Synthesis of (+)-8-deoxyvernolepin and its 11,13-dihydroderivative. A novel reaction initiated by sulfene elimination leads to the 2-oxa-cis-decalin skeleton

Article abstract of DOI:10.1021/jo0256538

The title compounds are interesting candidates for antifungal screening. This paper describes the enantiospecific synthesis of these compounds starting from (+)-costunolide isolated from a commercially available extract. We used two novel reactions as key synthetic steps in this work: the acid-induced cyclization of an δ,ε-epoxy ester, which stereoselectively gave a hydroxymethyl-substituted δ-lactone, with the hydroxyalkyl group in the desired β-equatorial disposition, and a reaction cascade, initiated by a base-promoted sulfene elimination, which led to a 10-oxiranyl-2-oxa-cis-decalin from the mesylate of a trans-fused δ-lactone. We also found that the reaction between selenium dioxide and the 1,5-diene system of elemanolides gave selenadecalins analogous to natural eudesmanolides. Our results prove that the synthetic strategy employed, on the basis of biomimetic concepts, is a useful procedure for the enantiospecific preparation of (+)-vernolepin-related compounds from accessible germacrolides.

Full text of DOI:10.1021/jo0256538

Syn th esis of (+)-8-Deoxyver n olep in a n d Its
11,13-Dih yd r od er iva tive. A Novel Rea ction In itia ted by Su lfen e
Elim in a tion Lea d s to th e 2-Oxa -cis-d eca lin Sk eleton
Alejandro F. Barrero,* J . Enrique Oltra,* M´ıriam AÄ lvarez,† and Antonio Rosales
Department of Organic Chemistry, Faculty of Sciences, University of Granada, Campus Fuentenueva s/ n,
E-18071 Granada, Spain
joltra@goliat.ugr.es; afbarre@goliat.ugr.es
Received February 25, 2002
The title compounds are interesting candidates for antifungal screening. This paper describes the
enantiospecific synthesis of these compounds starting from (+)-costunolide isolated from a
commercially available extract. We used two novel reactions as key synthetic steps in this work:
the acid-induced cyclization of an δ,ꢀ-epoxy ester, which stereoselectively gave a hydroxymethyl-
substituted δ-lactone, with the hydroxyalkyl group in the desired â-equatorial disposition, and a
reaction cascade, initiated by a base-promoted sulfene elimination, which led to a 10-oxiranyl-2-
oxa-cis-decalin from the mesylate of a trans-fused δ-lactone. We also found that the reaction between
selenium dioxide and the 1,5-diene system of elemanolides gave selenadecalins analogous to natural
eudesmanolides. Our results prove that the synthetic strategy employed, on the basis of biomimetic
concepts, is a useful procedure for the enantiospecific preparation of (+)-vernolepin-related
compounds from accessible germacrolides.
In tr od u ction
and an antifungal activity comparable to that of ampho-
tericin B.9 As far as 8-deoxyvernolepin is concerned, the
racemic form exhibits higher cytotoxic potency than
vernolepin against human lymphoblastic leukemia cells
in culture,3 and the (+)-enantiomer (3) also shows
antitumor activity in vivo.10 Moreover, as it is believed
that the antifungal activity of sesquiterpene lactones is
inversely related to their polarity,11 3 (which is less polar
than 1 and 2) is an interesting candidate for antifungal
screening.
(+)-Vernolepin (1) and (+)-vernodalin (2), from the
Ethiopian plants Vernonia hymenolepis1 and V. amigdali-
na,2 respectively, and the synthetic derivative 8-deoxy-
vernolepin3 (3) form a small group of sesquiterpene
lactones (elemanolides)4 with a unique 10-vinyl-2-oxa-
cis-decalin skeleton. These substances have interesting
(4) For
a classical review on the biogenesis and chemistry of
sesquiterpene lactones and the main skeletal types found in nature
(germacrolides, melampolides, elemanolides, eudesmanolides, etc.),
see: Fischer, N. H.; Olivier, E. J .; Fischer, H. D. Prog. Chem. Org.
Nat. Prod. 1979, 38, 47.
(5) Antitumor capacity both in vitro and in vivo: (a) Kupchan, S.
M.; Hemingway, R. J .; Werner, D.; Karim, A. J . Org. Chem. 1969,
34, 3903. (b) Kupchan, S. M.; Eakin, M. A.; Thomas, A. M. J . Med.
Chem. 1971, 14, 1147. (c) J isaka, M.; Ohigashi, H.; Takegawa, K.;
Huffman, M. A.; Koshimizu, K. Biosci. Biotechnol. Biochem. 1993, 57,
833.
(6) Antiaggregating, deaggregating, and spasmolytic activities: Vli-
etinck, A. J . In Biologically Active Natural Products; Hostettmann,
R., Led, P. J ., Eds.; Clarendon Press: Oxford, 1987; p 42.
(7) Antiparasitic properties: Robles, M.; Aregullin, M.; West, J .;
Rodr´ıguez, E. Planta Med. 1995, 61, 199.
(8) Al Magboul, A. Z. I.; Bashir, A. K.; Khalid, S. A.; Farouk, A.
Fitoterapia 1997, 68, 83.
biological properties,5-7 and thus, 1 and 2 showed stron-
ger antibacterial potency than ampicilin and neomycin8
(9) Amphotericin B is considered the “gold standard” for the treat-
ment of invasive fungal infections. See: Georgopapadakou, N. H.;
Walsh, T. J . Science 1994, 264, 371.
(10) Watanabe, M.; Yoshikoshi, A. J . Chem. Soc., Perkin Trans. 1
1987, 2833.
* To whom correspondence should be addressed. Fax: (A.F.B.) 34
958 24 33 18; (J .E.O.) 34 958 24 84 37.
† Present address: Organic Chemistry Area, University of Almer´ıa,
Almer´ıa, Spain.
(1) Kupchan, S. M.; Hemingway, R. J .; Werner, D.; Karim, A.;
McPhail, A. T.; Sim, G. A. J . Am Chem. Soc. 1968, 90, 3596.
(2) Kupchan, S. M.; Hemingway, R. J .; Karim, A.; Werner, D. J . Org.
Chem. 1969, 34, 3908.
(3) Grieco, P. A.; Noguez, J . A.; Masaki, Y. J . Org. Chem. 1977, 42,
495.
(11) (a) Inoue, A.; Tamogami, S.; Kato, H.; Nakazato, Y.; Akiyama,
M.; Kodama, O.; Akatsuka, T.; Hashidoko, Y. Phytochemistry 1995,
39, 845. (b) Barrero, A. F.; Oltra, J . E.; AÄ lvarez, M.; Raslan, D. S.;
Sau´de, D. A.; Akssira, M. Fitoterapia 2000, 71, 60. (c) Skaltsa, H.;
Lazari, D.; Panagouleas, C.; Georgiadou, E.; Garc´ıa, B.; Sokovic, M.
Phytochemistry 2000, 55, 903.
10.1021/jo0256538 CCC: $22.00 © 2002 American Chemical Society
Published on Web 07/11/2002
J . Org. Chem. 2002, 67, 5461-5469
5461

Barrero et al.
