Synthesis of the carbocyclic core of zoanthenol: Implementation of an unusual acid-catalyzed cyclization

Article abstract of DOI:10.1002/anie.200700430

(Chemical Equation Presented) A smokin' hot cyclization! When the racemic cyclization precursor is heated in neat trifluoroacetic acid, an unusual Friedel-Crafts-type cyclization forms the carbocyclic core of the marine alkaloid zoanthenol containing two

Full text of DOI:10.1002/anie.200700430

Angewandte
Chemie
DOI: 10.1002/anie.200700430
Natural Products Synthesis
Synthesis of the Carbocyclic Core of Zoanthenol: Implementation of an
Unusual Acid-Catalyzed Cyclization**
Douglas C. Behenna, Jennifer L. Stockdill, and Brian M. Stoltz*
The complex and elegant architecture of the zoanthamine
alkaloids has captivated synthetic chemists for well over a
decade.[1,2] Although numerous research groups have pub-
lished efforts toward the zoanthamines,[1] the only completed
synthesis of any member of this class was that of norzoanth-
amine by Miyashita et al. which appeared in 2004[3] (Figure 1).
the greatest stereochemical challenge with five contiguous
stereocenters, three of which are all-carbon quaternary
centers. Our overarching strategy was to generate one
quaternary center in an enantioselective fashion and then
derive the remaining stereocenters diastereoselectively.
Another design feature was to convergently unite the A and
C rings by a two-carbon tether and subsequently forge the
B ring. We planned to introduce all the functionality of the
heterocyclic C1 to C8 fragment in a single operation (i.e., 2!
3 + 4; Scheme 1). Previous work by the Kobayashi and
Williams groups demonstrated that the complicated hemi-
aminals forming the DEFG rings are thermodynamically
favored.[7] Thus, the DEFG heterocycles could be retrosyn-
thetically unraveled to give triketone 2 (Scheme 1). Discon-
ꢀ
nection of the C8 C9 bond and removal of the methyl groups
at C9 and C19 affords ketone 3 and enone 4. We envisioned
the cleavage of the tricyclic core structure 3 by scission of the
ꢀ
C12 C13 bond employing an intramolecular conjugate addi-
Figure 1. Selected members of the zoanthus family of alkaloids.
tion of the A ring into a C ring enone (i.e., 5).[8] We reasoned
that this type of intramolecular Friedel–Crafts reaction would
require a highly electron-rich arene for effective cyclization;
therefore, oxygenation was incorporated at C16 of enone 5 to
increase the nucleophilicity of the A ring. Enone 5 could arise
from 1,2-addition of Grignard reagent 6 to enal 7, which in
turn could be derived from ketoester 8, ultimately available
by enantioselective decarboxylative allylation of b-ketoester
9.
The zoanthamines exhibit a host of biological activities
highlighted by the anti-osteoporotic activity of norzoanth-
amine hydrochloride.[4] Our interest in the zoanthamines was
piqued by zoanthenol (1),[5] the sole family member possess-
ing an aromatic A ring. Zoanthenol retains the major
stereochemical challenges of the zoanthamines, while offering
the opportunity to explore unique retrosynthetic possibilities.
Herein, we describe an unusual SN’ cyclization to form the
carbocyclic core of zoanthenol. An asymmetric stereoselec-
tive route to this core structure is enabled by asymmetric
alkylation methodology recently developed in our laborato-
ries.[6]
In order to determine the feasibility of the 6-exo conjugate
addition, the target enal was synthesized as a racemate
(Scheme 2). Known dimethyl ketone 10[9] was deprotonated
and alkylated to give ketoester 8 in excellent yield as a
mixture of diastereomers. Deprotonation of methyl ketone 8
and quenching with PhNTf2 afforded enol triflate 11. After
significant optimization to accommodate the steric challenges
of the substrate, an efficient one-step reductive carbonylation
of triflate 11 was developed. Treatment of triflate 11 under an
atmosphere of CO with Pd(OAc)2, 1,4-bis(dicyclohexylphos-
phino)butane as a ligand, and TES-H as a reducing agent
afforded the desired enal 7 in good yield. To our knowledge,
this is the first time that such a hindered vinyl triflate was
carbonylated directly to the enal oxidation state. Addition of
Grignard reagent 6[10] to enal 7 produced allylic alcohol 12 in
high yield and diastereoselectivity. Use of methylene chloride
as a cosolvent for the addition reaction was critical, resulting
in the formation of the anti diastereomer as the sole product,
as confirmed by X-ray crystallography of the corresponding
lactone (i.e., 16, Figure 2).[10]
With seven rings and nine stereocenters confined to a 30-
carbon framework, zoanthenol is a densely functionalized,
topographically complex target molecule. The C ring poses
[*] D. C. Behenna, J. L. Stockdill, Prof. B. M. Stoltz
The Arnold and Mabel Beckman Laboratories
of Chemical Synthesis
Division of Chemistry and Chemical Engineering
California Institute of Technology
1200 East California Boulevard, MC 164-30
Pasadena, CA 91125 (USA)
Fax: (+1)626-564-9297
E-mail: stoltz@caltech.edu
[**] The authors wish to thankthe Fannie and John Hertz Foundation
(predoctoral fellowship to D.C.B), Novartis (predoctoral fellowship
to J.L.S.), Merck, Pfizer, and Lilly for financial support.
With the A and C rings joined, we could begin to
investigate the 6-exo cyclization by exposing allylic alcohol 12
to TFA at reflux (Scheme 3). We anticipated that loss of
Supporting information for this article (including information on
the X-ray structures) is available on the WWW under http://
www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2007, 46, 4077 –4080
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4077

