A New Strategy toward the Total Synthesis of Stachyflin, A Potent Anti-Influenza A Virus Agent: Concise Route to the Tetracyclic Core Structure

Article abstract of DOI:10.1021/ol0271032

(Matrix Presented) A new strategy directed toward the total synthesis of stachyflin, a potent and novel anti-influenza A virus agent isolated from a microorganism, has been presented through the enantioselective synthesis of the tetracyclic core structure. The synthetic method features a BF₃·Et₂O-induced domino epoxide-opening/rearrangement/cyclization reaction as the key step.

Full text of DOI:10.1021/ol0271032

ORGANIC
LETTERS
2002
Vol. 4, No. 25
4483-4486
A New Strategy toward the Total
Synthesis of Stachyflin, A Potent
Anti-Influenza A Virus Agent: Concise
Route to the Tetracyclic Core Structure
Mari Nakatani, Masahiko Nakamura,† Akiyuki Suzuki,‡ Munenori Inoue, and
Tadashi Katoh*
Sagami Chemical Research Center, Hayakawa 2743-1, Ayase,
Kanagawa 252-1193, Japan
takatoh@alles.or.jp
Received October 14, 2002
ABSTRACT
A new strategy directed toward the total synthesis of stachyflin, a potent and novel anti-influenza A virus agent isolated from a microorganism,
has been presented through the enantioselective synthesis of the tetracyclic core structure. The synthetic method features a BF3‚Et2O-
induced domino epoxide-opening/rearrangement/cyclization reaction as the key step.
Stachyflin (1, Figure 1) is a novel sesquiterpenoidal alkaloid
isolated from the culture broth of Stachybotrys sp. RF-7260
by Minagawa et al. at the Shionogi research group in 1997.1
This secondary metabolite was found to exhibit potent
antiviral activity against influenza A/WSN/33 (H1N1) virus
in vitro with an IC50 value of 0.003 µM with a novel
mechanism of action.1-3 Stachyflin has been found to inhibit
the fusion process between the viral envelop and the
endosome constituting the cell membrane, which is an
essential step in the entry of the influenza virus into host
cells.4 This mechanism is quite different from that of the
known anti-influenza virus agents such as amantadine,5
zanamivir,6 and oseltamivir.7 Therefore, 1 is anticipated to
be a highly promising agent for novel influenza therapeutics.
The structure of stachyflin (1) was revealed by means of
extensive spectroscopic studies and X-ray crystallographic
† Graduate student from Department of Electronic Chemistry, Tokyo
Institute of Technology, Nagatsuta, Yokohama 226-8502, Japan.
‡ Undergraduate student from Department of Chemistry, Kitasato
University, Kitasato, Sagamihara, Kanagawa 228-0829, Japan.
(1) (a) Kamigauchi, T.; Fujiwara, T.; Tani, H.; Kawamura, Y.; Horibe,
I (Shionogi & Co., Ltd, Japan) PCT WO 9711947 A1, April 3, 1997. (b)
Minagawa, K.; Kouzuki, S.; Yoshimoto, J.; Kawamura, Y.; Tani, H.; Iwata,
T.; Terui, Y.; Nakai, H.; Yagi, S.; Hattori, N.; Fujiwara, T.; Kamigauchi,
T. J. Antibiot. 2002, 55, 155. (c) Minagawa, K.; Kouzuki, S.; Kamigauchi,
T. J. Antibiot. 2002, 55, 165.
(2) It is reported that the anti-influenza A virus activity of stachyflin (1)
is 1760 times more active than that of amantadine (IC50 ) 5.3 µM) and is
250 times more active than that of zanamivir (IC50 ) 0.75 µM); see ref 1b.
Figure 1. Structures of stachyflin (1) and the tetracyclic core 2.
10.1021/ol0271032 CCC: $22.00 © 2002 American Chemical Society
Published on Web 11/16/2002

analysis to have a novel pentacyclic 3H-naphtho[1′,8′a:5,6]-
pyrano[2,3-e]isoindol-3-one skeleton (ABCDE ring system)
with five asymmetric carbons,1a,b in which cis-fused AB rings
and BC rings and an ether bond at the bridgehead of the AB
ring junction are particularly characteristic features. The
absolute configuration of 1 was determined by its circular
dichroism (CD) spectrum as depicted in Figure 1.1b
tetracyclic cores 2 and 3, respectively, in one step, where
we envisioned that this acid-induced domino reaction would
proceed stereoselectively to install the requisite cis-fused AB
ring junction (cf. 5f[IfIIfIII]f3, Scheme 3). To the best
of our knowledge, this domino sequence is hitherto unknown,
and hence it involves an interesting possibility at a synthetic
chemical level. The cyclized compound 3 having the
epimeric C3 R-hydroxy group would be converted to target
compound 2 through inversion of the C3 hydroxy group.
Epoxides 4 and 5, in turn, would be elaborated from
intermediate 6 by sequential manipulation of the C4 exo-
olefin moiety and deprotection of the phenolic O-methyl
group, or vice versa. Intermediate 6, which possesses the
trans-fused decalin system, would be prepared through the
stereocontrolled reductive alkylation of the known enone 711
with the known bromide 812 applying the related protocols
previously described in the literature.13
At first, as shown in Scheme 2, we pursued the synthesis
of epoxides 4 and 5, precursors of the key acid-induced
domino reaction. The synthesis commenced with the reduc-
tive alkylation of the known enantiomerically pure enone
711 (>99% ee) with 2-methoxybenzyl bromide (8),12 readily
prepared from commercially available 2-methoxybenzyl
alcohol. Thus, reduction of the enone 7 by treatment with
lithium (3 equiv) in liquid ammonia, followed by trapping
the intermediary lithium enolate through reaction with the
bromide 8 (6 equiv), furnished the expected coupling product
9 as the sole diastereomer in 72% yield. Subsequent Wittig
methylenation of the sterically hindered carbonyl group in
9 was best achieved by employing a combination of
methyltriphenylphosphonium bromide and potassium tert-
butoxide14 in refluxing benzene, providing exo-olefin 10 in
86% yield.
Its remarkable biological properties as well as its unique
structural features make 1 an exceptionally intriguing and
timely target for total synthesis. To date, only one total
synthesis of racemic (()-1 has been reported by Mori et al.
at Shionogi & Co., Ltd., in 1998,8 wherein the ABCDE ring
system was built step-by-step starting from 2,3-dimethyl-
cyclohexanone, which corresponds to the B ring of 1. We
embarked on a project directed at the total synthesis of
optically active (+)-1 and its analogues with the aim of
exploring structure-activity relationships.9 In this Letter, we
report our preliminary results concerning a highly concise
method for the synthesis of model compound 2, which
represents the tetracyclic core structure (ABCD ring system)
having the requisite substituents and asymmetric carbons
contained in 1.
The synthetic plan for the tetracyclic core 2 is outlined in
Scheme 1, which was designed on the basis of our previous
Scheme 1. Synthetic Plan for the Tetracyclic Core 2
For elaborating the C8 stereocenter, the ethylene acetal
moiety in 10 was first removed by acid treatment (97%),
and the resulting ketone derivative 11 was subjected to
hydrogenation (1 atm) over 10% Pd/C in 50:1 triethylamine-
methanol at ambient temperature, which afforded a mixture
of diastereomers that were separated by column chromatog-
(6) Hayden, F. G.; Osterhaus, A. D.; Treanor, J. J.; Fleming, D. M.;
Aoki, F. Y.; Nicholson, K. G.; Bohnen, A. M.; Hirst, H. M.; Keene, O.;
Wightman, K. N. Engl. J. Med. 1997, 337, 874.
(7) Lew, W.; Chen, X.; Kim, C. U. Curr. Med. Chem. 2000, 7, 663.
(8) Taishi, T.; Takechi, S.; Mori, S. Tetrahedron Lett. 1998, 39, 4347.