The research groups of Grieco,3,12 Danishefsky,13 and
others14 have reported the total synthesis of 1 and/or 3
in racemic form, and more recently, the asymmetric
synthesis of one of Danishefsky’s intermediates has
allowed the enantioselective preparation of (+)-vernol-
epin.15 These syntheses require numerous steps, however,
and provide low overall yields. Some procedures for the
synthesis of (+)-8-deoxyvernolepin (3) from (-)-R-santo-
nin (in 15 or more steps) have been reported also.16
Santonin, however, lacks any functional group at C-8,
and this is a serious drawback if one wants to extend
this strategy to the synthesis of the C-8 functionalized
compounds 1 and 2. Moreover, the procedures employed
to build the ∆3 exocyclic double bond of 3 always produced
isomeric mixtures and a concomitant fall in overall
yield.10,17
SCHEME 1
Elemanolides are believed to be synthesized from
germacrolides in nature.4 So, encouraged by this idea,
we have developed a new strategy that might facilitate
the synthesis of elemanolides 1-3, starting from acces-
sible germacrolides.18 The aim of the present work has
been to prove the value of this strategy for the syntheses
of (+)-8-deoxyvernolepin (3) and the 11,13-dihydroge-
nated derivative 4. Further screening of these compounds
might help us find new antifungal drugs and provide
information about the potential relationships, if any,
between the antifungal activity of sesquiterpene lactones
and the R-methylene-γ-lactone group, which is mainly
responsible for the toxic activity of these compounds
against mammalian cells.19
lactone 10 (step d) to afford 2-oxa-cis-decalin 11. The
remaining transformations of Scheme 1 can be achieved
by conventional methods, especially by using Grieco’s
procedure to build the ∆11(13) double bond of 3.21
Costunolide, the raw material proposed in Scheme 1,
can be obtained in racemic form by total synthesis.22
Nevertheless, we chose to start with (+)-costunolide (5)
isolated from commercially available Costus Resinoid,20
in order to confer an enantiospecific character to our
process. The synthesis began with the selective reduction
of the conjugate double bond of 5 to avoid potential
complications with a reactive Michael acceptor group.
This double bond can easily be restored at the end of the
synthetic sequence.21 Thus, catalytic hydrogenation of 5,
under mild conditions, took place regio- and stereoselec-
tively, giving 11â,13-dihydrocostunolide23 (12) at a yield
of 93% (Scheme 2). Similar selectivity degrees were
previously observed for related sesquiterpene lactones
under the same conditions.18b
With the germacrolide 12 in our hands, we envisaged
the first key step of our synthesis, which was the
rearrangement of the germacrane skeleton to the el-
emane one. Our previous work showed that 15-oxoger-
macrolides undergo Cope rearrangement in toluene at
reflux, giving elemanolides in medium-to-high yields.18b
It might, therefore, be thought that, to facilitate the
subsequent rearrangement, the transformation of 12 into
elemanolides such as 6 (Scheme 1) should start with the
allylic oxidation of the C-15 methyl group. In the present
case, however, this was not feasible because the allylic
oxidation of germacrolides occurs with a concomitant
isomerization of the neighboring double bond, generally
Resu lts a n d Discu ssion
Our synthetic planning (Scheme 1) was based on the
following key transformations: first, the Cope rearrange-
ment of the germacrane skeleton of 5 to the elemane
skeleton of 6 (step a), previously tested by us under both
thermal and catalytic conditions;18b,20 second (step c), the
cyclization of the δ,ꢀ-epoxy ester 7 to the δ-lactone 8,
which should leave the hydroxymethyl group of 8 in the
most stable equatorial position required to perform the
long-range functionalization of C-14, which would allow
the attachment of a suitable leaving group (Y) at this
carbon atom; and third (step e), the intramolecular
displacement of the Y leaving group by the carboxylate
ion derived from selective saponification of the enol
(12) Grieco, P. A.; Nishizawa, M.; Oguri, T.; Burke, S. D.; Marinovic,
N. J . Am. Chem. Soc. 1977, 99, 5773.
(13) Danishefsky, S.; Kitahara, T.; Schuda, P. F.; Etheredge, S. J .
J . Am. Chem. Soc. 1977, 99, 6066.
(14) For a review, see: Heathcock, C. H.; Graham, S. L.; Pirrung,
M. C.; Plavac, F.; White, C. T. In The Total Synthesis of Natural
Products; ApSimon, J ., Ed.; Wiley: New York, 1983; Vol. 5, p 93.
(15) Ohrai, K.; Kondo, K.; Sodeoka, M.; Shibasaki, M. J . Am. Chem.
Soc. 1994, 116, 11737.
(16) For a recent review on the synthesis of 8-deoxyvernolepin and
other bioactive sesquiterpenes from santonin, see: Blay, G.; Cardona,
L.; Garc´ıa, B.; Pedro, J . R. In Studies in Natural Products Chemistry;
Atta-ur-Rahman, Ed.; Elsevier: Amsterdam, 2000; Vol. 24, p 53.
(17) Fujimoto, Y.; Miura, H.; Shimizu, T.; Tatsuno, T. Tetrahedron
Lett. 1980, 21, 3409.
(18) (a) Barrero, A. F.; Oltra, J . E.; Barraga´n, A. Tetrahedron Lett.
1995, 36, 311. (b) Barrero, A. F.; Oltra, J . E.; Barraga´n, A.; AÄ lvarez,
M. J . Chem. Soc., Perkin Trans. 1 1998, 4107. (c) Barrero, A.; Oltra,
J . E.; AÄ lvarez, M. Tetrahedron Lett. 2000, 41, 7639.
(19) Hoffmann, H. M. R.; Rabe, J . Angew. Chem., Int. Ed. Engl.
1985, 24, 94.
(21) . In contrast with the poor results obtained in the formation of
the exocyclic double bond at ∆3, the ∆11(13) double bond of 3 has been
built in excellent yields using Grieco’s procedure via oxidation of the
selenide formed from the corresponding enolate.10,17
(22) For a review of the synthesis of costunolide and other germac-
rane sesquiterpenes, see: Minnaard, A. J .; Wijnberg, J . B. P. A.; de
Groot, A. Tetrahedron 1999, 55, 2115.
(20) Barrero, A. F.; Oltra, J . E.; AÄ lvarez, M. Tetrahedron Lett. 1998,
39, 1401.
(23) Grieco, P. A.; Nishizawa, M. J . Org. Chem. 1977, 42, 1717.
5462 J . Org. Chem., Vol. 67, No. 16, 2002

Synthesis of (+)-8-Deoxyvernolepin
SCHEME 2
SCHEME 3
effect, however, but resulted in an inseparable mixture
(roughly 1/1 ratio) of alcohol 14 and a tricyclic organose-
lenium compound 19, together with a low yield of 15
(32%). Nevertheless, the chemical and pharmacological
interest aroused in recent years by organoselenium
compounds27 prompted us to spend some effort in con-
firming the chemical structure of the unexpected selena-
decalin 19. After conventional acetylation of the alcohol
mixture, the derivative 20 could be isolated and analyzed
by spectroscopic techniques. The MS showed several
molecular ions in relative proportions in accordance with
those of Se isotopes,28 and exhaustive NMR experiments
confirmed the chemical structure of 20. We could thus
infer the structure and stereochemistry of 19, including
the R-axial disposition of the hydroxymethyl group (it is
noteworthy that we did not detect the presumably more
stable 1â-hydroxymethyl epimer of 19). Although SeO2
is a widely used reagent in organic synthesis,29 reactions
of this compound with diene systems to give selenahet-
erocycles are still scarcely documented.30 In fact, none of
the mechanisms proposed so far can account for the
structure and stereochemistry of the selenacyclohexane
19. A plausible mechanism was suggested to us by the
following observations: the ene reaction proposed for the
first step of the allylic hydroxylation of alkenes31 and the
reduction of selenoxides to selenides via the exchange of
the oxygen atom between selenoxides and species of
Se(II).32 Therefore, selenide 19 might be formed via the
mechanism depicted in Scheme 3. Thus, the initial ene
leading to melampolides,24 and these derivatives are not
prone to undergo Cope rearrangements.25 Therefore, the
rearrangement of 12 was attempted directly under
thermal conditions. Thus, an equilibrium mixture of the
germacrolide 12 and the elemanolide 13 was made at a
ratio of 1/1.4 (12/13). The remaining germacrolide was
recovered from the mixture and reused, giving an 88%
yield of 13 after three turnovers. The subsequent allylic
oxidation of 13 with SeO2/t-BuOOH gave a mixture of
the alcohol 14 and the aldehyde 15 from which the
alcohol could be easily separated and reoxidized to
produce 15 at a 79% overall yield (from 13).
(27) (a) Ishihara, H.; Koketsu, M.; Fukuta, Y.; Nada, F. J . Am. Chem.
Soc. 2001, 123, 8408. (b) Organoselenium Chemistry: A Practical
Approach; Back, T. G., Ed.; Oxford University Press: Oxford, U.K.,
1999.
