Total synthesis and biological assessment of (-)-Exiguolide and analogues

Article abstract of DOI:10.1002/chem.201003135

We describe herein an enantioselective total synthesis of (-)-exiguolide, the natural enantiomer. The methylene bis(tetrahydropyran) substructure was efficiently synthesized by exploiting olefin cross-metathesis for the assembly of readily available acycl

Full text of DOI:10.1002/chem.201003135

DOI: 10.1002/chem.201003135
Total Synthesis and Biological Assessment of (À)-Exiguolide and Analogues
Haruhiko Fuwa,*[a] Takaya Suzuki,[b] Hiroshi Kubo,[b] Takao Yamori,[c] and
Makoto Sasaki[a]
Abstract: We describe herein an enan-
tioselective total synthesis of (À)-exi-
guolide, the natural enantiomer. The
methylene bis(tetrahydropyran) sub-
structure was efficiently synthesized by
exploiting olefin cross-metathesis for
the assembly of readily available acy-
clic segments and intramolecular oxa-
conjugate cyclization and reductive
etherification for the formation of the
tetrahydropyran rings. The 20-mem-
bered macrocyclic framework was con-
structed in an efficient manner by
means of Julia–Kocienski coupling and
Yamaguchi macrolactonization. Finally,
the (E,Z,E)-triene side chain was intro-
duced stereoselectively via Suzuki–
Miyaura coupling to complete the total
synthesis. Assessment of the growth in-
hibitory activity of synthetic (À)-exi-
cancer cell lines elucidated for the first
time that this natural product is an ef-
fective antiproliferative agent against
the NCI-H460 human lung large cell
carcinoma and the A549 human lung
adenocarcinoma cell lines. Moreover,
we have investigated structure–activity
relationships of (À)-exiguolide, which
elucidated that the C5-methoxycarbo-
nylmethylidene group and the length
of the side chain are important for the
potent activity.
guolide against
a panel of human
Keywords: antiproliferative activi-
ty · macrolides · natural products ·
olefin metathesis · total synthesis
Introduction
Macrolide natural products originated from marine inverte-
brates and symbiotic bacteria represent a rich source of
structurally novel and potent anticancer chemotherapeutic
agents.[1,2] (À)-Exiguolide (1, see below) was isolated from
the methanol extract of the marine sponge Geodia exigua
Thiele (order Astrophorida, family Geodiidae), collected off
Amami-Oshima, Japan, by Ohta, Ikegami, and co-workers.[3]
The gross structure including relative stereochemistry of this
naturally occurring substance was established by the combi-
nation of extensive 2D NMR studies and conformational
3
analysis on the basis of JH,H values and NOE correlations.
The assigned relative stereochemistry was also confirmed by
J-based configurational analysis.[4] In 2008, Lee and co-work-
ers reported the first total synthesis of the unnatural enan-
tiomer, (+)-1, which unambiguously determined the abso-
lute configuration of this natural product.[5] The molecular
architecture of (À)-1, characterized by the 20-membered
macrolide framework embedded with the methylene bis(te-
trahydropyran) substructure, resembles to that of antineo-
plastic marine natural products, the bryostatins.[6] In addi-
tion, it has been reported that (À)-1 specifically inhibits fer-
tilization of sea urchin (Hemicentrotus pulcherrimus) game-
tes but not embryogenesis of the fertilized egg. These bio-
logical and structural aspects of (À)-1 have led to an
assumption that (À)-1 represents a structurally simplified
analogue of the bryostatins by Nature.[7] However, further
details on the biological activity of (À)-1 await elucidation.
Motivated by the elaborated structure and intriguing bio-
logical activity, we embarked on the enantioselective total
synthesis of (À)-1, the full details of which are described
herein.[8–10] Moreover, we report on assessment of the cell
growth inhibitory activity and structure–activity relation-
ships of (À)-1 against a panel of human cancer cell lines.
[a] Prof. Dr. H. Fuwa, Prof. Dr. M. Sasaki
Graduate School of Life Sciences, Tohoku University
2-1-1 Katahira, Aoba-ku, Sendai 980-8577 (Japan)
Fax : (+81)22-217-6214
E-mail: hfuwa@bios.tohoku.ac.jp
[b] Dr. T. Suzuki, Prof. Dr. H. Kubo
Tohoku University Graduate School of Medicine
2-1 Seiryo-machi, Aoba-ku, Sendai 980-8575 (Japan)
[c] Dr. T. Yamori
Division of Molecular Pharmacology
Cancer Chemotherapy Center
Japanese Foundation for Cancer Research
3-8-31 Ariake, Koto-ku, Tokyo 135-8550 (Japan)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201003135.
2678
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2678 – 2688

FULL PAPER
Results and Discussion
etherification[17] of 6 would then afford methylene bis(tetra-
hydropyran) 5 stereoselectively.
Our initial synthetic blueprint toward (À)-1 is summarized
in Scheme 1. We envisioned that the apparently sensitive
(E,Z,E)-triene side chain should be introduced at a late
The synthesis of hydroxy olefin 8, illustrated in Scheme 2,
commenced with MPM protection of the known homoallylic
alcohol 10,[18] readily prepared in multi-gram quantities. Oxi-
Scheme 2. Synthesis of hydroxy olefin 8. MPM=p-methoxyphenylmethyl,
Tf=trifluoromethanesulfonyl, RT=room temperature, NMO=N-meth-
ylmorpholine-N-oxide, d.r.=diastereomer ratio, DDQ=2,3-dichloro-5,6-
dicyano-1,4-benzoquinone.
dative cleavage of the double bond then gave aldehyde 11
in 69% yield for the two steps. Chelation-controlled diaste-
reoselective allylation of 11 (allylSiMe3, MgBr2·OEt2,
CH2Cl2, 08C) produced homoallylic alcohol 12 in 84% yield
1
as a single stereoisomer (judged by 600 MHz H NMR). Si-
lylation of the resultant alcohol with TIPSOTf/2,6-lutidine
followed by oxidative removal of the MPM group delivered
hydroxy olefin 8 in 97% yield for the two steps.
