps. Photolysis of 4, did indeed produce tricycloilli-
cinone as described. However, in our hands this reaction
proceeded in only 10%-15% yields.
Thus, we decided to focus on the Mn(III)-based oxidative
radical reaction to generate a radical at C2. Cyclization of 32
was conducted through the action of Mn(OAc)3 and Cu(OAc)2
at room temperature.11 32 was smoothly converted to 3-oxytri-
cycloillicinone (9) in 80% yield. Presumably, during this
reaction, the silyl enol ether was cleaved. The resultant
â-diketone then underwent oxidative cyclization as shown in
Scheme 2.
With the entire framework in place, all that remained to
complete the total synthesis of 1 was 4-electron reduction of
the ketone at C3 to the corresponding methylene group. In
practice, this goal turned out to be an extremely difficult task.
We first attempted deoxygenation directly from ketone 9 through
thioketal24 or hydrazone25 derivatives. Unfortunately, these
classical methods failed at the level of the syntheses of the
required intermediates. Accordingly, we evaluated the prob-
ability of the Barton-McCombie deoxygenation protocol for
Alkylation of the C3 phenol had now provided the proper
setting to liberate the C1-based phenolic hydroxyl group from
its precursor tertiary alcohol. Accordingly, 27 was exposed to
the action of concentrated H2O2 in acidic conditions for a short
(20) (a) Julia, M.; Paris, J.-M. Tetrahedron Lett. 1973, 14, 4833. (b)
Kocienski, P.; Lythgoe, B.; Ruston, S. J. Chem. Soc., Perkin Trans. 1 1978,
829. (c) Kocienski, P. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Fleming, I., Eds; Pergamon Press: Oxford, New York, Seoul, Tokyo; 1991;
Vol. 6, p 975.
(23) (a) Lindlar, H. HelV. Chim. Acta. 1952, 35, 446. (b) McEwen, A.
B.; Guttieri, M. J.; Maier, W. F.; Laine, R. M.; Shyo, Y. J. Org. Chem.
1983, 48, 4436.
(21) Hlubucek, J.; Ritche, E.; Tayler, W. C. Tetrahedron Lett. 1969, 1369.
(22) (a) Godfrey, J. D., Jr.; Mueller, R. H.; Sedergran, T. C.; Soundarara-
jan, N.; Colandrea, V. J. Tetrahedron Lett. 1994, 35, 6405. (b) Bell, D.;
Davies, M. R.; Geen, G. R.; Mann, I. S. Synthesis 1995, 707.
(24) Corey, E. J.; Tius, M. A.; Das, J. J. Am. Chem. Soc. 1980, 102,
7612.
(25) (a) Wolff, L. Justus Liebigs Ann. Chem. 1912, 394, 86. (b) Huang,
M. L. J. Am. Chem. Soc. 1946, 68, 2487.

6164 J. Am. Chem. Soc., Vol. 122, No. 26, 2000
Pettus et al.
Scheme 6.a Completion of the Total Synthesis of
Tricycloillicinone
Scheme 7. Synthetic Strategy of Bicycloillicinone Aldehyde
(3)
for constructing aldehyde 3, which is also an acid-induced
degradation product of the natural product 2. It is well to note
that tricycloillicinone has a cis relationship between C2 and C11;
3 has a trans relationship. A new departure for the ring
construction was thus appropriate.
Our retrosynthetic logic to reach 3 is suggested in Scheme
7. We planned to construct the five-membered ring by nucleo-
philic cyclization of kinetically generated C2 enolate via the
epoxide linkage in 36. Thus, ring closure would concurrently
expose the tertiary alcohol. Compound 36 might arise from a
regioselective 1,4-addition of a one-carbon unit to C3 over C5
in an axial manner, either with or without trapping of the
resultant enolate. In consideration of the potential hindrance at
C3, we chose to rely on the Nagata reagent (Et2AlCN) for this
purpose.27 This nucleophile would hopefully be relatively
nonvulnerable to steric blockade. It was anticipated that 37 could
be prepared from a synthetic intermediate on the route to
tricycloillicinone.
a Reagents and conditions: (a) (i) L-Selectride, THF, 0 °C, 3 h; then
p-TsOH, 0 °C, 2 h; (ii) Ac2O, DMAP, 0 °C, 3 h, 82% overall; (b) neat
DBU, room temperature, 1 h, 65%; (c) degassed Et2O, 450-W Hg lamp,
Pyrex tube, 1 h, 10-15% (see ref 10); (d) Mn(OAc)3‚2H2O,
Cu(OAc)2‚H2O, HOAc, room temperature, overnight, 80%; (e) Li-
AlH(O-t-Bu)3, THF, 78%; (f) KHMDS, THF, -78 °C, then DMAP,
PhOC(S)Cl, -20 °C, overnight, 90%; (g) Bu3SnH, AIBN, benzene,
80 °C, overnight, 42%.
our needs.26 Selective reduction of the isolated ketone, over the
R,â-unsaturated ketone, was achieved via the action of LiAlH-
(Ot-Bu)3 (Scheme 6). However, the secondary alcohol in 34
proved to be extremely resistant to activation. For instance,
under standard conditions (excess PhOC(S)Cl and DMAP), the
thioformate 35 failed to form. After extensive experimentation,
we were able to reach 35 under more forcing conditions.
Compound 34 was first converted to its potassium salt using
KHMDS. The latter, in turn reacted with PhOC(S)Cl and
DMAP, providing the desired 35 in 90% yield.