(28) Pretsch, E.; Clrerc, T.; Seibl, J .; Simon, W. Tabellen zur
Structuraufkla¨rung Organischer Verbindungen mit Spektroskopischen
Metoden, 3rd ed.; Springer-Verlag: Berlin, 1990.
(29) Bulman Page, P. C.; McCarthy, T. J . In Comprehensive Organic
Sintesis; Trost, B. M., Fleming, J ., Ley, S. V., Eds.; Pergamon Press:
Oxford, 1991; Vol. 7, p 83.
The allylic oxidation of elemanolide 13 was also as-
sayed with SeO2 in boiling anhydrous dioxane26 to try to
reach the aldehyde 15 in only one step and achieve a
higher yield. This procedure did not have the desired
(30) (a) San Feliciano, A.; Medarde, M.; Lo´pez, J . L.; Pereira, J . A.
P.; Caballero, E.; Perales, A. Tetrahedron 1989, 45, 5073. (b) Nguyen,
T. M.; Lee, D. Org. Lett. 2001, 3, 3161.
(31) (a) Sharpless, K. B.; Lauer, R. F. J . Am. Chem. Soc. 1972, 94,
7153. (b) Arigoni, D.; Vasella, A.; Sharpless, K. B.; J ensen, H. P. J .
Am. Chem. Soc. 1973, 95, 7917. (c) Singleton, D. A.; Hang, C. J . Org.
Chem. 2000, 65, 7554.
(24) (a) Haruna, M.; Ito, K. J . Chem. Soc., Chem. Commun. 1981,
483. (b) Barrero, A. F.; Oltra, J . E.; Rodr´ıguez-Garc´ıa, I.; Barraga´n,
A.; AÄ lvarez, M. J . Nat. Prod. 2000, 63, 305.
(25) Several assays intended to rearrange melampolides, carried out
in our laboratory under different experimental conditions, proved
unsuccessful.
(26) Bestmann, H. J .; Schobert, R. Angew. Chem., Int. Ed. Engl.
1985, 24, 791.
(32) Reich, H. J .; Renga, J . M.; Reich, I. L. J . Am. Chem. Soc. 1975,
97, 5434.
J . Org. Chem, Vol. 67, No. 16, 2002 5463

Barrero et al.
SCHEME 4
reaction between the disubstituted double bond of 13 and
SeO2 would lead to the allylseleninic acid 21. This
seleninic acid may undergo an intramolecular [2 + 2]
cycloaddition to give the oxaselenetane 22. (It is known
that the relatively poor π-overlap in CdSe bonds makes
them quite reactive in cycloadditions,27b and therefore,
the possibility that a OdSe bond works in a similar way
cannot be ruled out.) The subsequent opening of the
strained oxaselenetane 22, facilitated by the assistance
provided by the neighboring OH group, would lead to the
selenoxide 23. Finally, the reduction of 23 by a Se(II)
species (derived from the usual allylic oxidation process)
would give the cyclic selenide 19. We could not isolate
any of the proposed intermediates 21-23, but whatever
the exact mechanism, product 19 provides new evidence
to support the formation of a C-Se bond during the
allylic hydroxylation mediated by SeO2, as proposed by
Sharpless and Lauer.31a Our results also indicate that
the reaction between SeO2 and the 1,5-diene system of
elemanolides can be used for the synthesis of substituted
selenacycles, analogous to the case for natural eudes-
manolides.
Coming back to the synthesis of 3, we addressed the
preparation of the key intermediate 7 foreseen in Scheme
1. The oxidation of 15 with sodium chlorite followed by
the addition of CH2N2 afforded a 91% yield of methyl
ester 16 (Scheme 2). This ester proved to be an unusually
inert olefin and thus was recovered unchanged after
treatment with different oxidative reagents. Finally, the
reaction of 16 with dimethyldioxirane furnished us with
an isomeric mixture of the epoxides 7a (77% yield) and
7b (11% yield). We were pleased to find that both isomers
could be easily isolated by flash chromatography and
analyzed by spectroscopic techniques. Except for the ABX
system corresponding to the hydrogen atoms of the
NOEs observed between H-1 and H-5, as well as between
H-2 and H3-14. The reaction took place with a complete
inversion of the C-1 configuration, and thus, no formation
of the 1S isomer 24 was detected. In contrast, the reaction
of epoxide 7b under the same conditions was considerably
slower and gave a mixture of epimers 8 and 24 at a ratio
of 1/2, respectively. The stereochemical outcome of these
reactions can be rationalized in terms of a concerted
mechanism for 7a (Scheme 4) and a dual one for 7b,
which are consistent with the configurations proposed for
these epoxides.
In the case of 7a , the incipient secondary carbocation
developing at C-1 by the electrophilic opening of the
epoxide would be attacked by the nucleophilic methoxy-
carbonyl group, leading directly to the intermediate II
with a complete inversion of the C-1 configuration. This
intermediate would then evolve toward 8 (Scheme 4). On
the other hand, with epoxide 7b the reaction is slower,
presumably because the transition state (TS) leading to
24 by a concerted mechanism is higher in energy than
the corresponding TS leading to II. Thus, a stepwise
mechanism leading to a mixture of 8 and 24, via a
secondary carbocation, may compete with the concerted
one. Despite this, the main product obtained was the
epimer 24, with the hydroxymethyl group in the (pre-
sumably less stable) R-axial disposition; this fact suggests
that the concerted mechanism might work simulta-
neously with the stepwise one.
1
monosubstituted oxirane ring, the H NMR spectra of 7a
and 7b showed closely related signals, thus confirming
the C-1 epimeric relationship existing between these
compounds. The 1S configuration of 7a was assigned by
comparing the signals of its ABX system with those of
the corresponding protons of (1S)-1,2-epoxysaussurea
lactone.33 Consequently, the opposite stereochemistry,
1R, was attributed to the epimer 7b.
The configurations assigned to the epoxides 7a and 7b
were confirmed by the stereochemistry of the rearranged
products obtained by the acid-promoted opening of their
oxirane rings. In this way, the electrophilic opening of
the epoxide 7a stereoselectively gave the δ-lactone 8 (93%
yield) foreseen in Scheme 1, with the hydroxymethyl
group in the desired â-equatorial position. The 1R con-
figuration of 8 was suggested by the chemical shift of
C-14 (δ 13.9)18c (cf. δ value for C-14 of the 1S epimer 24
When we had dilactone 8 in our possesion, the syn-
thesis of 3 from arsantin (a natural eudesmanolide) via
the hydrogenated derivative 18 was reported by Astudillo
et al.34 Therefore, we addressed the preparation of 18
from 8 in order to complete the synthesis of 3 in a formal
sense (Scheme 2). In this way, the catalytic hydrogena-
tion of 8 followed by base-promoted isomerization of 17
gave acceptable yields of 18 (65% from 8). Thus, the
formal synthesis of the target molecule (+)-8-deoxyver-
nolepin (3) was achieved, proving the utility of the
in the Experimental Section) and was confirmed by the
(33) Blay, G.; Cardona, L.; Garc´ıa, B.; Pedro, J . R. Can. J . Chem.
1992, 70, 817.
(34) Astudillo, L.; Gonza´lez, A.; Galindo, A.; Mansilla, H. Tetrahe-
dron Lett. 1997, 38, 6737.