Communications
isolated aliphatic CH3 signals, indicating
that the reaction generated a product
containing the two desired quaternary
centers. However, the spectrum also con-
tains an olefinic resonance. Upon stand-
ing in CDCl3, the major product (15)
formed crystals suitable for X-ray diffrac-
tion. Interestingly, cyclization of allylic
alcohol 12 had occurred, but by 6-endo SN’
cyclization to give acid 15.[11,12] Addition-
ally, the solid-state structure confirmed
the anti disposition of the methyl groups
at C12 and C22 in 15.
The SN’ Friedel–Crafts reaction to
produce carboxylic acid 15 achieved the
important goal of generating the quater-
nary stereocenter at C12 with the desired
relative configuration. To better under-
stand the reaction pathway, we evaluated
a number of parameters. The choice of
acid in the reaction is crucial, as trifluoro-
acetic acid was unique in promoting SN’
cyclization. Both stronger acids (e.g.,
triflic acid) and weaker acids (e.g., acetic
acid) failed to produce tricycle 15. Even
the dilution of neat TFA with methylene
chloride, benzene, or acetic acid caused
the cyclization to fail.
Scheme 1. Retrosynthetic analysis of zoanthenol (1). Boc=tert-butyloxycarbonyl, TBS=tert-
butyldimethylsilyl.
Interestingly, both lactone 16 and
allylic acetate 17 underwent cyclization in TFA to give
acid 15 with similar yields and diastereoselectivities
(Figure 2).[13] Furthermore, C16 desoxy arene 18 failed
to generate any cyclized products, confirming the
importance of the nucleophilicity imparted by oxygen-
ation at C16. Finally, the C20 epimers of the substrates
do not readily undergo cyclization. The unique ability of
TFA to mediate the reaction suggests that its properties
as a strong acid and a dehydrating agent are important
to the reaction mechanism. The selectivity of the system
indicates that all three substrates (12, 16, and 17) may
proceed through an intermediate lactone (i.e., allylic
Scheme 2. Preparation of the cyclization substrate. c-Hex=cyclohexyl,
DMA=-N,N-dimethylacetamide, HMDS=hexamethyldisilazane, HMPA=
hexamethyl phosphoramide, Tf=trifluromethanesulfonyl, TES=triethylsilyl.
protecting groups and olefin migration would afford enone
13, which would undergo 6-exo conjugate addition to form
keto-alcohol 14. To our delight, the 1H NMR spectrum of the
ꢀ
major product contains a single aromatic C H as well as two
Figure 2. Other cyclization substrates.
4078 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 4077 –4080

