Convergent and enantioselective total synthesis of (-)-amphidinolide O and (-)-amphidinolide P

Article abstract of DOI:10.1021/ol401357k

A convergent and enantioselective total synthesis of (-)-amphidinolide O (1) and P (2), 15-membered macrolides with seven chiral centers along with many functional groups, is described. The key reactions include enantioselective Brown allylation, anti- and syn-selective aldol reactions, (E)-selective olefin metathesis, conformation-controlled stereoselective epoxidation, and selective introduction of the exomethylene group. Assignments of the absolute stereochemistries of the natural (+)-amphidinolide O (ent-1) and P (ent-2) are also discussed in detail.

Full text of DOI:10.1021/ol401357k

ORGANIC
LETTERS
XXXX
Vol. XX, No. XX
000–000
Convergent and Enantioselective Total
Synthesis of (ꢀ)-Amphidinolide O and
(ꢀ)-Amphidinolide P
Min-ho Hwang, Seo-Jung Han, and Duck-Hyung Lee*
Department of Chemistry, Sogang University, Organic Chemistry Research Center
(OCRC), Shinsu-dong 1, Mapo-gu, Seoul 121-742, Korea
dhlee@sogang.ac.kr
Received May 15, 2013
ABSTRACT
A convergent and enantioselective total synthesis of (ꢀ)-amphidinolide O (1) and P (2), 15-membered macrolides with seven chiral centers along
with many functional groups, is described. The key reactions include enantioselective Brown allylation, anti- and syn-selective aldol reactions,
(E)-selective olefin metathesis, conformation-controlled stereoselective epoxidation, and selective introduction of the exomethylene group.
Assignments of the absolute stereochemistries of the natural (þ)-amphidinolide O (ent-1) and P (ent-2) are also discussed in detail.
Amphidinolides AꢀHandJꢀY, isolated from laboratory-
cultured Okinawan marine dinoflagelate amphidinolium
sp. by Kobayashi and co-workers, have attracted much
attention from the synthetic community because of their
biogenetically unusual structural features and cytotoxic
activities against various cancer cell lines.1 Among them,
(þ)-amphidinolide O (ent-1) and (þ)-amphidinolide P
(ent-2) have shown in vitro cytotoxicity against murine
lymphoma L1210 (IC50 = 1.7 and 3.6 μg/mL, respectively)
and human epidermoid carcinoma KB cells (IC50 = 1.6
and 5.8 μg/mL, respectively).2
Amphidinolide O (1) and P (2) have many structural
features in common such as a novel 15-membered macro-
lide with an epoxide at C8ꢀC9, one double bond at
C12ꢀC13, one exocyclic double bond at C5, and one
6-membered ring bridged hemiacetal moiety. They have
a different functional group only at the C11 position. In
other words, amphidinolide O (1) has a C11 carbonyl
group whereas amphidinolide P (2) has a C11 exomethyl-
ene group (Scheme 1).2 So far, two total syntheses and one
formal synthesis of amphidinolide P have been reported by
three groups.3 However, their synthetic schemes do not
allow the transformation of amphidinolide P into amphi-
dinolide O because the C11 exocyclic double bond moiety
was installed at an early stage of the synthesis. We have
already published five preliminary papers in relation to the
convergent enantioselective total synthesis of (þ)-amphi-
dinolide O (ent-1) and P (ent-2),4 and we report herein the
(3) (a) Williams, D. R.; Myers, B. J.; Mi, L. Org. Lett. 2000, 2, 945–
948. (b) Chakraborty, T. K.; Das, S. Tetrahedron Lett. 2001, 42, 3387–
3390. (c) Trost, B. M.; Papillon, J. P. N. J. Am. Chem. Soc. 2004, 126,
13618–13619. (d) Trost, B. M.; Papillon, J. P. N.; Nussbaumer, T. J. Am.
