Total synthesis of resveratrol-based natural products: A chemoselective solution

Article abstract of DOI:10.1002/anie.200703333

Despite the attention paid to resveratrol (1) owing to its potent biological activity, little effort has been devoted to studying resveratrol-based oligomers (such as 2-4). The first general synthetic approach is outlined for accessing the carbogenic diversity possessed by this family of compounds. (Chemical Equation Presented).

Full text of DOI:10.1002/anie.200703333

Communications
DOI: 10.1002/anie.200703333
Natural Products
Total Synthesis of Resveratrol-Based Natural Products:
A Chemoselective Solution**
Scott A. Snyder,* Alexandros L. Zografos, and Yunqing Lin
The past decade has witnessed tremendous interest in the
relatively small natural product resveratrol (1, Figure 1)
primarily because of its promising and selective array of
in vitro and in vivo activity against a collection of disease
states, including inflammation, heart disease, aging, and
cancer.[1] In fact, its truly unique biochemical profile, coupled
with its relatively high concentration in red wine (ca. 100 mm)
and near absence in white varietals and grape juice, has led to
the popularly held notion that resveratrol is the main
protagonist for the so-called “French paradox”.[2] Amazingly,
however, virtually no effort has been devoted to the large
family of resveratrol-based oligomers (such as 2–8)[3] pro-
duced combinatorially by plants throughout the world in
response to environmental stress; initial screening suggests
these compounds should have similar, if not superior, activity
profiles to resveratrol itself.[4] Herein, we provide a means to
begin this exploration by outlining the first general synthetic
approach capable of accessing all the carbogenic diversity
posed by this family, a solution fueled by a new idea for the
selective generation of natural product structures in instances
in which nature abandons discrimination to achieve evolu-
tionary advantage.[5]
tivity problem, one empowered by the identification of
hidden relationships between their seemingly divergent
architectures. Such an answer arose when we considered
natural products such as paucifloral F (7) and diptoindone-
sin A (8), isolates whose structures are incongruent with the
notion of direct resveratrol oligomerization since they possess
To date, all attempts to prepare resveratrol-based natural
products have derived from strategies that parallel their
presumed biogenesis, that is, the generation of radicals or
carbocations from 1 through its exposure to different
chemicals or enzymes.[6–11] Typically, mixtures of compounds
were observed[6–8] and, in those rare instances when selectivity
was achieved, solely nonnatural products resulted.[9,10] Thus
far, only a highly engineered resveratrol fragment has proven
capable of leading to an actual dimeric natural product within
this class.[11] Given this state of affairs, we wondered if a
structural solution might exist for this general chemoselec-
[*] Prof. Dr. S. A. Snyder, Dr. A. L. Zografos, Y. Lin
Department of Chemistry
Columbia University
Havemeyer Hall, MC 3129
3000 Broadway, New York, NY 10027 (USA)
Fax: (+1)212-932-1289
E-mail: sas2197@columbia.edu
Homepage: http://www.columbia.edu/cu/chemistry/groups/
snyder/
[**] We thank Drs. J. Decatur and Y. Itagaki for assistance with NMR
spectroscopy and mass spectrometry, respectively. We also thank
the National Science Foundation (CHE-0619638) for the acquisition
of an X-ray diffractometer. Financial support for this work was
provided by Columbia University, Amgen, and Eli Lilly. S.A.S. is a
Camille and Henry Dreyfus New Faculty Awardee.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Figure 1. Selected examples of polyphenolic natural products pre-
sumed to arise from the union of resveratrol monomers.
8186
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 8186 –8191

Angewandte
Chemie
three, instead of four, aromatic rings.[12] Indeed, these
compounds led us to the idea that there are two highly
similar building blocks well removed from resveratrol,
each with three aryl groups around the same core structure,
which can controllably lead to every family member simply
by altering the reagents and reaction conditions to which
they are exposed.
