Total asymmetric synthesis of the putative structure of the cytotoxic diterpenoid (-)-sclerophytin A and of the authentic natural sclerophytins A and B

Article abstract of DOI:10.1021/ja011285y

An enantioselective synthetic route to the thermodynamically most stable diastereomer of the structure assigned to sclerophytin A (5) has been realized. The required tricyclic ketone 33 was prepared by sequential Tebbe-Claisen rearrangement of lactones 29 and 30, which originated from the Diels-Alder cycloaddition of Danishefsky's diene to (5S)-5-(d-menthyloxy)-2(5H)-furanone (14). An allyl and a cyano group were introduced into the resulting adduct by means of stereocontrolled allylindation under aqueous Barbier-like conditions and by way of cyanotrimethylsilane, respectively. Following stereocontrolled nucleophilic addition of a methyl group to 33, ring A was elaborated by formation of the silyl enol ether, ytterbium triflate-catalyzed condensation with formaldehyde, O-silylation, and Cu(I)-promoted 1,4-addition of isopropylmagnesium chloride. The superfluous ketone carbonyl was subsequently removed and the second ether bridge introduced by means of oxymercuration chemistry. Only then was the exocyclic methylene group unmasked via elimination. An alternative approach to the α-carbinol diastereomer proceeds by initial α-oxygenation of 37 and ensuing 1,2-carbonyl transposition. Neither this series of steps nor the Wittig olefination to follow induced epimerization at C10. Through deployment of oxymercuration chemistry, it was again possible to elaborate the dual oxygen-bridge network of the target ring system. Oxidation of the organomercurial products with O₂ in the presence of sodium borohydride furnished 72, which was readily separated from its isomer 73 after oxidation to 61. Hydride attack on this ketone proceeded with high selectivity from the β-direction to deliver (-)-60. Comparison of the high-field ¹H and ¹³C NMR properties and polarity of synthetic 5 with natural material required that structural revision be made. Following a complete spectral reassessment of the structural assignments to many sclerophytin diterpenes, a general approach to sclerophytin A, three diastereomers thereof, and of sclerophytin B was devised. The presence of two oxygen bridges as originally formulated was thereby ruled out, and absolute configurations were properly determined. Key elements of the strategy include dihydroxylation of a medium-ring double bond, oxidation of the secondary hydroxyl in the two resulting diols, unmasking of an exocyclic methylene group at C-11, and stereocontrolled 1,2-reduction of the α-hydroxy ketone functionality made available earlier.

Full text of DOI:10.1021/ja011285y

J. Am. Chem. Soc. 2001, 123, 9021-9032
9021
Total Asymmetric Synthesis of the Putative Structure of the Cytotoxic
Diterpenoid (-)-Sclerophytin A and of the Authentic Natural
Sclerophytins A and B
Patrick Bernardelli,† Oscar M. Moradei,† Dirk Friedrich,†,1 Jiong Yang,† Fabrice Gallou,†
Brian P. Dyck,† Raymond W. Doskotch,*,‡ Tim Lange,† and Leo A. Paquette*,†
Contribution from the Departments of Chemistry, Medicinal Chemistry, and Pharmacognosy,
The Ohio State UniVersity, Columbus, Ohio 43210
ReceiVed May 25, 2001
Abstract: An enantioselective synthetic route to the thermodynamically most stable diastereomer of the structure
assigned to sclerophytin A (5) has been realized. The required tricyclic ketone 33 was prepared by sequential
Tebbe-Claisen rearrangement of lactones 29 and 30, which originated from the Diels-Alder cycloaddition of
Danishefsky’s diene to (5S)-5-(d-menthyloxy)-2(5H)-furanone (14). An allyl and a cyano group were introduced
into the resulting adduct by means of stereocontrolled allylindation under aqueous Barbier-like conditions and
by way of cyanotrimethylsilane, respectively. Following stereocontrolled nucleophilic addition of a methyl
group to 33, ring A was elaborated by formation of the silyl enol ether, ytterbium triflate-catalyzed condensation
with formaldehyde, O-silylation, and Cu(I)-promoted 1,4-addition of isopropylmagnesium chloride. The
superfluous ketone carbonyl was subsequently removed and the second ether bridge introduced by means of
oxymercuration chemistry. Only then was the exocyclic methylene group unmasked via elimination. An
alternative approach to the R-carbinol diastereomer proceeds by initial R-oxygenation of 37 and ensuing 1,2-
carbonyl transposition. Neither this series of steps nor the Wittig olefination to follow induced epimerization
at C10. Through deployment of oxymercuration chemistry, it was again possible to elaborate the dual oxygen-
bridge network of the target ring system. Oxidation of the organomercurial products with O2 in the presence
of sodium borohydride furnished 72, which was readily separated from its isomer 73 after oxidation to 61.
Hydride attack on this ketone proceeded with high selectivity from the â-direction to deliver (-)-60. Comparison
1
of the high-field H and 13C NMR properties and polarity of synthetic 5 with natural material required that
structural revision be made. Following a complete spectral reassessment of the structural assignments to many
sclerophytin diterpenes, a general approach to sclerophytin A, three diastereomers thereof, and of sclerophytin
B was devised. The presence of two oxygen bridges as originally formulated was thereby ruled out, and absolute
configurations were properly determined. Key elements of the strategy include dihydroxylation of a medium-
ring double bond, oxidation of the secondary hydroxyl in the two resulting diols, unmasking of an exocyclic
methylene group at C-11, and stereocontrolled 1,2-reduction of the R-hydroxy ketone functionality made
available earlier.
Introduction
Scheme 1
Marine invertebrates are widely recognized to be an extraor-
dinarily rich source of oxygenated 2,11-cyclized cembranoids.2-5
The considerable structural diversity of this class of natural
products has resulted in the offering of a biogenetic proposal6
that rationalizes formation of the four related classes (Scheme
1). Of these, the cladiellins (also known as eunicellins)7 and
sarcodictyins8 appear to be primary metabolites generated by
† Department of Chemistry.
‡ College of Pharmacy.
(1) Present address: Aventis Pharmaceuticals, Inc., 1041 U.S. Route 202/
206, P.O. Box 6800, Bridgewater, NJ 08807-0800.
(2) Fenical, W. In Marine Natural Products, Chemical and Biochemical
PerspectiVes; Scheuer, P. J., Ed.; Academic Press: New York, 1978; pp
174-242.
(3) (a) Coll, J. C. Chem. ReV. 1992, 92, 613. (b) Faulkner, D. J. Nat.
Prod. Rep. 1997, 14, 259 and prior reports in this series.
(4) Wahlberg, I.; Eklund, A.-M. In Progress in the Chemistry of Organic
Natural Products; Springer-Verlag: New York, 1992; Vol. 60, pp 1ff.
(5) Bernardelli, P.; Paquette, L. A. Heterocycles 1998, 49, 531.
(6) Stierle, D. B.; Carte´, B.; Faulkner, D. J.; Tagle, B.; Clardy, J. J. Am.
Chem. Soc. 1980, 102, 5088.
carbon-carbon bond formation across positions C2 and C11
in cembrane precursors. They differ in the relative positioning
10.1021/ja011285y CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/22/2001

9022 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
Bernardelli et al.
of the tetrahydrofuranyl oxygen. The briarellins9 are the result
of further oxygenation to introduce an oxepane ring, and the
asbestinins10 possess a translocated methyl substituent.
In line with their natural role to ward off predation, many of
these metabolites are ichthyotoxic and molluscicidal.8,11 Sig-
nificantly, however, the range of inherent biological activity is
far more expansive and includes potent antitumor activity, insect
growth inhibition, antiinflammatory effects, and cell division
inhibition.5,11,12 Recently, the spotlight has been focused on
eleutherobin (1)13 whose capacity for inducing tubulin polym-
erization and microtubule stabilization14 has attracted consider-
able synthetic interest.15 Several other compounds in this large
class have, on preliminary examination, exhibited cytotoxic
properties that qualify them as opportune targets for de novo
synthesis. Included in this group are labiatin B (2),16,17 briarellin
A (3),10a and sclerophytin A (4).18 For example, 4 exhibits in
vitro cytotoxicity against the L1210 cell line at a concentration
of 1 ng/mL.
compounds have made clear the fact that the resident stereo-
centers C1, C2, C3, C9, C10, and C14 possess the R configu-
ration in common. Thus, one objective became to fuse the six-
membered ring properly in cis fashion to the larger cyclodecanol
unit. Beyond this, the detailed spectroscopic analysis reported
by Sharma and Alam18 for sclerophytin A was not able to define
the configuration of C7 with absolute certainty because of its
relatively remote location. The data can be construed to be
equally compatible with the 7R configuration as in 5 and the
absence of a second oxygenated bridge. The epimeric formula-
tion 5 is significantly less strained than 420 and conforms
additionally to the low-temperature boron trifluoride-catalyzed
cyclization of cladiellin 6 to 7 previously reported by Hochlo-
wski and Faulkner.23 Overman and Pennington have recognized
the same structural ambiguities and have independently achieved
a total synthesis of 5.24 The above considerations led us to seek
a solution to the challenge of setting the two oxygen bridges
anti as in 5, as the strategy should also enable an independent
synthesis of 7 to be accomplished. At the least, the structure
of sclerophytin A would be established by a process of elimina-
tion.
Retrosynthetic Planning
Initial disconnection of 5 at the O-C7 bond with provision
The total synthesis of any member of this select group
represents a significant challenge to contemporary synthetic
chemistry. Syntheses of eleutherobin (1)15 and 7-deacetoxyal-
cyonin acetate19 have been completed. More recently, we gained
interest in developing an enantiocontrolled approach to the
sclerophytin family that held the prospect of serving as a general
route to this class of diterpenes.20
for late-stage introduction of the methylenecyclohexane func-
(13) Lindel, T.; Jensen, P. R.; Fenical, W.; Long, B. H.; Casazza, A.M.;
Carboni, J.; Fairchild, C. R. J. Am. Chem. Soc. 1997, 119, 8744.
(14) Long, B. H.; Casazza, A. M.; Carboni, J.; Wasserman, A. J.; Cornell,
L.; Fairchild, C. R.; Lindel, T.; Jensen, P. R.; Fenical, W. Cancer Res.
1998, 58, 1111.
(15) (a) Nicolaou, K. C.; van Delft, F.; Ohshima, T.; Vourloumis, D.;
Xu, J.; Hosokawa, S.; Pfefferkorn, J.; Kim, S.; Li, T. Angew. Chem., Int.
Ed. Engl. 1997, 36, 2520. (b) Chen, X.-T.; Zhou, B.; Bhattacharya, S. K.;
Gutteridge, C. E.; Pettus, T. R. R.; Danishefsky, S. J. Angew. Chem., Int.
Ed. 1998, 37, 789. (c) Nicolaou, K. C.; Ohshima, T.; Hosokawa, S.; van
Delft, F. L.; Vourloumis, D.; Xu, J. Y.; Pfefferkorn, J.; Kim, S. J. Am.
Chem. Soc. 1998, 120, 8674. (d) Chen, X.-T.; Bhattacharya, S. K.; Zhou,
B.; Gutteridge, C. E.; Pettus, T. R. R.; Danishefsky, S. J. J. Am. Chem.
Soc. 1999, 121, 6563.
(16) Roussis, V.; Fenical, W.; Vagias, C.; Kornprobst, J.-M.; Miralles,
J. Tetrahedron 1996, 52, 2735.
(17) Ortega, M. J.; Zub´ıa, E.; Salva´, J. J. Nat. Prod. 1997, 60, 485.
(18) Sharma, P.; Alam, M. J. Chem. Soc., Perkin Trans. 1 1988, 2537.
(19) MacMillan, D. W. C.; Overman, L. E. J. Am. Chem. Soc. 1995,
117, 10391.
In line with this goal, X-ray crystallographic7,21 and circular
dichroism measurements22 performed on structurally related
(7) Eunicellin was the first cladiellane to be described: Kennard, O.;
Watson, D. G.; Riva di Sanseverino, L.; Tursch, B.; Bosmans, R.; Djerassi,
C. Tetrahedron Lett. 1968, 2879.
(8) Sarcodictyins A-F were the first members of this subgroup to be
isolated: (a) D’Ambrosio, M.; Guerriero, A.; Pietra, F. HelV. Chim. Acta
1987, 70, 2019. (b) D’Ambrosio, M.; Guerriero, A.; Pietra, F. HelV. Chim.
Acta 1988, 71, 964.
(9) At 10 members, the briarellin class is a significantly smaller group
than the other three: (a) Rodr´ıguez, A. D.; Co´bar, O. M. Tetrahedron 1995,
51, 6869. (b) Rodr´ıguez, A. D.; Co´bar, O. M. Chem. Pharm. Bull. 1995,
43, 1853.
(10) For early reports, consult: (a) Rodr´ıguez, A. D.; Co´bar, M. O.
Tetrahedron 1993, 49, 319. (b) Morales, J. J.; Lorenzo, D.; Rodr´ıguez, A.
D. J. Nat. Prod. 1991, 54, 1368.
(20) Paquette, L. A.; Moradei, O. M.; Bernardelli, P.; Lange, T. Org.
Lett. 2000, 2, 1875.
(21) (a) Kazlauskas, R.; Murphy, P. T.; Wells, R. J.; Scho¨nholzer, P.
Tetrahedron Lett. 1977, 4643. (b) Alam, M.; Sharma, P.; Zektzer, A. S.;
Martin, G. E.; Ji, X.; van der Helm, D. J. Org. Chem. 1989, 54, 1896. (c)
Su, J.; Zheng, Y.; Zeng, L.; Pordesimo, E. O.; Schmitz, F. J.; Hossain, M.
B.; Van der Helm, D. J. Nat. Prod. 1993, 56, 1601.
(11) (a) Kusumi, T.; Uchida, H.; Ishitsuka, M. O.; Yamamoto, H.;
Kakisawa, H. Chem. Lett. 1988, 1077. (b) Ochi, M.; Yamada, K.; Kataoka,
K.; Kotsuki, H.; Shibata, K. Chem. Lett. 1992, 155.
(12) Some representative reports are: (a) Ochi, M.; Yamada, K.;
Futatsugi, K.; Kotsuki, H.; Shibata, K. Heterocycles 1991, 32, 29. (b) Ochi,
M.; Yamada, K.; Shirase, K.; Kotsuki, H.; Shibata, K. Heterocycles 1991,
32, 19. (c) Lin, Y.; Bewley, C. A.; Faulkner, D. J. Tetrahedron 1993, 49,
7977. (d) Yamada, K.; Ogata, N.; Ryu, K.; Miyamoto, T.; Komori, T.;
Higuchi, R. J. Nat. Prod. 1997, 60, 393.
(22) (a) Bowden, B. F.; Coll, J. C.; Dai, M. C. Aust. J. Chem. 1989, 42,
665. (b) Ochi, M.; Yamada, K.; Futatsugi, K.; Kotsuki, H.; Shibata, K.
Chem. Lett. 1990, 2183. (c) Miyamoto, T.; Yamada, K.; Ikeda, N.; Komori,
T.; Higuchi, R. J. Nat. Prod. 1994, 57, 1212.
(23) Hochlowski, J. E.; Faulkner, D. J. Tetrahedron Lett. 1980, 21, 4055.
(24) Overman, L. E.; Pennington, L. D. Org. Lett. 2000, 2, 2683.

Synthesis of Structure of Sclerophytins A and B
J. Am. Chem. Soc., Vol. 123, No. 37, 2001 9023
Scheme 2
Scheme 3
tionality was envisioned to lead back to 8 (Scheme 2). Our
expectation was that the remaining ether oxygen in this advanced
intermediate would find it possible to steer the approach of Hg2+
to the R-surface of the unsaturated linkage by means of
coordination and thereby exercise a high degree of stereose-
lectivity.25 This prospectus had the useful feature that control
of the course of the introduction of the C3 methyl group into 9
would extend subsequently to C7. Of comparable significance
was the recognition that the global ensemble of structural
features resident in 9 might be installed by Claisen rearrange-
ment of 10. If this were so, one could easily imagine that
chemoselective vinyl anion addition to the ketone carbonyl in
11 with accompanying ring closure would deliver a tricyclic
lactone, Tebbe olefination26 of which would provide 10. This
choice of sequential steps would allow us to examine quickly
the critical ring expansion that is expected to provide 9. We
were mindful from the outset, however, that the proper con-
fluence of conformational factors had to be met to realize this
objective.
of Feringa,28 the heating of 13 with 14 in toluene resulted in a
completely regioselective cycloaddition. The appearance of two
products in a ratio of 2:1, diasteromeric at the carbon bearing
the methoxy group, is due to a modest endo stereoselectivity.
In practice, the formation of a mixture at this point is
inconsequential. As a result of the chiral auxiliary, elaboration
of the cis-fused bicyclic framework occurs to set stereogenicity
properly in either case, and both adducts undergo conversion
to 15 under controlled conditions. The latter step, recognized
to be demanding,28 could be efficiently and reproducibly
accomplished by application of Vorndam’s protocol (Scheme
3).29 These mild conditions rely on a catalytic quantity of
trimethylsilyl triflate in the presence of pyridine or lutidine as
the proton scavenger, instead of treatment with dilute aqueous
acidic solutions.
Although we shall leave unspecified for the moment which
of several options was to be implemented for introduction of
the necessary functionality in 11, it was recognized that a Diels-
Alder cycloaddition involving Danishefsky’s diene (13)27 and
(5S)-5-(d-menthyloxy)-2-(5H)-furanone (14)28 could generate 12
as the matrix upon which additional structural complexity could
be appended.
First Generation Synthesis. Preparation of Hydroxy Ester
27. The convergent assembly of 27 began with exploration of
a direct route to keto lactone 15. In line with the earlier report
There is no known direct method for conversion of the
uncommon γ-alkoxy-γ-lactone unit of the type resident in 15
into the corresponding allylated γ-lactone as found in 18 and
19. For our purposes, it was envisioned that addition of the
allylindium reagent to a γ-hydroxy-γ-lactone of general formula
A might well generally provide products such as C selectively
via chelation-controlled 1,2-addition to the ring-opened â-car-
boxy aldehyde tautomer B and subsequent lactonization. When
(25) (a) Henbest, H. B.; Nicholls, B. J. Chem. Soc. 1959, 227. (b)
Henbest, H. B.; McElhinney, R. S. J. Chem. Soc. 1959, 1834. (c)
Thaisrivongs, S.; Seebach, D. J. Am. Chem. Soc. 1983, 105, 7407. (d)
Chamberlain, P.; Whitham, G. H. J. Chem. Soc. B 1970, 1382. (e) Matsuki,
Y.; Kodama, M.; Itoˆ, S. Tetrahedron Lett. 1979, 2901. (f) Giese, B.;
Bartmann, D. Tetrahedron Lett. 1985, 26, 1197. (g) Pougny, J.-R.; Nassr,
M. A. M.; Sinay¨, P. J. Chem. Soc., Chem. Commun. 1981, 375. (h) Nicotra,
F.; Perego, R.; Ronchetti, F.; Russo, G.; Toma, L. Gazz. Chim. Ital. 1984,
114, 193. (i) Paquette, L. A.; Bolin, D. G.; Stepanian, M.; Branan, B. M.;
Mallavadhani, U. V.; Tae, J.; Eisenberg, S. W. E.; Rogers, R. D. J. Am.
Chem. Soc. 1998, 120, 11603.
(28) (a) de Jong, J. C.; van Bolhuis, F.; Feringa, B. L. Tetrahedron:
Asymmetry 1991, 2, 1247. (b) de Jong, J. C.; Jansen, J. F. G. A.; Feringa,
B. L. Tetrahedron Lett. 1990, 31, 3047.
(26) Review: Pine, S. H. Org. React. 1993, 43, 1.
(27) Danishefsky, S. Acc. Chem. Res. 1981, 14, 400.
(29) Vorndam, P. E. J. Org. Chem. 1990, 55, 3693.

9024 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
Bernardelli et al.
Scheme 4
preliminary studies involving congeners of 15 were found to
be highly responsive to this neighboring carboxyl effect,30 our
interest turned to assessing the response of 17 to these Barbier-
like aqueous allylation conditions. To this end, 15 was reduced
under Luche conditions31 to give predominantly (ratio of 9.6:
1) the R-carbinol 16, which was O-silylated and subsequently
subjected to mild acidic hydrolysis. In line with previous
observations, the allylindation of 17 served well to deliver a
13:1 mixture of 18 and 19. The need to clearly distinguish these
diastereomers was satisfied by means of NOE analysis. For the
major product, the absence of any observable interaction
between the two vicinal protons shown in the structural formula
unequivocally defined it to be the required â-allyl derivative.
With multigram quantities of 18 in hand, attention was turned
to effecting homologation at C2 as a prelude to implementation
of the key Tebbe-Claisen sequence. Thus, this lactone was
subjected to Dibal-H reduction, acetylation, and ultimately
treatment with trimethylsilyl cyanide (TMSCN) and boron
trifluoride etherate in CH2Cl2 at -78 °C. A 1:1 mixture of the
chromatographically separable nitriles 20 and 21 was thereby
obtained in excellent yield. A change to acetonitrile as solvent
did not alter the stereoselectivity. Activation of the lactol in
the form of its acetate prior to cyanation was found to be far
more efficient than direct treatment of the lactol with TMSCN
under the same conditions.32
At this juncture, verification that 21 was the â-isomer was
offered by NOE studies and reinforced by chemical equilibra-
tion. Although introduction of the cyano substituent proved not
to be stereoselective, the discovery that 20 could be efficiently
converted into 21 by exposure to potassium tert-butoxide in
tert-butyl alcohol constituted a major step forward.
cerium trichloride36 afforded 25, which proved to be a most
useful precursor to 27. Therefore, this propitious hydrolysis
sequence37 proceeds efficiently without interference from
neighboring keto or hydroxyl functionality. Where 27 is
concerned, its structural framework is constituted of all the
features required to advance to lactones 29 and 30.
The Tandem Tebbe-Claisen Ring Expansion. The con-
venient formation of 27 permitted its exploitation as a precursor
to 33. The requisite series of transformations, summarized in
Scheme 5, began with the generation of carboxylic acid 28 via
saponification of 27 with lithium hydroxide and subsequent
macrolactonization according to Yonemitsu’s modification of
the Yamaguchi protocol.39 These relatively efficient steps
As a direct consequence of the sensitivity of these nitriles,
the task of transforming them into their respective esters without
epimerization required the avoidance of conventional conditions.
Ultimately, recourse was made to a one-pot two-step procedure
that consisted of imidate formation with sodium methoxide in
methanol33 and mild acidic hydrolysis of the imino ethers.34
This protocol nicely skirted the risk of potential epimerization,
and proceeded in notably high yield when the CN substituent
resided on the sterically noncongested â-surface of the bicyclic
framework.
(36) The CeCl3-mediated addition [Dimitrov, V.; Bratovanov, S.; Simova,
S.; Kostova, K. Tetrahedron Lett. 1994, 35, 6713; Dimitrov, V.; Kostova,
K.; Genov, S. Tetrahedron Lett. 1996, 37, 6787] gave better yields than
the procedure involving preformation of the vinylcerate at -78 °C [Imamoto,
T.; Takiyama, N.; Nakamura, K.; Hatajima, T.; Kamiya, Y. J. Am. Chem.
Soc. 1989, 111, 4392.
(37) Attempts to bring about the direct hydrolysis of 25 to the carboxylic
acid with 10% KOH and H2O2 in DMSO38 invariably stopped at the amide
stage to give i.
These findings prompted two complementary studies (Scheme
4). In the first, 21 was transformed by Wacker oxidation35 into
the keto nitrile 24, a process which gave rise to the ester 26
through comparable processing of 23. More significantly,
addition of vinylmagnesium bromide to 24 in the presence of
(30) Bernardelli, P.; Paquette, L. A. J. Org. Chem. 1997, 62, 8284.
(31) Gemal, A. L.; Luche, J.-L. J. Am. Chem. Soc. 1981, 103, 5454.
(32) (a) Bru¨ckner, C.; Holzinger, H.; Reissig, H.-U. J. Org. Chem. 1988,
53, 2450. (b) Bru¨ckner, C.; Lorey, H.; Reissig, H.-U. Angew. Chem., Int.
Ed. Engl. 1986, 25, 556.
(33) (a) Koert, U.; Stein, M.; Harms, K. Tetrahedron Lett. 1993, 34,
2299. (b) Koert, U. Tetrahedron Lett. 1994, 35, 2517.
(34) Le Borgne, J. F.; Cuvigny, T.; Larcheveˆque, M.; Normant, H.
Synthesis 1976, 238.
(38) (a) Katritzky, A.; Pilarski, B.; Urogdi, L. Synthesis 1989, 949. (b)
Sasaki, S.; Nakashima, S.; Nagatsugi, F.; Tanaka, Y.; Hisatome, M.; Maeda,
M. Tetrahedron Lett. 1995, 36, 9521.
(35) Tsuji, J. Synthesis 1984, 369.

Synthesis of Structure of Sclerophytins A and B
J. Am. Chem. Soc., Vol. 123, No. 37, 2001 9025
Scheme 5
tography. Despite the strain release that serves as a component
of the driving force for these isomerizations, it was clearly
obvious that 32 was transformed into 33 more rapidly than was
31. In fact, after heating a 1:1 mixture of 31 and 32 at 140-
150 °C for 1 h, only 31 remained unreacted. Upon closer
scrutiny, the discovery was made that 32 rearranges quantita-
tively into 33 in just 3 h upon being heated in neat condition at
55 °C. By comparison, 31 required higher temperatures (e.g.,
130 °C) for a longer period. Even under these circumstances,
the conversion to 33 was incomplete, and some decomposition
was evident. Ultimately, a compromise position was taken to
reconcile this kinetic imbalance. Recourse to NaBF4, or less
preferably pyridinium camphorsulfonate, as promoter43 allowed
the sigmatropic process to be performed efficiently in refluxing
toluene during a few hours.
The interesting rate difference in question can be rationalized
by comparing the two chairlike transition states D and E that
are presumably involved (Scheme 5). Although the adoption
of these chair geometries guarantees that both enol ethers will
give rise to product having a Z double bond, the sterically and
electronically disfavored interactions that are incurred in D but
not in E elevate the energy barrier with observable conse-
quences.
The structural assignment to 33 rests on two constituent
spectroscopic features. When H6 was irradiated, the signal
corresponding to the neighboring methyl group showed a 10.8%
NOE enhancement. The phenomenon was reciprocal. Further-
more, according to 13C NMR correlations,10b,44 the chemical shift
at δ 25.4 is uniquely characteristic of a methyl substituent that
is geminal to one alkyl chain and trans to another.
Setting Stereochemistry at C3. The next issue to be
addressed was 1,2-addition of a methyl group to the C3 carbonyl
in 33 as a prelude to a more advanced functionalization of the
A ring (see 8). From the conformational vantage point, it was
considered reasonable that the carbonyl group in its medium-
sized ring would prefer to be oriented to the â-face as in F.
Note that this arrangement avoids the unfavorable electronic
interactions between the two oxygenated centers that must
clearly operate in the “carbonyl-down” alternative G. On this
basis, we predicted that addition of methyllithium to 33 would
result in predominant attack from the molecular exterior to
deliver 34 as the predominant or exclusive product. A strong
diastereofacial bias was indeed seen, such that 34 was the only
tertiary carbinol detected (Scheme 6). The important piece of
stereochemical identification was once again derived from NOE
delivered a chromatographically separable mixture of lactones
29 and 30, which were readily distinguished by NOE methods.
Individual submission of these advanced intermediates to Tebbe
methylenation,26 with proper provision for the complete removal
of AlMe3 and AlMe2Cl during preparation of the Tebbe reagent,
made available the enol ethers 31 and 32, respectively.
In pilot experiments, the planned Claisen rearrangements40-42
were performed by heating in p-cymene at 130-140 °C with
monitoring of the progress of reaction by thin-layer chroma-
(42) (a) Kinney, W. A.; Coghlan, M. J.; Paquette, L. A. J. Am. Chem.
Soc. 1985 107, 7352. (b) Paquette, L. A., Kang, H.-J. J. Am. Chem. Soc.
1991, 113, 2610. (c) Paquette, L. A.; Sweeney, T. J. Tetrahedron 1990,
46, 4487. (d) Friedrich, D.; Paquette, L. A. J. Org. Chem. 1991, 56, 3831.
(f) Paquette, L. A., Friedrich, D.; Rogers, R. D. J. Org. Chem. 1991, 56,
3841. (g) Paquette, L. A.; Philippo, C. M. G.; Vo, N. H. Can. J. Chem.
1992, 70, 1356.
(43) For a discussion of the catalysis of the Claisen rearrangement,
consult: Lutz, R. P. Chem. ReV. 1984, 84, 205.
(44) (a) Selover, S. J.; Crews, P.; Tagle, B.; Clardy, J. J. Org. Chem.
1981, 46, 964. (b) Crews, P.; Kho-Wiseman, E. J. Org. Chem. 1977, 42,
2812. (c) de Haan, J. W.; van de Ven, L. J. M. Tetrahedron Lett. 1971,
3965.
(39) (a) Hikota, M.; Sakurai, Y.; Horita, K.; