Synthesis of (+)-madindoline A and (+)-madindoline B

Article abstract of DOI:10.1021/ol062919e

The allene ether version of the Nazarov cyclization was used to construct the cyclopentane dione portion of madindolines A and B. The racemic cyclopentane dione from the Nazarov cyclization was converted to an enol ether that was combined with the chiral,

Full text of DOI:10.1021/ol062919e

ORGANIC
LETTERS
2007
Vol. 9, No. 4
647-650
Synthesis of (
+)-Madindoline A and
(+
)-Madindoline B
Lifeng Wan and Marcus A. Tius*
Department of Chemistry, 2545 The Mall, UniVersity of Hawaii,
Honolulu, Hawaii 96822, and The Cancer Research Center of Hawaii,
1236 Lauhala Street, Honolulu, Hawaii 96813
tius@gold.chem.hawaii.edu
Received December 2, 2006
ABSTRACT
The allene ether version of the Nazarov cyclization was used to construct the cyclopentane dione portion of madindolines A and B. The
racemic cyclopentane dione from the Nazarov cyclization was converted to an enol ether that was combined with the chiral, nonracemic
hydroxyfuroindoline in a Mannich reaction. Deprotection and oxidation led to (+)-madindoline A and (+)-madindoline B.
In 1996, Ohmura and co-workers isolated two small mol-
ecules, madindolines A and B, of unique structure from the
fermentation broth of Streptomyces nitrosporeus K93-0711.1,2
Both compounds have extraordinary activity and are selective
inhibitors of interleukin-6 (IL-6).3,4 Because the overproduc-
tion of IL-6 is associated with, inter alia, cancer cachexia,5
multiple myeloma,6 and rheumatoid arthritis,7 the madindo-
lines could serve as pharmaceutical lead compounds. As
sometimes happens, the microorganism stopped producing
these two metabolites. Because these compounds have
become unavailable from their original source, chemical
synthesis provides the only means of acquiring material for
biological study.
A number of syntheses have been developed. A summary
of the successful synthetic strategies that have been employed
appears in Scheme 1. The first synthesis by Ohmura, Smith,
and co-workers8,9 formed the cyclopentene through the RCM
reaction of diene 1. Kobayashi’s synthesis made use of a
regioselective aldol condensation of triketone 3.10,11 In both
the Kobayashi and the Ohmura-Smith syntheses, an asym-
(1) Hayashi, M.; Kim, Y.-P.; Takamatsu, S.; Enomoto, A.; Shinose, M.;
Takahashi, Y.; Tanaka, H.; Komiyama, K.; Ohmura, S. J. Antibiot. 1996,
49, 1091.
(2) Takamatsu, S.; Kim, Y.-P.; Enomoto, A.; Hayashi, M.; Tanaka, H.;
Komiyama, K.; Ohmura, S. J. Antibiot. 1997, 50, 1069.
(3) Hayashi, M.; Rho, M.-C.; Enomoto, A.; Fukami, A.; Kim, Y.-P.;
Kikuchi, Y.; Sunazuka, T.; Hirose, T.; Komiyama, K.; Ohmura, S. Proc.
Natl. Acad. Sci. U.S.A. 2002, 99, 14728.
(4) Ohmura, S.; Hayashi, M.; Tomoda, H. Pure Appl. Chem. 1999, 71,
1673.
(8) Sunazuka, T.; Hirose, T.; Shirahata, T.; Harigaya, Y.; Hayashi, M.;
Komiyama, K.; Ohmura, S.; Smith, A. B., III J. Am. Chem. Soc. 2000, 122,
2122.
(5) Strassmann, G.; Masui, Y.; Chizzonite, R.; Fong, M. J. Immunol.
1993, 150, 2341.
(6) Zhang, X. G.; Bataille, R.; Jourdan, M.; Saeland, S.; Banchereau, J.;
Mannoni, P.; Klein, B. Blood 1990, 76, 2599.
(7) Takagi, N.; Mihara, M.; Moriya, Y.; Nishimoto, N.; Yoshizaki, K.;
Kishimoto, T.; Takeda, Y.; Ohsugi, Y. Arthritis Rheum. 1998, 41, 2117.
(9) Hirose, T.; Sunazuka, T.; Yamamoto, D.; Kojima, N.; Shirahata, T.;
Harigaya, Y.; Kuwajima, I.; Ohmura, S. Tetrahedron 2005, 61, 6015.
(10) Hosokawa, S.; Sekiguchi, K.; Enemoto, M.; Kobayashi, S. Tetra-
hedron Lett. 2000, 41, 6429.
10.1021/ol062919e CCC: $37.00
© 2007 American Chemical Society
Published on Web 01/18/2007