(9) Preliminary structure-activity studies of stachyflin derivatives have
been performed by the Shionogi research group; see ref 1c.
(10) Nakamura, M., Suzuki, A.; Nakatani, M.; Fuchikami, T.; Inoue, M.;
Katoh, T. Tetrahedron Lett. 2002, 43, 6929.
(11) Hagiwara, H.; Uda, H. J. Org. Chem. 1988, 53, 2308.
(12) Kelly, J. L.; Linn, J. A.; Selway, J. W. T. J. Med. Chem. 1989, 32,
1757.
work.10 The key feature in this plan is a biogenetic-type acid-
induced domino epoxide-opening/rearrangement/cyclization
reaction of epoxides 4 and 5 to construct the desired
(13) This type of reductive alkylation, originally developed by Stork et
al. in the 1960s, has been widely utilized as a critical step in the total
synthesis of marine sesquiterpene quinones and hydroquinones. For recent
examples, see: (a) Ling, T.; Poupon, E.; Rueden, E. J.; Kim, S. H.;
Theodorakis, E. A. J. Am. Chem. Soc. 2002, 124, 12261. (b) Stahl, P.;
Kissau, L.; Mazitschek, R.; Huwe, A.; Furet, P.; Giannis, A.; Waldmann,
H. J. Am. Chem. Soc. 2001, 123, 11586. (c) Poigny, S.; Guyot, M.; Samadi,
M. J. Org. Chem. 1998, 63, 5890. (d) Locke, E. P.; Hecht, S. M. Chem.
Commun. 1996, 2717. (e) An, J.; Wiemer, D. F. J. Org. Chem. 1996, 61,
8775. (f) Bruner, S. D.; Radeke, H. S.; Tallarico, J. A.; Snapper, M. L. J.
Org. Chem. 1995, 60, 1114. (g) Sarma, A. S.; Chattopadhyay, P. J. Org.
Chem. 1982, 47, 1727.
(3) In addition to the in vitro activity, the in vivo anti-influenza virus
activity of stachyflin (1) and its derivatives was also extensively studied
by the Shionogi research group; see: (a) Yoshimoto, J.; Yagi, S.; Ono, J.;
Sugita, K.; Hattori, N.; Fujioka, T.; Fujiwara, T.; Sugimoto, H.; Hashimoto,
N. J. Pharm. Pharmacol. 2000, 52, 1247. (b) Yagi, S.; Ono, J.; Yoshimoto,
J.; Sugita, K.; Hattori, N.; Fujioka, T.; Fujiwara, T.; Sugimoto, H.; Hirano,
K.; Hashimoto, N. Pharm. Res. 1999, 16, 1041.
(4) (a) Yoshimoto, J.; Kakui, M.; Iwasaki, H.; Sugimoto, H.; Fujiwara,
T.; Hattori, N. Microbiol. Immunol. 2000, 44, 677. (b) Yoshimoto, J.; Kakui,
M.; Iwasaki, H.; Fujiwara, T.; Sugimoto, H.; Hattori, N. Arch. Virol. 1999,
144, 865.
(14) This reagent system is known to be effective for sterically hindered
ketones; see: Fitjer, L.; Quabeck, U. Synth. Commun. 1985, 15, 855.
(5) Pinto, L. H.; Holsinger, L. J.; Lamb, R. A. Cell 1992, 69, 517.
4484
Org. Lett., Vol. 4, No. 25, 2002

expected that R-epoxide 5 would be converted to the target
compound 2 through the subsequent key acid-induced
domino reaction followed by inversion of the C3 hydroxy
group. Therefore, at this stage, we did not so much consider
the stereoselectivity of this epoxidation reaction, and we
proceeded with the projected synthesis to ascertain the
feasibility of our devised synthetic strategy.
