A Baylis-Hillman approach to the synthesis of C1-C11 fragment of caribenolide I

Article abstract of DOI:10.1016/j.tetlet.2006.04.071

Stereoselective synthesis of C₁-C₁₁ fragment of caribenolide I, a potent antitumour macrolide isolated from a marine dinoflagellate Amphidinium sp. is described. The key steps rely on asymmetric aldol reactions, to control the absolu

Full text of DOI:10.1016/j.tetlet.2006.04.071

Tetrahedron Letters 47 (2006) 4175–4180
A Baylis–Hillman approach to the synthesis
of C1–C11 fragment of caribenolide I
Matar Seck, Xavier Franck,*, Blandine Seon-Meniel,
*
`
Reynald Hocquemiller and Bruno Figadere
Laboratoire de Pharmacognosie associe´ au CNRS (BioCIS), Faculte´ de Pharmacie, Universite´ Paris-Sud, rue Jean-Baptiste,
ˆ
Cle´ment 92296 Chatenay Malabry, France
Received 15 March 2006; revised 10 April 2006; accepted 11 April 2006
Abstract—Stereoselective synthesis of C1–C11 fragment of caribenolide I, a potent antitumour macrolide isolated from a marine
dinoflagellate Amphidinium sp. is described. The key steps rely on asymmetric aldol reactions, to control the absolute configurations
of C2, C3 and C10 stereogenic centres.
Ó 2006 Elsevier Ltd. All rights reserved.
Amphidinolides are members of a family of marine nat-
ural products that possess potent cytotoxic property
against a number of tumour cells (e.g., murine lym-
phoma L1210, human epidermoid carcinoma KB and
human colon tumour cells HCT116). They are isolated
from marine dinoflagellates, Amphidinium sp., that are
symbiotic of Okinawan marine flatworm Amphiscolops
sp. Since the first isolation of an amphidinolide,1 today
more than 35 amphidinolides have been isolated and
characterized.2 Caribenolide I is a 26-membered macro-
lactone of this family,3 and possesses an important
in vitro cytotoxicity against human colon tumour cells
of wild type as well as against those that have shown a
multi-drug resistance phenotype (IC50 = 0.001 lg/mL
or 1.6 nM, against HCT116/WT or HCT 116/VM 46).
It is worth noting that this cytotoxicity is 100 times
higher than that observed for amphidinolide B (IC50/
HCT116/WT = 0.122 lM), which was then considered
as the most active amphidinolide.4–7 Most importantly,
caribenolide I shows an important in vivo activity
against P388 tumour grafted mice (T/C = 150% at a
dose of 0.03 mg/kg). As most of the amphidinolides,
caribenolide I was isolated in minute amounts from
dried cells (0.026% yield). If a few total syntheses of
amphidinolides have appeared in the literature,4 to the
best of our knowledge, nothing has been reported
concerning total or partial synthesis of caribenolide I,
except a brief report by us.8
In continuation of our efforts to contribute to the struc-
tural elucidation of absolute and/or relative configura-
tions of caribenolide I, along with the desire to possess
a large quantity of such an extremely cytotoxic natural
product, we decided to study the total synthesis of carib-
enolide I. We wish to report herein our results concern-
ing the synthesis of the C1–C11 skeleton of caribenolide I.
Our retrosynthetic strategy is described in Scheme 1, and
shows that caribenolide I could be obtained by a conver-
gent approach from C12–C29 and C1–C11 fragments.
HO
24
HO
25
29
HO
15
19
21 O
O
OMe
O
1
O
O
O
O
1
11
11
OH
OH
X
OP
OR
O
6
6
7
O
7
OMe
1
O
OP
OP
OR
O
O
6
7
RO
O
OMe
*
Corresponding authors. E-mail: bruno.figadere@cep.u-psud.fr
New Address: UMR-CNRS 6014, IRCOF/LHO, Rue Tesniere,
76131 Mont Saint Aignan Cedex, France.
Scheme 1.