One of these compounds is biaryl alcohol 11
(Scheme 1), which was synthesized in 71% yield through
an aldol reaction between the lithiated form of 9[13] and 3,5-
dimethoxybenzaldehyde (10). As shown in Scheme 1,
when this key intermediate was treated with a stoichio-
metric amount of TFA under carefully controlled con-
ditions (ꢀ30!ꢀ208C) in CH2Cl2, a cascade sequence
featuring cation generation, regio- and stereoselective
cyclization (in relative terms), and stereoselective cation
capture, afforded intermediate 14 after 4 h. Quenching
under basic conditions (K2CO3, MeOH) then completed
the one-pot synthesis of intermediate 15 from 9 in 75%
yield; compound 15 proved to be just two steps away from
paucifloral F (7). Those operations, alcohol oxidation by
Dess–Martin periodinane and BBr3-induced global deme-
thylation in CH2Cl2 at 08C,[14] proceeded smoothly in 84%
overall yield. However, exposure of 11 to a proton source
with a nonnucleophilic counterion such as that possessed
by TsOH arrested the sequence at cation 13 prior to b-
hydride elimination and allowed access to entirely differ-
ent cyclic products. Indeed, if a nucleophile such as p-
methoxy-a-toluenethiol was added at ꢀ308C after 11 had
been exposed to TsOH for 5 h and the reaction medium
was then concentrated to near dryness, sulfide 16 was
obtained in 57% overall yield.[15] This new tetraaryl
intermediate could then be converted into the natural
product ampelopsin D (2) through a highly selective
Ramberg–Bäcklund reaction[16] under Meyerꢀs modified
conditions[17] that afforded permethylated ampelopsin D
along with its chromatographically separable Z-olefin
isomer in a 5:1 ratio (40% and 7% yield over two steps,
respectively), followed by Lewis acid mediated phenol
deprotection using BBr3.[18] Subsequent treatment of 2 with
five equivalents of HCl in MeOH at 808C effected olefin
isomerization to give isoampelopsin D (17) in near quan-
titative yield.[19]
Scheme 1. Total synthesis of three dimeric resveratrol-based natural
products (2, 7, and 17) from key building block 11: a) nBuLi
(1.0 equiv), THF, ꢀ788C, 20 min; then 10 (1.0 equiv), ꢀ78!258C,
4 h, 71%; b) for 15: TFA (1.0 equiv), CH2Cl2, ꢀ30!ꢀ208C, 5 h;
then K2CO3 (10 equiv), MeOH, 258C, 5 min, 75%; for 16: TsOH
(1.0 equiv), CH2Cl2, ꢀ30!ꢀ208C, 5 h; p-methoxy-a-toluenethiol
(3.0 equiv), then concentration to near dryness, 258C, 12 h, 57%;
c) Dess–Martin periodinane (1.2 equiv), NaHCO3 (5.0 equiv),
CH2Cl2, 258C, 3 h, 97%; d) BBr3 (1.0m in CH2Cl2, 10 equiv), CH2Cl2,
08C, 6 h, 86%; e) mCPBA (3.0 equiv), NaHCO3 (10 equiv), CH2Cl2,
0!258C, 3 h, 78%; f) tBuOH/H2O/CCl4 (5/1/5), KOH (20 equiv),
808C, 12 h, 52%; g) BBr3 (1.0m in CH2Cl2, 12 equiv), CH2Cl2, 258C,
6 h, 76% of 2, 13% of 17; h) conc. HCl (5 equiv), MeOH, 808C,
2 h, 96%. TFA=trifluoroacetic acid, TsOH=p-toluenesulfonic acid,
mCBPA=m-chloroperoxybenzoic acid.
Angew. Chem. Int. Ed. 2007, 46, 8186 –8191
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8187

Communications
The other building block is 20 (Scheme 2), which differs
from biaryl alcohol 11 architecturally (in terms of the
positioning of two of its three aromatic rings), but behaves
As shown in Scheme 3, that conjecture proved to be
correct if bromine was utilized as the activating species.[22] In
the event, exposure of permethylated quadrangularin A (23)
Scheme 2. Total synthesis of two resveratrol-based natural products
(21 and 22) from key building block 20. a) nBuLi (1.0 equiv), THF,
ꢀ788C, 20 min; then 19 (1.0 equiv), ꢀ78!258C, 4 h, 71%; b) TFA
(1.0 equiv), CH2Cl2, ꢀ30!ꢀ208C, 5 h; then K2CO3 (10 equiv), MeOH,
258C, 5 min, 93%; c) Dess–Martin periodinane (1.2 equiv), NaHCO3
(5.0 equiv), CH2Cl2, 258C, 3 h, 98%; d) 9-I-BBN (1.0m in hexanes,
10 equiv), CH2Cl2, 408C, 30 min, 72%; e) TsOH (1.0 equiv), CH2Cl2,
ꢀ30!ꢀ208C, 5 h; p-methoxy-a-toluenethiol (3.0 equiv), then concen-
tration to near dryness, 258C, 12 h, 65%; f) mCPBA (3.0 equiv),
NaHCO3 (10 equiv), CH2Cl2, 0!258C, 3 h, 70%; g) tBuOH/H2O/CCl4
(5/1/5), KOH (20 equiv), 808C, 12 h, 55%; h) BBr3 (1.0m in CH2Cl2,
12 equiv), CH2Cl2, 258C, 6 h, 75% of 21, 14% of internal alkene
isomer. 9-I-BBN=9-iodo-9-borabicyclo[3.3.1]nonane.