5464 J . Org. Chem., Vol. 67, No. 16, 2002

Synthesis of (+)-8-Deoxyvernolepin
SCHEME 5
SCHEME 6
accessible germacrolide 5 as an alternative raw material
for the enantiospecific synthesis of bioactive elemanolides
such as 3.
equate precursor (like 9, with different leaving groups
at C-2 and C-14) of the enol lactone 10 foreseen in
Scheme 1. Therefore, mesylate 26 was prepared (75%
yield) by a reaction between 25 and methanesulfonyl
chloride. Unexpectedly, basic treatment of 26 resulted
in a mixture of cis-fused δ-lactone 27 and tetrahydrofuran
derivative 30 instead of a trans-fused enol lactone such
as 10. The reaction was carried out under different
experimental conditions (temperatures ranging from -10
to 100 °C, reaction times from 1 to 24 h, and different
solvents), but the expected enol lactone was never
detected. In contrast, the interesting product 27 could
be isolated at a moderate 43% yield by working in THF
Two reports, however, deal with the difficulties in-
volved in building the ∆3 exocyclic double bond of ver-
nolepin-related compounds while avoiding mixtures with
an endocyclic regioisomer and a concomitant decrease in
yield.10,17 Therefore, we decided to preserve the corre-
sponding exocyclic double bond of 8, as was foreseen in
the original synthetic plan (see Scheme 1), to facilitate
the synthesis of our second target molecule 4 (Scheme
5). We tried transforming 8 into 25, taking advantage of
the suitable spatial disposition of the primary alcohol of
8 to perform the remote functionalization of the C-14
methyl group.35 For this purpose, we assayed several
reagents under photochemical conditions, obtaining dif-
ferent results. The irradiation of 8 in the presence of a
mixture of iodine and either (diacetoxyiodo)benzene36 or
lead tetraacetate gave tetracyclic derivative 28,37 whereas
1
at 25 °C. In the H NMR spectrum of 27,18c the signals
corresponding to the exocyclic methylene (δ 6.76, t, J )
1.0 Hz, 1H; δ 5.98, t, J ) 1.0 Hz, 1H) as well as the
diastereotopic protons H-14a (δ 4.55, d, J ) 12.3 Hz, 1H)
and H-14b (δ 4.14, dd, J ) 12.3, 1.6 Hz, 1H) showed
chemical shifts, multiplicities, and coupling constant
values characteristic of the 2-oxa-cis-decalin skeleton of
vernolepin-related compounds.41 Additionally, in the 13C
NMR spectrum of 27,18c the chemical shift and multiplic-
ity of C-14 (δ 70.0, t) supported the cis-fused δ-lactone
unit.8,36 We did not have direct evidence for the stereo-
chemistry of the oxirane ring of 27, but we tentatively
proposed a 1S configuration because this was consistent
with a plausible mechanism to explain the formation of
27 from 26 (see Scheme 6b and discussion below). We
did not detect the formation of the potential 1R epimer
of 27 in any of our experiments.
The methanesulfonyl group of 26 plays a critical role
in the reaction leading to 27, and thus, when iodohydrin
25 was treated under the same conditions, only the
tetrahydrofuran derivative 30 was obtained (Scheme 6a).
A simple inspection of structures 26 (C16H21IO7S) and 27
(C15H18O5) suggests sulfene (CH2SO2) and HI elimina-
tions instead of the expected elimination of methane-
sulfonic acid, which would have led to an enol lactone
such as 10. Sulfene elimination from organic mesylates
is an unusual reaction, although evidence of sulfene
formation from methanesulfonyl chloride in the presence
of pyridine has been reported.42 With this idea in mind,
8 was recovered unchanged when the irradiation was
performed in the presence of diphenylhydroxyselenium
acetate38 and iodine. Finally, under the condition sum-
marized in Scheme 5, a 30% yield of 25 was obtained but
60% of 8 could be recovered and reused, increasing the
yield of iodohydrin to 49%. Longer reaction times trans-
formed 25 into the transposed iodoformate 29,39 and thus,
the yield of iodohydrin was not improved.
In any case, iodohydrin 25 showed suitable enough
functionalization to be transformed easily into an ad-
(35) For an update on remote intramolecular functionalizations,
see: Majetich, G.; Wheles, K. Tetrahedron 1995, 51, 7095.
(36) Herna´ndez, R.; Vela´zquez, S. M.; Sua´rez, E.; Rodr´ıguez, M. S.
J . Org. Chem. 1994, 59, 6395.
(37) Experimental details and spectroscopic data of 28 are described
in the Supporting Information available.
(38) Dorta, R. L.; Francisco, C. G.; Freire, R.; Sua´rez, E. Tetrahedron
Lett. 1988, 29, 5429.
(39) Regioselective â-scission of alkoxy radicals (obtained from
alcohols by a multistep sequence) to iodoformates has already been
used for the syntheses of 18- and 19-norsteroids,40 but to the best of
our knowledge, reactions directly leading to transposed iodoformates
from alcohols (presumably via 1,4-halohydrins) have not been reported
before. For experimental details and the spectroscopic properties of
29, see the Supporting Information available.
(40) (a) Suginome, H.; Senboku, H.; Yamada, S. J . Chem. Soc.,
Perkin Trans. 1 1990, 2199. (b) Suginome, H.; Nakayama, Y.; Senboku,
H. J . Chem. Soc., Perkin Trans. 1 1992, 1837.
(41) Abegaz, B. M.; Keige, A. W.; D´ıaz, J . D.; Herz, W. Phytochem-
istry 1994, 37, 191.
(42) Crossland, R. K.; Servis, K. L. J . Org. Chem. 1970, 35, 3195.
J . Org. Chem, Vol. 67, No. 16, 2002 5465

Barrero et al.
we propose the mechanism depicted in Scheme 6b to
rationalize the formation of 27 from 26. Thus, the
selective abstraction of a relatively acidic hydrogen from
the methanesulfonyl group would promote sulfene elimi-
nation and the resultant oxirane formation process
leading to intermediate 31. (The preference for the
hydrogen of the methanesulfonyl group might derive from
the steric hindrance around H-1, which would obstruct
the approach of the bulky base employed.) The iodine
atom of 31 would subsequently be displaced by the
nucleophilic carboxylate group in a second intramolecular
substitution to give the cis-fused δ-lactone 27. Despite
the moderate yield obtained (43%), transformation of a
trans-fused δ-lactone into a cis-fused one bearing an
oxirane ring attached to a quaternary carbon atom is an
interesting reaction from a synthetic point of view. The
chemical building of quaternary centers is generally a
difficult task,43 whereas oxiranes are present in many
natural products and are one of the most versatile
intermediates in organic synthesis.44
Having obtained the epoxide 27, we only needed to
deoxygenate the oxirane ring to reach the target molecule
4 (Scheme 5). There are several reagents described for
performing oxirane deoxygenations,45 and therefore, we
decided to assay some of them, using a less elaborate
model compound than 27, to determine the optimum
experimental conditions for the deoxygenation reaction.
We eventually chose epoxide 7a because of the close
relationship between its functional groups and those of
27. The deoxygenation of epoxide 7a , however, turned
out to be no trivial task. After conventional treatment
with lithium diphenylphosphide and methyl iodide,46 7a
was recovered unchanged, whereas more recently de-
scribed reagents, such as Cp2TiCl2/Mn,47 Me3SiCl/NaI,48
and Amberlyst/NaI,49 afforded cyclization products 8 or
3250 instead of the expected vinyl derivative 16.
avoid the formation of byproducts. The combination of
t-BuOH and water at a ratio of 9:1 provided the best
results, giving a virtually quantitative yield of 16. Once
the optimum experimental conditions were established
for the model compound, they were applied to the
deoxygenation of epoxide 27 (Scheme 5). Thus, the 10-
vinyl-2-oxa-cis-decalin 4 was obtained at an acceptable
yield of 71%. In this way, the synthesis of (+)-11â,13-
dihydro-8-deoxyvernolepin (4) from (+)-costunolide (5)
was completed in 10 steps with an overall yield of almost
5%.