0040-4039/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2006.04.071

4176
M. Seck et al. / Tetrahedron Letters 47 (2006) 4175–4180
In this letter, we describe our efforts toward the prepara-
tion of the C1–C11 fragment of caribenolide I, with a
good control of the absolute configurations of four ste-
reogenic centres (C2, C3, C7 and C10), among the 13 ones
of caribenolide I, as shown in Scheme 1, with methods
that could provide either epimer, for later comparisons
with the natural product. The key reactions of this strat-
egy rely on the use of asymmetric aldolisations to con-
trol the relative and absolute configurations of C2 and
C3. Then, on a later stage of the synthesis the Sharpless
asymmetric epoxidation of the so formed allylic alcohol
will allow us to control the configurations of C4 and C5.
The exo double bond is generated by a Baylis–Hillman
reaction.
was obtained in 80% isolated yield as a 1:1 (E) and
(Z) mixture,11 after 5 h of reaction at room temperature.
However, the compound so obtained is the un-wanted
a-branched derivative (branching at C4, caribenolide
numbering), and none of the c-alkylated products
(branching at C6, caribenolide numbering) was formed.
By changing the amount of base, time of the reaction,
solvent, no improvements in yield and/or regioselectivity
were observed. Since the desired c-regioisomer was not
obtained, whatever the reaction conditions were, we
did not apply this transformation to our synthon, and
decided to build up step by step the required diene frag-
ment, as shown in Scheme 3. However, the scope and
limitations of this interesting reaction are now under
studies in our laboratories.
In a first attempt, we tried to introduce the diene moiety
of our fragment possessing an allylic hydroxyl by a
Baylis–Hillman reaction. This required to study the
coupling reaction of an aldehyde with ethyl penta-2,4-
dienoate (Scheme 2). To our knowledge, there are no
reports in the literature on the Baylis–Hillman reaction
with ethyl penta-2,4-dienoate as substrate.9 Thus we
decided to study a model reaction with 3-hydroxyquinu-
clidine (20% equiv) and 4-nitrobenzaldehyde (1 equiv) in
a 1:1 dioxane and water mixture with ethyl penta-2,4-
dienoate (2 equiv).10 Under these reaction conditions,
we were delighted to observe that a coupling product
In the absence of solvent, 3-paramethoxylbenzyloxy-
propanal 1 with methyl acrylate and racemic 3-hydroxy-
quinuclidine led to the desired racemic allylic alcohol 2
in 62% isolated yield.10 Then, protection of the latter
as a tert-butyldimethylsilyl ether was performed under
usual conditions (TBDMSCl, DMAP, imidazole,
DMF),12 and afforded the expected silyl ether 3 in
90% yield. Then for introduction of a two carbon unit
through a Wittig–Horner reaction, methyl ester 3 was
converted into its corresponding aldehyde in two steps;
first DIBAL reduction in dichloromethane13 afforded
OH
O
4
OEt
α
OH O2N
+
O
O
1:1 (E)/(Z ) : 80%
O
N
OEt
+
O2N
dioxane/H2O
4
OEt
OH
6
γ
O2N
not observed
Scheme 2.
OH
PMBO
O
OH
TBDMSCl, DMAP,
imidazole, DMF,
O
TBSO
O
N
rt, 16 h
OMe
OMe
O
PMBO
OMe
O
90%
2
3
62%
PMBO
1
OMe
O
TBSO
1) Dibal,CH2Cl2, 1 h
TBSO
(EtO)2OP
CO2Me
PMBO
2) MnO2, CH2Cl2, reflux 4 h
PMBO
BuLi, THF
83%
75% (2 steps)
4
5
1) Dibal,CH2Cl2, 1 h
2) MnO2, CH2Cl2, reflux 4 h
TBSO
O
PMBO
91% (2 steps)
6
Scheme 3.

M. Seck et al. / Tetrahedron Letters 47 (2006) 4175–4180
4177
the desired primary alcohol which was oxidized by man-
ganese oxide (MnO2) treatment,14 affording the required
aldehyde 4 in 75% yield for the two steps. The latter was
then treated by methyl diethylphosphonoacetate, to give
the (E) unsaturated ester 5 in 83% yield.15 Then, reduc-
tion of the latter (DIBAL in dichloromethane) and man-
ganese oxide oxidation of the so-obtained allylic alcohol
afforded the desired aldehyde 6 in 91% overall yield
(Scheme 3).