Scheme 3. Sequential, cascade-based halogenation to access pallidol
(3): a) Br2 (2.0 equiv), CH2Cl2, ꢀ788C, 2 h, then slow warming to
258C, 1 h, 81%; b) H2, Pd/C (20%, 0.2 equiv), MeOH, 258C, 24 h,
76%; c) BBr3 (1.0m in CH2Cl2, 12 equiv), CH2Cl2, 08C, 4 h, then 258C,
20 h, 83%.
in the same manner chemically. Indeed, as indicated in
Scheme 2, when this intermediate was subjected to the
reaction sequences outlined above, what resulted were total
syntheses of quadrangularin A (21) and isopaucifloral F
(22),[20] the structures of which, as expected, display the
opposite array of pendant phenol ring systems as those
accessed from 11.[21] Consequently, it would appear, on the
basis of these collated results, that any resveratrol-derived
structure possessing a single cyclopentane ring system can be
obtained cleanly from appropriate triaryl precursors.
What, though, about more complex intermediates such as
pallidol (3) and ampelopsin F (4, Figure 1), which possess an
additional ring appended onto a cyclopentane core? Prior
explorations with naturally derived materials had established
that their complexity could arise alongside several other
architectures by treating dihydrofuran-bearing substrates,
such as vaticanol C (5, Figure 1) and hopeaphenol (6,
Figure 1), with strong acid.[19] We wondered whether electro-
philic activation of the olefins within both ampelopsin D (2)
and quadrangularin A (21), followed by a Friedel–Crafts
alkylation onto the resultant quinone methide, could accom-
plish the same objective in a controlled manner.
to two equivalents of molecular bromine in CH2Cl2 at ꢀ788C
and subsequent slow warming to ambient temperature over
several hours accomplished
a highly selective cascade
sequence that provided bicycle 27 in 81% yield. On the
basis of a series of control experiments leading to the isolation
of both 24 and 25, the course of events is known to involve the
initial halogenation of the C-14b position, followed by
bromination of the second 3,5-dimethoxybenzene ring
system. Although both of these halogens are extraneous in
terms of the goal structure,[23] each served a critical role in
ensuring that the terminating ring closure leading to 27 was
stereoselective. Indeed, as revealed by molecular models, the
C-10a bromine atom provided a significant amount of steric
bulk to its ring system, forcing the third bromine atom to be
added solely from the opposite side of the molecule; the C-
14b bromide then prevented rotation of the newly formed
quinone methide (26) away from its initial, perfect positioning
for the final closure, thereby assuring that only 27 was formed.
From this key intermediate (27), pallidol (3) was then
accessed in 63% overall yield by hydrogenative replacement
8188
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 8186 –8191

Angewandte
Chemie
of all three bromides by using a catalytic amount of activated
Pd/C, followed by BBr3-induced cleavage of all six methyl
ethers. As documented in Scheme 4, the same sequence of
Scheme 4. Sequential, cascade-based halogenation to access ampelop-
sin F (4): a) Br2 (2.0 equiv), CH2Cl2, ꢀ788C, 2 h, then slow warming to
258C, 1 h, 53%; b) (TMS)3SiH (9.0 equiv), AIBN (1.0 equiv), toluene,
1008C, 8 h, 89%; c) BBr3 (1.0m in CH2Cl2, 12 equiv), CH2Cl2, 08C, 4 h,
then 258C, 15 h, 90%. TMS=trimethylsilyl, AIBN=2,2’-azobisisobutyr-
onitrile.
events with permethylated ampelopsin D (28) selectively
afforded ampelopsin F (4). In this case, radical conditions
[(TMS)3SiH, AIBN] were used to replace the three bromine
atoms left by the cascade sequence.[24] Of course, although an
ideal synthesis of any molecule would avoid the addition of
extra atoms, in these two cases the absence of atom economy
would appear to have the benefit of such potential access to
even greater molecular complexity in the resveratrol class.
Indeed, the aryl halides within intermediate 30 are positioned
perfectly to attempt construction of the dihydrofuran rings
that would lead to vaticanol C (5, Figure 1).