In conclusion, we describe here the syntheses of (+)-
8-deoxyvernolepin (3) and (+)-11â,13-dihydro-8-deoxy-
vernolepin (4) from (+)-costunolide (5) isolated from a
commercially available extract. During these syntheses,
the following novel reactions were used as key steps: the
stereoselective cyclization of a δ,ꢀ-epoxy ester to a hy-
droxymethyl-substituted δ-lactone and a cascade reac-
tion, initiated by sulfene elimination, which led to a 10-
oxiranyl-2-oxa-cis-decalin from the mesylate of a trans-
fused δ-lactone. We also found that the reaction of
selenium dioxide with the 1,5-diene system of elemano-
lides gave selenacycles analogous to natural eudesmano-
lides. Our results indicate that the strategy employed,
on the basis of biomimetic concepts, should provide a
useful alternative for the enantiospecific preparation of
(+)-vernolepin-related compounds from accessible ger-
macrolides. At the moment, we are using this strategy
to synthesize (+)-vernolepin (1), (+)-vernodalin (2), and
their corresponding 11,13-dihydroderivatives from (+)-
salonitenolide, an 8-hydroxygermacrolide easily isolated
in (multi)gram quantities from Centaurea calcitrapa and
other plants widespread in the Mediterranean area.11b,18b
We trust that this work will facilitate further studies
about the structure-activity relationships and the physi-
ological mechanisms involved in the antifungal activity
of vernolepin-related compounds.
Exp er im en ta l Section
Gen er a l Det a ils. The NOE-difference (NOE-dif) tech-
nique was employed to observe nuclear Overhauser effects.
13C NMR peak assignments were made with the aid of
DEPT and 2D NMR (COSY, HMQC, and HMBC) experiments.
The numbering used in the NMR assignments corresponds to
the germacrane system and not the IUPAC nomenclature.
Other general experimental details have been reported
elsewhere.18b
Costu n olid e (5). This compound was obtained in (multi)-
gram quantities from the commercially available extract
Costus Resinoid (Pierre Chauvet S. A., Seillans, France) as
described elsewhere.20
11â,13-Dih yd r ocostu n olid e (12). 10% Pd/C (34 mg) was
added to 5 (1.3 g, 5.5 mmol) in MeOH (40 mL), and the
suspension was slowly stirred for 30 min under H2 at atmo-
spheric pressure. The mixture was then filtered and the
solvent removed in vacuo from the filtrate, affording 12 (1.2
g, 93% yield). White solid: mp 70-71 °C; [R]25D +107° (c 1.08,
Finally, the treatment of 7a with KSeCN in aqueous
MeOH, as described by Behan et al.,51 gave a promising
mixture of 16 and some byproducts derived from the
Michael addition of MeOH to the conjugated double bond.
Therefore, the reaction was assayed in other solvents to
(43) For recent reviews of the enantioselective construction of
quaternary stereocenters, see: (a) Corey, E. J .; Guzman-Perez, A.
Angew. Chem., Int. Ed. 1998, 37, 388. (b) Christoffers, J .; Mann, A.
Angew. Chem., Int. Ed. 2001, 40, 4591.
(44) For a recent overview of the syntheses of selected natural
products containing elegant examples concerning the usefulness of
oxiranes in organic synthesis, see: Nicolaou, K. C.; Vourloumis, D.;
Winssinger, N.; Baran, P. S. Angew. Chem., Int. Ed. 2000, 39, 44.
(45) For a review on stereospecific deoxygenations of epoxides, see:
Wong, H. N. C.; Fok, C. C. M.; Wong, T. Heterocycles 1987, 26, 1345.
(46) Vedejs, E.; Snoble, K. A. J .; Fuchs, P. L. J . Org. Chem. 1973,
38, 1178.
(47) (a) RajanBabu, T. V.; Nugent, W. A. J . Am. Chem. Soc. 1994,
116, 986. (b) Gansa¨uer, A.; Bluhm, H.; Pierobon, M. J . Am. Chem. Soc.
1998, 120, 12849.
(48) Caputo, R.; Mangoni, L.; Neri, O.; Palumbo, G. Tetrahedron Lett.
1981, 22, 3551.
(50) Reaction of 7a with Cp2TiCl2/Mn gave 8, whereas reactions with
either Me3SiCl/NaI or Amberlyst 15/NaI gave iodo-derivative 32. These
results confirm that the stereoselective cyclization of δ,ꢀ-epoxy esters
is a versatile reaction useful for the synthesis of δ-alkyl-substituted
δ-lactones with different functional groups attached to the ꢀ-position.
For experimental details and the spectroscopic properties of 32, see
Supporting Information available.
(49) Righi, G.; Bovicelli, P.; Sperandio, A. Tetrahedron 2000, 56,
1733.
(51) Behan, J . M.; J ohnstone, R. A. W.; Wright, M. J . J . Chem. Soc.,
Perkin Trans. 1 1975, 1216.
5466 J . Org. Chem., Vol. 67, No. 16, 2002

Synthesis of (+)-8-Deoxyvernolepin
TABLE 1. 13C NMR Da ta (CDCl3, 100 MHz) of
tr a n s-F u sed Heter od eca lin s 17, 18, 20, a n d 24-26 a n d
cis-F u sed Heter od eca lin 4
CHCl3) (lit.23 mp 74-75 °C, [R]24 +108°); IR and 1H NMR
D
spectra in ref 23; 13C NMR (CDCl3, 75 MHz) δ 178.6 (s), 140.3
(s), 136.9 (s), 127.4 (d), 127.1 (d), 81.5 (d), 54.7 (d), 42.3 (d),
41.2 (t), 39.6 (t), 28.6 (t), 26.2 (t), 17.2 (q), 16.2 (q), 13.3 (q).
HRCIMS m/z calcd for C15H23O2: 235.1698. Found 235.1705.
Anal. Calcd for C15H22O2: C, 76.88; H, 9.46. Found: C, 76.59;
H, 9.85.
C
17
18
20a
24
25
26b
4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
90.2
61.1
14.9
35.1
46.9
78.5
52.4
22.9
35.8
37.3
41.0
178.7
12.5
14.9
173.3
89.0
61.1
18.3
38.0
51.2
81.7
52.2
22.9
34.4
37.7
40.7
178.5
12.5
13.6
172.9
45.5
65.1
24.8
140.8
49.2
79.2
51.9
24.0
36.7
43.4
41.4
178.9
12.5
23.2
111.3
88.3
61.9
130.2
134.0
44.3
80.5
51.7
22.6
33.1
38.1
40.4
178.6
12.5
20.2
165.1
90.9
61.6
130.9
132.3
49.9
78.9
51.5
22.4
35.6
40.1
40.5
178.0
12.4
2.6
87.0
67.4
131.9
131.7
50.1
78.7
51.6
22.2
35.0
40.5
40.4
177.6
12.5
1.6
141.0
116.3
134.9
130.9
47.1
Sa su ssu r ea La cton e (13). Compound 12 (95 mg) was
heated to 205-210 °C for 5 min under an Ar atmosphere in a
sealed flask. Subsequent flash chromatography (hexane/
t-BuOMe 9:1) of the mixture obtained gave 12 (39 mg, 41%)
and 13 (53 mg, 56% yield). Compound 13, colorless needles:
80.8
50.6
23.2
33.2
40.7
41.5
mp 147-148 °C; [R]25 +76.5° (c 0.69, CHCl3) (lit.52 mp 146-
D
147 °C, [R]22 +65.4°); IR and 1H NMR spectra in ref 52; 13C
D
NMR (CDCl3, 100 MHz) δ 179.3 (s), 147.8 (d), 141.0 (s), 115.2
(t), 111.5 (t), 81.3 (d), 55.7 (d), 52.2 (d), 42.7 (s), 41.6 (d), 39.4
(t), 24.0 (q), 23.6 (t), 18.3 (q), 12.6 (q). HRFABMS m/z calcd
for C15H22O2Na: 257.1518. Found: 257.1517.
178.1
12.6
70.6
163.2
162.2
163.6
8-Deoxym eliten sin (14) a n d Ald eh yd e 15. Compound 13
(550 mg, 2.35 mmol) in CH2Cl2 (10 mL) was added to a solution
of SeO2 (130 mg, 1.17 mmol) and t-BuOOH (1.8 mL of a 5-6
M solution in decane) in CH2Cl2 (10 mL) at 0 °C. The mixture
was allowed to reach rt and stirred under an Ar atmosphere
for 60 h. The mixture was then diluted with CH2Cl2 and
washed with brine. The organic layer was dried over anhyd
Na2SO4, and the solvent was removed. The residue was
submitted to flash chromatography (hexane/t-BuOMe 1:1),
affording 14 (235 mg, 40% yield) and 15 (250 mg, 43% yield).
