Synthesis of the schweinfurthin hexahydroxanthene core through Shi epoxidation

Article abstract of DOI:10.1016/j.tetlet.2007.11.086

Use of a Shi epoxidation for introduction of chirality in a key epoxide intermediate, together with revised protecting group tactics, has allowed an efficient synthesis of the hexahydroxanthene subunit common to the natural schweinfurthin F and the synthe

Full text of DOI:10.1016/j.tetlet.2007.11.086

Available online at www.sciencedirect.com
Tetrahedron Letters 49 (2008) 516–519
Synthesis of the schweinfurthin hexahydroxanthene core
through Shi epoxidation
Jeffrey D. Neighbors, Nolan R. Mente, Kelly D. Boss,
*
Donald W. Zehnder, II, David F. Wiemer
Departments of Chemistry and Pharmacology, University of Iowa, Iowa City, IA 52242-1294, USA
Received 10 October 2007; revised 13 November 2007; accepted 14 November 2007
Available online 21 November 2007
Abstract
Use of a Shi epoxidation for introduction of chirality in a key epoxide intermediate, together with revised protecting group tactics, has
allowed an efficient synthesis of the hexahydroxanthene subunit common to the natural schweinfurthin F and the synthetic analogue
3-deoxyschweinfurthin B.
Ó 2007 Elsevier Ltd. All rights reserved.
OR
The schweinfurthins are a small group of natural prod-
ucts that currently includes eight examples.1–3 Most of
these compounds contain a hexahydroxanthene substruc-
ture (e.g., schweinfurthins A, B, E, F, and G, 1–5, Fig. 1)
with a hydroxyl group at C-2 that suggests biosynthesis
through a cascade cyclization. This family of natural prod-
ucts has attracted some attention due to a unique antican-
cer profile displayed in the National Cancer Institute’s 60
cell line bioassay. In this panel several schweinfurthins, as
well as the related compound vedelianin (6), display mean
cytotoxicity values <1 lM (GI50). More importantly, the
pattern of activity they display across the cell lines is not
clearly related to any currently used agent.1,4,5 Indeed some
schweinfurthins show 10 nM activity against CNS, renal,
and prostate cancer cell lines, while showing up to four
orders of magnitude less activity against cell lines of lung
or ovarian tissue origin. Potent and unique activity in this
assay system could indicate a novel mechanism of action,
which would make the schweinfurthins a valuable lead
for a new type of chemotherapeutic agent.
HO
HO
O
3
2
OH
H
1 R = H
Schweinfurthin A
2 R = CH3 Schweinfurthin B
OH
OR
R'
O
OH
HO
H
3 R = CH3 R' = OH Schweinfurthin E
4 R = CH3 R' = H Schweinfurthin F
OH
5 R =
6 R =
H
H
R' = H Schweinfurthin G
R' = OH Vedelianin
Fig. 1. Hexahydroxanthenes of the schweinfurthin family.
have undertaken an effort aimed at their total synthesis.
This effort has led to the preparation of both enantiomers
of schweinfurthin F (4), which allowed assignment of the
natural absolute configuration, and the synthesis of
approximately 20 analogues that have been tested in the
60 cell screen.6–9 Studies of the natural and synthetic
schweinfurthins10 have illuminated the importance of the
right-half resorcinol substructure in the biological activity
of this family of compounds, but also confirm the impor-
tance of the hexahydroxanthene system.
For the reasons outlined above, as well as the difficulty
of obtaining schweinfurthins from the natural sources, we
*
Corresponding author. Tel.: +1 319 335 1365; fax: +1 319 335 1270.
E-mail address: david-wiemer@uiowa.edu (D.F. Wiemer).