Synthesis of Bicycloillicinone Aldehyde 3
The synthesis began with 33 (Scheme 8). Selective epoxi-
dation of the prenyl moiety was achieved using mCPBA at 0
°C. The reaction gave an inseparable 2:1 mixture of epoxides
38, favoring the desired isomer in 72% yield. DBU-induced
â-elimination of the acetoxy group provided precursor 39 in
60% yield. The crucial 1,4-addition of cyanide to 39 was
conducted using 2 equiv of Et2AlCN in THF at 0 °C. Not only
did the cyano nucleophile add to the C3 position of 39 selectively
but the resultant kinetically generated metaloenolate attacked
the epoxide, thereby furnishing cyclized products 41 (87% from
the â-epoxide) and 42 (81% from the R-epoxide). Fortunately,
41 and 42 were separable by silica gel column chromatography,
and the structures of these compounds were unambiguously
determined by NMR spectroscopy. Interestingly, only the
undesired isomer underwent a second 1,4-addition, leading to
bis(nitrile) 42.28 To our knowledge, reaching 41 in this way
constitutes the first example of using a Nagata reaction in such
Radical deoxygenation of 35 was carefully conducted on a
small scale using excess Bu3SnH to provide 1 in 42% yield.
The synthetic tricycloillicinone (racemate) was identical to the
1
authentic natural product by the criteria of H, 13C NMR and
IR spectroscopy. In addition to producing the natural product
in a more pleasing fashion, this oxidative radical route provides
access to a wealth of analogue structures.
Synthetic Plan for Bicycloillicinone Aldehyde (3)
Having successfully developed a flexible strategy for the
synthesis of tricycloillicinone, we were in a position to pursue
the construction of other neurotrophic illicinones. In this
connection, bicycloillicinone asaronacetal (2) attracted our
attention. Its structure is different from that of 1 in that it
contains an additional carbon branch at C3. The formyl group
at this position is actually encumbered in the form of an unusual
catechol acetal (see structure 2 in Figure 1). To pursue the total
synthesis of 2, we decided to develop a potentially novel solution
(27) (a) Nagata W.; Yoshikawa, M. Organic Reaction; Dauben, W. G.,
Ed.; Wiley and Sons: New York, 1984; Vol. 25, p 225. For a recent
application of Nagata reaction, see: (b) Ihara, M.; Katsumata, A.; Egashira,
M.; Suzuki, S.; Tokunaga, Y.; Fukumoto, K. J. Org. Chem. 1995, 60, 5560.
(28) Selective monocyanide addition to 41 could be explained as the
formation of aluminum acetal iii after the reaction.
(26) Barton, D. H. R.; McCombie, S. W. J. Chem. Soc., Perkin Trans.
1 1975, 1574.

Synthetic Route to the Neurotrophic Illicinones
J. Am. Chem. Soc., Vol. 122, No. 26, 2000 6165
Scheme 8..a Synthesis of Bicycloillicinone Aldehyde (3)
a Reagents and Conditions: (a) mCPBA, CH2Cl2-5% aqueous NaHCO3, 0 °C to room temperature, 3 h, 73%; (b) DBU, THF, 0 °C to room
temperature, 2.5 h, 60%; (c) Et2AlCN, THF, 0 °C, 40 min, 58% (41), 27% (42); (d) TMSOTf, i-Pr2NEt, CH2Cl2, 0 °C, 40 min; (e) DIBAL,
hexane-CH2Cl2 (2:1), -90 °C, 30 min; AcOH, CH3OH, room temperature, 2 h, 69% from 41; (f) TMSOMe, TMSOTf, CH2Cl2, -78 °C to 0 °C,
40 min, 49%.
a tandem sequence.29 While an epoxide opening by an aluminum
keto-enolate had been reported by Schreiber,30 such metal-
mediated additions are certainly less developed31 relative to their
ester enolate counterparts.32
Scheme 9.a Synthesis of the Catechol Moiety
With 41 in hand, the only remaining step to reach 3 was that
of controlled reduction of the nitrile function to produce the
aldehyde. However, in practice, achievement of this transforma-
tion turned out to be no trivial matter. Standard reducing agents
resulted in 1,2- or 1,4-reduction of the R,â-unsaturated ketone
but failed to reduce the nitrile. Given these results, a method
for protecting the ketone in 41 was pursued. Recognition that
the tertiary alcohol in 41 is located “underneath” the ketone
function suggested the possibility of protection of the ketone
by formation of an internal ketal. Gratifyingly, treatment of 41
with TMSOTf in the presence of base selectively led to the
TMS acetal 43. Without purification, this intermediate was
submitted to DIBAL reduction in hexane-CH2Cl2. Indeed
following acidic workup, the desired aldehyde 3 was obtained
in 69% overall yield. It is worth noting that the hexane content
is crucial for success of the DIBAL reduction. Thus, the starting
material was completely recovered using other solvent systems
(THF, toluene, CH2Cl2). The fully synthetic aldehyde 3 was
a Reagents and Conditions: (a) Pb(OAc)4, PhH, 60 °C, 4 h (see ref
36); (b) AcOH-H2O-THF, 0 °C to room temperature, 1.5 h; (c)
MeOH, room temperature, 1 h, 75% (three steps); (d) HN[Si(CH3)3]2,
cat. H2SO4, THF, reflux, 30 min.
