Toward a total synthesis of amphidinolide X and Y. The tetrahydrofuran- containing fragment C12-C21

Article abstract of DOI:10.1021/ol063035y

(Chemical Equation Presented) The key THF derivative (9a) for an enantioseiective synthesis of amphidinolide X/Y was obtained from 1a via a selenoetherification reaction. In fact, among the cyclization methods investigated, the highest yield and stereocon

Full text of DOI:10.1021/ol063035y

ORGANIC
LETTERS
2007
Vol. 9, No. 6
989-992
Toward a Total Synthesis of
Amphidinolide X and Y. The
Tetrahydrofuran-Containing Fragment
C12−C21
Carles Rodr´ıguez-Escrich, Anna Olivella, Fe`lix Urp´ı,* and Jaume Vilarrasa*
Departament de Qu´ımica Orga`nica, Facultat de Qu´ımica, UniVersitat de Barcelona,
AV. Diagonal 647, 08028 Barcelona, Catalonia, Spain
jVilarrasa@ub.edu
Received December 15, 2006
ABSTRACT
The key THF derivative (9a) for an enantioselective synthesis of amphidinolide X/Y was obtained from 1a via a selenoetherification reaction.
In fact, among the cyclization methods investigated, the highest yield and stereocontrol were achieved at 78
C with PhSeCl/EtiPr2N from
−
°
diols 1a (anti-Z) and 1b (anti-E) and with PhSeCl/ZnBr2 from diols 1c (syn-Z) and 1d (syn-E). Also, surprisingly, use of protecting groups (on
the allylic OH) was detrimental in the cases studied. The diverse THF-tetrasubstituted stereoisomers will provide a series of amphidinolide X/Y
analogues.
Amphidinolide X and its supposed biogenetic precursor
amphidinolide Y were isolated by Kobayashi et al.1 from a
Scheme 1
marine dinoflagellate of the genus Amphidinium sp. Among
the members of the amphidinolide family,2 many of which
exhibit a strong cytotoxic activity against several cancer cell
lines,2 amphidinolide X drew our attention both by its
unusual nondimeric macrodiolide structure and by the
synthetic challenge posed by the quaternary stereocenter of
the tetrahydrofuran ring. As shown in Scheme 1, where the
numbering of amphidinolide X was adapted to that of
amphidinolide Y for the sake of simplification, our synthetic
(1) (a) Tsuda, M.; Izui, N.; Shimbo, K.; Sato, M.; Fukushi, E.; Kawabata,
J.; Katsumata, K.; Horiguchi, T.; Kobayashi, J. J. Org. Chem. 2003, 68,
5339. (b) Tsuda, M.; Izui, N.; Shimbo, K.; Sato, M.; Fukushi, E.; Kawabata,
J.; Katsumata, K.; Horiguchi, T.; Kobayashi, J. J. Org. Chem. 2003, 68,
9109.
(2) Reviews: (a) Chakraborty, T. K.; Das, S. Curr. Med. Chem.: Anti-
Cancer Agents 2001, 1, 131. (b) Kobayashi, J.; Tsuda, M. Nat. Prod. Rep.
2004, 21, 778. For a review on oxacyclic macrodiolides, see: (c) Kang, E.
strategy was based on two disconnections that led to tetra-
J.; Lee, E. Chem. ReV. 2005, 105, 4348.
substituted oxolane derivatives.
10.1021/ol063035y CCC: $37.00
© 2007 American Chemical Society
Published on Web 02/13/2007

Thus, we envisaged the formation of the five-membered
ring by an electrophile-mediated cyclization of 1a. An alter-
native approach to the key five-membered ring has been
reported by Fu¨rstner et al. in their outstanding total syntheses
of amphidinolide X and Y;3 construction of the THF deriv-
ative relied upon the conversion of a chiral epoxy-alkyne
to a hydroxy-allene, which was cyclized with AgNO3 to
afford an 8:1 mixture of cis/trans-dihydrofurans, separated
after a subsequent bromoesterification. Dai et al. have re-
ported another interesting approach to this THF ring, through
an acid-catalyzed 5-endo cyclization of a vinyl epoxide.4
We have synthesized 1a as indicated in Scheme 2.
Nonnatural Evans’ chiral auxiliary was converted to the tert-
anti/syn mixture, which was purified by chromatography.
Cleavage of the TBS ether under standard conditions
(Bu4N+F-‚3H2O, THF) afforded the desired enantiopure 4,5-
dihydroxy-7-methyl-6-decenenitrile 1a.
Analogously, but starting from the E-iodoalkene (5b)10
shown in Scheme 2, we obtained the decenenitrile 6b, again
as an anti/syn mixture (93:7), which was deprotected and
purified to give 1b. The Swern oxidation of 6a, followed by
reduction with L-Selectride (lithium tri-sec-butylborohydride)
in THF at -78 °C, afforded in ca. 60% overall yield the
syn-diol 6c, which was deprotected to 1c. Similarly, a sample
of the anti-E derivative 6b was transformed to the syn-E 6d
and 1d. In this way, four different series of enediols (a )
anti-Z, b ) anti-E, c ) syn-Z, and d ) syn-E) were finally
in our hands. In principle, it was expected that the cyclization
of some of these derivatives would afford a THF-containing
fragment with the absolute configuration of amphidinolides
X/Y, whereas the other series would give rise eventually to
diverse stereoisomers (analogues), whose cytotoxicities
would deserve to be evaluated as well.
Scheme 2
Cycloetherification of homoallylic alcohols (formation of
tetrahydrofurans by 5-endo processes) and cyclization be-
tween more distant hydroxy groups and olefinic carbon atoms
(5-exo vs 6-endo, etc.) have been extensively investigated,11
but competition studies involving di-OH or tri-OH olefins
are scarce.12 In our case (“R,â-di-OH trisubstituted olefins”),
as regioselective electrophilic attacks giving rise to inter-
mediates with some tertiary carbenic ion character were
plausible, it was thought that the oxolane would predominate
over the two possible oxetanes and over the oxirane.
Therefore, the unprotected diols 1a and 1b were subjected
to various electrophile-mediated cyclization reactions.13
(9) (Z)-1-Iodo-2-methyl-1-pentene (5a) was prepared from propyne,
propylmagnesium chloride, and CuBr, followed by iodination, according
to: (a) Hoffmann, R. W.; Schlapbach, A. Liebigs Ann. Chem. 1990, 1243.
For the preparation of the zincate and its addition to aldehyde 4, see:
(b) Marshall, J. A.; Eidam, P. Org. Lett. 2004, 6, 445 and references
therein.
(10) (E)-1-Iodo-2-methyl-1-pentene (5b) was prepared from 1-pentyne
(with AlMe3 and Cp2ZrCl2), followed by iodination. See: Negishi, E.; Horn,
D. E. V.; Yoshida, T. J. Am. Chem. Soc. 1985, 107, 6639.
butyldimethylsilyloxy derivative 2, the titanium enolate of
which was treated with acrylonitrile,5 to give enantiopure 3
in 89% overall yield.6 By the method of Fukuyama et al.,7
3 was reduced to aldehyde 4. Reaction of 48 with the
appropriate alkenylzincate (from 5a)9 led to 6a as a 92:8
(11) For recent reviews, see: (a) Gruttadauria, M.; Meo, P. L.; Noto, R.
Targets Heterocycl. Syst. 2001, 5, 31. (b) Silva, F. M.; Jones, J.; Mattos,
M. C. S. Curr. Org. Synth. 2005, 2, 393. For relevant papers, see: (c)
Lipshutz, B. H.; Gross, T. J. Org. Chem. 1995, 60, 3572 and references
therein. (d) Bravo, F.; Kassou, M.; Castillo´n, S. Tetrahedron Lett. 1999,
40, 1187. (e) Bew, S. P.; Barks, J.; Knight, D. W.; Middleton, R. J.
Tetrahedron Lett. 2000, 41, 4447. (f) Senda, Y.; Takayanagi, S.; Sudo, T.;
Itoh, H. J. Chem. Soc., Perkin Trans. 1 2001, 270. (g) Kang, S. H.; Kang,
S. Y.; Park, C. M.; Kwon, H. Y.; Kim, M. Pure Appl. Chem. 2005, 77,
1269. (h) Alonso, F.; Mele´ndez, J.; Soler, T.; Yus, M. Tetrahedron 2006,
62, 2264 and references therein.
(12) Recent examples: (a) Bravo, F.; Castillo´n, S. Eur. J. Org. Chem.
2001, 507 (â,γ-di-OH). (b) Gruttadaria, M.; Aprile, C.; Riela, S.; Noto, R.
Tetrahedron Lett. 2001, 42, 2213 (â,δ-di-OH). (c) Weghe, P. V.; Bourg,
S.; Eustache, J. Tetrahedron 2003, 59, 7365 (R,â′-di-OH, Z olefin). (d)
Fettes, A.; Carreira, E. M. J. Org. Chem. 2003, 68, 9274 (R,δ′-di-OH) and
references therein. (e) Li, D.-R.; Zhang, D.-H.; Sun, C.-Y.; Zhang, J.-W.;
Yang, L.; Chen, J.; Liu, B.; Su, C.; Zhou, W.-S.; Lin, G.-Q. Chem.-Eur.
J. 2006, 12, 1185 (â,γ-di-OH). (f) Nicolaou, K. C.; Pihko, P. M.; Bernal,
F.; Frederick, M. O.; Qian, W.; Uesaka, N.; Diedrichs, N.; Hinrichs, J.;
Koftis, T. V.; Loizidou, E.; Petrovic, G.; Rodr´ıquez, M.; Sarlah, D.; Zou,
N. J. Am. Chem. Soc. 2006, 128, 2244 (R,γ,δ-tri-OH).
(3) (a) Lepage, O.; Kattnig, E.; Fu¨rstner, A. J. Am. Chem. Soc. 2004,
126, 15970. (b) Fu¨rstner, A.; Kattnig, E.; Lepage, O. J. Am. Chem. Soc.
2006, 128, 9194.
(4) Chen, Y.; Jin, J.; Wu, J.; Dai, W.-M. Synlett 2006, 1177.
(5) (a) Evans, D. A.; Bilodeau, M. T.; Somers, T. C.; Clardy, J.; Cherry,
D.; Kato, Y. J. Org. Chem. 1991, 56, 5750. (b) Mas, G.; Gonza´lez, L.;
Vilarrasa, J. Tetrahedron Lett. 2003, 44, 8805.
(6) Reactions with alternative Michael acceptors and drawbacks with
the enolate of the O-Bn derivative will be described later. The absolute
configuration of 3 and 4 was confirmed by reduction to the alcohol, removal
of the TBS group, and formation of the isopropylidene acetal, a known
compound ([R]D -27.3, c 1.1, CHCl3, cf. Buchanan, J. G.; Craven, D. A.;
Wightman, R. H.; Harnden, M. R. J. Chem. Soc., Perkin Trans. 1 1991,
195).
(7) (a) Miyazaki, T.; Han-ya, Y.; Tokuyama, H.; Fukuyama, T. Synlett
2004, 477. (b) For a review: Fukuyama, T.; Tokuyama, H. Aldrichimica
Acta 2004, 37, 87.
(13) Reviews: (a) Tiecco, M. Top. Curr. Chem. 2000, 208, 7. (b) Wirth,
T. Angew. Chem., Int. Ed. 2000, 39, 3742. (c) Petragnani, N.; Stefani, H.
A.; Valduga, C. J. Tetrahedron 2001, 57, 1411. Also see: (d) Khokhar, S.
S.; Wirth, T. Angew. Chem., Int. Ed. 2004, 43, 631 and references therein.
(8) Compound 4 was used after a simple filtration on Celite. Further
purification of this racemization-prone aldehyde, to remove the coproduct
C12H25SSiEt3, proved to be unnecessary.
990
Org. Lett., Vol. 9, No. 6, 2007

In practice, iodoetherification (with I2/NaHCO3 in
CH3CN) as well as treatment with Hg(OAc)2 gave complex
final mixtures. Cyclization by small amounts of TfOH was
unsuccessful (decomposition products). Also, Me3SiOTf was
inefficient. Heating with AgOTf as reported14 gave dehy-
drated products (probably due to the instability of the allylic
system to Lewis acids or when Bro¨nsted acids are generated
in the medium), whereas at room temperature other reactions
were noted.15 Cu2(OTf)2 behaved as AgOTf.15 Only the
PhSeCl-induced cyclizations gave clean results.
bonds) have been given11-13 so that the main arguments will
not be repeated. We are dealing, however, with the special
case of a trisubstituted alkene with allylic and homoallylic
hydroxy groups. The major compound of the PhSeδ+-
promoted cyclization of anti-Z diol 1a may be accounted
for by a predominant envelope or chairlike conformation of
the complex and intermediates leading to 7a (see the upper
part of Scheme 3, where X means Cl, ZnBr2Cl, or any
As shown in Table 1 for the reaction of 1a with PhSeCl
under quite standard conditions,16 a compound largely
Scheme 3
Table 1. Reaction of 1a with PhSeCl
entry
conditionsa
yield of 7a (%) yield of 8a (%)
1
2
3
4
5
ZnBr2, rt, 15 min
60
75
85
92
92
18
10
8
3
3
counterion of PhSe+), in which both the allylic OH group
and the CH2CH2CN chain are in an equatorial-like arrange-
ment. In the lower part of Scheme 3, the less favorable
conformation leading to 8a is depicted. For more details,
see the Supporting Information.
K2CO3, rt, 30 min
ZnBr2, -78 °C, 30 min
K2CO3, -78 °C, 6 h
EtiPr2N, -78 °C, 8 h
a
To 1.0 mmol of 1a in 10 mL of THF at the indicated temperature
were added 1.5 mmol of PhSeCl and 1.0 mmol of ZnBr2 in 4 mL of THF
via canula, and stirring was maintained for several minutes or hours. In
entries 2, 4, and 5, the PhSeCl solution was added to the suspension or
solution of 1a with the base (1.0 mmol).
The results obtained by the PhSeCl-promoted cyclization
of 1b are summarized in Table 2. A THF derivative largely
Table 2. Reaction of 1b with PhSeCl
predominated (see 7a) but another isomer (see 8a) of higher
Rf (with 95:5 CH2Cl2/EtOAc as the eluent) was also isolated.
1
Apart from the relevant H and 13C chemical shifts and
3JHH values, which agree with those reported for related
oxolanes,17 NOESY confirmed the configurations of these
products. The maximum stereocontrol in favor of the desired
compound 7a (see entries 4 and 5) was obtained at -78 °C
and when no Lewis acid was used to activate PhSeCl, but a
base was introduced to remove HCl as soon as generated.
Thus, under the conditions of entries 4 and 5, 92% of 7a
could be obtained after column chromatography.
entry
conditionsa
yield of 7b (%) yield of 8b (%)
1
2
3
4
5
ZnBr2, rt, 15 min
73
76
80
90
90
6
6
5
4
4
ZnBr2, -78 °C, 30 min
K2CO3, rt, 30 min
K2CO3, -78 °C, 4 h
EtiPr2Nb -78 °C, 6 h
Sound explanations for the stereocontrolled PhSeX-
induced cyclizations (anti stereospecific additions to double
a
See footnote a in Table 1. In THF, CH2Cl2, and toluene, at -78 °C
(14) Yang, C.-G.; Reich, N. W.; Shi, Z.; He, C. Org. Lett. 2005, 7, 4553.
(15) Such as the addition of the homoallylic OH group to the CN group,
to afford five-membered lactones after the workup. For example, 1b stirred
in CH2Cl2 with AgOTf for 4 days at room temperature gave 35% of a
butyrolactone. When MEM-1b (the 5-O-MEM derivative of 1b, see the
main text) was treated in CH2Cl2 with 0.5 equiv of AgOTf overnight, 21%
of the corresponding O-MEM butyrolactone was isolated (together with
43% of recovered starting material). When MEM-1b was stirred with Cu2-
(OTf)2 for 1 day, ca. 70% of the same butyrolactone was isolated.
(16) Kang, S. H.; Hwang, T. S.; Kim, W. J.; Lim, J. K. Tetrahedron
Lett. 1990, 31, 5917. Because stereoselectivities obtained with commercially
available PhSeCl were encouraging in the first trials, more sterically hindered
reagents such as Lipshutz’s 2,4,6-triisopropylphenylselenyl bromide (see
refs 11c and 12d) were not investigated.
for 30 min, without any additive, the product ratios were similar, with
b
isolation of 7b in ca. 80% yields.
butylpyridine, the reactions were slower, but the product ratios were similar.
With lutidine and 2,6-di-tert-
predominated, to which structure 7b could be attributed on
the basis of its NMR spectra (especially the high 3JHH values
between the ring protons, which indicate that they are in a
trans-trans relationship, and NOE experiments).
From 6b, we prepared its O-MEM derivative (protection
of the allylic OH) and removed the O-TBS protecting group
(deprotection of the homoallylic OH) by standard procedures,
to obtain the so-called MEM-1b. With PhSeCl at room
(17) (a) Tiecco, M.; Testaferri, L.; Santi, C. Eur. J. Org. Chem. 1999, 4,
797. (b) Reitz, A. B.; Nortey, S. O.; Maryanoff, B. E.; Liotta, D.; Monahan,
R. J. Org. Chem. 1987, 52, 4191.
Org. Lett., Vol. 9, No. 6, 2007
991

temperature, without any additive, MEM-1b gave a mixture
of two oxolanes in a 75:25 ratio. Cyclization of MEM-1b
with PhSeCl/EtiPr2N at -78 °C gave a similar crude product,
with the two oxolanes in an 80:20 ratio. The major isomer
was in all cases MEM-7b, as expected, and could be obtained
in ca. 70% yield after column chromatography. The second
isomer, isolated in ca. 15% yield, correlated with MEM-8b.
The relevant fact is that the stereocontrol was much lower
from MEM-1b than from 1b. In other words, with this
substrate it is better to discard the use of protecting groups,18
not only to save steps but also to improve the selectivity.
We explain this fact by a destabilizing gauchelike interac-
tion between the OMEM and CH2CH2CN chains of MEM-
1b, which may change the relative percentages of the
conformers (see Supporting Information).
Removal of the SePh group of oxolanes 7a, 7b, 8a, and
8b was carried out with (Me3Si)3SiH/AIBN19 in yields
>95%, as summarized in Scheme 5. Deselenylation of 7a
Scheme 5
When syn-Z diol 1c and syn-E diol 1d were subjected to
diverse selenoetherification conditions such as those shown
in Tables 1 and 2, almost equimolar mixtures of oxolanes
7c and 8c and, on the other hand, 7d and 8d were obtained
at room temperature. However, the main products at low
temperatures (see Scheme 4) were 8c and 8d, respectively;
gave 9a, the same product as that of 8b, and deselenylation
of 7b afforded 9b, the same compound as that of 8a,
confirming the structural assignments we had previously
carried out.
In summary, compound 9a, the key fragment in our total
synthesis of amphidinolide X/Y, was achieved in excellent
yields by a cyclization of an anti dihydroxy (allylic and
homoallylic), unprecedented trisubstituted-alkene substrate
(1a), followed by deselenylation. Other isomers of 9a, such
as 9b (and those arising from 7c and 8d and from 8c and
7d, to be reported in a future full work), will be used for the
syntheses and SAR studies of amphidinolide X/Y analogues.
As we obtained the C1-C6 synthon some time ago, our
present efforts are focused on the elaboration of the C23/
C7-C11 segment and the assembly of the three fragments.
Scheme 4
Acknowledgment. Financial support from the Spanish
Ministerio de Educacio´n y Ciencia, through the grant
SAF2002-02728 and a studentship to C.R.E. (May 2003-
April 2007), is acknowledged. The Generalitat de Catalunya
contributed partially by means of the grant 2001-SGR065
(Grup de S´ıntesi Estereoselectiva d’Antibio`tics i Antiv´ırics).
The MEC grant CTQ-2006-15393 is also acknowledged.
Thanks are due to Prof. Xavier Solans and Dr. Merce` Font-
Bardia (Facultat de Geologia, Universitat de Barcelona) for
the X-ray structure of 8c.
i.e., they had the OH and SePh groups in a cis relationship,
in contrast to that observed in series a and b. The configura-
tions were assigned by NMR and were confirmed by the
X-ray structure of 8c (see Supporting Information). By
contrast again to 1a and 1b, under basic conditions at -78
°C, 1c gave a 7c/8c ratio of only 25:75 and 1d afforded only
a 40:60 7d/8d ratio. To our surprise, the highest ratios were
reached in the presence of 1 equiv of ZnBr2 at -78 °C or,
even slightly better, at -100 °C (see in Scheme 4 the highest
isolated yields of 8c and 8d).
Supporting Information Available: Experimental data
for compounds 1a-d, 7a, 7b, 8c, 8d, 9a, and 9b, their NMR
spectra, the X-ray structure of 8c, and mechanistic interpreta-
tions. This material is available free of charge via the Internet
at http://pubs.acs.org.
OL063035Y
(18) The benzyl (Bn) and p-methoxybenzyl (PMB) groups were not
convenient as protecting groups because we obtained many byproducts when
various cyclizations were attempted with them instead of MEM.
(19) Classical review: (a) Chatgilialoglu, C. Acc. Chem. Res. 1992, 25,
188. Also see: (b) Parsons, A. Chem. Br. 2002, 38, 42. (c) Perchyonok, V.
T. Tetrahedron Lett. 2006, 47, 5163 and references therein.
992
Org. Lett., Vol. 9, No. 6, 2007