Enantioselective total synthesis of (+)-stachyflin: A potential anti-influenza a virus agent isolated from a microorganism

Article abstract of DOI:10.1002/ejoc.201100173

A novel and potent hemagglutinin inhibitor, (+)-stachyflin, was efficiently synthesized in an enantioselective manner starting from the (+)-5-methyl-Wieland-Miescher ketone. The synthetic method features a BF ₃·Et₂O-induced cascade epoxide-opening/ rearrangement/cyclization reaction tostereoselectively construct the requisite pentacyclic ring system in one step. In order to rationalize the mechanism of the cascade reaction, quantum chemical calculations of the possible intermediary carbocations and transition states in the model synthesis were carried out. An alternative approach to synthesize (+)-stachyflin by employing a similar cascade reaction was also described. A potential anti-influenza A virus agent, (+)-stachyflin, a novel hemagglutinin inhibitor, has been synthesized in an enantioselective manner for the first time starting from (+)-5-methyl-Wieland- Miescher ketone. The method features a BF₃·Et ₂O-induced cascade epoxide-opening/rearrangement/cyclization reaction to construct the requisite pentacyclic ring system in one step.

Full text of DOI:10.1002/ejoc.201100173

FULL PAPER
DOI: 10.1002/ejoc.201100173
Enantioselective Total Synthesis of (+)-Stachyflin: A Potential Anti-Influenza
A Virus Agent Isolated from a Microorganism
Junji Sakurai,[a] Takuya Kikuchi,[a] Ohgi Takahashi,[b] Kazuhiro Watanabe,[a] and
Tadashi Katoh*[a]
Keywords: Natural products / Total synthesis / Stachyflin / Influenza A / Antiviral agents / Hemagglutinin inhibitor
A novel and potent hemagglutinin inhibitor, (+)-stachyflin,
was efficiently synthesized in an enantioselective manner
starting from the (+)-5-methyl-Wieland–Miescher ketone.
The synthetic method features a BF3·Et2O-induced cascade
tem in one step. In order to rationalize the mechanism of the
cascade reaction, quantum chemical calculations of the pos-
sible intermediary carbocations and transition states in the
model synthesis were carried out. An alternative approach
to synthesize (+)-stachyflin by employing a similar cascade
reaction was also described.
epoxide-opening/rearrangement/cyclization
reaction
to
stereoselectively construct the requisite pentacyclic ring sys-
fection.[6] This mechanism is quite different from that of the
M2 protein inhibitors, such as amantadine (Symmetrel®)[7]
and rimantadine (Flumadine®),[8] and the NA inhibitors
such as zanamivir (Relenza®),[9] oseltamivir (Tamiflu®)[10]
and peramivir (Rapiacta®).[11] The anti-influenza A virus
activity of 1 has been reported to be 1760 times superior to
that of amantadine (IC50 = 5.3 μm) and 250 times superior
to that of zanamivir (IC50 = 0.75 μm).[4b] Stachyflin, there-
fore, is anticipated to be a promising candidate or new lead
for innovative molecularly targeted anti-influenza A virus
agents.[12]
Introduction
Influenza is a long-standing health problem. The world-
wide spread of the H5N1 avian influenza and the recent
outbreak of the new type of H1N1 human influenza have
raised public concerns about global influenza pandemics.[1]
Two classes of anti-influenza virus agents, matrix-2 (M2)
protein inhibitors and neuraminidase (NA) inhibitors, are
currently available for clinical use.[2] However, adverse side
effects and the emergence of drug-resistance to these agents
have been reported and constitute serious clinical barriers.[3]
Thus, there is a need to discover and develop new anti-
influenza agents exhibiting alternative mechanisms of ac-
tions.
In 1997, scientists from Shionogi & Co., Ltd., Japan, re-
ported the isolation and structural elucidation of stachyflin
(1, Figure 1) from the culture broth of Stachybotrys sp. RF-
7260.[4] This secondary metabolite was found to exhibit po-
tent antiviral activity against the influenza A/WSN/33
(H1N1) virus in vitro (IC50 = 3 nm) with a novel mechanism
of action.[4b] It has also been reported that phosphate ester
derivatives (water-soluble prodrugs) of 1 exhibited in vivo
anti-influenza virus activity in animal models.[5] The antivi-
ral activity of 1 has been shown to result from interference
with the low-pH induced hemagglutinin (HA) conforma-
tion change, which is indispensable for the virus–cell mem-
brane fusion process that occurs during influenza viral in-
Figure 1. Structure of (+)-stachyflin (1) (in the 3D structure, hydro-
gen atoms are omitted for clarity).
The gross structure and relative stereochemistry of 1 has
been determined by extensive spectroscopic studies, includ-
ing 2D NMR experiments (COSY, HMBC and NOESY
spectra) and X-ray crystallographic analysis.[4a,4b] This
natural product possesses a novel pentacyclic 3H-naphtho-
[1Ј,8Јa:5,6]pyrano[2,3-e]isoindol-3(2H)-one skeleton (ABCDE
ring system) involving five asymmetric carbons, with the
special characteristic features of cis-fused AB and BC rings,
an ethereal bond at the bridgehead of the AB ring junction,
a pentasubstituted arene (D ring) and a five-membered lac-
[a] Laboratory of Synthetic Medicinal Chemistry, Faculty of
Pharmaceutical Sciences, Tohoku Pharmaceutical University,
4-4-1 Komatsushima, Aoba-ku, Sendai 981-8558, Japan
Fax: +81-22-727-0135
E-mail: katoh@tohoku-pharm.ac.jp
[b] Laboratory of Pharmaceutical Physical Chemistry, Faculty of
Pharmaceutical Sciences, Tohoku Pharmaceutical University,
4-4-1 Komatsushima, Aoba-ku, Sendai 981-8558, Japan
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201100173.