Chem. Soc. 2005, 127, 17921–17937. (e) Williams, D. R.; Myers, B. J.;
Mi, L.; Binder, R. J. J. Org. Chem. 2013, 78, 4762–4778. (f) Williams,
D. R.; Myers, B. J.; Mi, L. Org. Lett. 2013, 15, 2070–2070.
(4) (a) Pang, J. H.; Lee, D. H. Bull. Korean Chem. Soc. 2002, 23, 1173–
1176. (b) Pang, J. H.; Ham, Y. J.; Lee, D. H. Bull. Korean Chem. Soc.
2003, 24, 891–892. (d) Jang, M. Y.; Kim, J. W.; Lee, D. H. Bull. Korean
Chem. Soc. 2005, 26, 1497–1498. (d) Kim, J. W.; Kong, S. J.; Kim, Y. J.;
Lee, D. H. Bull. Korean Chem. Soc. 2008, 29, 297–298. (e) Joo, H. W.;
Jung, H. J.; Hwang, M. H.; Lee, D. H. Bull. Korean Chem. Soc. 2009, 30,
2201–2202.
(1) (a) Kobayashi, J.; Ishibashi, M. Chem. Rev. 1993, 93, 1753–1769.
(b) Chakraborty, T. K.; Das, S. Curr. Med. Chem.: Anti-Cancer Agents
2001, 1, 131–149. (c) Kobayashi, J.; Tsuda, M. Nat. Prod. Rep. 2004, 21,
77–93. (d) Kobayashi, J.; Kubota, T. J. Nat. Prod. 2007, 70, 451–460. (e)
Morris, J. C.; Phillips, A. J. Nat. Prod. Rep. 2009, 26, 245–65. (f)
Fuerstner, A. Isr. J. Chem. 2011, 51, 329–345.
(2) Ishibashi, M.; Takahashi, M.; Kobayashi, J. J. Org. Chem. 1995,
60, 6062–6066.
r
10.1021/ol401357k
XXXX American Chemical Society

first convergent and enantioselective total synthesis of
(ꢀ)-amphidinolide O (1) and (ꢀ)-amphidinolide P (2) as
well as the confirmation of the absolute stereochemistry of
the natural products.
provided compound 8 (96% yield). Selective deprotection
of the primary TBS group using PPTS in methanol (67%
yield) followed bySwernoxidation of the resulting primary
alcohol produced the (E)-conjugated aldehyde 6. Anti-
selective aldol reaction between aldehyde 6 and Masamune’s
norephedrine-derived auxiliary 10 yielded a secondary
alcohol (67% yield over two steps),5 which was subse-
quently protected by TESCl to give an ester 9 in 94% yield.
Treatment of the ester 9 with MeMgBr in THF (75% yield)
and deprotection of the secondary TES protecting group
by TBAF (84% yield) produced the C9ꢀC17 fragment 4.
Scheme 1. Synthesis of C9ꢀC17 Fragment 4
Figure 1. Retrosynthetic analysis of (ꢀ)-amphidinolide O (1)
and (ꢀ)-amphidinolide P (2).
The retrosynthetic analysis for compounds 1 and 2 is
depicted in Figure 1. From their structural similarity, we
envisaged that 1 and 2 might be constructed from the
common intermediate 3. The intermediate 3 could be
available from the esterification reaction between the diol
4 and the carboxylic acid 5 followed by the ring-closing
metathesis (RCM). In order to perform the regioselective
epoxidation at the C8ꢀC9 double bond, two exo-methylene
groups at C5ꢀC19 and C16ꢀC17 should be installed
after the C8ꢀC9 epoxide formation. Intermediate 4 could
be prepared from anti-aldol reaction of 6,5 which in turn
might be obtained from (Z)-2-butene-1,4-diol via oxida-
tion and Brown asymmetric allylation.6 Carboxylic acid 5
couldbeaccessiblefromL-asparticacidvia Evanssyn-aldol
reaction7 and acid catalyzed cyclization as key steps.