a reductive amination was attractive. However, Kobayashi
found that aldehyde 8 was very unstable and underwent rapid
decomposition upon standing at room temperature.10 De-
formylation is likely to be the dominant reaction pathway
in any attempt to conduct the reductive amination of a
substrate such as 8 that has the adjacent keto groups intact.
By contrast, the reductive amination of formylcyclopentene
9 by Ohmura and Smith took place in high yield but
necessitated reduction, protection, deprotection, and reoxi-
dation steps in the synthesis.8
Scheme 1
This was the backdrop against which our synthesis was
designed. From the outset, there were two goals. The first
was to use the allene ether version15 of the Nazarov
cyclization16 for an efficient preparation of the cyclopenten-
one. The second was to install the heterocyclic fragment by
means of a Mannich reaction. The ready availability of chiral,
nonracemic hydroxyfuroindoline prompted us to develop the
stereodivergent synthesis of (+)-madindolines A and B that
is summarized in Scheme 2. Hexanamide 10 was formylated
and converted to E-silyl enol ether 11 in 50% yield for the
two steps. The geometry of the double bond depends on the
base that is used in the second step (triethylamine led to a
3:1 E/Z mixture) and is important because our earlier work
had shown that the Nazarov cyclization proceeds poorly from
substrates derived from Z-enamides.17 Treatment of 11 with
methyllithium led to enone 12 in 70% yield. The substrate
for the Nazarov cyclization (13) was formed by addition of
1-lithio-1-(methoxy)methoxyallene18 to 12. Upon exposure
to trifluoroacetic anhydride and 2,6-lutidine, cyclization of
13 to 14 took place in 88% yield over the two steps.18 No
loss of the silyl ether protecting group was observed. The
exocyclic double bond in 14 was saturated quantitatively and
selectively by hydrogenation over palladium on carbon.
Conversion of cyclopentenone 15 to triethylsilyl enol ether
16 turned out to be straightforward when the enolate was
generated at -78 °C and trapped at -50 °C. Success of this
step could not have been predicted from the outset because
there are three sites where deprotonation of 15 could have
occurred. Also, [1,5]-hydrogen shifts might have converted
16 to three isomeric silyl enol ethers.
metric Evans aldol reaction was employed. Van Vranken
published a very clever synthesis of racemic madindolines
A and B through a Moore ring contraction of 5 that led to
6.12 Ohmura’s second-generation approach coupled a fluo-
rodesilylation reaction of 7 with an intramolecular Claisen
condensation to produce (+)-madindoline A.13,14 Fluorode-
silylation-Claisen condensation of a structural isomer of 7
gave (+)-madindoline B.
Fortunately, these concerns proved to be unfounded. In
the key Mannich reaction, the coupling of cyclopentane
fragment 16 with chiral, nonracemic hydroxyfuroindoline
fragment 27 took place in dichloromethane in the presence
of ZnBr2 at -30 °C.19 Enol ether 16 was added to the solution
last. The solution was heterogeneous but became homoge-
neous upon warming to 0 °C. It should be noted that the
Mannich reaction failed to take place in THF. In both of the
two major products, diastereomers 17 and 18, the relative
Since Ohmura and Smith had shown that the hydroxy-
furoindoline ring could be prepared easily and enantio-
selectively from commercially available tryptophol,8,9 joining
heterocycle and cyclopentane dione fragments by means of
(15) Tius, M. A. Acc. Chem. Res. 2003, 36, 284.
(16) For reviews of the Nazarov reaction, see: (a) Habermas, K. L.;
Denmark, S.; Jones, T. K. In Organic Reactions; Paquette, L. A., Ed.; John
Wiley & Sons, Inc.: New York, 1994; Vol. 45, pp 1-158. (b) Harmata,
M. Chemtracts 2004, 17, 416. (c) Frontier, A.; Collison, C. Tetrahedron
2005, 61, 7577. (d) Pellissier, H. Tetrahedron 2005, 61, 6479. (e) Tius, M.
A. Eur. J. Org. Chem. 2005, 2193.
(17) Tius, M. A.; Kwok, C.-K.; Gu, X.-q.; Zhao, C. Synth. Commun.
1994, 24, 871.
(18) Tius, M. A.; Astrab, D. P. Tetrahedron Lett. 1984, 25, 1539.
(19) Agami, C.; Bihan, D.; Puchot-Kadouri, C. Tetrahedron 1996, 52,
9079.
(11) Hosokawa, S.; Sekiguchi, K.; Hayase, K.; Hirukawa, Y.; Kobayashi,
S. Tetrahedron Lett. 2000, 41, 6435.
(12) McComas, C. C.; Perales, J. B.; Van Vranken, D. L. Org. Lett. 2002,
4, 2337.
(13) Hirose, T.; Sunazuka, T.; Shirahata, T.; Yamamoto, D.; Harigaya,
Y.; Kuwajima, I.; Ohmura, S. Org. Lett. 2002, 4, 501.
(14) Hirose, T.; Sunazuka, T.; Yamamoto, D.; Kaji, E.; Ohmura, S.
Tetrahedron Lett. 2006, 47, 6761.
648
Org. Lett., Vol. 9, No. 4, 2007