2948
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 2948–2957

Enantioselective Total Synthesis of (+)-Stachyflin
tam (E ring).[4a,4b] The absolute configuration of 1 was de-
termined by circular dichroism (CD) spectroscopy and was
assigned as depicted in Figure 1.[4b]
The intriguing biological properties and unique struc-
tural features prompted us to undertake a project directed
towards the total synthesis of optically active 1. To date,
the Shionogi research group, in 1998, are the only research-
ers to have reported the total synthesis of racemic (Ϯ)-1.[13]
Through this synthetic work, they discovered that the anti-
influenza A viral activity of (Ϯ)-1 is about half that of natu-
ral (+)-1.[13] An enantioselective approach to the tetracyclic
core structure (ABCD ring system) has been disclosed by
Cramer et al.[14] We have previously reported a novel and
efficient method for the construction of the tetracyclic
model compound (ABCD ring system).[15] Recently, our as-
siduous endeavours culminated in the completion of the
first enantioselective total synthesis of (+)-1.[16] In this pa-
per, we describe in full detail, our total synthesis of (+)-1
using a novel acid-induced cascade epoxide-opening/re-
arrangement/cyclization reaction as the key step.
Results and Discussion
Basic Strategy toward (+)-Stachyflin (1)
Our basic strategy for the synthesis of (+)-stachyflin (1)
is outlined in Scheme 1. The key element of this approach
is the acid-induced sequential epoxide-opening/rearrange-
ment/cyclization reaction of I to stereoselectively construct
the desired pentacyclic core V, having cis-fused AB- and
BC-ring junctions, in one step. We anticipated that this cas-
cade reaction would proceed in a stepwise manner through
the agency of carbocation intermediates II, III and IV.
Thus, the initial coordination–activation between an appro-
priate acid and the epoxide moiety of I would induce epox-
ide ring-opening and the formation of intermediate II,
which would proceed to intermediate III following mi-
gration of the C5 methyl group to the C4 carbocation cen-
tre. Intermediate III would then undergo a 1,2-hydride shift
from the C10 position to the C5 carbocation centre on the
α-face of the molecule to deliver intermediate IV, wherein
the C10 carbocation centre would be trapped by the phen-
olic hydroxy group to produce the desired cyclized product
V. Intermediate V, in turn, would be converted into target
molecule 1 by the establishment of the desired stereochem-
istry at C3 followed by removal of two protective groups
(N-P and O-Me). In order to explore the feasibility of this
devised cascade sequence (IǞ[IIǞIIIǞIV]ǞV), we ini-
tially performed model studies involving the construction
of tetracyclic core structure 2b (Scheme 2) which possesses
the requisite substituents and asymmetric carbons but lacks
the lactam moiety of the E ring and the methoxy group of
the D ring.
Scheme 1. Proposed cascade for the rapid construction of the
pentacyclic core V (ABCDE ring system) of (+)-stachyflin (1). P =
protective group.
Scheme 2. Synthetic approach to model compound 2b.
Synthesis of Tetracyclic Model Compound 2b
struction of initial target 2b is outlined in Scheme 2. The
Synthetic plan: We pursued the synthesis of tetracyclic key acid-induced cascade reaction of epoxides 3a and 3b
compound 2b as a model study. Our synthetic plan for con- would deliver desired tetracyclic cores 2a and 2b, respec-
Eur. J. Org. Chem. 2011, 2948–2957
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2949

J. Sakurai, T. Kikuchi, O. Takahashi, K. Watanabe, T. Katoh
To elaborate the C8 stereogenic centre, the ethylene ace-
FULL PAPER
tively, in one step (Scheme 1). The cyclized product 2a pos-
sessing the epimeric C3 α-hydroxy group would be con- tal moiety in 8 was first removed by acid treatment
verted into target compound 2b by inversion of the hydroxy (98%),[21] and resulting ketone 9 was subjected to hydrogen-
group. Epoxides 3a and 3b, in turn, would be accessible ation [H2 (1 atm), 10% Pd/C, Et3N/MeOH 50:1, r.t.], which
from intermediate 4 following sequential manipulation of afforded desired product 10 (91%) and its C8 epimer (7%)
the C4 exo-olefin moiety and deprotection of the phenolic after separation by silica gel column chromatography.
O-Me group, or vice versa. Intermediate 4, having the trans- Ketone 10 was further converted to endo-olefin 12 (86%
fused decalin system, would be produced by stereocon- overall yield) using a three-step sequence involving the Wit-
trolled reductive alkylation of known optically active enone tig methylenation, rhodium-catalysed olefin isomerisation
5[17] with known bromide 6.[18]
(RhCl3·3H2O, EtOH, reflux)[21a,21b,22] of the resulting exo-
olefin 4 and removal (nBuSLi, HMPA, 100 °C)[23] of the
phenolic O-methyl moiety in 11. Finally, the epoxidation of
12 with mCPBA led to the formation of a diastereomeric
mixture of desired epoxides 3a and 3b, which were sepa-
rated by silica gel column chromatography to afford α-epox-
ide 3a (76%) and β-epoxide 3b (21%). The stereochemistry
of 3a and 3b was validated by nuclear Overhauser effect
spectroscopy (NOESY). Clear NOE interactions between
the signals of the C4- and C5-methyl groups were observed
in 3a. In contrast, no NOE correlations between these
Synthesis of Epoxides 3a and 3b
We initially investigated the synthesis of epoxides 3a and
3b, precursors of the key cascade reaction (Scheme 3). The
synthesis commenced with reductive alkylation[19] of the
known enantiomerically pure enone 5 (Ͼ 99% ee)[17] with
2-methoxybenzyl bromide (6),[18] readily prepared from
commercially available 2-methoxybenzyl alcohol, producing
expected coupling product 7 (75%) as a single diastereomer.
Subsequent Wittig methylenation (Ph3P+CH3Br–, tBuOK,
benzene, reflux)[20] of 7 provided exo-olefin 8 (89%).
Table 1. Screening of acids for the cascade reaction of epoxides 3a
and 3b.