The synthesis of enone 9 started with Brown asymmetric
allylation[19] of the known aldehyde 13,[20] prepared from
(S)-Roche ester in five steps (Scheme 3). The resultant ho-
moallylic alcohol was homologated via olefin cross-metathe-
sis with methyl acrylate using the Grubbs second-generation
catalyst (G-II)[21] to give enoate 14 in 72% yield for the two
steps. The double bond within 14 was saturated by hydroge-
nation and the ester functionality was transformed into
Weinreb amide (Me3Al, MeONHMe·HCl).[22] Subsequent si-
lylation of the hydroxy group afforded 15 in 92% yield
(three steps). Treatment of 15 with vinyl lithium,[23] generat-
Scheme 1. Initial synthetic blueprint toward (À)-1. Bn=benzyl,
TBDPS=tert-butyldiphenylsilyl, TIPS=triisopropylsilyl, TES=triethyl-
silyl.
stage of the total synthesis and this would be achieved by
means of Suzuki–Miyaura coupling[11] of vinyl boronate 2
and vinyl iodide 3. We thought that it would be possible to
forge the C16–C17[12] double bond via chemoselective ring-
closing olefin metathesis (RCM)[5,13] of triene 4, which could
ed in situ from tetra
in 96% yield.
ACHTUNGTRENN(UNG vinyl)tin and MeLi, furnished enone 9
With the requisite segments available, we focused our at-
tention to the construction of methylene bis(tetrahydropyr-
an) 5 (Scheme 4). After several preliminary experiments, we
found that hydroxy olefin 8 and enone 9 could be assembled
efficiently through olefin cross-metathesis using 10 mol% of
the Hoveyda–Grubbs second-generation catalyst (HG-II)[24]
in CH2Cl2 at 358C, affording hydroxy enone 7 in 93% yield
as a single stereoisomer. Exposure of 7 to 0.2 equiv of
KOtBu in THF at 08C for 1 h cleanly effected intramolecu-
lar oxa-conjugate cyclization to deliver silyloxy ketone 6 in
95% yield as a single stereoisomer (judged by 600 MHz
be accessible from methylene bis(tetrahydropyran)
5
through esterification. We planned to build the methylene
bis(tetrahydropyran) substructure of (À)-1 in a convergent
manner from readily available acyclic segments. Thus, olefin
cross-metathesis[14] of hydroxy olefin 8 and enone 9 would
produce hydroxy enone 7, which would undergo intramolec-
ular oxa-conjugate cyclization[15,16] to deliver silyloxy ketone
6
under thermodynamic control. Subsequent reductive
Chem. Eur. J. 2011, 17, 2678 – 2688
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2679

H. Fuwa et al.
domino olefin cross-metathesis/intramolecular oxa-conju-
gate cyclization.[26] Thus, microwave heating of a mixture of
8 (1.5 equiv) and 9 in the presence of HG-II (10 mol%) in
CH2Cl2 at 1008C for 30 min generated silyloxy ketone 6,
which without isolation was reacted with BF3·OEt2 and
Et3SiH (À60 to À158C over 50 min) to afford 5 in 89%
yield as an approximately 10:1 mixture of diastereomers at
the C9 stereogenic center. Thus, by taking advantage of the
superb chemoselectivity and bond-forming ability of olefin
metathesis, we were able to build up the complex methylene
bis(tetrahydropyran) substructure of (À)-1 from readily
available acyclic segments 8 and 9 in a very rapid and effi-
cient manner.
Having completed the methylene bis(tetrahydropyran)
substructure, triene 4 was next synthesized as depicted in
Scheme 6. Deprotection of the benzyl group within 5 by hy-
drogenolysis gave alcohol 16 in 90% yield. At this stage, the
minor C9 epimer could be removed by flash chromatogra-
phy on silica gel. Oxidation of 16 with Dess–Martin periodi-
nane[27] followed by Wittig methylenation of the derived al-
dehyde 17 provided olefin 18 in 87% yield for the two steps.
Selective cleavage of the TBDPS group under basic condi-
tions led to alcohol 19 in 90% yield. A two-stage oxidation
of 19 followed by esterification of the resultant carboxylic
acid 20 with alcohol 21 under Yamaguchi conditions[28] gave
rise to triene 4 in 95% yield for the three steps.
Scheme 3. Synthesis of enone 9. Ipc=isopinocampheyl, DMAP=4-dime-
thylaminopyridine.
We investigated RCM of triene 4 under a number of con-
ditions (Table 1). We quickly recognized that triene 4 is a
challenging substrate for RCM; the presence of the C15 and
C18 allylic methyl groups renders 4 less reactive toward
RCM,[29] while the vinyl iodide moiety reacts with rutheni-
um methylidene species, the actual active catalyst of the
Table 1. Screening of the reaction conditions for the RCM of triene 4.
Entry Catalyst
[mol%]
Conditions
Yield [%]
22 23
Recovered
4 [%]
Scheme 4. Construction of methylene bis(tetrahydropyran) 5.
1
2
3
4
5
G-II [30]
CH2Cl2 (2.5mm),
reflux, 2 d
toluene (5mm),
808C, 2 d
trace n.d.[a] n.d.[a]
1H NMR), presumably under thermodynamic control.[25] Fi-
nally, treatment of 6 with BF3·OEt2 in Et3SiH/CH2Cl2 1:5 at
À60 to À258C furnished methylene bis(tetrahydropyran) 5
in 98% yield as an approximately 10:1 mixture of diastereo-
mers at the C9 stereogenic center. At this stage, we estab-
lished the newly generated stereogenic centers of the major
diastereomer by NOE experiments as shown.
