A cascade cyclization approach to schweinfurthin B

Article abstract of DOI:10.1021/ol0266368

(graph presented) A strategy for synthesis of the hexahydroxanthene moiety of the natural products schweinfurthin A, B, and D is described. The relative stereochemistry in the key cationic cyclization step is established through the preference of the phen

Full text of DOI:10.1021/ol0266368

ORGANIC
LETTERS
2002
Vol. 4, No. 21
3639-3642
A Cascade Cyclization Approach to
Schweinfurthin B
Edward M. Treadwell, Jeffrey D. Neighbors, and David F. Wiemer*
Department of Chemistry, UniVersity of Iowa, Iowa City, Iowa 52242
daVid-wiemer@uiowa.edu
Received July 30, 2002
ABSTRACT
A strategy for synthesis of the hexahydroxanthene moiety of the natural products schweinfurthin A, B, and D is described. The relative
stereochemistry in the key cationic cyclization step is established through the preference of the phenylselenide substituent for an equatorial
orientation.
The schweinfurthins (Figure 1, 1-4) are a small set of doubly
prenylated stilbenes isolated from the African plant Macar-
display significant activity in the NCI’s 60-cell line anticancer
assay with GI50 values less than 0.5 µM.1,2 Their profile of
activity does not match that of any clinically used anticancer
agent, which suggests that these compounds may act either
by a novel mechanism or at an unknown site. The schwein-
furthins have been isolated in low and varying amounts from
the natural source, and their absolute stereochemistry has
yet to be elucidated. For these reasons, as well as their
interesting biological activity, we have undertaken a total
synthesis that ultimately should allow assignment of the
schweinfurthins’ absolute stereochemistry and provide a
reliable source for further biological testing.
We have demonstrated the feasibility of a conver-
gent approach to the schweinfurthins through synthesis of
schweinfurthin C (3), the inactive congener.3 In that syn-
thesis, the central stilbene olefin was prepared by a Horner-
Wadsworth-Emmons condensation of a benzylic phospho-
nate (compound 5) and a complementary aldehyde. The
phosphonate was prepared in eight steps from commercially
available 3,5-dihydroxybenzoic acid (6) employing a directed
ortho metalation for introduction of the geranyl substituent.
Phosphonate 5 also could be used to advantage in preparation
of the more complex schweinfurthins, provided preparation
Figure 1. Structures of the schweinfurthins.
(1) Beutler, J. A.; Shoemaker, R. H.; Johnson, T.; Boyd, M. R. J. Nat.
Prod. 1998, 61, 1509-1512.
(2) Beutler, J. A.; Jato, J.; Cragg, G. M.; Boyd, M. R. Nat. Prod. Lett.
2000, 14, 399-404.
anga schweinfurthii Pax. by Beutler et al. at the National
Cancer Institute. Schweinfurthins A (1), B (2), and D (4)
(3) Treadwell, E. M.; Cermak, S. C.; Wiemer, D. F. J. Org. Chem. 1999,
64, 8718-8723 and references therein.
10.1021/ol0266368 CCC: $22.00 © 2002 American Chemical Society
Published on Web 09/19/2002

Scheme 1. Retrosynthetic Analysis of Schweinfurthin B
Figure 2. Possible transition states for cyclization of hydroxy-
selenide 8.
application to enantiopure material was not attempted.7 With
this aim in mind, racemic â-hydroxyselenide 8 was viewed
as a cyclization precursor that would allow evaluation of the
viability of such an approach.
The synthesis began with preparation of the known
benzaldehyde derivative 108 (Scheme 2) from commercially
Scheme 2. Initial Synthesis of Hydroxyselenide 16
of a tricyclic aldehyde (7, Scheme 1) could be achieved. The
methylated version of this tricyclic aldehyde was targeted
initially because the requisite phenolic methyl ether could
be carried along the sequence from the aromatic starting
material, bromovanillin 9.
One approach to the hexahydroxanthene core could be
based on an acid-catalyzed cyclization to assemble both the
A- and B-rings on an aromatic C-ring in a single reaction.
Previous reports on cyclizations of geranylated phenols are
known, but often the cyclizations occurred in low yield with
numerous byproducts observed.4,5 We hypothesized that a
substituent R to the incipient carbocation could help stabilize
the terminal cation, thereby possibly increasing the yield and
providing an opportunity for stereocontrol. There is substan-
tial precedent for stabilization of adjacent cations by phen-
ylthio substituents, and some precedent for stabilization by
phenylselenyl groups.6 As shown in Figure 2, one transition
state would place the phenylselenide substituent in an
equatorial position with a pseudochair conformation in the
incipient B-ring, while the other would require an axial
phenylselenide group with a pseudoboat conformation. The
use of hydroxyselenides for similar reactions has been
described in two seminal papers by Kametani et al., though
available vanillin. Reduction of the aldehyde and subsequent
protection of the alcohol as the triethylsilyl ether afforded
3640
Org. Lett., Vol. 4, No. 21, 2002

Scheme 3. Revised Synthesis of Hydroxyselenide 16
the fully protected arene 12, and halogen-metal exchange
A second synthetic strategy was developed to address this
problematic deprotection issue. Because the silyl ether could
be readily removed, it appeared attractive to protect the
phenolic functionality as a silyl ether as well. However,
introduction of the phenolic silyl ether would have to follow
the alkylation step in the synthetic sequence, because
migration of the silyl group from the oxygen to the adjacent
ortho carbon has been observed in similar reactions.10
Therefore, an ethoxyethyl-protected phenol was envisioned
for the sequence up to and including the alkylation step, at
which point it would be removed and a silyl ether installed
in its place.11
Direct protection of the phenol as the ethoxyethyl ether
was not successful under acidic conditions, so an indirect
route was employed. The known alcohol 17,12 also available
from vanillin, was disilylated and then selectively cleaved
to the free phenol 19 by treatment with 1.0 equiv of
tetrabutylammonium fluoride13 (Scheme 3). An acid-
catalyzed reaction of compound 19 with ethyl vinyl ether
gave the fully protected aryl bromide 20. This intermediate
can be prepared in multigram quantities in an overall yield
of 68% from vanillin without need for a chromatographic
separation. Application of the halogen-metal exchange
protocol and reaction with geranyl bromide afforded the
analogous geranylated arene, which upon acidic workup gave
the free phenol 21. After silylation of the free phenol, the
material was subjected to oxidation, and epoxide opening
analogous to that used on arene 13 delivered the protected
R-hydroxyselenide 24. The deprotected target 16 could be
obtained in 84% yield by treatment of the disilylated material
with excess TBAF.
followed by reaction with geranyl bromide allowed instal-
lation of the geranyl chain in 74% yield. An mCPBA
epoxidation of compound 13 initially afforded a 1:1 mixture
of the regioisomeric 6,7- and 2,3-epoxides in 55% yield along
with the diepoxide (7%). Even though careful column
chromatography could separate the two regioisomers, the low
yield of the desired product was unattractive. When the
reaction was conducted at lower temperatures with slow
addition of the oxidant, the yield of the desired 6,7-epoxide
14 increased to 53% along with only 8% of the 2,3-epoxide
and significant recovery of the starting material (32%).
Epoxide 14 reacted smoothly with phenylselenide anion
generated in situ9 to give the hydroxyselenide 15 in 83%
yield.
The only transformations remaining prior to cyclization
were removal of the two protecting groups, but in the best
case scenario this was done through a two-step procedure.
Initial treatment with 0.5 M HCl hydrolyzed the silyl ether,
and subsequent treatment with 1.0 M HCl hydrolyzed the
MOM acetal in an overall yield of 33%. Despite numerous
attempts, all efforts at removing both protecting groups in a
single step gave either incomplete deprotection or lower
yields with greater byproduct formation.
(4) (a) Barua, A. K.; Banerjee, S. K.; Basak, A.; Bose, P. K. J. Indian
Chem. Soc. 1976, 53, 638-639. (b) Manners, G.; Jurd, L.; Stevens, K.
Tetrahedron 1972, 28, 2949-2959. (c) Trammell, G. L. Tetrahedron Lett.
1978, 1525-1528.
(5) Mechoulam and Yagen have reported cyclization of geranylolivetol
in 88% yield, but this required heating with concentrated H2SO4 in
nitromethane. Mechoulam, R.; Yagen, B. Tetrahedron Lett. 1969, 5349-
5352.
(6) (a) For a review, cf.: Harring, S. R.; Edstrom, E. D.; Livinghouse,
T. In AdVances in Heterocyclic Natural Product Synthesis; Pearson, H. W.,
Ed.; Jai Press: Greenwich, CT, 1992; Vol 2., pp 299-376. For more recent
examples, cf.: (b) Branchaud, B. P.; Blanchette, H. S. Tetrahedron Lett.
2002, 43, 351-353. (c) Toshimitsu, A.; Hirosawa, C.; Tamao, K. Synlett
1996, 465-467 and references therein.
(7) (a) Kametani, T.; Suzuki, K.; Kurobe, H.; Nemoto, H. J. Chem. Soc.,
Chem. Commun. 1979, 1128-1129. (b) Kametani, T.; Kurobe, H.; Nemoto,
H.; Fukumoto, K. J. Chem. Soc., Perkin Trans. 1 1982, 1085-87.
(8) Boger, D. L.; Jacobson, I. C. J. Org. Chem. 1991, 56, 2115-2122.
(9) Sharpless, K. B.; Lauer, R. F. J. Am. Chem. Soc. 1973, 95, 2697-
2699.
(10) When treated with n-butyllithium, both compound 18 and the TIPS
analogue show a 1,3 O-C silyl migration in the only isolable products.
(11) The EE group was not carried throughout the sequence to avoid
introduction of diastereomers and because the phenolic EE group was readily
cleaved upon silica gel column chromatography.
(12) (a) Brink, M. Acta Chem. Scand. 1965, 19, 255-256. (b) Claus,
P.; Schilling, P.; Gratzl, J. S.; Kratzl, K. Monatsh. Chem. 1972, 103, 1178-
1193.
(13) Collington, E. W.; Finch, H.; Smith, I. J. Tetrahedron Lett. 1985,
26, 681-684.
Org. Lett., Vol. 4, No. 21, 2002
3641

To induce the desired cationic cyclization, the tertiary
alcohol 16 was treated with acid under various conditions.
Treatment of compound 16 with TFA afforded a single
hexahydroxanthene system as the labile trifluoroacetate 25.
Purification of this product by column chromatography gave
both the trifluoroacetate 25 and the parent alcohol 26 in 43%
combined yield.
The relative stereochemistry of the hexahydroxanthene was
assigned after extensive NMR spectroscopy on the trifluoro-
acetate 25. Analysis of the coupling constants observed for
the C-2 hydrogen (schweinfurthin numbering) suggested an
axial disposition and hence an equatorial orientation for the
phenylselenide group. The bridgehead methine hydrogen (C-
9a) also appeared to be in an axial orientation on the basis
of analysis of the coupling constants with the benzylic
hydrogens at C-9. In this case, a COSY spectrum nicely
displayed the H-9ax, H-9eq, H-9a spin system, indicative of
a trans-decalin skeleton. Furthermore, the chemical shifts
of the methyl groups compared favorably to those reported
for a related trans-fused system but did not agree with those
of a related cis-fused structure.14 Finally, a NOESY spectrum
revealed correlations (Figure 3) of the bridgehead methyl
group with axial hydrogens at C-3 and C-9 and to the axial
methyl group at C-1. On the other face of the molecule,
complementary correlations were observed between the
equatorial methyl group at C-1 and the axial hydrogen at
C-9a, as well as from the axial hydrogen at C-2 to both the
C-1 equatorial methyl group and the C-9 equatorial hydrogen.
The NMR data make clear that the phenylselenide sub-
stituent was successful in providing a single diastereomer
of the hexahydroxanthene and may facilitate the cyclization.
The equatorial disposition of the phenylselenide moiety in
the final product is encouraging in that this single substituent
appears to effectively govern the stereochemistry of the
Figure 3. Selected NOESY correlations for compound 25 shown
on a SPARTAN minimized structure (PM3 level).
bridgehead centers, as expected from consideration of the
transition states (Figure 2).
Preparation of the tricycle 26 should allow elaboration of
racemic schweinfurthin B after introduction of the A-ring
hydroxyl groups and coupling with phosphonate 5. Alter-
natively, now that the viability of this cyclization strategy
has been shown, preparation of the epoxide 23 in nonracemic
form should allow preparation of nonracemic schweinfurthin
B (2). Our efforts to prepare the nonracemic epoxide, as well
as to complete preparation of the natural products themselves,
will be reported in due course.
Acknowledgment. Financial support from the DOD
Breast Cancer Research Program (DAMD17-01-1-0276 and
DAMD17-02-1-0423) is gratefully acknowledged.
Supporting Information Available: Experimental pro-
cedures and spectral data for compounds 16-26. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(14) Rouessac, A.; Rouessac, F. Tetrahedron 1987, 37, 4165-4170.
OL0266368
3642
Org. Lett., Vol. 4, No. 21, 2002