Entry Substrate Conditions
Product Yield[a] [%]
1
2
3
4
5
6
7
8
9
3a
3b
3b
3b
3b
3b
3b
3b
3b
BF3·Et2O (5 equiv.),
CH2Cl2,
–50 °C to r.t., 5 h
BF3·Et2O (5 equiv.),
CH2Cl2,
–50 °C to r.t., 5 h
TiCl4 (2 equiv.),
CH2Cl2,
–78 to –20 °C
AlCl3 (2 equiv.),
CH2Cl2,
–78 to –20 °C
MeAlCl2 (2 equiv.),
CH2Cl2,
–78 to –20 °C
Me2AlCl (2 equiv.),
CH2Cl2,
–78 to –20 °C
TFA[c] (5 equiv.),
CH2Cl2,
–50 °C to r.t., 12 h
TsOH[d] (5 equiv.),
CH2Cl2,
2a
72
2b
76
dec.[b]
dec.[b]
dec.[b]
dec.[b]
2b
43
n.r.[e]
n.r.[e]
Scheme 3. Synthesis of epoxides 3a and 3b. a) Li, liq. NH3/THF,
–78 to –30 °C; isoprene, H2O, 6, –30 °C to r.t.. 75%; b)
Ph3P+CH3Br–, tBuOK, benzene, reflux, 89%; c) 4 m HCl, THF,
r.t., 98%; d) H2 (1 atm), 10% Pd/C, Et3N/MeOH 50:1, r.t., 91%;
e) Ph3P+CH3Br–, tBuOK, benzene, reflux, 100%; f) RhCl3·3H2O,
EtOH, reflux, 100%; g) nBuSLi, HMPA, 100 °C, 86%; h) mCPBA,
–50 °C to r.t., 12 h
CSA[f] (5 equiv.),
CH2Cl2,
–50 °C to r.t., 12 h
[a] Isolated yield. [b] Decomposition. [c] TFA = trifluoroacetic
NaHCO3, CH2Cl2, 0 °C, 76% for 3a, 21% for 3b. HMPA = hexa- acid. [d] TsOH = p-toluenesulfonic acid. [e] n.r. = no reaction.
methylphosphoramide, mCPBA = m-chloroperbenzoic acid. [f] CSA = camphorsulfonic acid.
2950
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 2948–2957

Enantioselective Total Synthesis of (+)-Stachyflin
methyl groups were detected in 3b. As mentioned pre- tained with BF3·Et2O (Table 1, Entries 1 and 2). Thus, ep-
viously, we expected that major epoxidation product 3a oxides 3a and 3b were treated with BF3·Et2O (5 equiv.) in
would be converted into target compound 2b through appli- CH2Cl2 from –50 °C to r.t. over the course of 5 h, which
cation of the key cascade reaction followed by inversion of successfully led to the formation of requisite cyclized prod-
the C3 hydroxy group (Table 1 and Scheme 2). Therefore,
at this stage, we gave little consideration to the stereoselec-
tivity of the epoxidation reaction, and forwarded the syn-
thetic approach to ascertain its feasibility.
Synthesis of Model Compound 2b using the Key Cascade
Reaction
After obtaining epoxides 3a and 3b, we next investigated
their crucial acid-induced cascade epoxide-opening/re-
arrangement/cyclization reactions by using several Lewis
and Brønsted acids (Table 1). The best results were ob- Table 2.
Scheme 4. Conversion of 2a to 2b by stereoselective reduction of
ketone 13. a) Dess–Martin periodinane, CH2Cl2, r.t., 93%; b) see
Scheme 5. Minimum-energy structures VI, VIIa–c, VIIIa and VIIIb and transition states TS1 and TS2 calculated by the B3LYP/6-31G*
method in the cascade reaction of 3b leading to 2b. The values in parentheses are ZPE-corrected relative energies with respect to initial
carbocation VI. Selected interatomic distances are shown in Å units.
Eur. J. Org. Chem. 2011, 2948–2957
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2951

J. Sakurai, T. Kikuchi, O. Takahashi, K. Watanabe, T. Katoh
FULL PAPER
ucts 2a (72%) and 2b (76%). The use of other Lewis acids
In order to continue the synthesis, the conversion of 2a
such as TiCl4, AlCl3, MeAlCl2 and Me2AlCl resulted only to 2b by inversion of the C3 hydroxy group was next studied
in the decomposition of material (Table 1, Entries 3–6). By (Scheme 4).[26] The sequence involved Dess–Martin oxi-
employing trifluoroacetic acid (TFA), a lower yield of 2b dation (93%) followed by the stereoselective reduction of
(43%) was produced (Table 1, Entry 7). The use of p-tolu- resulting ketone 13. In the reduction step, several reducing
enesulfonic acid (TsOH) or camphorsulfonic acid (CSA) led agents such as NaBH4, l-selectride®, LiAlH(tBuO)3 and
only to recovery of starting material 3b (Table 1, Entries 8 H2/PtO2 were examined (Table 2). Among these reductants,
and 9). Structural elucidation of cascade products 2a and the use of LiAlH(tBuO)3 gave the best result (Table 2, En-
2b was carried out using ketone 13 (Scheme 4) due to the try 3), and desired β-hydroxy compound 2b was produced
1
observed overlapping of crucial signals in the 400 MHz H in 82% yield with high stereoselectivity (2a/2b 1:16). When
NMR spectra of 2a and 2b.
the reduction was performed using other reducing agents
In order to rationalize the mechanism of the cascade re- (Table 2, Entries 1, 2 and 4), the stereoselectivity was appar-
action (3bǞ2b, Table 1), we undertook quantum chemical ently reduced (2a/2b 2:1–1:2). The structure and stereo-
calculations of possible intermediary carbocations VI, VII chemistry of cascade products 2a and 2b were unambigu-
and VIII (Scheme 5). Density functional theory (DFT) cal- ously confirmed by NOESY experiments using ketone 13,
culations were conducted at the B3LYP and 6-31G* basis as depicted in Figure 2.[27]
set by using the Spartan Ј08 program.[24] The calculated
minimum-energy structures (VI, VIIa–c, VIIIa and VIIIb)
and transition states (TS1 and TS2) are shown in Scheme 5. Entry Conditions
Vibrational frequencies were calculated for all stationary
Table 2. Results of stereoselective reduction of ketone 13.