Angewandte
Chemie
23 in 73% yield. Treatment of ketoester 23 with p-toluene-
sulfonic acid produced diketone 3, which was characterized by
X-ray crystallography. The solid-state structure confirmed the
desired stereochemistry at C21 from the hydride shift.
Although racemic material was useful for exploratory
studies, our goal from the outset was an asymmetric synthesis
of zoanthenol. Toward this end, we were delighted to find that
our recently developed asymmetric decarboxylative allylation
methodology[6] was a reliable and efficient method to produce
a-quaternary ketone (ꢀ)-24 in excellent yield and high
enantioselectivity on 25-mmol scale (Scheme 5). Oxidative
olefin cleavage and esterification gave tert-butyl ester (+)-25
in 51% yield over two steps. Subsequent methylation
provided a good yield of methyl ketoester 8, an intermediate
in our C-ring synthesis (vide supra), allowing entry into a
catalytic enantioselective synthesis of zoanthenol.
In conclusion, a concise method for the construction of the
zoanthenol carbocyclic skeleton was developed. The synthesis
is highlighted by an unusual diastereoselective SN’ cyclization
of allylic alcohol 12 producing tricycle 15 bearing all-carbon-
substituted quaternary centers at C12 and C22 in the desired
anti configuration. This key step is flanked in our route by a
number of interesting observations and discoveries. Most
notably, we demonstrate an unusual palladium-catalyzed
formylation of a hindered vinyl triflate, a highly diastereose-
lective Grignard addition to a congested enal, and the
incorporation of the C20 ketone by use of the pendant C24
carboxylate by means of iodolactonization and subsequent
Scheme 3. 6-endo SN’ cyclization of the B ring. TBAF=tetrabutylammo-
nium fluoride.
alcohol 12 and acetate 17 may be converted to the lactone in
situ), and that the reactions proceed by a partially concerted
displacement relying on the directing ability of a carboxylate
leaving group and not via a full allylic cation.
With an efficient route in hand to construct a
zoanthenol carbocyclic ring system containing two
of the three quaternary stereocenters, we turned
our attention to the completion of our proposed
intermediate 3. Following diazomethane-mediated
esterification, deoxygenation of the C16 phenol
was accomplished by formation of aryl triflate 19
and subsequent treatment with [PdCl2(PPh3)2] and
formic acid (Scheme 4).[14]
Owing to our serendipitous discovery of the SN’
reaction, we had not anticipated the reoxygenation
of the olefin in our retrosynthetic planning. As
such, significant experimentation was required to
find a synthetic strategy to convert the D20,21 double
bond of ketoester 20 into the desired C20 ketone.[15]
The X-ray structure in Scheme 3 illustrates the
pseudo-axial orientation of the methyl groups
surrounding the olefin, which partially block the p
bond and hinder the approach of typical oxidants.
Thus, we chose to pursue an alternative, intra-
molecular method of olefin oxygenation. Our
approach began with saponification of ketoester
20 followed by ketalization (Scheme 4). Treatment
of the crude product with KI, I2, and base gave
iodolactone 21 in 85% yield over three steps after
recrystallization. Lactone methanolysis under basic
conditions afforded smooth conversion to epoxide
22. Hydride migration from C20 was accomplished
by heating epoxide 22 in toluene with MgCl2,[16]
Scheme 4. Functionalization of the zoanthenol tricyclic core. DPPP=propane-1,3-
diylbis(diphenylphosphine), Tol=toluene, Ts=toluenesulfonyl.
providing clean conversion to rearranged ketoester
Angew. Chem. Int. Ed. 2007, 46, 4077 –4080
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4079