A plausible mechanistic rationalization of the formation of this
compound is shown. These preliminary investigations have
revealed that aldehyde 3 is relatively unstable to strongly acidic
media. Completion of total synthesis of 2 would require
avoidance of acidic conditions. Several schemes intended to
reach this goal have failed and the total synthesis of 2, bearing
the catechol acetal, still awaits realization.35
1
identical to the hydrolysis product of 2 by H, 13C NMR and
IR spectroscopic criteria.
For a preliminary study directed to the possibility of the
catechol acetalization of 3 required to synthesize 2, the catechol
47 and bis-TMS catechol 48 were prepared from the known
compound 4533 (see Scheme 9). However, under a variety of
conditions, attempted coupling of aldehyde 3 and derivatives
47 or 48 resulted in significant decomposition. A common
byproduct (vide infra) was isolated from each reaction. We
found that methyl acetal 44 could be prepared in 49% yield by
using Noyori conditions (Scheme 8).34 Even in this reaction,
the same byproduct was obtained as a minor component.
Eventually, its structure was elucidated to be 49 (Scheme 10).
Summary
A concise total synthesis of tricycloillicinone has been
achieved from commercially available methyl 3,4,5-trihydroxy-
benzoate in 15 steps. This synthesis employed new strategies
to control the regiochemistry of two key Claisen rearrangements.
A cumyl hydroperoxide rearrangement reaction simplified the
protecting group management issue. Application of the Corey-
Snider oxidative cyclization and the Barton-McCombie deoxy-
(32) For representative examples of aluminum ester enolate addition to
epoxide, see: (a) Danishefsky, S.; Kitahara, T.; Tsai, M.; Dynak, J. J. Org.
Chem. 1976, 41, 1669. (b) Taylor, S. K.; Fried, J. A.; Grassl, Y. N.;
Marolewski, A. E.; Pelton, E. A.; Poel, T.-J.; Rezanka, D. S.; Whittaker,
M. R. J. Org. Chem. 1993, 58, 7304.
(33) Compound 45 was synthesized in three steps from commercially
available sesamol: (a) Alexander, B. H.; Gertler, S. I.; Brown, R. T.; Oda,
T. A.; Beroza, M. J. Org. Chem. 1959, 4, 49. (b) Schuda, P. F.; Price, W.
A. J. Org. Chem. 1987, 52, 1972.
(29) Further work to study the generality of this aluminum enolate
epoxide opening reaction is in progress.
(30) Schreiber, S. L. J. Am. Chem. Soc. 1980, 102, 6165.
(31) Metal-mediated reaction of ketone enolate to epoxide has been
extensively studied by Crotti and associates: (a) Chini, M.; Crotti, P.;
Favero, L.; Pineschi, M. Tetrahedron Lett. 1991, 32, 7583. (b) Crotti, P.;
Di Bussolo, V.; Favero, L.; Pineschi, M. J. Org. Chem. 1996, 61, 9548. (c)
Crotti, P.; Di Bussolo, V.; Favero, L.; Macchia, F.; Pineschi, M. Tetrahedron
Lett. 1994, 35, 6537. (d) Crotti, P, Di Bussolo, V.; Favero, L.; Macchia,
F.; Pineschi, M.; Napolitano E. Tetrahedron 1999, 55, 5853.
(34) Tsunoda, T.; Suzuki, M.; Noyori, R. Tetrahedron Lett. 1980, 21,
1357.

6166 J. Am. Chem. Soc., Vol. 122, No. 26, 2000
Pettus et al.
1
1300 cm-1; H NMR (CDCl3, 300 MHz) δ 7.81 (2H, m), 7.65 (1H,
m), 7.55 (2H, m), 7.18 (1H, d, J ) 1.3 Hz), 7.08 (1H, d, J ) 1.3 Hz),
6.59 (1H, s), 6.22 (1H, s), 5.96 (2H, s), 4.84 (2H, t, J ) 1.3 Hz), 3.44
(3H, s); 13C NMR (CDCl3, 75.5 MHz) δ 166.4, 149.5, 146.4, 141.1,
139.9, 139.5, 134.2, 129.7, 128.6, 127.1, 125.0, 113.0, 105.0, 102.6,
66.5, 52.6; HRMS calcd for C18H16O7S [M]+ 376.0617, found 376.0629;
LRMS 376, 219, 203, 165, 141, 125.
Scheme 10. Possible Decomposition Pathway of 3 in Acidic
Media
Phenol 24. Allyl ether 23 (8.4 g, 22 mmol) in a pressure tube was
heated at 165 °C for 12 h. After being cooled to room temperature, the
crude product was used in the following reaction without further
purification: 1H NMR (CDCl3, 500 MHz) δ 7.89 (2H, d, J ) 7.5 Hz),
7.62 (1H, t, J ) 7.5 Hz), 7.54 (2H, t, J ) 7.5 Hz), 7.03 (1H, s), 6.34
(1H, s), 6.01 (1H, s), 5.92 (1H, s), 5.55 (1H, s), 3.95 (2H, s), 3.68 (3H,
s); 13C NMR (CDCl3, 75.5 MHz) δ 166.8, 150.4, 147.6, 139.9, 139.2,
138.4, 133.7, 129.4, 128.6, 124.6, 122.7, 122.0, 103.4, 102.5, 51.8,
27.1; HRMS calcd for C18H16O7S [M]+ 376.0617, found 376.0616.