Yield[a] [%]
Ratio
2a/2b
2a
2b
points, and the relative energies were corrected for the zero-
point energy (ZPE) differences based on starting carbo-
cation VI, as shown in Scheme 4. TS1 is the transition state
for the methyl migration from the first carbocation VI to
the second carbocation VII. The activation barrier for this
migration is as low as 4.7 kcal mol–1. The migration is
strongly coupled with a conformational change of the A
ring. TS1 is correlated to the structure of VIIa, which is a
conformer of VII and is lower in energy than VI by 6.2 kcal
mol–1. Carbocation VII was found to possess a large degree
of conformational flexibility. We have indeed located some
structures of VII with different conformations of the deca-
lin ring and/or with the twist angle of the 2-hydroxybenzyl
group within 4.2 kcal mol–1. Structure VIIb, which is lower
in energy than VI by 10.4 kcal mol–1, is the most stable
conformer found for VII. However, the lowest-energy tran-
sition state located for the 1,2-hydride shift, TS2, is con-
nected to another conformer VIIc, which is higher in energy
than VIIb by 3.2 kcal mol–1. Interestingly, in the structure
of VIIc, a hydrogen bond-like interaction between the C10-
H and the phenolic oxygen atom was observed. The H···O
distance is 1.96 Å whereas this distance is 2.88 Å for VIIa
and 2.58 Å for VIIb. TS2 is the transition state leading to
structure VIIIa, a conformer of the third carbocation VIII,
which is lower in energy than VIIc by 1.2 kcal mol–1. Al-
though VIIIa is higher in energy than VIIb, we have found
another conformer VIIIb that is more stable than VIIb by
1.2 kcal mol–1. By comparing energies for the most stable
conformer of each carbocation, the order of the cation sta-
bility was estimated to be VIϽVIIϽVIII. Moreover, the
interatomic distance between the C10 carbocation centre
and the phenolic oxygen atom in VIIIb is only 2.64 Å. This
value is much smaller than the sum (3.22 Å) of the van der
1
2
NaBH4 (2 equiv.), EtOH,
–20 to 0 °C, 2 h
l-Selectride® (2 equiv.),
EtOH,
–20 to 0 °C, 2 h
LiAlH(tBuO)3 (5 equiv.),
THF,
–20 to 0 °C, 2 h
H2 (1 atm), PtO2, EtOH/
CH2Cl2,
60
28
2:1
40
38
82
61
1:1
3
4
5
1:16
1:2
31
r.t., 12 h
[a] Isolated yield.
Figure 2. Selected NOESY correlations for ketone 13.
As previously mentioned, we demonstrated the feasibility
of our devised strategy toward (+)-stachyflin (1) through
the synthesis of tetracyclic model compound 2b. Thus, we
next undertook the total synthesis of (+)-1 relying on our
results from the model studies.
Total Synthesis of (+)-Stachyflin (1)
Synthetic Plan: Our synthetic plan for (+)-stachyflin (1)
Waals radii of the carbon (1.70 Å) and oxygen (1.52 Å) is outlined in Scheme 6. Target molecule 1 was to be derived
atoms.[25] The oxygen atom in VIIIb, therefore, is poised to from epoxide I through an acid-induced cascade reaction
attack the C10 carbocation centre. These calculations sug- (Scheme 1). Epoxide I was to be prepared through the cou-
gest that the cascade reaction proceeded smoothly in an pling reaction of 5 with bromide IX accessible from com-
almost unhindered manner
mercially available 3,5-dihydroxybenzoic acid (14).
2952
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 2948–2957

Enantioselective Total Synthesis of (+)-Stachyflin
Scheme 7. Synthesis of intermediate 20. a) CO2, KHCO3, glycerol,
180 °C; b) Me2SO4, KHCO3, acetone, reflux, 64% (2 steps); c)
NaBH4, THF/H2O, r.t.; d) 2,2-dimethoxypropane, pTsOH, r.t.; e)
MeI, K2CO3, acetone, r.t., 79% (3 steps); f) NBS, MeCN, 0 °C; g)
CuCN, DMF, 120 °C, 66% (2 steps); h) H2 (1 atm), PtO2, EtOH/
CHCl3, r.t.; i) NaOMe, MeOH, r.t., 98% (2 steps). NBS = N-bro-
mosuccinimide, DMF = N,N-dimethylformamide.
Scheme 6. Synthetic plan for (+)-stachyflin (1); P = protective
group.
Synthesis of Segment 26
We initially pursued the synthesis of segment 26 [corre-
sponding to IX (P = 3,4DMB)] the coupling partner of 5,
starting from resorcinol 14 (Scheme 7, Scheme 8, and
Scheme 9). The Kolbe–Schmitt reaction[28] of 14 followed
by methyl esterification provided dimethyl ester 15 (64%
yield in two steps). Compound 15 was then converted to
acetonide 17 (79% overall yield) via methyl ester 16 by se-
lective reduction of the C4 methyl ester function to its cor-
responding hydroxymethyl congener, acetonide formation
and methyl etherification. Subsequent regioselective bromi-
nation at C2 in 17 followed by cyanation with CuCN pro-
vided nitrile 18 (66% yield in two steps). Isoindolinone 20
was produced by hydrogenation [H2 (1 atm), PtO2, EtOH/
CHCl3, r.t.] of 18 and subsequent lactamization of the re-
sulting ammonium salt 19 following treatment with Na-
OMe in MeOH at r.t. (98% yield in two steps). Because
compound 20 showed low solubility in organic solvents, the
protection of the lactam amide moiety was necessary to ad-
vance the projected synthesis. Therefore, we undertook pre-
liminary studies to select a promising N-protective group
for 20. To this end, some benzyl-type protective groups such
as benzyl (Bn), p-methoxybenzyl (PMB) and 3,4-dimeth-
oxybenzyl (3,4DMB), all of which were expected to be re-
moved under almost neutral conditions, were nominated as
model substrates (22a–c, Scheme 8), and the deprotection
of these protective groups was examined. Three types of N-
protected model substrates 22a–c were prepared from 20
(70–72% overall yields) through the agency of acetonides
21a–c by N-protection followed by an exchange of the diol
protective group.