0040-4039/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2007.11.086

J. D. Neighbors et al. / Tetrahedron Letters 49 (2008) 516–519
517
OCH
3
OCH3
To date, synthetic efforts have centered on the C-2
HO
Br
TBSO
8 steps, 52%
hydroxy compounds, including the natural products
schweinfurthin F (4) and G (5), as well as the synthetic ana-
logue 3-deoxyschweinfurthin B (3dSB, 10). The synthetic
plan relies upon a penultimate Horner–Wadsworth–Emmons
(HWE) condensation to set the central stilbene olefin and
requires two suitable coupling partners as summarized in
Figure 2. For example, the coupling of aldehyde 7 represent-
ing the left-half of schweinfurthin F and 3dSB with the
right-half phosphonates 8 and 9 afforded reasonable access
to the stilbenes 4 and 10. As a part of our ongoing efforts at
further development of these agents it became necessary to
prepare gram scale quantities of our lead compounds.
Here, we report a new synthesis of the hexahydroxanthene
7 that exploits Shi epoxidation to obtain nonracemic inter-
mediates in both improved yield and high ee. This new
approach should facilitate the preparation of substantial
amounts of the materials of interest.
OR
H
HO
OTBS
O
11
12 (65 to 80% ee)
4 steps
46%
4 steps
41%
OCH3
OCH3
TBSO
O
O
H
HO
H
OTBS
O
7
13 (100% ee)
Scheme 1. First generation synthesis of hexahydroxanthene 7.
To explore these options (Scheme 2), a straightforward
reduction of bromovanillin (11) was followed by the treat-
ment of the known benzylic alcohol 14 with MOMCl and
Hunig’s base to afford the bismethoxymethyl ether 15.
¨
Our original synthesis of aldehyde 7 used a biomimetic
cascade cyclization to construct the hexahydroxanthene
ring system from an aryl epoxide precursor (Scheme 1).7
This approach builds upon extensive historical precedent
for faithful transmission of stereochemical information
through the cationic cascade manifold.7 After the introduc-
tion of the epoxide stereocenter via asymmetric dihydroxyl-
ation, the cyclization reaction afforded a single hexa-
hydroxanthene enantiomer. The geranyl precursor was
derived from bromovanillin (11) that already bears the
required methyl ether common to several targets. The first
generation route to aldehyde 7 required 16 steps and
proceeded in an overall yield of ꢀ10% from vanillin. The
goals for this synthesis were to preserve the central strategy
while more efficient tactics for the construction of non-
racemic cyclization precursors such as epoxide 13 were
explored.
As a first effort toward these goals, the use of different
protecting groups in the initial stages of the synthetic route
was explored. For example, the use of methoxymethyl
(MOM) ether protection could save numerous steps if the
protecting groups could be carried through the sequence
without removal until a very late stage, where the cascade
cyclization conditions might be employed for concurrent
deprotection.
Treatment with n-butyl lithium followed by geranyl bro-
mide to trap the intermediate anion gave the geranylated
intermediate 16 in excellent yield.
From this intermediate the route is parallel to that pre-
viously disclosed. Treatment of the arene 16 with an asym-
metric dihydroxylation reagent gave the diol 17 in good ee.
The use of the conditions already developed to resolve diol
17 via mandelate ester 18, followed by formation of epox-
ide 19 through the mesylate, gave material of high ee and
set the stage for the exploration of a cascade cyclization
in this system. Treatment of epoxide 19 with TFA under
optimized conditions afforded tricycle 20, with the benzylic
MOM protecting group still in place in moderate yields.
This result was disappointing because it was anticipated
that both protecting groups could be removed by the
acid-catalyzed cyclization conditions, thus making a final
deprotection unnecessary. The benzylic MOM group
proved intransigent to several other deprotection condi-
tions as well.11 However, after some experimentation we
found that treatment with 48% HBr in dichloromethane
would afford a modest yield of the desired diol 21. Oxida-
tion of diol 21 was quantitative.
At this point, examination of silyl ether protection at the
benzylic position became attractive, because this group
could be removed in high yield. The known arene 22 (avail-
able from bromovanillin)8 was protected as the benzylic
t-butyldimethylsilyl ether to give bromide 23, which was
subjected to a series of reactions to afford the desired tricycle
28 via intermediates 24–27. While these steps proceeded
in yields comparable to those observed with the corres-
ponding MOM compounds, silyl deprotection of com-
pound 28 to afford the diol 21 was virtually quantitative.