Sulfone Ether 25. The above phenol 24 was converted into sulfone
ether 25 by refluxing in acetone with K2CO3. After an aqueous workup,
the residue was subjected to flash column chromatography (hexane-
ether 90:10) to give sulfone ether 25 (7.9 g, 94% for two steps): IR
(neat) 2920, 2880, 1695, 1618, 1490, 1470, 1430, 1421, 1305 cm-1
;
1H NMR (CDCl3, 400 MHz) δ 7.91 (2H, m), 7.64 (1H, m), 7.55 (2H,
m), 7.13 (1H, s), 5.98 (2H, s), 4.58 (1H, dd, J ) 11.0, 3.3 Hz), 4.24
(1H, dd, J ) 11.0, 8.5 Hz), 3.81 (3H, s), 3.57 (1H, m), 3.47 (2H, m);
13C NMR (CDCl3, 75.5 MHz) δ 166.3, 146.6, 138.4, 138.1, 137.1,
134.2, 129.3, 129.0, 122.4, 118.2, 105.3, 102.5, 63.5, 57.9, 52.0, 24.6;
HRMS calcd for C18H16O7S [M]+ 376.0617, found 376.0618; LRMS
376, 234, 219, 203, 175, 145, 133, 104, 77.
genation sequence provided a direct route to tricycloillicinone.
For bicycloillicinone aldehyde 3, we employed a new tandem
reaction using the Nagata reagent to construct the cage structure
and the tertiary alcohol simultaneously. These flexible syntheses
of the polycyclic illicinones should provide access to analogous
structures for future biological and SAR studies.
Phenol 17. Sulfone ether 25 (7.9 g, 21 mmol) was dissolved in
EtOAc-MeOH (2:1) and cooled to -20 °C. Solid sodium amalgam
(10%, 40 g, 52 mmol) was added to this solution, and the resultant
mixture was stirred for 12 h at -20 °C. The reaction mixture was
quenched by solid citric acid, diluted with ether, and washed with
saturated aqueous NaHCO3. Drying over Na2SO4, concentration, and
flash column chromatography (hexane-ether 80:20) gave phenol 17
(4.2 g, 86%): IR (neat) 3300, 1675, 1425, 1260 cm-1; 1H NMR (CDCl3,
400 MHz) δ 7.06 (1H, s), 6.00 (2H, s), 5.99 (1H, m) 5.06 (1H, dd, J
) 7.2, 1.5 Hz), 5.03 (1H, s), 4.98 (1H, s), 3.83 (3H, s), 3.77 (2H, dt,
J ) 6.0, 1.6 Hz); 13C NMR (CDCl3, 75.5 MHz) δ 167.8, 146.8, 138.6,
138.1, 137.1, 125.5, 124.0, 115.7, 104.4, 102.6, 52.4, 31.0; HRMS calcd
for C12H12O5 [M]+ 236.0685, found 236.0686; LRMS 254, 237.
Tertiary Alcohol 26. Ester 17 (4.2 g, 18 mmol) was dissolved in
THF and cooled to -78 °C. MeLi (72 mL, 1 M in THF with LiBr)
was added dropwise to this solution at -78 °C. The solution was
allowed to warm to 0 °C and stirring was continued for 12 h. The
reaction was then quenched with saturated aqueous NH4Cl, and the
organic layer was dried over Na2SO4. Concentration and flash column
chromatography (hexane-ether 90:10) gave tertiary alcohol 26 (3.8 g,
90%): IR (neat) 3421, 1636 cm-1; 1H NMR (CDCl3, 500 MHz) δ 6.64
(1H, s), 6.07 (1H, m), 5.93 (2H, s), 5.11 (1H, dd, J ) 10.1, 1.6 Hz),
5.02 (1H, dd, J ) 17.0, 1.6 Hz), 4.96 (1H, s), 3.81 (2H, dt, J ) 5.2,
1.6 Hz), 1.73 (1H, s), 1.60 (3H, s), 1.59 (3H, s); 13C NMR (CDCl3,
75.5 MHz) δ 140.3, 139.1, 137.9, 119.2, 115.4, 102.7, 101.5, 73.8,
31.9, 31.5 (2 unresolved); HRMS calcd for C13H16O4 [M]+ 236.1048,
found 236.1042; LRMS 236, 221, 218, 203, 173, 145, 115.
Propargyl Ether 27. To a solution of 2-methyl-3-butyn-2-ol (970
µM, 10 mmol) in CH3CN (10 mL) at 0 °C was added DBU (1.7 mL,
11.3 mmol), followed by dropwise addition of (CF3CO)2O (1.4 mL,
10 mmol). To the above solution was added phenol 26 (1.5413 g, 6.5
mmol), DBU (2.4 mL, 16.0 mmol), and CuCl2‚H2O (6.7 mg, 0.039
mmol) in CH3CN (30 mL). After being stirred at room temperature
for 24 h, the solution was evaporated, diluted with EtOAc, and washed
with H2O, saturated aqueous NaHCO3, and brine. Flash column
chromatography (hexanes-EtOAc 90:10-70:30) gave propargyl ether
27 (1.344 g, 68%) along with recovered starting material 26 (0.178 g,
11%): IR (neat) 3480, 3281, 2957, 2116 cm-1; 1H NMR (CDCl3, 400
MHz) δ 6.76 (1H, s), 6.04 (1H, m), 5.87 (2H, s), 4.96 (1H, dd, J )
10.2, 1.8 Hz), 4.82 (1H, dd, J ) 17.0, 1.8 Hz), 3.82 (2H, dt, J ) 5.2,
1.8 Hz), 2.37 (1H, s), 1.95 (1H, s), 1.72 (6H, s), 1.58 (6H, s); 13C
NMR (CDCl3, 125 MHz) δ 146.5, 140.0, 139.7, 138.54, 138.45, 127.6,
114.3, 102.5, 100.7, 85.2, 75.8, 73.9, 73.6, 32.3, 32.0, 30.4; HRMS
Experimental Section
General Methods. All commercial materials (purchased from
Aldrich-Sigma) were used without further purification. The following
solvents were obtained from a dry solvent system (passed through a
column of alumina): THF, diethyl ether (Et2O), CH2Cl2, toluene, and
benzene. All reactions were performed under an atmosphere of argon.