Finally, the remaining element of carbogenic complexity
possessed by the resveratrol family, the seven-membered
rings of compounds such as diptoindonesin A (8, Figure 1),
could be obtained through an electrophilic activation/cycliza-
tion sequence similar to that just described. In this case, the
key starting material is ketone 31 (Scheme 5), the oxidized
form of building block 11, which afforded 33 in 50% yield of
isolated product following its exposure to bromine. Although
work with this highly sensitive intermediate is only in its
initial stages, the halogen handle within 33 is likely to be a key
tool for efforts to synthesize the carbon–carbon bond uniting
the two halves of hopeaphenol (6, Figure 1) and generate the
additional oxygen function of both diptoindonesin A (8,
Figure 1) and the related natural product hemsleyanol E
(38).[25]
Scheme 5. Alternate use of key intermediate 11 to access the unique
architectures of related, nonnatural natural products (such as 37)
through a bromonium-induced cascade sequence followed by an acid-
induced phenonium shift: a) Br2 (1 equiv), CH2Cl2, ꢀ788C, 1 h, then
258C, 12 h, 50%; b) AgOAc (3.0 equiv), AcOH, 258C, 4 h, 62%;
c) K2CO3 (10 equiv), MeOH, 258C, 12 h, 78%; d) Dess–Martin period-
inane (1.2 equiv), NaHCO3 (5.0 equiv), CH2Cl2, 258C, 1 h, 99%.
appear to be employed by nature in its construction of this
molecule class.[26] Indeed, exposure of bromide 33 to an excess
of AgOAc (3.0 equiv) in AcOH at 258C[27] led to the smooth
synthesis of acetate 36 in 62% yield. This unique structure
(confirmed by X-ray crystallographic analysis), in which the
pendant aryl ring has migrated, likely resulted from a
thermodynamically favored phenonium shift following the
generation of cation 34; the strategically positioned ortho-
and para-disposed alkoxy groups within the resultant inter-
mediate (35) then effected ring opening to provide a single,
and new, electrophilic site for a terminating acetate attack.[28]
Subsequent cleavage of the acetate within 36 (K2CO3,
Equally important, this halogen atom has already enabled
access to a collection of nonnatural analogues through a
molecular rearrangement that, despite its facility, does not
Angew. Chem. Int. Ed. 2007, 46, 8186 –8191
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8189

Communications
j) B. Supudompol, K. Likhitwitayawuid, P. J. Houghton, Phyto-
chemistry 2004, 65, 2589 – 2594; k) N. S. Aminah, S. A. Achmad,
N. Aimi, E. L. Ghisalberti, E. H. Hakim, M. Kitajima, Y. M.
Syah, H. Takayama, Filoterapia 2002, 73, 501 – 507.
MeOH) then provided a protected regioisomeric analogue of
hemsleyanol E in 78% yield, and oxidation of the resultant
alcohol led to the corresponding diptoindonesin A congener
as expressed by structure 37.
[4] For selected references, see: a) M. Ohyama, T. Tanaka, T. Ito, M.
Iinuma, K. F. Bastow, K.-H. Lee, Bioorg. Med. Chem. Lett. 1999,
9, 3057 – 3060; b) K. Ohguchi, T. Tanaka, T. Ito, M. Iinuma, K.
Matsumoto, Y. Akao, Y. Nozawa, Biosci. Biotechnol. Biochem.
2003, 67, 1587 – 1589; c) K. Ohguchi, Y. Akao, K. Matsumoto, T.
Tanaka, T. Ito, M. Iinuma, Y. Nozawa, Biosci. Biotechnol.
Biochem. 2005, 69, 353 – 356; d) T. Ito, Y. Akao, H. Yi, K.
Ohguchi, K. Matsumoto, T. Tanaka, M. Iinuma, Y. Nozawa,
Carcinogenesis 2003, 24, 1489 – 1497.
[5] For an insightful discussion on this topic, see: M. A. Fischbach, J.
Clardy, Nat. Chem. Biol. 2007, 3, 353 – 355.
[6] J. M. Aguirre, E. N. Alesso, G. Y. M. Iglesias, J. Chem. Soc.
Perkin Trans. 1 1999, 1353 – 1358.
[7] X.-C. Li, D. Ferreira, Tetrahedron 2003, 59, 1501 – 1507.
[8] Y. Takaya, K. Terashima, J. Ito, Y.-H. He, M. Takeoka, N.
Yamaguchi, M. Niwa, Tetrahedron 2005, 61, 10285 – 10290.