Scheme 2. Synthesis of Key Intermediates 4 and 5a
With epoxides 4 and 5 in hand, we next investigated the
crucial acid-induced domino epoxide-opening/rearrangement/
cyclization reaction as shown in Scheme 3. First, the major
product, R-epoxide 5, was subjected to the acid-induced
domino reaction; thus, 5 was allowed to react with BF3‚
Et2O18 (3 equiv) in dichloromethane at -30 °C for 30 min,
which successfully led to the formation of the requisite
cyclized compound 3 in a reasonable yield (41%).19 To our
knowledge, this is the first example of the direct construction
of such a tetracyclic benzo[d]xanthene skeleton having a C3
hydroxy group. We believe that this domino transformation
proceeds in a stepwise manner through carbocation inter-
mediates such as I-III, transformation of which follows that
described in our previous studies.10 Thus, initial coordina-
tion-activation between the Lewis acid and the epoxide
Scheme 3. Synthesis of Tetracyclic Core 2 through the Key
BF3‚Et2O-Induced Domino Epoxide-Opening/Rearrangement/
Cyclization Reaction of 4 and 5.a
a Reagents and conditions: (a) Li, liquid NH3, 2-methoxybenzyl
bromide (8), THF, 72%. (b) Ph3P+CH3Br-, t-BuOK, benzene,
reflux, 86%. (c) 4 M HCl, THF, rt, 97%. (d) H2 (1 atm), 10% Pd-
C, 50:1 Et3N-MeOH, rt, 80%. (e) Ph3P+CH3Br-, t-BuOK, benzene,
reflux, 100%. (f) RhCl3‚3H2O, EtOH, reflux, 100%. (g) n-BuSLi,
HMPA, 100 °C, 84%. (h) MCPBA, NaHCO3, CH2Cl2, 0 °C, 22%
for 4, 75% for 5.
raphy on silica gel to give the desired product 12 (80%) and
its C8 epimer (13%) in a ratio of ca. 6:1.15
Ketone 12 was further converted to the phenol derivative
14 having an olefinic bond at C3-C4 in 84% overall yield
via a three-step sequence involving Wittig methylenation of
the carbonyl group in 12, rhodium-catalyzed isomeriza-
tion13a,b,d,e of the resulting exocyclic olefin 6 to the more
stable endocyclic olefin 13, and demethylation of the
phenolic O-methyl protecting group in 13 by reaction with
lithium n-butylthiolate10,16 in hexamethylphosphoramide
(HMPA). Finally, treatment of endo-olefin 14 with m-
chloroperbenzoic acid (MCPBA) in the presence of sodium
hydrogen carbonate in dichloromethane at 0 °C led to the
formation of a diastereomeric mixture of the desired epoxides
4 and 5, which were separated by column chromatography
on silica gel to produce â-epoxide 4 in 22% yield and
R-epoxide 5 in 75% yield.17 As mentioned earlier, we
(15) When ethylene acetal 10 was used as a substrate for this hydrogena-
tion under the same reaction conditions, a considerable decrease in the
stereoselectivity at C8 was observed (â-Me:R-Me ) ca. 2:1); this is probably
due to an unfavorable conformation of the decalin ring. For similar examples
of this phenomenon, see ref 13b,c,d,e.
(16) Katoh, T.; Nakatani, M.; Shikita, S.; Sampe, R.; Ishiwata, A.;
Ohmori, O.; Nakamura. M.; Terashima, S. Org. Lett. 2001, 3, 2701.
a Reagents and conditions: (a) BF3‚Et2O, CH2Cl2, -30 °C, 41%
for 5f3, 15% for 4f2. (b) Dess-Martin periodinane, CH2Cl2, rt,
87% for 3f15, 85% for 2f15. (c) H2 (1 atm), PtO2, 5:1 EtOH-
CHCl3, rt, 42% for 2, 28% for 3.
Org. Lett., Vol. 4, No. 25, 2002
4485

moiety of 5 would lead to epoxide ring-opening and the
formation of intermediate I, which then would provide
intermediate II via migration of the C5 methyl group to the
C4 carbocation center. Intermediate II would undergo a 1,2-
hydride shift from the C10 position to the C5 cationic carbon
center to furnish intermediate III, wherein the C10 carboca-
tion center would be trapped by the inner phenolic hydroxy
group to deliver, after elimination of the Lewis acid, the
requisite cyclized product 3.