Scheme 8. Synthesis of model substrates 22a–c and examination of
their N-deprotection leading to 23. a) R–Cl (R = Bn, PMB or
3,4DMB), NaN(SiMe3)2, Bu4NI, THF, 0 °C to r.t., 82% for 21a,
79% for 21b, 78% for 21c; b) 1 m HCl, THF, r.t.; c) Ac2O, Et3N,
DMAP, CH2Cl2, 88% (2 steps) for 22a, 90% (2 steps) for 22b,c; d)
see Table 3. Bn = benzyl, PMB = p-methoxybenzyl, 3,4DMB = 3,4-
dimethoxybenzyl, Ac = acetyl, DMAP = 4-(dimethylamino)pyr-
idine.
Scheme 9. Synthesis of intermediate 26. a) 1 m HCl, THF, r.t.; b)
TBSCl, imidazole, DMF, r.t., 75% (2 steps); c) 48% aq. HF,
MeCN, 0 °C; d) CBr4, PPh3, CH2Cl2, r.t., 86% (2 steps). TBS =
tert-butyldimethylsilyl.
After preparing model substrates 22a–c, we next investi-
gated the N-deprotection of these compounds under several some oxidizing agents. The use of ammonium cerium(IV)
conditions (22a–cǞ23 in Scheme 8 and Table 3). Initial nitrate (CAN) afforded a moderate yield (53%) of the de-
attempts to achieve N-deprotection of 22a (P = Bn) under sired product 23 (Table 3, Entry 3), whereas poor results
standard conditions [H2 (1 atm), 10% Pd/C or Raney Ni in were obtained using phenyliodine(III) bis(trifluoroacetate)
MeOH or EtOH] were unsuccessful (Table 3, Entries 1 and (PIFA)
or
2,3-dichloro-5,6-dicyano-1,4-benzoquinone
2) and the starting material was recovered unchanged. The (DDQ) (Table 3, Entries 4 and 5). Eventually, the best result
N-deprotection of 22b (P = PMB) was next examined using was obtained when 22c (P = 3,4DMB) was treated with
Eur. J. Org. Chem. 2011, 2948–2957
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2953

J. Sakurai, T. Kikuchi, O. Takahashi, K. Watanabe, T. Katoh
FULL PAPER
PIFA (Table 3, Entry 7), producing desired lactam 23 in
high yield (83%). It is noteworthy that this is the first exam-
ple of 3,4DMB cleavage from a lactam amide moiety with
PIFA.[29] The use of other oxidizing agents such as CAN,
phenyliodine(III) diacetate (PIDA) or DDQ instead of
PIFA resulted in either a lower yield (55%) of 23 (Table 3,
Entry 6) or recovery of starting material (Table 3, Entries 8
and 9).
Table 3. Examination of N-deprotection of 22a–c.
Entry Substrate Conditions
Yield[a] [%]
of 23
1
2
3
4
5
6
7
8
9
22a
22a
22b
22b
22b
22c
22c
22c
22c
H2 (1 atm), 10% Pd/C, MeOH, r.t., 48 h n.r.[b]
H2 (1 atm), Raney Ni, EtOH, r.t., 48 h
CAN (5 equiv.), MeCN/H2O, r.t., 1 h
PIFA (10 equiv.), CH2Cl2, r.t., 24 h
n.r.[b]
53
5[c]
DDQ (5 equiv.), CH2Cl2/H2O, r.t., 24 h n.r.[b]
CAN (5 equiv.), MeCN/H2O, r.t., 1 h
PIFA (10 equiv.), CH2Cl2, r.t., 3 h
PIDA (10 equiv.), CH2Cl2, r.t., 24 h
55
83
n.r.[b]
DDQ (5 equiv.), CH2Cl2/H2O, r.t., 24 h n.r.[b]
[a] Isolated yield. [b] No reaction. [c] 88% yield of recovered 22b.
CAN = ammonium cerium(IV) nitrate, PIFA = phenyliodine(III)
bis(trifluoroacetate), DDQ = 2,3-dichloro-5,6-dicyano-1,4-benzo-
quinone, PIDA = phenyliodine(III) diacetate.
After selecting a reliable protective group for the lactam
amide moiety, we further investigated the synthesis of seg-
ment 26 (Scheme 9). Thus, intermediate 21c was trans-
formed into bis-TBS ether 24 (75% yield in two steps) by
acetonide cleavage and subsequent silylation of the two lib-
erated hydroxy groups. Regioselective desilylation (48% aq.
HF, MeCN, 0 °C) of 24 followed by bromination of re-
sulting benzylic alcohol 25 afforded requisite bromide 26
(86% yield in two steps), which represented key segment IX
(P = 3,4DMB) in Scheme 6.
Scheme 10. Synthesis of epoxide 34. a) Li, liq. NH3/THF, –78 to
–30 °C; isoprene, H2O, 26, –30 °C to r.t., 76%; b) TBAF, THF,
0 °C to r.t., 88%; c) Ph3P+CH3Br–, tBuOK, benzene, reflux, 86%;
d) 4 m HCl, THF, r.t., 98%; e) H2 (1 atm), 10% Pd/C, Et3N/MeOH
50:1, r.t., 96%; f) Ph3P+CH3Br–, tBuOK, benzene, reflux, 74%; g)
RhCl3·3H2O, EtOH, reflux, 100%; h) mCPBA, NaHCO3, CH2Cl2,
0 °C to r.t., 86% (α-/β-epoxide, 7:1). TBAF = tetrabutylammonium
fluoride.
Completion of Total Synthesis of (+)-Stachyflin (1)
With requisite segment 26 in hand, we next performed
the synthesis of (+)-stachyflin (1) (Scheme 10 and
Scheme 11) by employing a reaction sequence similar to
After synthesizing epoxide 34, our efforts were directed
that used to construct tetracyclic model compound 2b towards the completion of total synthesis of (+)-1. To this
(Scheme 3 and Scheme 4, Table 1 and Table 2). Thus, the end, as shown in Scheme 11, the key acid-induced cascade
crucial reductive alkylation of enone 5 (Ͼ 99% ee) with epoxide-opening/rearrangement/cyclization reaction of 34
26 under Birch conditions proceeded smoothly and cleanly, (α-/β-epoxide, 7:1) was performed under the same condi-
producing expected coupling product 27 in 76% yield as a tions optimized in model studies (Table 1, Entries 1 and 2),
single diastereomer. Cleavage of the TBS group in 27 fol- resulting in the formation of desired products 35a (C3 α-
lowed by the Wittig methylenation of hemiacetal 28 pro- OH, 66%) and 35b (C3 β-OH, 9%) which were readily sep-
vided exo-olefin 29 (79% yield in two steps). In order to arated by silica gel column chromatography. The subse-
establish the C8 stereochemistry, the ethylene acetal moiety quent inversion of configuration at C3 of 35a was achieved
in 29 was first removed and resulting ketone 30 was sub- using Dess–Martin oxidation (94%) followed by LiAlH-
jected to stereoselective hydrogenation, to afford desired (tBuO)3 reduction of the resulting ketone 36. This sequence
product 31 in 96% yield as a single stereoisomer. The subse- of steps furnished desired product 35b in 96% yield with
quent Wittig methylenation of 31 followed by the isomeris- complete stereoselectivity. Finally, compound 35b was suc-
ation of the olefinic double bond furnished endo-olefin 33 cessfully converted to (+)-stachyflin (1) using a two-step op-
(74% yield in two steps). Epoxidation of 33 afforded desired eration involving cleavage of the 3,4DMB group (54%) with
epoxide 34 in 86% yield as an inseparable mixture of α- and PIFA under previously described conditions (Table 3, En-
1
β-epoxides (7:1 by 400 MHz H NMR).
try 7) followed by cleavage of the O-methyl moiety in the
2954
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 2948–2957

Enantioselective Total Synthesis of (+)-Stachyflin
Scheme 12. Alternative synthetic plan for (+)-stachyflin (1).
Synthesis of Segment 40
As shown in Scheme 13, segment 40, the coupling part-
ner of 5, was prepared starting from intermediate 17 by
employing a reaction sequence similar to that previously
described (Scheme 9). Thus, acetonide cleavage of 17 fol-
lowed by the silylation of the two liberated hydroxy groups
in the resulting diol provided bis-TBS ether 41 (81% yield
in two steps). The subsequent regioselective desilylation of
41 (80%) and bromination of resulting alcohol 42 furnished
requisite bromide 40 (96%).
Scheme 11. Completion of the synthesis of (+)-stachyflin (1). a)
BF3·Et2O, CH2Cl2, –40 °C to r.t., 66% for 35a, 9% for 35b; b)
Dess–Martin periodinane, CH2Cl2, r.t., 94%; c) LiAlH(tBuO)3,
THF, –20 °C, 96%; d) PIFA, CH2Cl2, r.t., 54%; e) nBuSLi, HMPA,
110 °C, 80%.
resulting lactam 37 (80%). The spectroscopic properties
1
(IR, H and 13C NMR, HRMS) of synthetic 1 were iden-
tical to those of natural 1. The optical rotation of synthetic
1 {[α]2D4 = +133.3 (c = 0.46 in MeOH)} also showed good
agreement with that of natural 1 {ref.[4b] [α]D24 = +138.7 (c
= 1.00 in MeOH)}.
Alternative Approach to (+)-Stachyflin (1): Synthesis of
Shionogi’s Racemic Late Intermediate 38 in Optically Pure
Form
Scheme 13. Synthesis of intermediate 40. a) 1 m HCl, THF, r.t.; b)
TBSOTf, Et3N, CH2Cl2, 0 °C, 81% (2 steps); c) 48% aq. HF,
MeCN, 0 °C, 80%; d) CBr4, Ph3P, CH2Cl2, 0 °C, 96%.
Although we successfully achieved the enantioselective
Synthesis of Tetracyclic Compound 38: With segment 40
total synthesis of (+)-stachyflin (1), we further investigated in hand, we next investigated the synthesis of tetracyclic
an alternative approach to (+)-1 in order to illustrate the ketone 38 (Scheme 14) which represents an optically pure
flexibility of our devised cascade reaction.
version of Shionogi’s racemic late intermediate. The synthe-
Synthetic Plan: Our alternative synthetic plan is outlined sis of 38 was performed starting from 5 by employing a
in Scheme 12, which features the early construction of the reaction sequence similar to that described for the synthesis
tetracyclic core structure (ABCD ring) followed by the an- of 36 (Scheme 10 and Scheme 11). The coupling reaction of
nulation of the lactam moiety (E ring). This approach has 5 with 40 under Birch conditions proceeded smoothly and
the advantage that protection and deprotection steps for the resulted in the formation of hemiacetal 43 (66%) as a single
lactam amide moiety are not necessary. The key acid-in- diastereomer. In this reaction, conveniently, the O-TBS
duced cascade reaction of epoxide 39 was expected to de- group was removed presumably during the course of reac-
liver the tetracyclic core 38 upon C3 oxidation of the re- tion or work-up. Compound 43 was then converted to
sulting product. A racemic version of 38 is the key interme- ketone 46 using a three-step operation involving Wittig
diate for Shionogi’s total synthesis of (Ϯ)-stachyflin (1).[13] methylenation (61%), ethylene acetal hydrolysis (98%) and
Epoxide 39, in turn, was to be prepared through the hydrogenation (96%). After the Wittig methylenation of 46
stereocontrolled reductive alkylation of enone 5 with (74%), olefin isomerisation followed by epoxidation pro-
bromide 40.