[9] P. Langcake, R. J. Pryce, J. Chem. Soc. Chem. Commun. 1977,
208 – 210.
[10] M. Sako, H. Hosokawa, T. Ito, M. Iinuma, J. Org. Chem. 2004, 69,
2598 – 2600.
[11] W. Li, H. Li, Z. Hou, Angew. Chem. 2006, 118, 7771 – 7773;
Angew. Chem. Int. Ed. 2006, 45, 7609 – 7611.
In conclusion, we have established that the entire array of
carbogenic complexity posed by the resveratrol family of
natural products, along with several additional isosteres, can
be accessed smoothly and selectively from building blocks
quite distinct from the compound postulated for their
biosynthesis.[29] Apart from revealing previously hidden
structural relationships within the architectural diversity
possessed by this compound class, the efficiency of the
developed routes (four to seven steps from 9 and 18, each
natural product accessed in 7 to 46% overall yield from
commercial materials) ensures that the biochemical studies
needed to elucidate their full medicinal potential can finally
begin in earnest. Future efforts are focused not only on
achieving this critical objective, but also on synthesizing the
most complex members of this fascinating family of secondary
metabolites. We expect that the general principle illustrated
by this work, namely the use of common, but nonobvious,
precursors to access molecular diversity selectively through
reagent-induced cascades, will prove applicable to many
classes of polymeric molecules.
[12] In fact, there are many members of this family with an odd
number of aryl rings.
[13] Prepared in four steps in 80% overall yield from 3,5-dimethoxy-
benzaldehyde; see the Supporting Information for more details.
Inspiration for part of this sequence came from the following
total synthesis of resveratrol: L. Botella, C. Nꢁjera, Tetrahedron
2004, 60, 5563 – 5570.
[14] For a recent example of this deprotection in the context of a total
synthesis, see: P. S. Baran, N. Z. Burns, J. Am. Chem. Soc. 2006,
128, 3908 – 3909.
Received: July 24, 2007
Published online: September 21, 2007
Keywords: chemoselectivity · natural products · resveratrol ·
.
total synthesis
[15] Sulfide 16 could also be accessed in 82% yield from alcohol 15
upon its treatment with TsOH and p-methoxy-a-toluenethiol in
CH2Cl2 at 258C; see the Supporting Information for full details.
[16] For a recent review on this reaction, see: R. J. K. Taylor, Chem.
Commun. 1999, 217 – 227.
[17] C. Y. Meyers, A. M. Malte, W. S. Matthews, J. Am. Chem. Soc.
1969, 91, 7510 – 7512. The separable Z-alkene isomer produced
in this Ramberg–Bäcklund step is a fully protected form of the
natural product parthenocissin A: T. Tanaka, M. Iinuma, H.
Murata, Phytochemistry 1998, 48, 1045 – 1049.
[18] This final deprotection step produced a 5:1 mixture of both
ampelopsin D (2) and isoampelopsin D (17), which were
obtained in pure form in near quantitative yield by treating the
product mixture with Ac2O, chromatographically separating the
resultant acetates, and using KCN in MeOH to then effect ester
hydrolysis.
[1] For selected references, see: a) M. Jang, L. Cai, G. O. Udeani,
K. V. Slowing, C. F. Thomas, C. W. W. Beecher, H. H. S. Fong,
N. R. Farnsworth, A. D. Kinghorn, R. G. Mehta, R. C. Moon,
J. M. Pezzuto, Science 1997, 275, 218 – 220; b) K. T. Howitz, K. J.
Bitterman, H. Y. Cohen, D. W. Lamming, S. Lavu, J. G. Wood,
R. E. Zipkin, P. Chung, A. Kisielewski, L.-L. Zhang, B. Scherer,
D. A. Sinclair, Nature 2003, 425, 191 – 196; c) L. M. Szewczuk, L.
Forti, L. A. Stivala, T. M. Penning, J. Biol. Chem. 2004, 279,
22727 – 22737; d) J. A. Baur, K. J. Pearson, N. L. Price, H. A.
Jamieson, C. Lerin, A. Kalra, V. V. Prabhu, J. S. Allard, G.
Lopez-Lluch, K. Lewis, P. J. Pistell, S. Poosala, K. G. Becker, O.
Boss, D. Gwinn, M. Wang, S. Ramswamy, K. W. Fishbein, R. G.
Spencer, E. G. Lakatta, D. Le Couteur, R. J. Shaw, P. Navas, P.