As we expected, compound 3 could be converted to the
target compound 2 via inversion of the C3 hydroxy group.20
Thus, Dess-Martin oxidation of 3 provided the correspond-
ing ketone 15 in 87% yield, which upon hydrogenation (1
atm) over Pt2O8 in 5:1 ethanol-chloroform at ambient
temperature furnished, after separation by silica gel column
chromatography, the desired 2 in 42% yield along with its
C3 epimer 3 (28%).21 On the other hand, the acid-induced
domino reaction of the minor product, â-epoxide 4, under
the same reaction conditions as that described for the
conversion of 5 to 3 resulted in production of cyclized
product 2 in poor yield (15%).19,22
The structure and stereochemistry of cyclized products 2
and 3 were unambiguously confirmed by extensive spectro-
scopic analysis including NOESY experiment in the 500
MHz 1H NMR spectrum of the ketone 15 derived from both
2 and 3 via Dess-Martin oxidation.23 The selected NOESY
correlation of 15 is depicted in Figure 2, wherein two key
NOEs between Me-13 and H-6 and between H-5 and H-11
Figure 2. Selected NOESY correlation of the ketone 15.
(indicated with red-colored arrows) established the cis-fused
connectivity of the decalin ring (AB ring system).24
In summary, we have demonstrated the feasibility of our
novel strategy toward (+)-stachyflin (1) through an enantio-
selective synthesis of the tetracyclic core structure 2. The
synthetic method features a BF3‚Et2O-induced domino ep-
oxide-opening/rearrangement/cyclization reaction as the key
step. On the basis of the explored strategy, total synthesis
of optically active stachyflin and its analogues is currently
under investigation in our laboratory.
Acknowledgment. We are grateful to Prof. H. Miyaoka
and Dr. H. Mitome, Tokyo University of Pharmacy and Life
Science, for valuable suggestions. We also thank Dr. N.
Sugimoto and Dr. N. Kawahara, National Institute of Health
Science, for measurements of the high-resolution mass
spectra and assistance with NMR experiments, respectively.
(17) The stereochemistry of epoxides 4 and 5 was proven by NOESY
experiment in their 500 MHz 1H NMR spectra. Thus, significant NOE
interactions between the signals due to the C4 methyl group and the C5
methyl group were observed in R-epoxide 5. On the other hand, no NOE
correlation between the C4 methyl group and the C5 methyl group was
detected in â-epoxide 4.
(18) We preliminarily examined this acid-induced domino reaction using
several Lewis acids or Brønsted acids such as BF3‚Et2O, AlCl3, MeAlCl2,
TiCl4, TMSOTf, Sc(OTf)3, p-TsOH‚H2O, camphorsulfonic acid, and CF3-
SO3H; among them, BF3‚Et2O was found to give the best result.
(19) In this reaction, disappearance of the starting material could be
ascertained by TLC analysis, while unidentified byproducts were produced
along with the desired rearrangement/cyclization product.
(20) Attempts to invert the C3 hydroxy group by employing the
Mitsunobu protocol were unsuccessful, presumably due to steric factors.
(21) When ketone 15 was reduced with NaBH4, the undesired C3 epimer
3 was predominantly produced (3:2 ) ca. 4:1).
(22) Further investigation is required to optimize the yield of the cyclized
product 2; this is the subject of our current attention.
Note Added after ASAP Posting. In Scheme 3, formula-
tion II, a bond was missing in the version posted ASAP on
November 16, 2002; the corrected version was posted
November 20, 2002.
Supporting Information Available: Detailed experi-
mental procedures and full characterization data for all new
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
OL0271032
(23) Structural elucidation of cyclized products 2 and 3 was made at the
stage of ketone 15 because some crucial signals were unfortunately
overlapped in the 500 MHz H NMR spectra of 2 and 3.
(24) Similar NOESY correlation has been reported for the structural
determination of stachyflin (1); see ref 1b.
1
4486
Org. Lett., Vol. 4, No. 25, 2002