G-II [60]
trace n.d.[a] n.d.[a]
HG-II [60] toluene (5mm),
808C, 1 d
HG-II [90] toluene (5mm),
958C, 2 d
30
31
trace 18
19
trace 24–27
0
HG-II [60]
N
28–
52
Moreover, as illustrated in Scheme 5, we have examined
one-pot synthesis of 5 based on our recently developed
[a] n.d.=not determined.
2680
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2678 – 2688

(À)-Exiguolide and Analogues
FULL PAPER
toluene (5 mm) at 808C resulted
in decomposition of the catalyst
at an early stage of the reac-
tion, and very low conversion
(<5%) was observed under
these conditions (entries 1 and
2). By contrast, when 4 was
treated with more robust and
powerful HG-II (60 mol%) in
toluene (5 mm) at 808C for one
day, the desired macrocycle 22
was isolated in 30% yield along
with recovered 4 in 18% yield
(entry 3). The geometry of the
newly formed double bond was
confirmed to be E on the basis
of the large coupling constant
between H-16 and H-17 pro-
tons (3J16,17 =15.0 Hz). Under
more forcing the reaction con-
Scheme 5. One-pot synthesis of methylene bis(tetrahydropyran) 5.
ditions (90 mol% of HG-II, tol-
uene (5 mm), 958C), a significant amount of 23 was observed
as a byproduct (entry 4). Switching the solvent to 1,2-di-
chloroethane was beneficial to suppress the undesired col-
lapse of the vinyl iodide moiety. Thus, when the reaction
was carried out in the presence of HG-II (60 mol%) in 1,2-
dichloroethane (5 mm) at 758C for two days, macrocycle 22
was isolated in 52% yield and 4 was recovered in 24% yield
(entry 5). Unfortunately, this reaction was found to be capri-
cious; our attempts to reproduce the result under the opti-
mized conditions sometimes ended up in low mass recovery
and, for example, gave 22 in only 28% yield (4 was recov-
ered in 27% yield). We have also evaluated the use of Ti-
[30]
AHCTUNGTRENNUNG
(OiPr)4 or 2,6-dichloro-1,4-benzoquinone,[31] but these ad-
ditives did not improve the outcome of the present RCM.
We have also explored relay RCM[32] of triene 24 as
shown in Scheme 7. However, we found that triene 24 was
rapidly converted to triene 4 in 54% yield upon exposure to
G-II catalyst (30 mol%) in 1,2-dichloroethane (2 mm) at
608C, and only 8% of macrocycle 22 was isolated alongside.
Changing the catalyst to HG-II did not have any impact on
the outcome. These disappointing results could be explained
as summarized in Scheme 8. Thus, exposure of 24 to G-II or
HG-II would give ruthenium alkylidene A, which would un-
dergo RCM to deliver ruthenium alkylidene B via extrusion
of 3,4-dihydrofuran. Finally, RCM of the generated B would
afford macrocycle 22. However, the low reactivity of the
C16 double bond precluded this successful scenario at the
final stage. Actually, ruthenium alkylidene species B reacted
with triene 24 as a propagation catalyst itself to generate
ruthenium alkylidene A and triene 4, even though the reac-
tion was carried out under high-dilution conditions.[33]
Scheme 6. Synthesis of triene 4. DMP=Dess–Martin periodinane.
The unsatisfactory results of the RCM of 4 and 24 led us
to seek for an alternative scenario toward the macrocyclic
framework of (À)-1. We envisioned that the 20-membered
macrolactone 22 would be forged by means of macrolactoni-
zation of the corresponding hydroxy acid 25 (Scheme 9).
propagation step of RCM, to lose the iodine atom under
forcing conditions. Our initial attempts at the RCM of 4
using G-II (30–60 mol%) in CH2Cl2 (2.5 mm) at reflux or in
Chem. Eur. J. 2011, 17, 2678 – 2688
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2681

H. Fuwa et al.
Scheme 9. Revised synthesis plan toward macrolactone 22.
ence of propylene oxide[36] followed by exposure of the re-
sultant chloro-epoxide to LDA[37] provided propargylic alco-
hol 29 in 84% yield for the two steps. After conversion of
29 to the corresponding bromoalkyne with NBS in the pres-
ence of AgNO3, palladium-catalyzed hydrostannylation
Scheme 7. Unsuccessful relay RCM of triene 24.
(nBu3SnH, [PdACTHNUTRGNEUNG
(PPh3)4])[38] and subsequent iodolysis of the
The sterically encumbered C16–C17 double bond would be
formed via Julia–Kocienski olefination[34] of aldehyde 17 and
an anion derived from sulfone 26.
The synthesis of sulfone 26 commenced with Sharpless
asymmetric epoxidation of the known allylic alcohol 27,[35]
available in five steps from (S)-Roche ester in multi-gram
quantities (Scheme 10). This led to epoxy alcohol 28 in 89%
derived (E)-vinylstannane delivered (E)-vinyl iodide 30 as a
sole product in 94% yield for the three steps. Protection of
the hydroxy group within 30 as its MPM ether followed by
desilylation with TBAF gave alcohol 31 in 71% yield for
the two steps. Finally, Mitsunobu reaction[39] of 31 with 1-
phenyl-1H-tetrazole-5-thiol and subsequent molybdenum-
catalyzed peroxide oxidation[40] afforded sulfone 26 in 89%
yield (two steps).
yield as
a single stereoisomer (judged by 600 MHz
1H NMR). Chlorination of 28 with NCS/Ph3P in the pres-
Scheme 8. Plausible rationale for the outcome of the relay RCM of 24.
2682
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2678 – 2688

(À)-Exiguolide and Analogues
FULL PAPER
Table 2. Optimization of Julia–Kocienski olefination of aldehyde 17 and
sulfone 26.