Powdered molecular sieves (4 Å) were charged by flame drying under
high vacuum for 20 min. NMR (1H and 13C) spectra were recorded on
a Bruker AMX-400 MHz or Bruker DRX-500 MHz and referenced to
residual solvent unless otherwise noted. IR spectra were recorded with
a Perkin-Elmer 1600 series FTIR spectrometer. High-resolution mass
spectra were recorded on a JEOL JMS-DX-303 HF mass spectrometer,
and low-resolution mass spectra were recorded on a Nermag R10-10
mass spectrometer. Analytical TLC was performed on E. Merck silica
gel 60 F254 plates, and flash column chromatography was performed
using indicated solvents on E. Merck silica gel 60 (40-63 µm).
Allyl Ether 23. Phenol 21 (5 g, 25 mmol) was dissolved in acetone
(250 mL), and solid potassium carbonate (10.3 g, 75 mmol) was added.
The suspension was stirred for 10 min at room temperature and then
treated with 1,3-dibromo-2-phenylsulfonylpropane (9.6 g, 28 mmol)
and potassium iodide (100 mg, 0.60 mmol). The mixture was stirred
at room temperature for 12 h. The suspension was filtered through Celite
and concentrated. Flash column chromatography (hexane-ether 90:
10) gave allyl ether 23 (8.4 g, 87%): IR (neat) 2953, 1697, 1612, 1420,
(35) Methyl acetal 44 was successfully converted to mixed acetal 50
via bromination, followed by treatment with catechol 47 in the presence of
base. Unfortunately, all attempts to close the five-membered ring of iv to
2 was unsuccessful. For instance, acid-induced reactions of iv typically gave
aldehyde 3 and catechol 47.
(36) Ikeya, Y.; Taguchi, H.; Yoshioka, I. Chem. Pharm. Bull. 1981, 29,
2893.

Synthetic Route to the Neurotrophic Illicinones
J. Am. Chem. Soc., Vol. 122, No. 26, 2000 6167
calcd for C18H22O4 [M]+ 302.1518, found 302.1515; LRMS 302, 221,
218, 203.
this mixture, and the resultant solution was stirred for 3h at 0 °C.
Concentration and flash column chromatography (hexane-ether 95:5)
gave acetate 33 (125 mg, 82%): IR (neat) 1735, 1635 cm-1; 1H NMR
(CDCl3, 400 MHz) 5.64-5.58 (2H, m), 5.59 (1H, s), 5.55 (1H, s), 5.53
(1H, s), 5.18 (1H, m), 5.06 (1H, s), 5.01 (1H, d, J ) 5.1 Hz), 2.78-
2.51 (4H, m), 2.35 (1H, m), 2.17 (3H, s), 1.73 (3H, s), 1.69 (3H, s);
13C NMR (CDCl3, 75.5 MHz) 196.0, 173.4, 170.0, 137.3, 134.2, 118.7,
116.8, 101.2, 99.8, 85.8, 74.5, 48.4, 32.4, 30.7, 26.4, 21.3, 18.4; HRMS
calcd for C17H22O5 [M]+ 306.1467, found 306.1454.
Phenol 29. Tertiary alcohol 27 (3.5 g, 11.8 mmol) was dissolved in
CH2Cl2 (400 mL) and cooled to 0 °C. To this solution was added
dropwise a precooled (0 °C) mixture of concentrated sulfuric acid (45
mL) and 50% H2O2 (50 mL). After 10 min at 0 °C, the reaction mixture
was diluted with water, and the organic layer was washed successively
with saturated aqueous Na2S2O4 and saturated aqueous NaHCO3 and
dried over Na2SO4. Flash column chromatography (hexane-ether 90:
10) gave the desired phenol 29 (2.6 g, 85%): IR (neat) 3499, 3288,
Illicinone-A (4). To acetate 33 (62 mg, 0.21 mmol) at 0 °C was
added DBU (0.5 mL), and the reaction was warmed to room
temperature. The solution was stirred for 3 h and diluted with EtOAc.
The organic layer was washed with H2O and brine, dried over Na2-
SO4, and concentrated. Flash column chromatography (hexane-EtOAc
1
2985, 2250, 1633, 1461 cm-1; H NMR (CDCl3, 500 MHz) δ 6.28
(1H, s), 5.94 (1H, m), 5.84 (2H, s), 5.18 (1H, dd, J ) 17.2, 1.6 Hz),
5.15 (1H, dd, J ) 10.1, 1.6 Hz), 4.94 (1H, s), 3.46 (2H, dt, J ) 6.0,
1.6 Hz), 2.41 (1H, s), 1.69 (6H, s); 13C NMR (CDCl3, 75.5 MHz) δ
149.6, 147.0, 136.8, 136.7, 134.7, 116.2, 113.5, 100.5, 95.3, 85.3, 76.2,
73.7, 30.1, 29.3; HRMS calcd for C15H16O4 [M]+ 260.1049, found
260.1037.