Scheme 2
Scheme 3
18) for the Mannich reactions is probably not related to the
source of the iminium ion. Exposure of 16 to R-aminonitrile
23 in the presence of AgNO3 in acetonitrile led to amino-
cyclohexane 28 (Scheme 3) in 66% overall yield from 15.
The stereochemistry of 28 reflects the influence of the
silyloxy group and results from cis addition. This effect may
be due to an electrostatic attraction between the positively
charged iminium ion and the partial negative charge on the
silyloxy oxygen atom.20 Woerpel and co-workers21 have
demonstrated that strategically placed alkoxy substituents on
oxocarbenium ions can powerfully influence the stereochem-
ical course of nucleophilic additions in a contrasteric sense.
It remains to be seen whether the present work is related to
his. If the effect is general, it could be useful in a broader
context.
Fluorodesilylation of the mixture of 17 and 18 followed
by oxidation with pyridinium dichromate led to a 1:1 mixture
of (+)-madindolines A and B in 30% overall yield for the
four steps from 15. The two products were separated by flash
column chromatography and were found to have properties
matching the published data for the two natural products.
The preparation of 27 is summarized in Scheme 4. Twice
Scheme 4
stereochemistry for the Mannich reaction was cis. This was
ascertained by observing a positive NOE of the signal for
the cyclopentane methine proton upon irradiation of the
quaternary methyl group in 19 and 20. The stereochemistry
was not the one predicted on the basis of the model study
summarized in Scheme 3. The Mannich reaction of silyl enol
ether 22 with R-aminonitrile 23 took place in the presence
of AgNO3 in acetonitrile to produce aminocyclopentane 24
in good yield. The stereochemistry of product 24 was
consistent with addition of the electrophile trans to the phenyl
substituent and is the one expected for 17 and 18 (Scheme
2). The difference in stereochemical outcomes (24 vs 17 and
recrystallized hydroxyfuroindoline 258,9 (>98% ee) was
converted to the benzyloxy carbamate and then exposed to
triethylsilyl triflate and triethylamine. Hydrogenolytic cleav-
age of the carbamoyl group led to 26 in 82% overall yield.
Attempts to convert the hydroxyl group in 25 to the silyl
ether in the presence of the free amine invariably led to lower
(20) We thank a referee for suggesting this explanation.
(21) Larsen, C. H.; Ridgway, B. H.; Shaw, J. T.; Smith, D. M.; Woerpel,
K. A. J. Am. Chem. Soc. 2005, 127, 10879.
Org. Lett., Vol. 9, No. 4, 2007
649

yields of material that was difficult to purify. Heating an
ethanolic solution of 26 in the presence of thiophenol and
aqueous formaldehyde gave 27 in 93% yield as an isolable
and stable precursor of the iminium ion.
Acknowledgment. We thank the Department of Defense
Breast Cancer Research Program (DAMD17-03-1-0685) and
the National Institutes of Health (GM57873) for generous
support.
This concludes a brief (10 steps) enantiodivergent synthesis
of (+)-madindoline A and (+)-madindoline B in 9.2%
overall yield from 10. The use of cyclopentadiene 16 as a
key intermediate as well as the diastereoselective Mannich
reaction that produces 17 and 18 are noteworthy. Because
the Mannich reaction is diastereoselective, there is an
opportunity to use the asymmetric version of the Nazarov22
reaction to form 16, or its equivalent, for an asymmetric total
synthesis targeted solely at (+)-madindoline A, or at (+)-
madindoline B.
Supporting Information Available: Experimental pro-
cedures for 11-20, 26, 27, (+)-madindoline A, and (+)-
1
madindoline B; H and 13C NMR, HRMS, and IR data for
11, 12, 14, 15, 26, 27, (+)-madindoline A, and (+)-
1
madindoline B; and reproductions of H and 13C NMR
spectra for 11, 12, 14, 15, 26, 27, (+)-madindoline A, and
(+)-madindoline B. This material is available free of charge
via the Internet at http://pubs.acs.org.
(22) Harrington, P. E.; Murai, T.; Chu, C.; Tius, M. A. J. Am. Chem
Soc. 2002, 124, 10091.
OL062919E
650
Org. Lett., Vol. 9, No. 4, 2007