‘syn-non-Evans’ aldols, depending on the amount of
titanium tetrachloride (TiCl4) and amine used. To study
the influence of a remote protected hydroxy such as
OBn, a model reaction was first studied with two ox-
azolidin-2-thiones a and b20 and aldehyde 7, under the
reaction conditions described by Crimmins et al. In
order to attribute both the relative and absolute config-
urations of aldols 8a,b after separation by flash chroma-
tography on silica gel, they were reduced in high yields
by LiBH4 to afford the corresponding 1,3-diols 9, which
were then quantitatively protected as acetonides 10
(Scheme 4). Comparisons of the 1H and 13C NMR
data of diols 9, with the known syn diol,21 and anti
diol,22 allowed us to determine the syn and anti relative
configuration of aldols 8a,b. Whereas comparisons of
the specific rotations ([a]D) of acetonides 10 with the
known 10-anti acetonide23 allowed us to unambiguously
attribute the absolute configurations of the obtained
acetonides (since the specific rotation of diols 9 is low,
and thus subject to experimental errors). These mea-
Then, the two stereogenic centres at C2 and C3 (caribe-
nolide numbering) could be introduced by an asymmet-
ric aldol reaction. Indeed, the carbon–carbon bond
formation mediated by an aldol reaction is one of the
most powerful methods in organic synthesis for the con-
trol of the absolute configurations of the newly formed
stereogenic centres. Dibutylboron triflate enolates of
chiral N-acyl oxazolidinones and titanium enolates of
chiral N-acyl oxazolidin-2-thiones have been shown to
provide comparable levels of stereoselectivity.16–19 If
the Evans’ boron-mediated oxazolidinones give only
the ‘syn-Evans’ aldols, with Crimmins’ titanium-medi-
ated oxazolidin-2-thiones, it is possible to get separately
either one of the two syn products: the ‘syn-Evans’ and
1
sures, combined with the comparisons of the H NMR
of the acetonides, allowed us to unambiguously secure
the attribution of the absolute configurations of the
newly created stereogenic centres in 8a,b (Scheme 4).
O
S
N
O
a : R = H
b : R = Ph
R
R
Bn
aux. method d.r. (yield) diastereomers 8a,b
Base, Lewis acid
O
a
a
b
b
A
B
A
B
>98:2 (43) Syn Evans
70:30 (73) Syn non-Evans/anti
NR
BnO
75:25 (94) Syn non-Evans/anti
7
O
S
O
S
O
S
OH
OH
OH
BnO
BnO
BnO
+
+
N
O
N
O
N
O
R
R
R
Bn
Syn-Evans 8a,b
Bn
Bn
R
R
R
Syn-non-Evans 8a,b
Anti-non-Evans 8a,b
LiBH4
3.80-3.63
77.8
4.03
74.4
4.03
74.4
OH
OH
OH
OH
OH
OH
1.73
40.1
1.82
39.4
1.82
39.4
BnO
BnO
BnO
0.90
10.8
0.90
10.8
0.87
13.9
9-syn Evans
9-syn non-Evans
9-anti
(CH3)2C(OCH3)2
98.6
98.6
98.1
4.19
68.2
3.70-3.57
72.0
4.19
68.2
BnO
O
O
O
O
O
O
1.60
34.3
R
S
S
BnO
BnO
R
S
R
1.40
32.0
1.06
10.7
1.40
32.0
1.06
10.7
0.76
12.6
10-syn Evans
10-syn non-Evans
10-anti
[α]D = + 27 (c 0.83, CHCl3)
[α]D = - 27 (c 0.66, CHCl3)
[α]D = + 56 (c 0.98, CHCl3)
Scheme 4.