1
85:15) gave 4 (33 mg, 65%): IR (neat) 1665, 1630, 1610 cm-1; H
NMR (CDCl3, 400 MHz) δ 6.62 (1H, s), 5.81 (1H, m), 5.65 (1H, s),
5.62 (1H, s), 5.55 (1H, s), 5.13 (2H, m), 5.01 (1H, m), 3.07 (2H, m),
2.51 (1H, dd, J ) 17.4, 9.5 Hz), 2.41 (1H, dd, J ) 17.4, 10.0 Hz),
1.69 (3H, s), 1.53 (3H, s); 13C NMR (CDCl3, 125 MHz) δ 186.9, 173.7,
139.4, 137.7, 134.8, 134.4, 117.3, 115.9, 98.4, 98.0, 82.1, 34.8, 33.4,
25.9, 17.9; HRMS calcd for C15H18O3 [M]+ 246.1256, found 246.1267;
LRMS 246, 203, 188, 178, 150, 147.
Irradiation of Illicinone-A. A solution of illicinone-A (30 mg, 0.14
mmol) in a Pyrex test tube was degassed and exposed to light emitted
from a 450-W Hg lamp for 1 h. Flash column chromatography gave 1
(3 mg), the tetracyclic compound 6 (4 mg), and recovered illicinone-A
(20 mg). Prolonged exposure resulted in significant decomposition of
tricycloillicinone: 1H NMR (CDCl3, 500 MHz) δ 5.64 (1H, s), 5.45
(1H, s), 5.27 (1H, s), 4.85 (1H, s), 4.78 (1H, s), 3.32 (1H, dt, J ) 15.6,
2.1 Hz), 2.00-2.25 (4H, m), 1.84-1.93 (2H, m), 1.10 (3H, s), 1.07
(3H, s); HRMS calcd for C15H18O3 [M]+ 246.1256, found 246.1251;
LRMS 246, 231, 216, 201, 188, 173.
TBDPS Ether 30. Phenol 29 (2.6 g, 10 mmol) was dissolved in
CH2Cl2 (20 mL) and treated with TBDPSCl (2.8 mL, 10.8 mmol),
imidazole (735 mg, 10.8 mmol), and dimethylaminopyridine (25 mg).
After 12 h at room temperature, the solvent was removed and the residue
was directly subjected to flash column chromatography (hexane-ether
95:5) to give the desired silyl ether 30 (4.9 g, 97%): IR (neat) 2931,
1
2180, 1469 cm-1; H NMR (CDCl3, 500 MHz) δ 7.68 (4H, m), 7.39
(2H, m), 7.34 (4H, m), 6.02 (1H, m), 5.86 (1H, s), 5.69 (2H, s), 5.02
(1H, dd, J ) 17.1, 1.8 Hz), 4.98 (1H, dd, J ) 10.2, 1.8 Hz), 3.55 (2H,
dt, J ) 5.8, 1.5 Hz), 2.39 (1H, s), 1.70 (6H, s), 1.05 (9H, s); 13C NMR
(CDCl3, 75.5 MHz) δ 147.9, 145.9, 143.6, 137.3, 137.1, 135.4, 134.4,
132.7, 129.8, 127.7, 118.7, 114.3, 100.3, 97.7, 85.5, 75.7, 73.3, 30.2,
29.4, 26.5, 19.4; HRMS calcd for C31H34O4Si [M]+ 498.2226, found
498.2219.
Allyl Ether 31. Propargyl ether 30 (1 g, 2 mmol) was dissolved in
ethyl acetate (10 mL), and Lindlar’s catalyst (10% Pb-poisoned, 50
mg) and quinoline (50 µL) were added. The suspension was stirred
under hydrogen for 2 h, filtered through Celite, concentrated, and
subjected to flash column chromatography (hexane-ether 95:5) to give
the allyl ether 31 (990 mg, 99%): IR (neat) 2931, 1621 cm-1; 1H NMR
(CDCl3, 500 MHz) δ 7.67 (4H, dd, J ) 7.5, 1.1 Hz), 7.40 (2H, t, J )
7.5 Hz), 7.34 (4H, t, J ) 7.5 Hz), 6.11 (1H, dd, J ) 17.4, 10.8 Hz),
6.01 (1H, m), 5.83 (1H, s), 5.66 (2H, s), 5.18 (1H, d, J ) 17.4 Hz),
Oxytricycloillicinone (9). Mn(OAc)3‚2H2O (181 mg, 0.675 mmol)
and Cu(OAc)2‚H2O (66 mg, 0.330 mmol) were suspended in AcOH
(3.0 mL), and cyclohexadienone 32 (154 mg, 0.307 mmol) was added
to this suspension. The reaction mixture was stirred at room temperature
for 12 h and diluted with EtOAc. The organic layer was washed with
H2O, saturated NaHCO3, H2O, and brine and dried over Na2SO4.