Thus, after a final oxidation this reaction sequence gave
the aldehyde 7 in 13 steps and 9% overall yield from
vanillin.
OCH3
P(O)(OEt)2
O
OR
+
H
HO
HO
H
n
H
O
OR
7
8 n = 1
9 n = 2
OCH3
O
OH
H
With more efficient protecting group tactics in place a
more fundamental series of changes was explored. The lin-
ear nature of the sequences used to introduce the epoxide
stereocenter of compounds 19 and 27 is cumbersome, and
H
n
4
n = 1
10 n = 2
OH
Fig. 2. Retrosynthetic analysis of the schweinfurthins.

518
J. D. Neighbors et al. / Tetrahedron Letters 49 (2008) 516–519
OCH3
OCH3
n-BuLi, -78 o
geranyl bromide
C
OCH3
OCH3
DIPEA,
MOMCl
RO
Br
MOMO
MOMO
MOMO
Br
OR'
AD-mix-α
or
TBSCl
imid
HO
OH
OR
OR
OR
14 R = H
22 R = MOM
17 R = MOM, R' = H (63%, 80% ee)
25 R = TBS, R' = H (74%, 87% ee)
16 R = MOM (99%)
24 R = TBS (78%)
15 R = MOM (95%)
23 R = TBS (99%)
RCO2H, EDC
(68%)
RCO2H, EDC
(75%)
17
NaOH, EtOH
(97%)
18 R = MOM, R' =
25
26 R = TBS, R' =
S
-O-methylmandelate
O
O
O
NaOH, EtOH
(91%)
S-O-methylmandelate
O
H2O2, -5 o
C
O
CH3CN\CH2Cl2\EtOH (1:2:1)
buffer
O
1) MsCl, TEA
29
2) K2CO3, MeOH
OCH3
O
OCH3
O
OCH3
HBr
(58% from 20)
MOMO
MnO2
quant.
TFA, 0 o
C
O
7
or
TBAF
(99% from 28)
HO
HO
H
H
OH
OR
OR
21
20 R = MOM (41%)
28 R = TBS (35%)
19 R = MOM (82% from 17; 40%, 100% ee from16)
27 R = TBS (71% from 25; 48%, 98% ee from 24)
Scheme 2. Protecting group tactics and epoxidations used in synthesis of the hexahydroxanthene intermediate 21.
both a traditional resolution and a chromatographic sepa-
ration of the diastereomers proved necessary to achieve
high ee. If direct epoxidation could be accomplished in
good yield and high ee, the synthesis would be significantly
more efficient. To explore this possibility, the oxidation of
compounds 16 and 24 via the nonracemic sugar catalyzed
process pioneered by Shi et al.12 was explored. Because
enantioselective epoxidation on aromatic esters of isopre-
noids has been successful,12 it was reasonable to consider
epoxidation of these isoprenylated aromatic systems. Oxi-
dation of diene 16 proceeded in just 40% yield, along with
significant amounts of recovered starting material (38%)
and a trace of the regioisomeric epoxide, but did afford
the desired material in high ee (ꢀ100% by comparison of
rotation with authentic material7). Oxidation of diene 24
also proceeded in very good ee (ꢀ98%) and somewhat
better yield (48%).
If epoxidation could be removed from the primary syn-
thetic sequence, a more convergent and efficient approach
might result. In particular, the formation of the epoxide
prior to the attachment of the terpenoid chain to the arene
might improve the regio- and stereoselectivity of the olefin
epoxidation. Geraniol (6,7)-epoxide (32) has been synthe-
sized as either enantiomer by both a salen Mn(III) catalyst
method13 and via the Shi epoxidation.12 Similarly, both R-
and S-(6,7)-epoxygeranyl bromide (33) have been synthe-
sized.14,15 With the availability of these reagents, a reaction
sequence in which an aryl nucleophile would selectively
attack the allylic bromide of epoxygeranyl bromide was
of clear interest.