4178
M. Seck et al. / Tetrahedron Letters 47 (2006) 4175–4180
When the titanium enolate derived from oxazolidin-2-
thione a (1 equiv TiCl4, 2.5 equiv TMEDA, CH2Cl2,
ꢀ78 °C, method A)18 reacted with 1 equiv of aldehyde
7, the syn Evans aldol adduct 8a was obtained albeit in
a moderate 43% yield but with an excellent diastereo-
selectivity (>98:2). Again, the stereochemistry of the
compound was secured through the reduction forming
the intermediate diol 9 and preparation of acetonide
10. Surprisingly, adduct 8a was obtained with a low dia-
stereoselectivity (dr = 70:30), when the titanium enolate
derived from oxazolidin-2-thione a, prepared from
2 equiv TiCl4, and 1.1 equiv of DIEA in CH2Cl2, at
ꢀ78 °C (method B),18 reacted with 1 equiv of aldehyde
7. The major diastereomer was determined as the syn
non-Evans aldol 8a, after reduction to diol 9 and forma-
tion of acetonide 10. The minor diastereomer however
was surprisingly identified as the anti adduct 8a (and
not the syn Evans product as expected). The partial
formation of anti isomer 8a, under these reaction
conditions, is reported for the first time, as far as we
are aware, but can be related to the observed reverse
selectivity when other chiral enolates react with alde-
hydes bearing at the 3-position a chelating group such
as a benzyloxy function.24 However, when 3-tris-isopro-
pylsilyloxy propanal was treated by the titanium enolate
of a, prepared through method B, the syn non-Evans
aldol was now obtained as the sole product in 60%
yield (result not shown).
slightly higher diastereoselectivity (75:25) and the minor
diastereomer was again securely determined as the anti
adduct 8b. The combined yield was slightly better than
with chiral auxiliary a (94% vs 73%). Unfortunately,
the titanium enolate derived from oxazolidin-2-thione
b, prepared from 1 equiv of TiCl4, and 2.5 equiv of
TMEDA in CH2Cl2 at ꢀ78 °C (method A), did not react
with aldehyde 7 (despite several trials where the temper-
ature varied from ꢀ78 °C to rt).
Then, aldehyde 6 is now engaged in an asymmetric aldol
reaction using method B (2 equiv of TiCl4 and 1.1 equiv
of DIEA) to prepare the titanium enolate of the bulky
chiral N-propionyl oxazolidin-2-thione b.20 The reaction
affords only the syn non-Evans aldols 11 and 12 in 92%
chemical yield, and as a 1:1 mixture. The relative and
absolute configurations are supposed to be as depicted
in the scheme, based on the published data18,25 and on
the above observations (reduction of the amide function
to the corresponding alcohol and formation of the 1,3-
acetonides on related cases showed the syn non-Evans
relationship, see Ref. 8). At this stage the mixture of
the two diastereomers is due to the non-controlled stereo-
genic centre at C7 (caribenolide numbering), but fortu-
nately the two compounds could be separated by flash
chromatography. Both aldols 11 and 12 were then con-
verted into their corresponding methyl esters (by K2CO3
treatment in the presence of methanol) in high yield,20
and the chiral auxiliary recovered in typical 90–92%
yield, after purification by flash chromatography (only
conversion of 12 to 13 is represented in Scheme 5). Then
the free hydroxyl of 13 was protected as a triethyl silyl
ether by usual treatment (TESCl, DMAP, imidazole),26
to afford the expected silyl ether 14. Oxidative removal
Then, we were pleased to observe that when the titanium
enolate derived from oxazolidin-2-thione b (prepared
from 2 equiv TiCl4, 1.1 equiv of DIEA, in CH2Cl2, at
ꢀ78 °C, method B) reacted with 1 equiv of aldehyde 7,
the syn non-Evans adduct 8b was obtained with a
O
S
N
O
O
S
O
TBSO
TBSO
Ph
Ph
OH
Bn
Ph
S
N
O
PMBO
+
TBSO
11
12
O
Ph
b
Bn
O
Ph
PMBO
TiCl4 (2 equiv), iPr2NEt (1 equiv)
-78 C, CH2Cl2
N
6
°
OH
92%
Bn
Ph
PMBO
O
S
OMe
HN
Bn
O
TBSO
OMe
OH
+
K2CO3, MeOH
83%
Ph
12
Ph
PMBO
O
13
90%
TESCl, DMAP,
imidazole, DMF,16h,
82%
TBSO
OTES
PMBO
14
Scheme 5.