Entry Base
Conditions
Yield E/Z[a] Recovery
[%]
of 17 [%]
1
2
3
4
5
KHMDS THF, À788C (1h) to RT 14
(21h)
7:1
39[b]
KHMDS DME, À558C (1h) to RT 11
13:1
62
(1.5h)
KHMDS THF/HMPA 4:1, À788C 15
(2.5h) to RT (3.5h)
>20:1 80
>20:1 23
LHMDS THF/HMPA 4:1, À788C 63
(2.5h) to RT (3.5h)
LHMDS THF/HMPA 4:1, À788C 80
(0.5h) to 08C (2h)
15:1
<5
1
[a] Estimated by H NMR analysis (600 MHz) of a purified mixture of E/
Z isomers. [b] Recovered as a 1:1 mixture of diastereomers at the C15
stereogenic center. KHMDS=potassium bis(trimethylsilyl)amide,
Scheme 10. Synthesis of sulfone 26. DET=diethyl tartrate, MS=molecu-
lar sieves, NCS=N-chlorosuccinimide, LDA=lithium diisopropylamide,
NBS=N-bromosuccinimide, TBAF=tetra-n-butylammonium fluoride,
DEAD=diethyl azodicarboxylate.
DME=1,2-dimethoxyethane,
HMPA=hexamethylphosphoramide,
LHMDS=lithium bis(trimethylsilyl)amide.
Julia–Kocienski olefination of aldehyde 17 and an anion
generated from sulfone 26 was investigated under several
conditions, as summarized in Table 2. Our initial attempts
involving the use of KHMDS as a base gave only poor yield
of the desired (E)-olefin 32. Thus, deprotonation of sulfone
26 with KHMDS in THF at À788C followed by addition of
aldehyde 17 and warming the reaction mixture to room tem-
perature over 22 h gave olefin 32 in 14% yield as an approx-
imately 7:1 mixture of E/Z isomers (entry 1). Although alde-
hyde 17 was recovered in 39% yield alongside, complete ep-
imerization of the C15 stereogenic center was observed by
1H NMR analysis. Monitoring the reaction by TLC analysis
suggested that the reaction stalled after initial several hours
and extended reaction time only resulted in decomposition
of unreacted materials. When the reaction was performed in
DME (À558C for 1 h then warmed to room temperature
over 1.5 h), the E/Z selectivity was improved to ca. 13:1,
while the product yield and mass recovery remained poor
(entry 2). The stereoselectivity and mass recovery could be
improved by running the reaction in THF/HMPA 4:1
(À788C for 2.5 h then warmed to room temperature over
3.5 h), giving olefin 32 as a sole E isomer in 15% yield
along with 80% recovery of aldehyde 17 with no sign of epi-
merization at the C15 stereogenic center (entry 3). Gratify-
ingly, the product yield was significantly enhanced upon
switching the base to LHMDS (entry 4). Thus, reaction of a
sulfone anion, generated in situ from sulfone 26 and
LHMDS, with aldehyde 17 in THF/HMPA 4:1 (À788C for
2.5 h then warmed to room temperature over 3.5 h) fur-
nished (E)-olefin 32 in 63% yield as a single stereoisomer,
along with recovered 17 (23%). Although the yield could be
further improved to 80% under the reaction conditions de-
scribed for entry 5, a slightly declined stereoselectivity was
observed in this case (E/Z ca. 15:1).
With the sterically encumbered C16–C17 double bond se-
cured, we proceeded to forge the 20-membered macrocycle
and complete the total synthesis of (À)-1 (Scheme 11). Se-
lective deprotection of the TBDPS group within 32 under
basic conditions gave alcohol 33 in 94% yield, which was
oxidized to the corresponding carboxylic acid via a two-
stage oxidation and then esterified with TMSCHN2 to pro-
vide methyl ester 34 in 94% yield for the three steps. Cleav-
age of the MPM group within 34 was best achieved by treat-
ment of 34 with BF3·OEt2 in Et3SiH/CH2Cl2 1:2 at 08C,
giving alcohol 35 in 89% yield.[41] After saponification of 35
with TMSOK,[42] macrolactonization of the resultant hy-
droxy
acid
under
Yamaguchi
conditions
(2,4,6-
Cl3C6H2COCl, Et3N, THF, room temperature; then added to
DMAP, toluene (final concentration=0.5 mm), 808C over
2 h) smoothly proceeded to afford macrolactone 22 in 94%
yield for the two steps. We observed competitive formation
of the corresponding dimer under higher concentrations
(above 1 mm). Deprotection of the TIPS group of 22 with
HF·pyridine followed by oxidation of the resultant alcohol
36 with Dess–Martin periodinane provided ketone 37 in
quantitative yield. Horner–Wadsworth–Emmons reaction of
Chem. Eur. J. 2011, 17, 2678 – 2688
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2683

H. Fuwa et al.
37 with chiral phosphonate 38 developed by Fuji and co-
workers[5,43] led to enoate 3 in 94% yield as an approximate-
ly 5:1 mixture of Z/E isomers. Fortunately, the desired (Z)-
isomer 3 could be isolated in a geometrically pure form in
75% yield after flash chromatography on silica gel. Finally,
stereoselective introduction of the (E,Z,E)-triene side chain
was accomplished by means of Suzuki–Miyaura coupling of
3 with (Z)-vinyl boronate 2[44] under the influence of a [Pd2-
ACHTUNGTRENNUNG
(dba)3]/Ph3As catalyst system in the presence of Ag2O[45] in
THF at room temperature.[46] Under these exceptionally
mild conditions, (À)-exiguolide (1) was isolated in 73%
yield as
a single stereoisomer (judged by 600 MHz
1H NMR). The spectroscopic data (1H, 13C NMR, HRMS)
were in full accordance with those of the natural product.
The specific rotation value of the synthetic (À)-exiguolide
([a]2D4 = À121.5 (c=0.22 in CHCl3)) slightly differed from
that of the natural product (lit.[3] [a]D25 =À92.5 (c=0.069 in
CHCl3)) but matched that of the synthetic (+)-exiguolide,
except for the sign of the rotation (lit.[5] [a]D25 =+119 (c=
0.11 in CHCl3)). The present total synthesis proceeded in 23
longest linear steps from the known aldehyde 13 (28 longest
linear steps from commercially available (S)-Roche ester).