8-Deoxymelitensin (14), oil: [R]25D +50.1° (c 1.28, CHCl3) (lit.52
a
b
Chemical shifts for CH3CO2: δ 21.3 (q), 170.9 (s). Chemical
shift for CH3SO3: δ 38.2 (q).
) 10.6 Hz, 1H), 3.62 (d, J ) 11.8 Hz, 1H), 3.04 (d, J ) 11.8
Hz, 1H), 2.64 (dd, J ) 7.2, 4.6 Hz, 1H), 2.43 (d, J ) 10.7 Hz,
1H), 2.34 (dq, J ) 12.3, 6.9 Hz, 1H), 2.11 (s, 3H), 1.30 (s, 3H),
1.23 (d, J ) 6.9 Hz, 3H); NOE-dif experiments, δ proton
irradiated (δ NOEs observed) 2.64 (4.80, 4.64, 1.30), 2.43 (4.80,
4.64, 3.62), 1.30 (4.07, 2.64); 13C NMR in Table 1; CIMS m/z
375 (2) [M + H, 82Se]+, 373 (8), 371 (4), 313 (25), 161 (26), 81
(42), 61 (100). HRFABMS m/z calcd for C17H24O4Na80Se:
395.0737. Found: 395.0737.
[R]22 +52.4°); IR and 1H NMR spectra in ref 52; 13C NMR
D
(CDCl3, 75 MHz) δ 179.0 (s), 147.3 (d), 145.1 (s), 114.2 (t), 112.1
(t), 81.4 (d), 67.4 (t), 52.5 (d), 51.1 (d), 42.4 (s), 41.8 (d), 39.4
(t), 23.5 (t), 17.7 (q), 12.6 (q). HRFABMS m/z calcd for C15H22O3-
Na: 273.1467. Found: 273.1470. Aldehyde 15, white solid: mp
149-150 °C; [R]25D +13.9° (c 0.91, CHCl3); IR (film) 1770, 1689
Meth yl Ester 16. A solution of NaClO2 (3.191 g, 35.28
mmol) and NaH2PO4‚2H2O (3.761 g, 24.11 mmol) in H2O (31
mL) was dripped into a mixture of 15 (1.050 g, 4.23 mmol)
and 2-methyl-2-butene (18 mL) in t-BuOH (90 mL) for 30 min.
The mixture was then stirred for 4 h, and the volatile
compounds were removed in vacuo. H2O was added, the
organic compounds were extracted with t-BuOMe, and the
ethereal solution was dried over anhyd Na2SO4 and filtered.
A saturated solution of CH2N2/t-BuOMe (14 mL) was then
added to the filtrate at 0 °C and stirred until the yellow color
disappeared. The solvent was removed and the residue sub-
mitted to flash chromatography (hexane/t-BuOMe 8:2), giving
16 (1073 mg, 3.86 mmol, 91% yield). White solid: mp 105-
1
cm-1; H NMR (CDCl3, 400 MHz) δ 9.42 (s, 1H), 6.24 (s, 1H),
6.22 (s, 1H), 5.66 (dd, J ) 17.4, 10.7 Hz, 1H), 4.87 (d, J ) 10.7
Hz, 1H), 4.79 (d, J ) 17.4 Hz, 1H), 4.31 (dd, J ) 11.8, 10.4
Hz, 1H), 2.99 (d, J ) 11.8 Hz, 1H), 2.37 (dq, J ) 12.4, 6.9 Hz,
1H), 1.90 (dq, J ) 12.4, 3.2 Hz, 1H), 1.22 (d, J ) 6.9 Hz, 3H),
1.00 (s, 3H); 13C NMR (CDCl3, 100 MHz) δ 194.0 (d), 178.9 (s),
146.7 (d), 145.7 (s), 137.3 (t), 112.0 (t), 80.0 (d), 52.3 (d), 46.2
(d), 42.5 (s), 41.6 (d), 39.0 (t), 23.5 (t), 16.8 (q), 13.6 (q); EIMS
m/z 248 (3) [M]+, 219 (6), 179 (11), 175 (21), 123 (18), 107 (31),
81 (77), 55 (82), 41 (100). HRFABMS m/z calcd for C15H20O3-
Na: 271.1310. Found: 271.1309. Anal. Calcd for C15H20O3: C,
72.55; H, 8.12. Found: C, 72.11; H, 8.44.
106 °C; [R]25D +61.2° (c 1.33, CHCl3); IR (film) 1767, 1723 cm-1
;
1H NMR (CDCl3, 400 MHz) δ 6.34 (s, 1H), 5.70 (dd, J ) 17.4,
10.8 Hz, 1H), 5.49 (s, 1H), 4.86 (d, J ) 10.8 Hz, 1H), 4.82 (d,
J ) 17.4 Hz, 1H), 4.18 (dd, J ) 11.4, 10.9 Hz, 1H), 3.66 (s,
3H), 3.04 (d, J ) 11.4 Hz, 1H), 2.36 (dq, J ) 12.6, 6.9 Hz, 1H),
1.87 (dq, J ) 12.5, 3.1 Hz, 1H), 1.72 (dq, J ) 12.0, 3.0 Hz,
1H), 1.21 (d, J ) 6.9 Hz, 3H), 0.98 (s, 3H); 13C NMR (CDCl3,
100 MHz) δ 178.9 (s), 167.9 (s), 146.5 (d), 136.7 (s), 126.9 (t),
111.9 (t), 80.0 (d), 52.2 (d), 51.9 (q), 48.8 (d), 42.5 (s), 41.7 (d),
39.0 (t), 23.5 (t), 16.4 (q), 12.6 (q); EIMS m/z 278 (1) [M]+, 246
(11), 218 (10), 145 (17), 105 (23), 81 (51), 55 (65), 41 (100).
HRFABMS m/zcalcd for C16H22O4Na: 301.1416.Found: 301.1418.
Anal. Calcd for C16H22O4: C, 69.04; H, 7.97. Found: C, 68.74;
H, 7.87.
A sample of PDC (530 mg, 1.4 mmol) was added to 14 (235
mg, 0.94 mmol) in anhyd THF (10 mL) at 0 °C. The mixture
was allowed to reach rt and stirred under an Ar atmosphere
for 48 h. The mixture was then diluted with t-BuOMe and
filtered through a pad of Florisil. The solvent was removed,
and 15 (208 mg) was obtained.
Rea ction of 13 w ith SeO2 in Boilin g Dioxa n e. We added
SeO2 (398 mg, 3.58 mmol) to a solution of 13 (600 mg, 2.56
mmol) in anhyd dioxane (80 mL) under an Ar atmosphere and
heated the mixture to reflux for 40 min. The mixture was then
filtered and the solvent removed, leaving a residue that was
submitted to flash chromatography (hexane/t-BuOMe 6:4),
giving 13 (60 mg, 0.256 mmol, 10%), 15 (200 mg, 0.81 mmol,
32%), and an inseparable mixture (190 mg) of two alcohols.
Acetic anhydride (0.07 mL, 5 mmol) and DMAP (660 mg, 5.4
mmol) were added to this mixture dissolved in CH2Cl2, and
the solution was stirred at rt for 30 min. The solution was
then washed with 2 N HCl, 2 N NaOH, and brine. The organic
layer was dried over anhyd Na2SO4, and the solvent was
removed. Flash chromatography (t-BuOMe/AcOEt 6:4) of the
residue gave 20 (43 mg) and the acetylated derivative of 14
(53 mg, 0.18 mmol). Compound 20, colorless oil: [R]25D +227.9°
Ep oxid a tion of 16 w ith Dim eth yld ioxir a n e (DMDO).