The synthesis of C9ꢀC17 fragment 4 is summarized in
Scheme 1. Monosilylation of (Z)-2-buten-1,4-diol with
TBSCl (87% yield) and PCC oxidation of the remaining
primary alcohol resulted in concomitant isomerization of
the (Z)-conjugated aldehyde into the thermodynamically
more stable (E)-isomer 7.8 Brown asymmetric allylation of
aldehyde 7 was carried out to afford the (11S)-selective
secondary alcohol in a 62% two-step yield with 92% ee.6
Protection of the secondary hydroxyl group with TBSCl
Synthesis of the C1ꢀC8 fragment 5 is summarized
in Scheme 2. L-Aspartic acid was treated with KBrꢀ
NaNO2ꢀH2SO4 reagent to convert the amino group into
the corresponding bromide withretention of configuration
via a diazonium salt-mediated double inversion strategy.
Subsequently, two carboxylic acids were reduced with
borane to provide a diol in a 79% two-step yield. The diol
was then subjected to NaH-promoted epoxidation and in
situ protection of remaining alkoxide with TBSCl to afford
the epoxide 11 (74% yield).9 The epoxide ring of 11 was
treated with the lithium ylide derived from trimethylsulfo-
nium iodide and LHMDS to give the allyl secondary
alcohol in 88% yield,10 and the resulting secondary alcohol
was converted to the aldehyde 12 via a three-step sequence:
(a) TBS protection of the secondary alcohol (100% yield),
(b) selectivedeprotection of theprimary TBS group (82%),
and (c) Swern oxidation of the primary alcohol (90%
yield).
A syn-selectivealdol reactionof aldehyde 12using Evans
chiral oxazolidinone 18 in the presence of (ꢀ)-sparteine
afforded the aldol product 13 (83% yield).7 The chiral
auxiliary in compound 13 was converted to the Weinreb
(5) (a) Abiko, A.; Liu, J.; Masamune, S. J. Am. Chem. Soc. 1997, 119,
2586–2587. (b) Abiko, A. Org. Synth. 2003, 79, 109–115.
(6) Brown, H. C.; Jadhav, P. K. J. Am. Chem. Soc. 1983, 105, 2092–
2093.
(7) (a) Evans, D. A.; Bartroli, J.; Shih, T. L. J. Am. Chem. Soc. 1981,
103, 2127–2129. (b) Gage, J. R.; Evans, D. A. Org. Synth. 1990, 68, 77–
82. (c) Gage, J. R.; Evans, D. A. Org. Synth. 1990, 68, 83–91.
(8) Ramachandran, P. V.; Liu, H.; Ram Reddy, M. V.; Brown, H. C.
Org. Lett. 2003, 5, 3755–3757.
(9) (a) Volkmann, R. A.; Kelbaugh, P. R.; Nason, D. M.; Jasys, V. J.
J. Org. Chem. 1992, 57, 4352–4361. (b) Robinson, J. E.; Brimble, M. A.
Chem. Commun. 2005, 1560–1562.
(10) Alcaraz, L.; Harnett, J. J.; Mioskowski, C.; Martel, J. P.; Le
Gall, T.; Shin, D.; Falck, J. R. Tetrahedron Lett. 1994, 35, 5449–5452.
B
Org. Lett., Vol. XX, No. XX, XXXX

Scheme 2. Synthesis of C1ꢀC8 Fragment
Figure 2. Assignment of the relative stereochemistry of β-
hydroxylsilane 19.