With sufficient quantities of the synthetic material avail-
able, we evaluated the growth inhibitory activity of (À)-1
against a panel of 39 human cancer cell lines,[47,48] and a part
of the results is shown in Table 3 (for full details, see the
Table 3. The growth inhibitory activity of (À)-1 against selected human
cancer cell lines.[a]
[b]
[c]
Cell line
log GI50
GI50[b] [mm]
LC50 [mm]
NCI-H460
A549
SK-OV-3
MKN74
À8.00
À6.19
À6.15
À6.16
0.01
0.65
0.70
0.69
>100
>100
>100
>100
[a] For details of the assay procedure, see reference [47]. [b] GI50 =con-
centration that induces 50% growth inhibition. [c] LC50 =concentration
that induces 50% cell death.
Supporting Information). Importantly, (À)-1 effectively in-
hibited in vitro proliferation of the NCI-H460 human lung
large cell carcinoma, the A549 human lung adenocarcinoma,
the SK-OV-3 human ovarian carcinoma, and the MKN-74
human gastric carcinoma cell lines, with GI50 values below
submicromolar concentrations. Importantly, higher concen-
trations of (À)-1 did not completely abolish cell viability of
these sensitive cell lines (LC50 >100 mm). Other cancer cell
lines were found to be less sensitive to (À)-1. Significantly,
(À)-1 exhibited antiproliferative activity against several
human cancer cell lines with approximately 10 to 1000-fold
greater potency than that of bryostatin 1; the log GI50 values
of bryostatin 1 against the NCI-H460, A549, and SK-OV-3
cells are À5.6, À5.4, and À5.3, respectively.[49] On the basis
Scheme 11. Completion of the total synthesis of (À)-1. TMS=trimethyl-
silyl, NaHMDS=sodium bis(trimethylsilyl)amide, dba=dibenzylidenea-
cetone.
2684
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2678 – 2688

(À)-Exiguolide and Analogues
FULL PAPER
of the COMPARE analysis,[47,48] the fingerprint of (À)-1
showed only marginal similarities with those of DNA-relat-
ed agents listed in Table 4 (see below) and did not show any
significant correlation with those of more than 100 anticanc-
er agents, implying the possibility that (À)-1 may have a
unique biological mode-of-action.
to À178C)[51] provided exomethylene 41 quantitatively,
which was coupled with 2 under Suzuki–Miyaura conditions
to afford C5-methylene analogue 43 in 73% yield. Ana-
logues 45 and 46 with a simple alkyl side chain were pre-
pared from 36. Stille coupling[52] of 36 with the known (Z)-
vinylstannane 44[53] in the presence of a [Pd
2ACHTNUTRGNE(UGN dba)3]/Ph3As
catalyst system in DMF at room temperature proceeded
with partial erosion of the olefin geometry to give analogue
45 as an inseparable 5:1 mixture of Z/E isomers in 64%
yield. Acetylation of 45 afforded analogue 46 in 93% yield.
We have also prepared analogues with a truncated side
chain to probe the role of the triene side chain of (À)-1.
Suzuki–Miyaura coupling of 3 with commercially available
phenyl or vinyl pinacolboronate provided analogues 47
(93%) or 48 (94%), respectively.
The antiproliferative activity of (À)-exiguolide (1) and an-
alogues 39, 40, 42, 43, and 45–48 against the NCI-H460,
A549, and A172 human glioblastoma cell lines was evaluat-
ed in detail, and the results are summarized in Table 5. (À)-
1 displayed potent antiproliferative activity with submicro-
molar IC50 values (0.28, 0.59, and 0.47 mm against NCI-H460,
A549, and A172 cells, respectively). We found that C5-hy-
droxy analogue 39 showed about 10-fold less activity than
(À)-1. On the other hand, C5-acetoxy analogue 40 was inac-
tive at 10 mm against the A549 and A172 cell lines, indicating
that masking of the C5 hydroxy group of 39 was detrimental
for the activity. C5-Keto analogue 42 showed 10- to 100-fold
less potency than (À)-1. Importantly, the fact that C5-meth-
ylene analogue 43 was inactive indicated the striking effect
of the C5 methoxycarbonylmethylidene group on the potent
antiproliferative activity of (À)-1. Thus, only a limited reper-
toire of functionalities would be able to replace the C5 me-
thoxycarbonylmethylidene group of (À)-1. Interestingly, an-
alogue 45 was almost equipotent to analogue 39, implying
that the triene side chain of the natural product could be re-
placed with a simple alkyl chain without losing potency.
However, analogues with a truncated side chain displayed
diminished activity; phenyl analogue 47 was inactive and
vinyl analogue 48 was only marginally active compared to
(À)-1. These results suggested that the length of the side
chain of (À)-1 is important for exerting potent antiprolifera-
tive activity, while the terminal C27 methyl ester group
would not be essential.
Table 4. COMPARE analysis on (À)-1.
Rank
Compound
r[a]
Molecular targets/Drug type
1
2
3
pirarubicin
mitomycin C
SM-5887
0.561
0.556
0.537
DNA intercalater
DNA alkylating drugs
DNA topoisomerase II inhibitors
[a] r=correlation efficiency.
Encouraged by the results of the panel screening, we next
explored the structure–activity relationships of (À)-exiguo-
lide. Specifically, we focused our attention to the modifica-
tion of the C5 methoxycarbonylmethylidene group and the
triene side chain, because omission of these functionalities
would reduce the complexity of the molecule and the side
chain was found to be somewhat labile under acidic condi-
tions.[50] Thus, we designed and synthesized analogues 39, 40,
42, 43, and 45–48 as summarized in Scheme 12.
The synthesis of analogues lacking the C5 methoxycarbo-
nylmethylidene group started from intermediate 36 or 37.
C5-Hydroxy analogue 39 was synthesized by Suzuki–
Miyaura coupling of alcohol 36 with (Z)-vinylboronate 2
Conclusion
We have accomplished the total synthesis of (À)-exiguolide
(1), the naturally occurring enantiomer, for the first time.
Our strategy for the construction of the methylene bis(tetra-
hydropyran) substructure 5 of (À)-1 exploited the superb
chemoselectivity and bond-forming ability of olefin metathe-
sis reactions, which allowed for direct utilization of the pre-
existing functionalities within acyclic segments 8 and 9 in
subsequent ring-forming events. Thus, the readily available
segments 8 and 9 were assembled through olefin cross-meta-
thesis reaction, and the two tetrahydropyran rings were suc-
([Pd2ACHTUNGTRENNUNG(dba)3], Ph3As, Ag2O, THF/H2O 10:1, room tempera-
ture) in 96% yield. Acetylation (Ac2O, pyridine) of 39 af-
forded C5-acetoxy analogue 40 in 86% yield. Two addition-
al C5-modified analogues, 42 and 43, were synthesized.
Suzuki–Miyaura coupling of 37 with 2 gave C5-keto ana-
logue 42 in quantitative yield. Methylenation of 37 under
the modified Julia–Kocienski olefination conditions (1-tert-
butyl-5-methanesulfonyl-1H-tetrazole, NaHMDS, THF, À78
Chem. Eur. J. 2011, 17, 2678 – 2688
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2685

H. Fuwa et al.
cessively forged in a stereocon-
trolled manner via intramolecu-
lar oxa-conjugate cyclization
and reductive etherification. Al-
though our initial efforts on the
construction of the 20-mem-
bered macrocyclic framework
of (À)-1 via RCM ultimately
met with a limited success, we
eventually developed an effi-
cient strategy that hinges on
Julia–Kocienski olefination for
the stereoselective formation of
the C16–C17 double bond and
Yamaguchi macrolactonization
for the construction of the 20-
membered macrocycle. Finally,
the (E,Z,E)-triene side chain
was introduced in a highly ste-
reoselective
manner
via
Suzuki–Miyaura coupling under
exceptionally mild conditions to
complete the total synthesis of
(À)-1. Assessment of the
growth inhibitory activity of
synthetic (À)-1 against a panel
of 39 human cancer cell lines
elucidated for the first time that
(À)-1 exhibits potent antiproli-
ferative activity against the
NCI-H460 human lung large
cell carcinoma, the A549
human lung adenocarcinoma,
the SK-OV-3 human ovarian
carcinoma, and the MKN-74
human gastric carcinoma cells.
The COMPARE analysis sug-
gested the possibility that this
natural product may have
a
unique biological mode-of-
action, which would be of
worth investigating. In addition,
we have synthesized a series of
structural analogues and ex-
plored structure–activity rela-
tionships of (À)-1, which laid
the foundation for further struc-
ture optimization study. Future
studies will include synthesis
and evaluation of designed ana-
logues, mouse xenograft studies,
and target identification.
Scheme 12. Synthesis of structural analogues of (À)-exiguolide. DMF=N,N-dimethylformamide.
www.chemeurj.org ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2686
Chem. Eur. J. 2011, 17, 2678 – 2688

(À)-Exiguolide and Analogues
FULL PAPER
Table 5. Detailed evaluation of antiproliferative activity of (À)-1 and its
structural analogues against selected human lung cancer cell lines (IC50
values in mm).[a]
P. G. Bulger, D. Sarlah, Angew. Chem. 2005, 117, 4564–4601;
Angew. Chem. Int. Ed. 2005, 44, 4490–4527; d) A. Deiters, S. F.
Martin, Chem. Rev. 2004, 104, 2199–2238; e) A. Fꢁrstner, Angew.
Chem. 2000, 112, 3140–3172; Angew. Chem. Int. Ed. 2000, 39, 3012–
3043, and references therein.
Compound
NCI-H460
A549
A172
1
0.28
3.6
3.0
110
>100
6.5
>100
>100
24
0.59
2.9
>100
40
>100
2.7
>100
>100
>100
0.47
1.9
>100
46
>100
1.7
>100
>100
2.9
[14] For a review of olefin cross-metathesis, see: S. J. Connon, S. Ble-
chert, Angew. Chem. 2003, 115, 1944–1968; Angew. Chem. Int. Ed.
2003, 42, 1900–1923.
[15] For a review of oxa-conjugate reactions, see: C. F. Nising, S. Brꢂse,
Chem. Soc. Rev. 2008, 37, 1218–1228.
[16] For discussions on the stereochemical outcome of intramolecular
oxa-conjugate addition, see: a) J. M. Betancort, V. S. Martꢃn, J. M.
Padrꢄn, J. M. Palazꢄn, M. A. Ramꢃrez, M. A. Soler, J. Org. Chem.
1997, 62, 4570–4583; b) C. Schneider, A. Schuffenhauer, Eur. J.
Org. Chem. 2000, 73–82.
39
40
42
43
45
46
47
48
[a] For details of the assay procedure, see the Supporting Information.
[17] D. L. Lewis, J. K. Cha, Y. Kishi, J. Am. Chem. Soc. 1982, 104, 4976–
4978.
[18] L. V. Heumann, G. E. Keck, Org. Lett. 2007, 9, 4275–4278.
[19] H. C. Brown, P. K. Yadav, J. Am. Chem. Soc. 1983, 105, 2092–2093.
[20] A. K. Ghosh, Y. Wang, J. T. Kim, J. Org. Chem. 2001, 66, 8973–
8982.
[21] M. Scholl, S. Ding, C. W. Lee, R. H. Grubbs, Org. Lett. 1999, 1, 953–
956.
[22] A. Basha, M. Lipton, S. M. Weinreb, Tetrahedron Lett. 1977, 18,
4171–4172.
Acknowledgements
We thank Professor Shinji Ohta (Nagahama Institute of Bioscience and
Technology) for providing us with copies of 1H and 13C NMR spectra of
the authentic sample of (À)-exiguolide. We gratefully acknowledge the
Screening Committee of Anticancer Drugs supported by a Grant-in-Aid
for Scientific Research on Priority Area “Cancer” from the Ministry of
Education, Culture, Sports, Science and Technology (MEXT), Japan.
This work was supported in part by a Grant-in-Aid for Young Scientists
(B) and for Scientific Research on Innovative Areas “Organic Synthesis
Based on Reaction Integration” (No. 22106504) from MEXT, Japan.
[23] S. Nahm, S. M. Weinreb, Tetrahedron Lett. 1981, 22, 3815–3818.
[24] S. B. Garber, J. S. Kingsbury, B. L. Gray, A. H. Hoveyda, J. Am.
Chem. Soc. 2000, 122, 8168–8179.
[25] In a preliminary experiment that uses a model substrate, a small
amount (ꢀ15%) of 2,6-trans isomer was observed as the minor
product when the reaction was quenched within 15 min.
[26] H. Fuwa, K. Noto, M. Sasaki, Org. Lett. 2010, 12, 1636–1639.
[27] a) D. B. Dess, J. C. Martin, J. Org. Chem. 1983, 48, 4155–4156;
b) D. B. Dess, J. C. Martin, J. Am. Chem. Soc. 1991, 113, 7277–7287.
[28] J. Inanaga, K. Hirata, H. Saeki, T. Katsuki, M. Yamaguchi, Bull.
Chem. Soc. Jpn. 1979, 52, 1989–1993.
[29] For reports on the negative effect of allylic methyl group on olefin
metathesis reactions, see: a) M. Ulman, R. H. Grubbs, Organometal-
lics 1998, 17, 2484–2489; b) F. C. Courchay, T. W. Baughman, K. B.
Wagener, J. Organomet. Chem. 2006, 691, 585–594; c) I. C. Stewart,
C. J. Douglas, R. H. Grubbs, Org. Lett. 2008, 10, 441–444; d) Y.
Liang, R. Raju, T. Le, C. D. Taylor, A. R. Howell, Tetrahedron Lett.
2009, 50, 1020–1022.
[30] A. Fꢁrstner, K. Langemann, J. Am. Chem. Soc. 1997, 119, 9130–
9136.
[31] S. H. Hong, D. P. Sanders, C. W. Lee, R. H. Grubbs, J. Am. Chem.
Soc. 2005, 127, 17160–17161.
[32] T. R. Hoye, C. S. Jeffrey, M. A. Tennakoon, J. Wang, H. Zhao, J.
Am. Chem. Soc. 2004, 126, 10210–10211.
[33] For related examples, see: a) T. R. Hoye, M. E. Danielson, A. E.
May, H. Zhao, J. Org. Chem. 2010, 75, 7052–7060; b) S. Y. Yun,
E. C. Hansen, I. Volchkov, E. J. Cho, W. Y. Lo, D. Lee, Angew.
Chem. 2010, 122, 4357–4359; Angew. Chem. Int. Ed. 2010, 49, 4261–
4263.
[34] a) P. R. Blakemore, W. J. Cole, P. J. Kocienski, A. Morley, Synlett
1998, 26–28; b) P. R. Blakemore, J. Chem. Soc. Perkin Trans. 1 2002,
2563–2585; c) C. Aꢅssa, Eur. J. Org. Chem. 2009, 1831–1844.
[35] S. Chandrasekhar, S. R. Yaragorla, L. Sreelakshmi, C. R. Reddy, Tet-
rahedron 2008, 64, 5174–5183.
[1] Selected recent reviews on marine natural products: a) J. W. Blunt,
B. R. Copp, M. H. G. Munro, P. T. Northcote, M. R. Prinsep, Nat.
Prod. Rep. 2010, 27, 165–237; b) M. Kita, O. Ohno, C. Han, D.
Uemura, Chem. Rec. 2010, 10, 57–69; c) J. W. Blunt, B. R. Copp,
M. H. G. Munro, P. T. Northcote, M. R. Prinsep, Nat. Prod. Rep.
2009, 26, 170–244; d) T. F. Molinski, D. S. Dalisay. S. L. Lievens, J. P.
Saludes, Nat. Rev. Drug Discovery 2009, 8, 69–86; e) K. Nakamura,
M. Kitamura, D. Uemura, Heterocycles 2009, 78, 1–17, and referen-
ces therein.
[2] For a recent review on the synthetic aspects of marine macrolide
natural products, see: K.-S. Yeung, I. Paterson, Chem. Rev. 2005,
105, 4237–4313.
[3] S. Ohta, M. M. Uy, M. Yanai, E. Ohta, T. Hirata, S. Ikegami, Tetra-
hedron Lett. 2006, 47, 1957–1960.
[4] N. Matsumori, D. Kaneno, M. Murata, H. Nakamura, K. Tachibana,
J. Org. Chem. 1999, 64, 866–876.
[5] M. S. Kwon, S. K. Woo, S. W. Na, E. Lee, Angew. Chem. 2008, 120,
1757–1759; Angew. Chem. Int. Ed. 2008, 47, 1733–1735.
[6] G. R. Pettit, C. L. Herald, D. L. Doubek, D. L. Herald, E. Arnold, J.
Clardy, J. Am. Chem. Soc. 1982, 104, 6846–6848.
[7] J. Cossy, C. R. Chim. 2008, 11, 1477–1482.
[8] For a preliminary communication of this work, see: H. Fuwa, M.
Sasaki, Org. Lett. 2010, 12, 584–587.
[9] Roulland and co-workers reported another total synthesis of (À)-1
almost simultaneously with our report. See: C. Cook, X. Guinchard,
F. Liron, E. Roulland, Org. Lett. 2010, 12, 744–747.
[10] C. R. Reddy, N. N. Rao, Tetrahedron Lett. 2010, 51, 5840–5842.
[11] For reviews on Suzuki–Miyaura coupling, see: a) N. Miyaura, A.
Suzuki, Chem. Rev. 1995, 95, 2457–2483; b) S. R. Chemler, D. Trau-
ner, S. J. Danishefsky, Angew. Chem. 2001, 113, 4676–4701; Angew.
Chem. Int. Ed. 2001, 40, 4544–4568; c) S. Kotha, K. Lahiri, D. Ka-
shinath, Tetrahedron 2002, 58, 9633–9695; d) A. Suzuki, Chem.
Commun. 2005, 4759–4763.
[12] The carbon numbering in this paper corresponds to that of the natu-
ral product.
[13] For selected recent reviews on olefin-metathesis reactions, see:
a) A. H. Hoveyda, A. R. Zhugralin, Nature 2007, 450, 243–251;
b) A. Gradillas, J. Perez-Castells, Angew. Chem. 2006, 118, 6232–
6247; Angew. Chem. Int. Ed. 2006, 45, 6086–6101; c) K. C. Nicolaou,
[36] T. Nagamitsu, D. Takano, M. Seki, S. Arima, M. Ohtawa, K. Shiomi,
¯
Y. Harigaya, S. Omura, Tetrahedron 2008, 64, 8117–8127.
[37] S. Takano, K. Samizu, T. Sugihara, K. Ogasawara, J. Chem. Soc.
Chem. Commun. 1989, 1344–1345.
[38] H. X. Zhang, F. Guibe, G. Balavoine, J. Org. Chem. 1990, 55, 1857–
1867.
[39] O. Mitsunobu, Synthesis 1981, 1–28.
[40] H. S. Schultz, H. B. Freyermuth, S. R. Buc, J. Org. Chem. 1963, 28,
1140–1142.
Chem. Eur. J. 2011, 17, 2678 – 2688
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2687

H. Fuwa et al.
[41] Oxidative cleavage of the MPM group with DDQ (1.1 equiv) or
DDQ (1.1 equiv)/diallyl ether (excess) accompanied over-oxidation
of the liberated allylic alcohol as a side reaction. For the beneficial
effect of diallyl ether in DDQ oxidation, see: a) A. Hamajima, M.
Isobe, Angew. Chem. 2009, 121, 2985–2989; Angew. Chem. Int. Ed.
2009, 48, 2941–2945; b) H. Furuta, Y. Hasegawa, M. Hase, Y. Mori,
Chem. Eur. J. 2010, 16, 7586–7595.
[42] E. D. Laganis, B. L. Chenard, Tetrahedron Lett. 1984, 25, 5831–5834.
[43] a) K. Tanaka, Y. Ohta, K. Fuji, Tetrahedron Lett. 1993, 34, 4071–
4074; b) D. A. Evans, P. H. Csarter, E. M. Carreira, A. B. Charette,
J. A. Prunet, M. Lautens, J. Am. Chem. Soc. 1999, 121, 7540–7552;
c) K. Ohmori, Y. Ogawa, T. Obitsu, Y. Ishikawa, S. Nishiyama, S.
Yamamura, Angew. Chem. 2000, 112, 2376–2379; Angew. Chem. Int.
Ed. 2000, 39, 2290–2294.
[44] For the preparation of (Z)-vinylboronates from the corresponding
terminal alkynes by trans-hydroboration, see: T. Ohmura, Y. Yama-
moto, N. Miyaura, J. Am. Chem. Soc. 2000, 122, 4990–4991.
[45] a) Y. K. Reddy, J. R. Falck, Org. Lett. 2002, 4, 969–971; b) T. Gill-
mann, T. Weeber, Synlett 1994, 649–650; c) J. Uenishi, J.-M. Beau,
R. W. Armstrong, Y. Kishi, J. Am. Chem. Soc. 1987, 109, 4756–4758.
[46] We found it important to use moist THF as the solvent. The present
reaction did not proceed to an appreciable extent under strictly an-
hydrous conditions. Consequently, we used H2O as a co-solvent in
the synthesis of structural analogues.
Kohno, Y. Nakajima, H. Komatsu, T. Andoh, T. Tsuruo, Cancer Res.
1999, 59, 4042–4049.
[48] T. Yamori, Cancer Chemother. Pharmacol. 2003, 52ACTHNUTRGNE(UNG Suppl. 1), 74–
79.
[49] Information on the growth inhibition activity of bryostatin 1 against
a panel of 60 human cancer cell lines can be found on the National
Cancer Institute database at http://dtp.nci.nih.gov/branches/btb/
ivclsp.html.
[50] Upon standing an unpurified CDCl3 solution of (À)-1 at room tem-
perature for more than 5 h, ca. 5% of what-thought-to-be isomer-
1
ized product was observed by H NMR analysis.
[51] C. Aissa, J. Org. Chem. 2006, 71, 360–363.
[52] For selected recent reviews on Stille reaction, see: a) K. C. Nicolaou,
P. G. Bulger, D. Sarlah, Angew. Chem. 2005, 117, 4516–4563;
Angew. Chem. Int. Ed. 2005, 44, 4442–4489; b) P. Espinet, A. E.
Echavarren, Angew. Chem. 2004, 116, 4808–4839; Angew. Chem.
Int. Ed. 2004, 43, 4704–4734; c) T. C. Mitchell in Metal-Catalyzed
Cross-Coupling Reactions, 2nd ed. (Eds.: A. de Meijere, F. Dieder-
ich), Wiley-VCH, Weinheim, 2004; pp. 125–162; d) V. Farina, V.
Krishnamurthy, W. J. Scott, Org. React. 1997, 50, 1–652.
[53] T. Hosoya, K. Sumi, H. Doi, M. Wakao, M. Suzuki, Org. Biomol.
Chem. 2006, 4, 410–415.
[47] T. Yamori, A. Matsunaga, S. Sato, K. Yamazaki, A. Komi, K. Ishizu,
I. Mita, H. Edatsugi, Y. Matsuba, K. Takezawa, O. Nakanishi, H.
Received: November 1, 2010
Published online: January 24, 2011
2688
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2678 – 2688