Puigserver, D. K. Ingram, R. de Cabo, D. A. Sinclair, Nature
2006, 444, 337 – 342.
[2] For some early commentary, see: G. J. Soleas, E. P. Diamandis,
D. M. Goldberg, Clin. Biochem. 1997, 30, 91 – 113.
[19] Y. Takaya, K.-X. Yan, K. Terashima, Y.-H. He, M. Niwa,
Tetrahedron 2002, 58, 9265 – 9271.
[20] Isopaucifloral F (22) represents a likely, but not yet isolated,
natural product structure.
[3] a) Y. Oshima, Y. Ueno, K. Hisamachi, M. Takeshita, Tetrahedron
1993, 49, 5801 – 5804; b) Y. Oshima, Y. Ueno, Phytochemistry
1993, 33, 179 – 182; c) M. Niwa, J. Ito, K. Terashima, T. Koizumi,
Y. Takaya, K.-X. Yan, Heterocycles 2000, 53, 1475 – 1478; d) J.
Ito, T. Tanaka, M. Iinuma, K. Nakaya, Y. Takahashi, R. Sawa, J.
Murata, D. Darnaedi, J. Nat. Prod. 2004, 67, 932 – 937; e) H.-F.
Luo, L.-P. Zhang, C.-Q. Hu, Tetrahedron 2001, 57, 4849 – 4854;
f) M. A. Khan, S. G. Nabi, S. Prakash, A. Zaman, Phytochem-
istry 1986, 25, 1945 – 1948; g) H. A. Guebailia, K. Chira, T.
Richard, T. Mabrouk, A. Furiga, X. Vitrac, J.-P. Monti, J.-C.
Delaunay, J.-M. Merillon, J. Agric. Food Chem. 2006, 54, 9559 –
9564; h) T. Tanaka, T. Ito, K. Nakaya, M. Iinuma, S. Riswan,
Phytochemistry 2000, 54, 63 – 69; i) Y. Takaya, K.-X. Yan, K.
Terashima, J. Ito, M. Niwa, Tetrahedron 2002, 58, 7259 – 7265;
[21] S. A. Adesanya, R. Nia, M.-T. Martin, N. Boukamcha, A.
Montagnac, M. Paꢂs, J. Nat. Prod. 1999, 62, 1694 – 1695.
[22] The use of proton as an activating electrophile led in all cases to
internal alkenes, such as that possessed by isoampelopsin D (17,
Scheme 1), whereas all efforts to form epoxides led to intract-
able mixtures of compounds or unchanged starting material.
[23] For an interesting commentary on utilizing bromine as a
protective device, see: F. Effenberger, Angew. Chem. 2002,
114, 1775 – 1776; Angew. Chem. Int. Ed. 2002, 41, 1699 – 1700.
[24] With the opposite array of ring systems, the second bromine
atom attached to this molecule on the pendant 3,5-dimethoxy-
8190 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 8186 –8191

Angewandte
Chemie
benzene ring provided enough steric bulk to ensure that the third
bromine atom ultimately leading to quinone methide 29 came
from the same side of the molecule as the adjacent aromatic ring
system.
a pure enough form to verify its connectivity through NOE
experiments (owing to its sensitivity), we have substituted the
acetate within structure 36 for chloride with retention of
configuration. This compound, along with 36 and 37, possess
the four aryl protons of the 3,5-dimethoxybenzene rings tightly
and consistently grouped between d = 6.40 and 6.90 ppm; by
contrast, bromide 33 possesses one of these signals as an outlier
at d = 5.76 ppm, presumably the one that is in close proximity to
the halogen atom.
[25] T. Tanaka, T. Ito, K. Nakaya, M. Iinuma, Y. Takahashi, H.
Naganawa, S. Riswan, Heterocycles 2001, 55, 729 – 740.
[26] In fact, there are no molecules known with this particular array
of phenols attached to a seven-membered carbocycle.
[27] a) M. Harmata, S. Wacharasindhu, Org. Lett. 2005, 7, 2563 –
2565; b) M. Miesch, A. Cottꢃ, M. Frank-Neumann, Tetrahedron
Lett. 1993, 34, 8085 – 8086.
[29] In all cases, spectral data for synthetic materials perfectly match
those of the natural isolates. It should be noted that all molecules
reported in this manuscript are racemic.
[28] All evidence indicates that the aryl group has not shifted within
33. Although this compound has proven too difficult to obtain in
Angew. Chem. Int. Ed. 2007, 46, 8186 –8191
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8191