Concentration and flash column chromatography (hexane-ether 85:
15) gave 9 (66 mg, 76%): IR (neat) 3100, 3063, 2970, 1772, 1670,
1643 cm-1; 1H NMR (500 MHz, CDCl3) δ 5.80 (1H, s), 5.64 (1H, s),
5.62 (1H, s), 4.90 (1H, s), 4.78 (1H, s), 3.13 (1H, d, J ) 16.8 Hz),
2.87 (1H, dt, J ) 16.8, 2.2 Hz), 2.35 (1H, dd, J ) 12.5, 10.2 Hz), 2.18
(1H, dd, J ) 10.2, 6.1 Hz), 2.06 (1H, dd, J ) 12.5, 6.1 Hz), 1.07 (3H,
s), 0.90 (3H, s); 13C NMR (75 MHz, CDCl3) δ 199.7, 193.2, 174.8,
158.0, 105.7, 101.4, 98.7, 90.8, 67.5, 51.6, 45.4, 27.9, 27.1, 26.5, 24.0;
HRMS (FAB) calcd for C15H16O4 [M]+ 260.1049, found 260.1038.
Alcohol 34. 9 (42 mg, 0.161 mmol) was dissolved in THF (5 mL),
and LiAlH(O-t-Bu)3 (1.0 M solution in THF, 185 µL, 0.185 mmol)
was added at -78 °C. The reaction mixture was stirred at -78 to 0 °C
for 3 h and then was quenched with acetone. The organic layer was
diluted with Et2O, washed with H2O and brine, and dried over Na2-
SO4. Concentration and flash column chromatography (hexane-ether
70:30) gave alcohol 34 (40 mg, >95% yield): IR (neat) 3420, 3058,
5.01 (3H, m), 3.47 (2H, d, J ) 5.8 Hz), 1.46 (6H, s), 1.05 (9H, s); 13
C
NMR (CDCl3, 75.5 MHz) δ 148.1, 145.9, 143.6, 137.6, 137.2, 135.5,
132.7, 129.8, 127.7, 127.5, 118.3, 114.3, 112.7, 100.1, 97.1, 81.9, 31.5,
27.0, 26.5, 19.4; HRMS calcd for C31H36O4Si [M]+ 500.2383, found
500.2390.
Cyclohexadienone 32. Allyl ether 31 (250 mg, 0.5 mmol) was
dissolved in toluene (10 mL) and heated at 100 °C for 2 h.
Concentration gave the rearranged product 32 (250 mg, 100%), which
was used in the following reaction without further purification: IR (neat)
3073, 2960, 2931, 2859, 1684, 1654 cm-1; 1H NMR (CDCl3, 400 MHz)
δ 7.68 (4H, m), 7.44 (6H, m), 5.88 (1H, m), 5.41 (1H, s), 5.35 (1H, s),
5.11-4.95 (3H, m), 4.78 (1H, s), 3.26 (1H, dd, J ) 14.8, 6.2 Hz),
3.13 (1H, ddt, J ) 14.8, 6.2, 1.4 Hz), 2.58 (1H, dd, J ) 14.5, 7.5 Hz),
2.48 (1H, dd, J ) 14.5, 7.1 Hz), 1.67 (3H, s), 1.55 (3H, s), 1.05 (9H,
s); 13C NMR (CDCl3, 100 MHz) δ 196.1, 167.3, 165.2, 137.3, 136.5,
135.5, 132.3, 130.9, 128.7, 128.6, 128.5, 116.3, 114.9, 111.5, 99.3,
91.8, 87.9, 35.6, 27.3, 26.9, 26.3, 19.8, 18.3; HRMS calcd for C31H37O4-
Si [M + H]+ 501.2461, found 501.2452.
1
2958, 1668, 1645 cm-1; H NMR (500 MHz, CDCl3) δ 5.64 (1H, s),
5.47 (1H, s), 5.37 (1H, s), 4.87 (1H, s), 4.74 (1H, s), 3.82 (1H, d, J )
3.5 Hz), 3.30 (1H, dt, J ) 15.4, 2.2 Hz), 2.49 (1H, d, J ) 16.4 Hz),
2.43 (1H, d, J ) 3.5 Hz), 2.30 (1H, m), 2.06 (2H, m), 1.17 (3H, s),
1.07 (3H, s); 13C NMR (75 MHz, CDCl3) δ 199.1, 176.0, 162.6, 103.8,
100.0, 97.1, 92.2, 79.8, 63.3, 54.9, 43.5, 32.3, 30.7, 28.1, 24.5; HRMS
(FAB) calcd for C15H18O4 [M]+ 262.1205, found 262.1212.
Acetate 33. Cyclohexadienone 32 (250 mg, 0.5 mmol) was dissolved
in THF (10 mL) and cooled to 0 °C. Lithium tri-sec-butylborohydride
(1.0 M in THF, 0.5 mL, 0.5 mmol) was added to this mixture, and the
resultant solution was stirred for 3 h at 0 °C. The reaction mixture was
quenched by p-toluenesulfonic acid (250 mg) and stirred for 2 h at 0
°C. Concentration gave the crude alcohol, which was used in the
following reaction without further purification: IR (neat) 3406, 1651,
Xanthate 35. To a solution of alcohol 34 (14.0 mg, 0.053 mmol) in
THF (2 mL) was added KHMDS (0.5 M in toluene, 0.75 mL, 0.375
mmol) at -78 °C. The solution stirred at -78 °C for 30 min. This
solution was added to the salt of DMAP (52 mg, 0.42 mmol) and PhOC-
(S)Cl (59 µL, 0.42 mmol) in THF (4 mL) at -20 °C, and the resultant
mixture was stirred overnight at the same temperature. The reaction
mixture was diluted with EtOAc, washed with H2O and brine, and dried
over Na2SO4. Concentration and flash column chromatography (hex-
ane-ether 80:20) gave desired 35 (21 mg, >95% yield): IR (neat)
1
1643, 1636 cm-1; H NMR (CDCl3, 400 MHz) δ 5.73 (1H, m), 5.55
(1H, s), 5.53 (1H, s), 5.51 (1H, s), 5.25 (1H, m), 5.18-5.05 (2H, m),
4.14 (1H, dd, J ) 11.2, 3.7 Hz), 2.83 (2H, m), 2.61 (2H, m), 2.45 (1H,
m), 2.32 (1H, d, J ) 3.7 Hz), 1.75 (3H, s), 1.67 (3H, s).