Communications
Yazawa, Symposium Papers, 38th Symposium on the Chemistry
of Natural Products, 1996, 79 – 84; b) K. Yamaguchi, M. Yada, T.
Tsuji, M. Kuramoto, D. Uemura, Biol. Pharm. Bull. 1999, 22,
920 – 924.
[5] A. H. Daranas, J. J. Fernµndez, J. A. Gavín, M. Norte, Tetrahe-
dron 1999, 55, 5539 – 5546.
[6] a) D. C. Behenna, B. M. Stoltz, J. Am. Chem. Soc. 2004, 126,
15044 – 15045; b) J. T. Mohr, D. C. Behenna, A. M. Harned,
B. M. Stoltz, Angew. Chem. 2005, 117, 7084 – 7087; Angew.
Chem. Int. Ed. 2005, 44, 6924 – 6927; c) R. M. McFadden, B. M.
Stoltz, J. Am. Chem. Soc. 2006, 128, 7738 – 7739.
[7] a) N. Hikage, H. Furukawa, K. Takao, S. Kobayashi, Tetrahedron
Lett. 1998, 39, 6237 – 6240; b) N. Hikage, H. Furukawa, K. Takao,
S. Kobayashi, Tetrahedron Lett. 1998, 39, 6241 – 6244; c) D. R.
Williams, G. A. Cortez, Tetrahedron Lett. 1998, 39, 2675 – 2678.
[8] a) T. Matsumoto, T. Ohmura, Chem. Lett. 1977, 335 – 338;
b) L. S. Liebeskind, R. Chidambaram, S. Nimkar, D. Liotta,
Tetrahedron Lett. 1990, 31, 3723 – 3726; c) C. Schmidt, J.
Thazhuthaveetil, Can. J. Chem. 1973, 51, 3620 – 3624; d) T.
Matsumoto, S. Usui, T. Morimoto, Bull. Chem. Soc. Jpn. 1977,
50, 1575 – 1579; e) T. Matsumoto, S. Usui, Chem. Lett. 1978, 897 –
900; f) R. V. Stevens, G. S. Bisacchi, J. Org. Chem. 1982, 47,
2396 – 2399.
[9] a) M. Franck-Neumann, M. Miesch, F. Barth, Tetrahedron 1993,
49, 1409 – 1420; b) In our experience, dimethyl ketone 10 was
best prepared by oxidation of 2,6-dimethylphenol with iodoben-
zene diacetate in ethylene glycol to give a quinone monoketal
product, which was readily reduced to the ketone with Pd/C
under hydrogen (60 psi).
Scheme 5. Catalytic asymmetric alkylation for the enantioselective syn-
thesis of ketoester (+)-8. dba=dibenzylideneacetone, DMAP=4-dime-
thylaminopyridine.
epoxide rearrangement. Finally, application of our catalytic
asymmetric decarboxylative alkylation methodology allows
ready access to an enantioselective synthesis of zoanthenol.
Our ongoing efforts to advance an enantioenriched tricyclic
core to zoanthenol will be reported in due course.
Received: January 31, 2007
Published online: April 19, 2007
[10] See the Supporting Information for details.
[11] For an excellent review of SN’ reactions, see: L. A. Paquette,
C. J. M. Stirling, Tetrahedron 1992, 48, 7383 – 7423.
Keywords: alkaloids · asymmetric catalysis · cyclization ·
natural products
[12] For related cyclizations, see: S. Ma, J. Zhang, Tetrahedron Lett.
2002, 43, 3435 – 3438.
[13] The analogous trifluoroacetate underwent cyclization in TFA to
give yields and diastereoselectivities comparable to those of the
reaction of allylic alcohol 12.
[14] These conditions were specifically designed for highly congested
aryl triflates, see: J. M. Saµ, M. Dopico, G. Martorell, A. G.
García-Raso, J. Org. Chem. 1990, 55, 991 – 995.
[15] Typical epoxidation conditions such as m-CPBA, dimethyl
dioxirane, urea hydroperoxide, iron(III) acetylacetate and
hydrogen peroxide, hexafluoroacetone and hydrogen peroxide,
potassium permanganate and copper(II) sulfate, and methyl-
trioxorhenium and hydrogen peroxide gave no reaction or
oxidation at the benzylic C19 position. Hydroboration was also
unsuccessful.
.
[1] For the most recent publication from each group, see: a) M. Juhl,
T. E. Nelson, S. LeQuement, D. Tanner, J. Org. Chem. 2006, 71,
265 – 280; b) T. Irifune, T. Ohashi, T. Ichino, E. Sakia, K.
Suenaga, D. Uemura, Chem. Lett. 2005, 34, 1058 – 1059; c) D. R.
Williams, T. A. Brugel, Org. Lett. 2000, 2, 1023 – 1026; d) F.
Rivas, S. Ghosh, E. A. Theodorakis, Tetrahedron Lett. 2005, 46,
5281 – 5284; e) N. Hikage, H. Furukawa, K. Takao, S. Kobayashi,
Chem. Pharm. Bull. 2000, 48, 1370 – 1372; f) S. M. Moharram, H.
Oguri, M. Hirama, Egypt. J. Pharm. Sci. 2003, 44, 177 – 193.
[2] D. Tanner, P. G. Anderson, L. Tedenborg, P. Somfai, Tetrahedron
1994, 50, 9135 – 9144.
[16] a) Treatment with BF3·Et2O gave the desired hydride shift with
concomitant deketalization; however, the yields were lower than
the two-step protocol; b) S. M. Naqvi, J. P. Horwitz, R. Filler, J.
Am. Chem. Soc. 1957, 79, 6283 – 6286.
[3] M. Miyashita, M. Sasaki, I. Hattori, M. Sakai, K. Tanino, Science
2004, 305, 495 – 499.
[4] a) M. Kuramoto, H. Arimoto, K. Hayashi, I. Hayakawa, D.
Uemura, T. Chou, K. Yamada, T. Tsuji, K. Yamaguchi, K.
4080 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 4077 –4080