With key intermediates 4 and 5 in hand, we proceeded
with the synthesis of (ꢀ)-amphidinolide O (1) and (ꢀ)-
amphidinolide P (2) (Scheme 3). Intermediates 4 and 5
were coupled using the EDCI-DMAP protocol to afford
the ester in 94% yield, and the ester was subjected to olefin
metathesis to implement the (E)-selective C8ꢀC9 double
bond in 93% yield.13 Formation of the (E)-double bond
was deduced by measuring the coupling constant (J = 16
Hz) between C8ꢀH and C9ꢀH. Although the epoxidation
reaction of 20 by m-CPBA proceeded smoothly only at the
C8ꢀC9 double bond to give the epoxide 21 in 88% yield,
the relative stereochemistry at the C8ꢀC9 epoxide was
unclear at this point because the facial selectivity in the
epoxidation reaction could not be confirmed completely
from its 1D and 2D 1H NMR spectra.14
Dehydration of 21 with Martin sulfurane introduced the
C16ꢀC17 double bonds (71% yield).15 Peterson olefina-
tion of the β-hydroxysilane moiety was carried out success-
fully with PCC (64% yield),16 and deprotection of the
C11-OTBS group with TBAF produced the intermediate 3
(86% yield). Oxidationof the C11 hydroxyl groupin3 with
Dess-Martin periodinane furnished the ketone 22, a com-
mon key intermediate for the convergent synthesis of 1and 2.
(ꢀ)-Amphidinolide O (1) was obtained in a 93% two-
stepyield from 3 bysubjectionof the ketone 22to the acidic
acetal hydrolysis conditions. The 1H and 13C spectroscopic
data and HRMS data of 1 were identical in all respects
with the reported data for the natural (þ)-amphidinolide
O (ent-1). The issue of absolute stereochemistry will be
discussed soon.
amide (78% yield),11 the secondary hydroxyl group was
protected by TBS group (99% yield), and the Weinreb
amide wastreated byDIBAL-Htoprovidethe aldehyde14
(96% yield). Reaction of the lithium enolate derived from
ethyl acetate with the aldehyde 14 proceeded smoothly
to afford the β-hydroxy ester in 90% yield, and the
secondary hydroxyl group was oxidized by Dess-Martin
periodinane to provide the β-keto ester 15 in 83% yield.
Exposure of 15 with TsOH in MeOH resulted in the
deprotection of two TBS protecting groups followed by
concomitant cyclization and acetalyzation to afford the
cyclic acetal 16 in 73% yield. After the oxidation of the
C5 hydroxyl of 16 with Dess-Martin periodinane (92%
yield), 1,2-addition of TMSCH2MgCl to the ketone 17
was carried out successfully to afford the β-hydroxysilane
19 in 80% yield. Finally, the ethyl ester group was removed
under basic conditions to afford the crude C1ꢀC8 frag-
ment 5, which was used in the next step without further
purification.
(13) Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett. 1999,
1, 953–956.
(14) Prediction of the stereochemistry in the epoxidation reaction is
described in the SI.
(15) (a) Martin, J. C.; Arhart, R. J. J. Am. Chem. Soc. 1971, 93, 4327–
4329. (b) Arhart, R. J.; Martin, J. C. J. Am. Chem. Soc. 1972, 94, 5003–
5010.
(16) (a) Wang, Z.; Jimenez, L. S. Tetrahedron Lett. 1996, 37, 6049–
6052. (b) Several reaction conditions for the Peterson olefination were
investigated using the intermediate X as a reference material. The
starting material X remained intact with silica gel and decomposed by
treatment of KH, NaH, or t-BuOK. Fortunately, HF-pyridine and PCC
provided the desired product in 71% and 69% yield, respectively.
The stereochemical configuration at the C5 chiral center
can be explained by steric effects. In other words, the
nucleophile approaches from the least hindered face of
the ketone 17 (anti to both the equatorial C4-methyl group
and axial C3-methoxy group), and this relationship was
supported strongly by the 2D-NOE experiment of the
intermediate 19 (Figure 2).12
(11) (a) Basha, A.; Lipton, M.; Weinreb, S. M. Tetrahedron Lett.
1977, 18, 4171–4172. (b) Levin, J. I.; Turos, E.; Weinreb, S. M. Synth.
Commun. 1982, 12, 989–993.
(12) 2D-NOE data for 19 are available in the Supporting Informa-
tion (SI).
Org. Lett., Vol. XX, No. XX, XXXX
C

and (þ)-amphidinolide P (ent-2). In 2000, the Williams
group reported the first enantioselective total synthesis of
amphidinolide P (ent-2) along with the optical rotation
Scheme 3. RCM and Completion of Amphidinolide O (1) andP(2)
([R]20 ꢀ30° (c 0.09, MeOH)).3a However, an inconsis-
D
tency in the absolute stereochemistry was found unexpect-
edly when the Trost group published the synthesis of
amphidinolide P (2) and its optical rotation ([R]20D ꢀ27.4°
(c 0.17, MeOH)) in 2004.3b Recently, the synthesis and
structural correction for amphidinolide P was detailed in
a full paper by the Williams group.3e,f The optical rota-
tion of [R]20 ꢀ29.2° (c 0.31, MeOH) for our synthetic
D
compound 2 is in accord with the Trost result and the
recently revised assignment by Williams, confirming clearly
that the natural (þ)-amphidinolide P is ent-2 ([R]20D þ31°
(c 0.098, MeOH)).2
The optical rotation of our synthetic compound 1 was
measured tobe [R]23D ꢀ129° (c 0.13, MeOH), [R]23D ꢀ131°
(c 0.20, MeOH), and [R]23 ꢀ137° (c 0.25, MeOH).
D
In comparison with the literature value for the natural
(þ)-amphidinolide O (ent-1) ([R]23D þ65° (c 0.12, MeOH)),
the direction of rotation is opposite to each other and the
absolute value is about two times larger in our synthetic
compound 1. However, we are quite sure that the natural
(þ)-amphidinolide O should be ent-1 again.
In conclusion, we accomplished the first convergent
and enantioselective total synthesis of (ꢀ)-amphidinolide
O (1) and P (2). The key reactions include enantioselective
Brown allylation, anti- and syn-selective aldol reaction,
(E)-selective olefin metathesis for the C8ꢀC9 double bond
formation, conformation-controlled stereoselective epox-
idation at the C8ꢀC9 double bond, and selective conver-
sion of C5 ketone moieties into the exomethylene group.
The absolute stereochemistry of the natural (þ)-amphidi-
nolide O (ent-1) was determined unambiguously, and that
of (þ)-amphidinolide P (ent-2) was in accord with the
reported information by Trost and Williams.
Next, several approaches for (ꢀ)-amphidinolide P (2)
were investigated starting from the common intermediate
22, and the bestroute was determinedtobe very simple and
straightforward. The addition of methyl lithium to the
ketone moiety of 22, direct dehydration with Martin
sulfurane as above,17 and deprotection of the acetal group
by HF in H2OꢀCH3CN provided the compound 2 in a
42% four-step yield. Again, the 1H and 13C spectroscopic
dataand HRMSdataof 2 were identical in all respectswith
those of the natural (þ)-amphidinolide P (ent-2).
Acknowledgment. This research was assisted financially
by the National Research Foundation of Korea (NRF-
2010-0026141). The instrument facilities of the Organic
Chemistry Research Center (OCRC) in the chemistry
department were also helpful.
Now, we need to clarify the issue of absolute stereo-
chemistry for the natural (þ)-amphidinolide O (ent-1)
Supporting Information Available. Experimental pro-
cedures and full spectroscopic data for all new com-
pounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
(17) Initially, addition of TMSCH2Li to ketone 22 in THF at ꢀ78 °C
was followed by Peterson olefination/hemiacetal formation in aqueous
HF-CH3CN to produce the desired (ꢀ)-amphidinolide P 2 in 13% three-
step yield starting from 3. Second, Tebbe olefination of 22 in THF at
ꢀ78 °C and hemiacetal formation in aqueous HF-CH3CN afforded the
desired product 2 in 21% three-step yield starting from 3.
The authors declare no competing financial interest.
D
Org. Lett., Vol. XX, No. XX, XXXX