A solution of DMDO in acetone, previously prepared from 120
g of caroate,53 was added to 16 (800 mg, 2.88 mmol) in acetone
(15 mL). The mixture was stirred for 2 h at rt, and the solvent
was then removed. Flash chromatography (hexane/t-BuOMe
1:1) of the residue gave 7a (657 mg, 77% yield) and 7b (90
mg, 11% yield). Epoxide 7a , white solid: mp 135-136 °C; [R]25
D
+59.7° (c 1.33, CHCl3); IR (film) 1769, 1724 cm-1
;
1H NMR
(CDCl3 400 MHz) δ 6.45 (s, 1H), 5.58 (s, 1H), 4.18 (t, J ) 11.1
Hz, 1H), 3.73 (s, 3H), 3.16 (d, J ) 11.7 Hz, 1H), 2.77 (t, J )
1
(c 0.61, CHCl3); IR (film) 1775, 1738 cm-1; H NMR (CDCl3,
(52) Ando, M.; Tajima, K.; Takase, K. J . Org. Chem. 1983, 48, 1210.
(53) Adam, W.; Bialas, J .; Hadjiarapoglou, L. Chem. Ber. 1991, 124,
2377.
400 MHz) δ 5.21 (s, 1H), 4.84 (d, J ) 1.4 Hz, 1H), 4.80 (dd, J
) 11.7, 7.2 Hz, 1H), 4.64 (dd, J ) 11.7, 4.6 Hz, 1H), 4.07 (t, J
J . Org. Chem, Vol. 67, No. 16, 2002 5467

Barrero et al.
3.4 Hz, 1H), 2.45 (d, J ) 3.4 Hz, 2H), 2.36 (dq, J ) 12.6, 6.9
Hz, 1H), 1.92 (dq, J ) 12.6, 3.2 Hz, 1H), 1.23 (d, J ) 6.9 Hz,
3H), 0.70 (s, 3H); 13C NMR (CDCl3, 100 MHz) δ 178.7 (s), 167.6
(s), 136.2 (s), 128.4 (t), 79.7 (d), 58.8 (d), 52.2 (d), 52.1 (q), 47.5
(d), 43.8 (t), 41.7 (d), 39.9 (s), 35.6 (t), 23.1 (t), 12.6 (q), 12.1
(q); EIMS m/z 294 (0.3) [M]+, 262 (2), 244 (2), 189 (6), 161 (8),
121 (14), 91 (30), 67 (28), 55 (88), 41 (100); HRFABMS m/z
calcd for C16H22O5Na: 317.1365. Found: 317.1368. Anal. Calcd
for C16H22O5: C, 65.29; H, 7.53. Found: C, 64.96; H, 7.96.
in anhyd 1:1 CH2Cl2/CCl4 (50 mL) under an Ar atmosphere.
The mixture was heated to reflux and irradiated with two 100
W tungsten filament lamps for 8 h. The mixture was then
diluted with CH2Cl2 and washed with a Na2S2O3 saturated
solution. The organic layer was dried over anhyd Na2SO4 and
filtered, and the solvent was removed. Flash chromatography
(t-BuOMe) of the residue gave 25 (45 mg, 24% yield) and 8
(78 mg). Compound 25, pale yellow solid: mp 114-115 °C
(decompose); [R]25 +3.3 (c 0.9, Cl3CH); IR (film) 3470, 1780,
D
1
Epoxide 7b, white solid: mp 95-96 °C; [R]25 +61.3° (c 1.69,
1725 cm-1; H NMR (CDCl3, 300 MHz) δ 6.68 (d, J ) 2.6 Hz,
D
CHCl3); IR (film) 1776, 1719 cm-1; 1H NMR (CDCl3, 400 MHz)
δ 6.45 (s, 1H), 5.58 (s, 1H), 4.14 (dd, J ) 11.6, 10.5 Hz, 1H),
3.76 (s, 3H), 3.26 (d, J ) 11.6 Hz, 1H), 2.82 (dd, J ) 3.8, 3.1
Hz, 1H), 2.53 (dd, J ) 4.4, 3.8 Hz, 1H), 2.50 (dd, J ) 4.4, 3.1
Hz, 1H), 2.37 (dq, J ) 12.5, 6.9 Hz, 1H), 1.23 (d, J ) 6.9 Hz,
3H), 0.72 (s, 3H); 13C NMR (CDCl3, 100 MHz) δ 178.7 (s), 167.6
(s), 136.5 (s), 126.9 (t), 80.1 (d), 59.2 (d), 52.3 (q), 52.2 (d), 48.8
(d), 42.2 (t), 41.7 (d), 39.9 (s), 35.3 (t), 23.0 (t), 13.0 (q), 12.6
(q). HRFABMS m/z calcd for C16H22O5Na: 317.1365. Found:
317.1369.
δ-La cton e 8. BF3‚OEt2 (0.11 mL, 0.9 mmol) was added to
a solution of 7a (430 mg, 1.46 mmol) in anhyd CH2Cl2 (25 mL)
under an Ar atmosphere. The mixture was stirred at rt for 30
min, CH2Cl2 was added, and the organic solution was washed
with saturated NaHCO3 and brine. The organic layer was dried
over anhyd Na2SO4, and the solvent was removed. Flash
chromatography (t-BuOMe/AcOEt 6:4) of the residue gave 818c
(380 mg, 93% yield).
Acid -P r om oted Cycliza tion of Ep oxid e 7b. BF3‚OEt2 (10
µL, 0.09 mmol) was added to a solution of 7b (50 mg, 0.17
mmol) in anhyd CH2Cl2 (4 mL) under an Ar atmosphere. The
mixture was stirred at rt for 3.5 h, CH2Cl2 was added, and
the organic solution was washed with saturated NaHCO3 and
brine. The organic layer was dried over anhyd Na2SO4, and
the solvent was removed. Flash chromatography (t-BuOMe/
AcOEt 7:3) of the residue gave 8 (8 mg, 17% yield), 24 (16 mg,
34% yield), and 7b (4 mg, 8%). Compound 24, colorless oil:
1H NMR (CDCl3 300 MHz) δ 6.59 (d, J ) 2.6 Hz, 1H), 6.09 (d,
J ) 2.6 Hz, 1H), 4.05-3.95 (m, 4H), 3.33 (dt, J ) 10.3, 2.6 Hz,
1H), 2.33 (dq, J ) 12.6, 6.9 Hz, 1H), 1.24 (d, J ) 6.9 Hz, 3H),
1.11 (s, 3H); 13C NMR in Table 1.
Hyd r ogen a ted Der iva tive 17. 10% Pd/C was added to 8
(60 mg, 0.21 mmol) in MeOH (40 mL), and the suspension was
bubbled with H2 and vigorously stirred for 1.5 h. The mixture
was then filtered and the solvent removed in vacuo from the
filtrate, affording 17 (50 mg, 85% yield). Oil: 1H NMR (CDCl3
400 MHz) δ 4.17 (dd, J ) 6.2, 4.4 Hz, 1H), 3.92 (dd, J ) 11.3,
10.3 Hz, 1H), 3.77-3.73 (m, 2H), 3.05 (dq, J ) 8.1, 7.8 Hz,
1H), 2.35 (dq, J ) 12.4, 6.9 Hz, 1H), 2.15 (dd, J ) 11.3, 8.1
Hz, 1H), 1.94 (dq, J ) 13.0, 3.0 Hz, 1H), 1.71 (ddd, J ) 13.2,
4.0, 3.0 Hz, 1H), 1.43 (d, J ) 7.8 Hz, 3H), 1.38 (dt, J ) 13.0,
4.0 Hz, 1H), 1.22 (d, J ) 6.9 Hz, 3H), 1.06 (s, 3H). NOE-dif
experiments, δ proton irradiated (δ NOEs observed): 3.92
(2.35, 1.43, 1.06), 3.05 (2.15, 1.43), 2.15 (4.17, 3.05), 1.43 (3.92,
3.05). 13C NMR in Table 1.
1H), 6.14 (d, J ) 2.6 Hz, 1H), 4.46 (br d, J ) 8.7 Hz, 1H), 4.22
(dd, J ) 12.2, 8.7 Hz, 1H), 4.15 (br d, J ) 12.2 Hz, 1H), 3.97
(t, J ) 10.5 Hz, 1H), 3.81 (br s, 1H), 3.34 (d, J ) 12.0 Hz, 1H),
3.09 (dd, J ) 12.0, 1.1 Hz, 1H), 2.89 (dt, J ) 10.5, 2.6 Hz, 1H),
1.28 (d, J ) 6.8 Hz, 3H); 13C NMR in Table 1. HRCIMS m/z
calcd for C15H20O5I: 407.0355. Found: 407.0356.