The above alcohol was dissolved in CH2Cl2 (10 mL) and cooled to
0 °C. Acetic anhydride (0.25 mL) and DMAP (250 mg) were added to
1
2962, 1674, 1640, 1480 cm-1; H NMR (500 MHz, CDCl3) δ 7.40

6168 J. Am. Chem. Soc., Vol. 122, No. 26, 2000
Pettus et al.
(2H, t, J ) 7.6 Hz), 7.28 (1H, t, J ) 7.6 Hz), 7.08 (2H, d, J ) 7.6 Hz),
5.70 (1H, s), 5.66 (1H, s), 5.53 (1H, s), 5.49 (1H, s), 4.84 (1H, s), 4.72
(1H, s), 3.35 (1H, dt, J ) 16.8, 2.5 Hz), 2.53 (1H, d, J ) 16.8 Hz),
2.37 (1H, dd, J ) 12.1, 7.3 Hz), 2.22 (1H, dd, J ) 11.2, 2.7 Hz), 2.15
(1H, dd, J ) 16.5, 9.2 Hz), 1.17 (3H, s), 1.10 (3H, s); 13C NMR (75
MHz, CDCl3) δ 174.8, 160.8, 129.5, 126.7, 121.7, 102.6, 99.9, 97.4,
89.7, 87.1, 63.3, 54.6, 43.0, 33.1, 31.4, 29.5, 23.9; HRMS (FAB) calcd
for C22H23O5S [M + H]+ 399.1267, found 399.1281.
Acknowledgment. This work was supported by the NIH
(Grants HL25848 (S.J.D.)). T.R.R.P. gratefully acknowledges
the NSF for a Postdoctoral Fellowship (Grant CHE96-26443).
M.I. gratefully acknowledges the Uehara Memorial Foundation
for a postdoctoral fellowship. We are grateful to Vinka
Parmakovich and Barbara Sporer of the Columbia University
Mass Spectral Facility for their measurements and Dr. George
Sukenick of the Sloan Kettering Institutes NMR Core Facility
(NIH-CA08748 (S.K.I. Core Grant)) for NMR and mass spectral
analyses. We also thank Professor Fukuyama for kindly provid-
ing us a copy of the spectral data of 1-3.
Tricycloillicinone (1). To a solution of xanthate 35 (7.0 mg, 0.0176
mmol) in benzene (2 mL) were added Bu3SnH (35 mg, 0.12 mmol)
and AIBN (1.4 mg, 0.0085 mmol). The solution was degassed twice
and heated at 60 °C for 7 h. Concentration and flash column
chromatography (hexane-ether 85:15) gave 1 (1.8 mg, 42% yield):
1
IR (neat) 3060, 2962, 2924, 1670, 1646 cm-1; H NMR (500 MHz,
CDCl3) δ 5.64 (1H, s), 5.45 (1H, s), 5.27 (1H, s), 4.85 (1H, s), 4.78
(1H, s), 3.32 (1H, dt, J ) 15.6, 2.1 Hz), 2.00-2.25 (4H, m), 1.84-
Supporting Information Available: 1H and 13C spectra for
1, 3, 4, 9, 23-27, 29-35, 38, 39, 41-43, 47, and 49 and H
NMR for 44 and 48; experimental protocols leading to 3, 36,
39, 41-44, and 47-49 are available (PDF) (print/PDF). This
information is available free of charge via the Internet at
http://pubs.acs.org.
1
1.93 (2H, m), 1.10 (3H, s), 1.07 (3H, s); H NMR (500 MHz, C6D6,
1
δref ) 7.16) δ 5.43 (1H, s), 4.90 (1H, d, J ) 0.8 Hz), 4.77 (1H, s),
4.70 (1H, s), 4.69 (1H, s), 3.63 (1H, dt, J ) 15.5, 2.1 Hz), 1.90 (1H,
d, J ) 15.5 Hz), 1.83 (1H, d, J ) 10.3 Hz), 1.78 (1H, t, J ) 8.2 Hz),
1.64 (1H, dd, J ) 10.3, 0.8 Hz), 1.45 (2H, m), 0.93 (3H, s), 0.80 (3H,
s); 13C NMR (75 MHz, C6D6, δref ) 128.0) δ 198.3, 177.0, 159.9, 105.1,
98.8, 95.3, 87.6, 56.9, 53.8, 44.8, 42.8, 35.4, 31.7, 31.4, 24.4; HRMS
(FAB) calcd for [C15H18O3]+ 246.1256, found 246.1251; LRMS (CI)
246, 231, 216, 201, 188, 173.
JA000521M