A recent report by Gansaeuer et al.16 on the copper-
mediated displacement of the bromide of epoxy geranyl
bromide (33) appeared to indicate this should be possible
using an organometallic nucleophile derived from the aryl
bromide 23. Toward this end, the known p-nitrobenzoate
ester 3017 (Scheme 3) was treated with hydrogen peroxide
in the presence of carbohydrate catalyst 29, to afford the
epoxyester 31 in 80–88% yield and high ee (ꢀ93% ee by
HPLC analysis).18 A relatively low reaction temperature
(ca. À15 °C) and longer reaction times appeared to be
advantageous in this reaction. Under these conditions less
than 5% of the starting material was recovered along with
a trace of the diepoxide, and the regioisomeric mono
epoxide could not be isolated. Hydrolysis of the resulting
ester through reaction with base gave R-(6,7)-epoxygeraniol
(32) which was found to have an enantiomeric excess of
ꢀ100% based on the comparison of the optical rotation
to the literature value15 ([a] +8.5 (c 0.01, CH3OH) vs
H2O2, 29, CH3CN
NO2
NO2
buffer
O
8 hrs -15 o
C
O
O
O
O
31 (88%, 93% ee)
30
NaOMe
Bu4NI
87%
MeOH
1) MsCl, TEA
2) LiBr, THF
O
O
OH
Br
32
33
Scheme 3. Synthesis of R-(6,7)-epoxygeranyl bromide (33).

J. D. Neighbors et al. / Tetrahedron Letters 49 (2008) 516–519
519
1)
n
-BuLi, -78 o
C
C
Supplementary data
2) CuCN, -65 o
OCH3
O
Br
OCH3
MOMO
Experimental procedures and/or spectral data for com-
pounds 15–20, 23–28, and 31–33 are available. Supplemen-
tary data associated with this article can be found, in the
online version, at doi:10.1016/j.tetlet.2007.11.086.
33
MOMO
Br
O
67%
OTBS
OTBS
27
23
OCH3
References and notes
O
4
schweinfurthin F
10 3-deoxyschweinfurthin B
1. Beutler, J. A.; Shoemaker, R. H.; Johnson, T.; Boyd, M. R. J. Nat.
Prod. 1998, 61, 1509–1512.
HO
H
21
OH
2. Beutler, J. A.; Jato, J.; Cragg, G. M.; Boyd, M. R. Nat. Prod. Lett.
2000, 14, 399–404.
Scheme 4. Coupling of metallated arene 23 and epoxybromide 33.
3. Yoder, B. J.; Cao, S. G.; Norris, A.; Miller, J. S.; Ratovoson, F.;
Razafitsalama, J.; Andriantsiferana, R.; Rasamison, V. E.; Kingston,
D. G. I. J. Nat. Prod. 2007, 70, 342–346.
4. Paull, K. D.; Shoemaker, R. H.; Hodes, L.; Monks, A.; Scudiero, D.
A.; Rubinstein, L.; Plowman, J.; Boyd, M. R. J. Natl. Cancer Inst.
1989, 81, 1088–1092.
literature +8.5 (c 0.01, CH3OH)). Treatment of alcohol
32 with methanesulfonyl chloride followed by lithium
bromide gave the known epoxybromide 33.
To our great delight subjection of the aryl bromide 23 to
halogen metal exchange conditions followed by transmetal-
lation with copper cyanide or copper bromide–dimethyl
sulfide complex, gave upon reaction with epoxybromide
33 the desired product 27 in 67% yield (Scheme 4). Epoxide
27 was converted to the aldehyde 7 through the sequence
shown in Scheme 2, and then converted to 3-deoxyschwein-
furthin B (10) via known methods.7 Analysis of the enan-
tiomeric excess of the final product by HPLC18 indicated
material of 94% ee confirming the stereocontrol of the cas-
cade process. This represents the removal of eight steps
from the first generation sequence leading to tricyclic left-
half diol 21, leaving the complete route from vanillin to
aldehyde 7 with a total of just eight steps in the longest
linear sequence and 23% overall yield.
Use of a Shi epoxidation and new protecting group tac-
tics in this second generation synthesis of the hexahydro-
xanthene substructure of schweinfurthin F and 3dSB has
dramatically improved access to this series of agents. This
should allow synthesis of sufficient quantities of 3dSB to
initiate in vivo investigations in mice. These efforts will
be the subject of future communications in this area.
5. Paull, K. D. Prediction of biochemical mechanism of action from the
in vitro antitumor screen of the National Cancer Institute. In Cancer
Chemotherapeutic Agents; Foye, W. O., Ed.; American Chemical
Society: Washington, DC, 1995; pp 9–45.
6. Neighbors, J. D.; Salnikova, M. S.; Beutler, J. A.; Wiemer, D. F.
Biorg. Med. Chem. 2006, 14, 1771–1784.
7. Neighbors, J. D.; Beutler, J. A.; Wiemer, D. F. J. Org. Chem. 2005,
70, 925–931.
8. Treadwell, E. M.; Neighbors, J. D.; Wiemer, D. F. Org. Lett. 2002, 4,
3639–3642.
9. Mente, N. R.; Wiemer, A. J.; Neighbors, J. D.; Beutler, J. A.;
Hohl, R. J.; Wiemer, D. F. Biorg. Med. Chem. Lett. 2007, 17,
911–915.
10. Wiemer, D. F.; Neighbors, J. D.; Beutler, J. A.; Preparation of
Schweinfurthin analogues, WO 2005092878, 2005, University of Iowa
Research Foundation, USA; Government of the United States of
America as Represented by the Secretary of the Department of Health
and Human Services: USA. p 76.
11. Greene, T. W.; Wuts, P. G. M., 3rd ed. In Protective Groups in
Organic Synthesis; Wiley: New York, 1999; Vol. xxi, 779 p.
12. (a) Wang, Z. X.; Shi, Y. J. Org. Chem. 1998, 63, 3099–3104; (b) Shu,
L.; Shi, Y. Tetrahedron Lett. 1999, 40, 8721–8724.
13. Garcia, M. A.; Meou, A.; Brun, P. Synlett 1996, 1049–1050.
14. Corey, E. J.; Noe, M. C.; Shieh, W. C. Tetrahedron Lett. 1993, 34,
5995–5998.
15. Meier, H.; Uebelhart, P.; Eugster, C. H. Helv. Chim. Acta 1986, 69,
106–123.
16. Gansaeuer, A.; Justicia, J.; Rosales, A.; Rinker, B. Synlett 2005,
1954–1956.
Acknowledgments
17. Meou, A.; Garcia, M. A.; Brun, P. J. Mol. Cat. Chem. 1999, 138,
221–226.
Financial support from the Roy J. Carver Charitable
Trust, the Children’s Tumor Fund, the Breast Cancer
Research Program (DAMD17-01-1-0276 and DAMD17-
02-1-0423), an Oncology Research Training Award from
the Holden Comprehensive Cancer Center’s Institutional
National Research Service Award (2 T32 CA79445) and
the Predoctoral Training Program in the Pharmacological
Sciences (2 T32 GM067795), is gratefully acknowledged.
18. The HPLC analyses were performed on a Shimadzu LC-20AT
instrument with a Chiralcel OD-H column. The enantiomeric excess
of the epoxidation was determined by elution with a 99:1 mixture of
hexanes and 2-propanol. Retention times for the enantiomers of
epoxide 31 were 22.6 min (3.5%) and 24.7 min (96.5%), respectively.
For compound 10, the analysis was conducted with a solvent mixture
of 84:16 hexanes and 2-propanol, and retention times of 48.9 min
(2.9%) and 63.3 min (97.1%) were observed.