Mesyla te 26. Methanesulfonyl chloride (20 µL, 0.25 mmol)
and Et3N (80 µL, 0.55 mmol) were added to iodohydrin 25 (45
mg, 0.11 mmol) in anhyd CH2Cl2 (2 mL) at 0 °C. The mixture
was stirred at this temperature for 11 h and at 25 °C for a
further 6 h. The mixture was then diluted with CH2Cl2 and
washed with saturated solutions of KHSO4 and NaHCO3. The
organic layer was dried over anhyd Na2SO4, and the solvent
was removed. Flash chromatography (t-BuOMe) of the residue
gave mesylate 26 (40 mg, 75% yield). White solid: mp 103-
104 °C; [R]25 +20.5° (c 1.27, CHCl3); 1H NMR (CDCl3, 300
D
MHz) δ 6.69 (d, J ) 2.5 Hz, 1H), 6.17 (d, J ) 2.5 Hz, 1H),
4.66-4.58 (m, 3H), 3.95 (t, J ) 10.6 Hz, 1H), 3.35 (dd, J )
12.2, 1.4 Hz, 1H), 3.13 (s, 3H), 3.12 (d, J ) 12.2 Hz, 1H), 2.92
(dt, J ) 10.6, 2.5 Hz, 1H), 2.43 (dt, J ) 13.3, 3.0 Hz, 1H), 2.38
(dq, J ) 12.6, 6.9 Hz, 1H), 2.01 (dq, J ) 13.3, 3.1 Hz, 1H),
1.45 (dt, J ) 13.3, 3.0 Hz, 1H), 1.28 (d, J ) 6.9 Hz, 3H); 13C
NMR in Table 1. HRCIMS m/z calcd for C16H22O7IS: 485.0131.
Found: 485.0132.
Rea ction betw een 26 a n d DBU a t 25 °C. DBU (0.2 mL,
1.34 mmol) was added to mesylate 26 (40 mg, 0.083 mmol) in
THF (2 mL) under an Ar atmosphere, and the mixture was
stirred at 25 °C for 24 h. The mixture was then diluted with
CH2Cl2 and washed with 1 N HCl and H2O. The organic layer
was dried over anhyd Na2SO4, and the solvent was removed.
Flash chromatography (t-BuOMe) of the residue gave 2-oxa-
cis-decalin 2718c (10 mg, 43%) and tetracyclic ether 30 (11
mg, 47%). Compound 30, oil: 1H NMR (CDCl3, 300 MHz) δ
6.57 (d, J ) 2.2 Hz, 1H), 6.01 (d, J ) 2.2 Hz, 1H), 4.70 (d,
J ) 3.1 Hz, 1H), 4.01 (d, J ) 10.8 Hz, 1H), 3.89 (dd, J ) 10.8,
3.1 Hz, 1H), 3.83 (dd, J ) 9.2, 1.5 Hz, 1H), 3.71 (t, J ) 10.6
Hz, 1H), 3.61 (d, J ) 9.2 Hz, 1H), 2.99 (dt, J ) 10.6, 2.2 Hz,
1H), 2.35 (dq, J ) 12.4, 6.9 Hz, 1H), 1.96 (dt, J ) 13.2, 3.2 Hz,
1H), 1.28 (d, J ) 6.9 Hz, 3H); 13C NMR (CDCl3, 100 MHz) δ
177.9 (s), 132.3 (s), 128.9 (t), 88.2 (d), 80.4 (d), 73.2 (t), 72.0
(t), 51.1 (d), 47.7 (s), 45.2 (d), 40.5 (d), 34.3 (t), 23.5 (t), 12.6
(q). HRFABMS m/z calcd for C15H18O5Na: 301.1052. Found:
301.1055.
11â,13-Dih yd r o-8-d eoxyver n olep in (4). KSeCN (12 mg,
0.083 mmol) was added to a solution of 27 (18 mg, 0.063 mmol)
in 9:1 t-BuOH/H2O (3 mL), and the mixture was stirred at 65
°C for 6 h. Then, t-BuOMe was added, the solution was filtered,
and the solvent was removed from the filtrate. Flash chroma-
tography (t-BuOMe/AcOEt, 85:15) of the residue afforded 4 (12
Ba se-P r om oted Isom er iza tion of 17. Diazabicyclounde-
cane (DBU) (15 µL, 0.04 mmol) was added to a solution of 17
(50 mg, 0.18 mmol) in toluene (2 mL) under an Ar atmosphere.
The mixture was heated to 60 °C and stirred for 2 h. Then
CHCl3 was added, and the solution was washed with 1 N HCl
and H2O. The organic layer was dried over anhyd Na2SO4, and
the solvent was removed. Flash chromatography of the residue
(t-BuOMe/AcOEt 85:15) afforded 1834 (40 mg, 80% yield). Oil:
mg, 71%). Oil: [R]25 +28.0° (c 0.25, CHCl3); IR (film) 1779,
D
1720 cm-1; 1H NMR (CDCl3, 300 MHz) δ 6.67 (dd, J ) 1.1, 0.7
Hz, 1H), 5.86 (t, J ) 1.1 Hz, 1H), 5.69 (dd, J ) 17.2, 11.2 Hz,
1H), 5.24 (d, J ) 11.2 Hz, 1H), 5.22 (d, J ) 17.2 Hz, 1H), 4.55
(d, J ) 12.1 Hz, 1H), 4.20 (dd, J ) 12.1, 1.8 Hz, 1H), 3.94 (t,
J ) 10.8 Hz, 1H), 2.81 (br d, J ) 11.0, 1H), 2.37 (dq, J ) 12.4,
6.8 Hz, 1H), 1.25 (d, J ) 6.8 Hz, 3H); 13C NMR in Table 1;
CIMS m/z 263 [M + H]+ (13), 217 (10), 141 (41), 123 (100).
HRCIMS m/z calcd for C15H19O4: 263.1283. Found: 263.1284.
[R]25 +28.0 (c 0.4, Cl3CH); IR (film) 3454, 1779, 1725 cm-1
;
D
1H NMR (CDCl3, 400 MHz) δ 4.19 (t, J ) 5.3 Hz, 1H), 3.80 (t,
J ) 10.5 Hz, 1H), 3.74 (d, J ) 5.3 Hz, 2H), 2.62 (dq, J ) 10.8,
7.1 Hz, 1H), 2.28 (dq, J ) 12.3, 6.9 Hz, 1H), 1.51 (d, J ) 7.1
Hz, 3H), 1.22 (d, J ) 6.9 Hz, 3H), 1.02 (s, 3H); 13C NMR in
Table 1. HRFABMS m/z calcd for C15H22O5Na: 305.1364.
Found: 305.1365.
Ack n ow led gm en t. This research was supported by
the Spanish DGICYT (Project PB 98-1365) and by the
Spanish Ministerio de Educacio´n y Cultura (grants
Iod oh yd r in 25. HgO (201 mg, 0.93 mmol) and I2 (118 mg,
o.46 mmol) were added to a solution of 8 (130 mg, 0.46 mmol)
5468 J . Org. Chem., Vol. 67, No. 16, 2002

Synthesis of (+)-8-Deoxyvernolepin
Copies of the 13C NMR spectra of compounds 4, 7b, 17, 20,
24-26, 29, 30, and 32. A copy of the H NMR spectrum of 28.
This material is available free of charge via the Internet at
http://pubs.acs.org.
provided to M.A. and A.R.). We thank our English
colleague A. L. Tate for revising our English text, and
our colleague J . J usticia for his collaboration.
1
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails and spectroscopic data for compounds 28, 29, and 32.
J O0256538
J . Org. Chem, Vol. 67, No. 16, 2002 5469