M. Seck et al. / Tetrahedron Letters 47 (2006) 4175–4180
4179
O
O
OMe
OTES
OMe
1) DDQ, CH2Cl2, H2O, 74%
TBSO
TBSO
OTES 2) PDC, PTA, CH2Cl2, reflux, 76%
PMBO
O
14
15
O
S
O
OMe
OTES
b
N
O
OTBS
S
O
OH
Ph
TiCl4 (2 equiv), DIEA (1 equiv)
Bn
Ph
O
N
+ isomer
CH2Cl2, -78°C
Ph
O
77%
Ph Bn
16
OMe
OTES
oxidation, protection,then
reductive removal of auxilliary
OTBS
O
O
O
H
Scheme 6.
of the paramethoxybenzyl group (DDQ, CH2Cl2,
H2O)27 of 14, and PDC oxidation of the primary alcohol
so obtained28 gave the expected aldehyde 15 in 56%
overall yield.
References and notes
1. Kobayashi, J.; Ishibashi, M.; Nakamura, H.; Ohizumi, Y.;
Yamasu, T.; Sasaki, T.; Hirata, Y. Tetrahedron Lett. 1986,
27, 5755.
2. Norcross, R. D.; Paterson, I. Chem. Rev. 1995, 95, 2041.
3. Bauer, I.; Maranda, L.; Young, K. A.; Shimizu, Y.;
Fairchild, C.; Cornell, L.; MacBeth, J.; Huang, S. J. Org.
Chem. 1995, 60, 1084.
A second aldol reaction using method B conditions with
the N-propionyl oxazolidin-2-thione b and aldehyde 15,
as described above, afforded unexpectedly a mixture of
the major syn non-Evans aldol 16 with an isomer
(dr = 70:30, supposed to be the anti isomer, as depicted
in the model study in Scheme 4, may be due to the pres-
ence of a protected hydroxyl at b-position of carbonyl)
in 77% combined yield (Scheme 6). However, oxidation
of the carbinol of both isomers will give rise to the same
oxo derivative. Indeed, oxidation of alcohol 16 followed
by protection of the carbonyl and reductive removal of
the chiral auxiliary (e.g., DIBAL reduction) will give
rise to the expected aldehyde that is required for the
coupling reaction with the C12–C29 fragment of
caribenolide I.
4. For an overview of synthetic efforts, see: A¨ıssa, C.;
Riveiros, R.; Ragot, J.; Furstner, A. J. Am. Chem. Soc.
¨
2003, 125, 15512.
5. Bauer, I.; Maranda, L.; Shimizu, Y.; Peterson, R. W.;
Cornell, L.; Steiner, J. R.; Clardy, J. J. Am. Chem. Soc.
1994, 116, 2657.
6. Ishibashi, M.; Ohizumi, Y.; Hamashima, M.; Nakamura,
H.; Hirata, Y.; Sasaki, T.; Kobayashi, J. J. Chem. Soc.
Chem. Commun. 1987, 1127.
7. Kobayashi, J.; Ishibashi, M.; Nakamura, H.; Ohizumi, Y.;
Yamasu, T.; Hirata, Y.; Sasaki, T.; Ohta, T.; Nozoe, S. J.
Nat. Prod. 1989, 52, 1036.
8. Seck, M.; Seon-Meniel, B.; Jullian, J.-C.; Franck, X.;
`
Hocquemiller, R.; Figadere, B. Lett. Org. Chem. 2006, 3,
In conclusion, we have described herein a very efficient
stereoselective synthesis of the enantiopure C1–C11 skel-
eton of caribenolide I, in 13 steps with a good control of
the configurations of the newly built stereogenic centres.
In principle, all diastereomers are accessible, depending
on the reaction conditions used for the aldol reactions,
using our bulky oxazolidin-2-thione b.
390.
9. Basavaiah, D.; Rao, A. J.; Satyanarayana, T. Chem. Rev.
2003, 103, 811.