duced epoxide 39 (α-/β-epoxide, 8:1) in 86% overall yield
Eur. J. Org. Chem. 2011, 2948–2957
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2955

J. Sakurai, T. Kikuchi, O. Takahashi, K. Watanabe, T. Katoh
FULL PAPER
as an inseparable mixture. The key acid-induced cascade an innovative acid-induced cascade epoxide-opening/re-
reaction of 39 (α-/β-epoxide, 8:1), under the previously de- arrangement/cyclization reaction to stereoselectively con-
termined conditions, led to the formation of desired prod- struct the requisite pentacyclic ring system in one step
ucts 49a (C3 α-OH, 75%) and 49b (C3 β-OH, 11%) follow- (IǞ[IIǞIIIǞIV]ǞV, Scheme 1; 34Ǟ35a and 35b,
ing separation by silica gel column chromatography. Subse- Scheme 11). DFT calculations support the rationalization
quent Dess–Martin oxidation of 49a and 49b provided of the observed cascade reaction (Scheme 5). An alternative
ketone 38 in 94% and 92% yields, respectively. A racemic synthetic approach to (+)-1 was also investigated to demon-
version of 38 was previously converted to (Ϯ)-stachyflin (1) strate the flexibility of our devised cascade reaction
by the Shionogi group[13] using a five-step sequence of reac- (39Ǟ49a and 49b, Scheme 14). On the basis of the present
tions.
study, we are currently synthesizing stachyflin analogues
with the aim of exploring structure–activity relationships
and will report our results in the future.
Experimental Section
Supporting Information (see footnote on the first page of this arti-
cle): For information on experimental procedures and characteriza-
1
tion data for all new compounds along with copies of H and 13C
NMR spectra, see the Supporting Information.
Acknowledgments
We are grateful to Dr. Kazuyuki Minagawa, Shionogi & Co., Ltd.,
for providing us with natural (+)-stachyflin (1). We also thank Prof.
Masayuki Inoue, University of Tokyo, for useful discussions and
suggestions. This work was supported by the Ministry of Educa-
tion, Culture, Sports, Science and Technology of Japan (MEXT):
a Grant-in-Aid for Scientific Research on Priority Area “Creation
of Biologically Functional Molecules” (grant numbers 17035073,
17035073 and 18032065), a Grant-in-Aid for Scientific Research
(C) (grant numbers 18590013 and 21590018) and a Grant-in-Aid
for Matching Fund Subsidy for Private Universities.
[1] a) G. Neumann, T. Noda, Y. Kawaoka, Nature 2009, 459, 931–
939; b) J. Cohen, M. Enserink, Science 2009, 324, 1496–1497.
[2] a) P. E. Aldrich, E. C. Hermann, W. E. Meier, M. Paulshock,
W. W. Prichard, J. A. Snyder, J. C. Watts, J. Med. Chem. 1971,
14, 535–543; b) E. De Clercq, Nat. Rev. Drug Discovery 2006,
5, 1015–1025.
[3] a) M. Schmidtke, R. Zell, K. Bauer, A. Krumbholtz, C.
Schrader, J. Suess, P. Wutzler, Intervirology 2006, 49, 286–293;
b) A. Moscona, N. Engl. J. Med. 2009, 360, 953–956.
[4] a) T. Kamigauchi, T. Fujiwara, H. Tani, Y. Kawamura, I. Hor-
ibe (Shionogi & Co., Ltd., Japan), PCT WO 9711947 A1, 03
April, 1997; b) K. Minagawa, S. Kouzuki, J. Yoshimoto, Y.
Kawamura, H. Tani, T. Iwata, Y. Terui, H. Nakai, S. Yagi, N.
Hattori, T. Fujiwara, T. Kamigauchi, J. Antibiot. 2002, 55, 155–
164; c) K. Minagawa, S. Kouzuki, T. Kamigauchi, J. Antibiot.
2002, 55, 165–171.
[5] a) J. Yoshimoto, S. Yagi, J. Ono, K. Sugita, N. Hattori, T. Fu-
jioka, T. Fujiwara, H. Sugimoto, N. Hashimoto, J. Pharm.
Pharmacol. 2000, 52, 1247–1255; b) S. Yagi, J. Ono, J. Yoshim-
oto, K. Sugita, N. Hattori, T. Fujioka, T. Fujiwara, H. Sugim-
oto, K. Hirano, N. Hashimoto, Pharm. Res. 1999, 16, 1041–
1046.
Scheme 14. Alternative synthesis of (+)-stachyflin (1). a) Li, liq.
NH3/THF, –78 to –30 °C; isoprene, H2O, 40, –30 °C to r.t., 66%;
b) Ph3P+CH3Br–, tBuOK, benzene, reflux, 61%; c) 1 m HCl, THF,
r.t., 98%; d) H2 (1 atm), 10% Pd/C, Et3N/MeOH 50:1, r.t., 96%;
e) Ph3P+CH3Br–, tBuOK, benzene, reflux, 74%; f) RhCl3·3H2O,
EtOH, reflux, 100%; g) mCPBA, NaHCO3, CH2Cl2, 0 °C to r.t.,
86% (α-/β-epoxide, 8:1); h) BF3·Et2O, CH2Cl2, –40 °C to r.t., 75%
for 49a, 9% for 49b; i) Dess–Martin periodinane, CH2Cl2, r.t., 94%
for 49a, 92% for 49b.
[6] a) J. Yoshimoto, M. Kakui, H. Iwasaki, H. Sugimoto, T. Fuji-
wara, N. Hattori, Microbiol. Immunol. 2000, 44, 677–685; b) J.
Yoshimoto, M. Kakui, H. Iwasaki, T. Fujiwara, H. Sugimoto,
N. Hattori, Arch. Virol. 1999, 144, 865–878.
[7] R. G. Douglas Jr., N. Eng. J. Med. 1990, 322, 443–450.
[8] S. M. Wintermeyer, M. C. Nahata, Ann. Pharmacother. 1995,
29, 299–310.
Conclusions
We accomplished the first enantioselective total synthesis
of (+)-stachyflin (1), a potential anti-influenza A virus
agent exhibiting a unique mechanism of action, in a highly
efficient and convergent manner starting from the (+)-5-
methyl Wieland–Miescher ketone 5. The method features
[9] M. von Itzstein, W. Y. Wu, G. B. Kok, M. S. Pegg, J. C. Dya-
son, B. Jin, T. Van Phan, M. L. Smythe, H. F. White, S. W. Oli-
2956
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 2948–2957

Enantioselective Total Synthesis of (+)-Stachyflin
ver, P. M. Colman, J. N. Varghese, D. M. Ryan, J. M. Woods,
R. C. Bethell, V. J. Hotham, J. M. Cameron, C. R. Penn, Na-
ture 1993, 363, 418–423.
C. U. Kim, W. Lew, M. A. Williams, H. Liu, L. Zhang, S. Swa-
minathan, N. Bischofberger, M. S. Chen, D. B. Medel, C. Y.
Tai, W. G. Laver, R. C. Stevens, J. Am. Chem. Soc. 1997, 119,
681–690.
Y. S. Babu, P. Chand, S. Bantia, P. Kotian, A. Dehghani, Y. El-
Kattan, T. H. Lin, T. L. Hutchison, A. J. Elliot, C. D. Parker,
S. L. Ananth, L. L. Horn, G. W. Laver, A. Montgomery, J.
Med. Chem. 2000, 43, 3482–3486.
H.-P. Hsieh, J. T.-A. Hsu, Curr. Pharm. Des. 2007, 13, 3531–
3542.
T. Taishi, S. Takechi, S. Mori, Tetrahedron Lett. 1998, 39, 4347–
4350.
D. T. Ngoc, M. Albicker, L. Schneider, N. Cramer, Org. Biomol.
Chem. 2010, 8, 1781–1784.
[20] This reagent system has been known to be effective for steri-
cally hindered ketones, see: L. Fitjer, U. Quabeck, Synth. Com-
mun. 1985, 15, 855–864.
[21] When ethylene acetal 8 was subjected to hydrogenation, the
stereoselectivity at C8 was reduced (β-/α-Me: ca. 2:1); this was
probably caused by an unfavourable conformational change of
the decalin ring. Similar examples of this phenomenon have
been reported, see: a) P. Stahl, L. Kissau, R. Mazitschek, A.
Huwe, P. Furet, A. Giannis, H. Waldmann, J. Am. Chem. Soc.
2001, 123, 11586–11593; b) E. P. Locke, S. M. Hecht, Chem.
Commun. 1996, 2717–2718.
[10]
[11]
[22] a) T. Ling, E. Poupon, E. J. Rueden, S. H. Kim, E. Theodo-
rakis, J. Am. Chem. Soc. 2002, 124, 12261–12267; b) J. An,
D. F. Wiemer, J. Org. Chem. 1996, 61, 8775–8779.
[23] a) P. A. Aristoff, A. W. Harison, A. M. Huber, Tetrahedron
Lett. 1984, 25, 3955–3958; b) S. S. Welch, A. S. C. P. Pao, Tetra-
hedron Lett. 1977, 18, 505–508; c) G. I. Feutrill, R. N. Mirring-
ton, Aust. J. Chem. 1972, 25, 1719–1723; d) P. A. Bartlett, W. S.
Johnson, Tetrahedron Lett. 1970, 11, 4459–4462.
[12]
[13]
[14]
[15] M. Nakatani, M. Nakamura, A. Suzuki, M. Inoue, T. Katoh,
Org. Lett. 2002, 4, 4483–4486.
[24] Spartan ’08, Wavefunction, Inc., Irvine, CA.
[16] K. Watanabe, J. Sakurai, H. Abe, T. Katoh, Chem. Commun.
2010, 46, 4055–4057.
[25] B. Douglas, D. McDaniel, J. Alexander, Concepts and Models
of Inorganic Chemistry, 3rd ed., Wiley, New York, 1994.
[26] Initial attempts to invert the C3 hydroxy group by employing
the Mitsunobu procedure were unsuccessful, presumably due
to steric factors.
[17] T. Oguchi, K. Watanabe, K. Ohkubo, H. Abe, T. Katoh, Chem.
Eur. J. 2009, 15, 2826–2845.
[18] J. L. Kelly, J. A. Linn, J. W. T. Selway, J. Med. Chem. 1989, 32,
1575–1763.
[27] Similar NOESY correlations have been reported for the struc-
[19] This type of reductive alkylation, originally developed by Stork
et al. in the 1960s (for a representative work; see: G. Stork, P.
Rosen, N. Goldman, R. V. Coombs, J. Tsuji, J. Am. Chem. Soc.
1965, 87, 275–286), has been successfully utilized as a crucial
step in our total synthesis of marine sesquiterpene quinones
and hydroquinones; see: J. Sakurai, T. Oguchi, K. Watanabe,
H. Abe, S. Kanno, M. Ishikawa, T. Katoh, Chem. Eur. J. 2008,
14, 829–837.
tural determination of stachyflin (1); see ref.[4b]
[28] C. M. Harris, J. J. Kibby, J. R. Fehlner, A. B. Raabe, T. A. Bar-
ber, T. M. Harris, J. Am. Chem. Soc. 1979, 101, 437–445.
[29] General application of this deprotection method is now under
investigation in our laboratory and will be reported in due
course.
Received: February 6, 2011
Published Online: April 18, 2011
Eur. J. Org. Chem. 2011, 2948–2957
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2957