10. Franck, X.; Figadere, B. Tetrahedron Lett. 2002, 43, 1449.
`
11. Selected spectroscopic data of the 1:1 mixture of (E) and
1
(Z) Baylis–Hillman adduct: H NMR (400 MHz, CDCl3)
d: 8.19 (d, J = 8.8 Hz, 2HaroZ), 8.18 (d, J = 8.8 Hz,
2HaroE), 7.56 (d, J = 8.6 Hz, 2HaroZ), 7.54 (d,
J = 8.9 Hz, 2HaroE), 7.40 (d, J = 11.6 Hz, 1H3E), 7.30
(ddd, J = 10.1, 11.1, 16.9 Hz, 1H4Z), 6.86 (ddd, J = 10.1,
11.4, 16.6 Hz, 1H4E), 5.89 (br d, J = 9.3 Hz, 1H30E), 5.88
(br s, 1H30Z), 5.83 (d, J = 16.2 Hz, 1H5E), 5.70 (d,
J = 10.6 Hz, 1H5E), 5.60 (d, J = 16.6 Hz, 1H5Z), 5.57 (d,
J = 10.2 Hz, 1H5Z), 4.16 (m, OCH2), 1.25 (t, J = 7.1 Hz,
CH3E), 1.23 (t, J = 7.1 Hz, CH3Z); 13C NMR (75 MHz,
CDCl3) d: 167.1, 166.2, 150.1, 144.5, 141.3, 132.8, 130.5,
130.3, 128.8, 127.1, 127.0, 126.1, 123.5, 74.6, 69.1, 61.3,
61.1, 14.0.
Diastereoselective epoxidation of the so obtained allylic
alcohol will be performed at a later stage, because of the
high chemical reactivity of such an allylic epoxide, and
the connection of this fragment with other parts of the
molecule is now under study in our laboratories.
Acknowledgement
12. Corey, E. J.; Venkateswarlu, A. J. Am. Chem. Soc. 1972,
94, 6190.
13. Parsons, W.; Schlessinger, R. H.; Quesada, M. L. J. Am.
Chem. Soc. 1980, 102, 889.
`
`
We wish to thank the Ministere des Affaires Etrangeres
for a fellowship to M.S.

4180
M. Seck et al. / Tetrahedron Letters 47 (2006) 4175–4180
14. Greene, A. E.; Le Drian, C.; Crabbe, P. J. Am. Chem. Soc.
1980, 102, 7583.
21. Ghosh, A. K.; Liu, C. J. Am. Chem. Soc. 2003, 125,
2374.
15. Wadsworth, W. S.; Emmons, W. D. J. Am. Chem. Soc.
1961, 83, 1733.
22. Chattopadhyay, S. K.; Pattenden, G. J. Chem. Soc.,
Perkin Trans. 1 2000, 2429.
16. Evans, D. A.; Bartoli, J.; Shih, T. L. J. Am. Chem. Soc.
1981, 103, 2127.
23. Rychnovsky, S. D.; Hoye, R. C. J. Am. Chem. Soc. 1994,
116, 1753.
17. Evans, D. A.; Nelson, J. V.; Taber, T. R. Top. Stereochem.
1982, 13, 1.
18. Crimmins, M. T.; King, B. W.; Tabet, E. A.; Chaudhary,
K. J. Org. Chem. 2001, 66, 894.
19. Evans, D. A.; Kim, A. S. In Handbook of Reagents for
Organic Synthesis: Reagents, Auxiliaries and Catalysts for
C–C Bonds; Coates, R. M., Denmark, S. E., Eds.; John
Wiley & Sons: New York, 1999; p 91.
24. (a) Ghosh, A. K.; Fidanze, S.; Onishi, M.; Hussain, K. A.
Tetrahedron Lett. 1997, 38, 7171; (b) Gennari, C.; Cozzi,
P. G. J. Org. Chem. 1988, 53, 4015.
25. Crimmins, M. T.; King, B. W.; Tabet, E. J. Am. Chem.
Soc. 1997, 119, 7883.
26. Oppolzer, W.; Snowden, R. L.; Simmons, D. P. Helv.
Chim. Acta 1981, 64, 2002.
27. Oikawa, Y.; Yoshioka, T.; Yonemitsu, O. Tetrahedron
Lett. 1982, 23, 885.
20. Jalce, G.; Seck, M.; Franck, X.; Hocquemiller, R.;
`
Figadere, B. J. Org. Chem. 2004, 69, 3240.
28. Corey, E. J.; Schmidt, G. Tetrahedron Lett. 1979, 399.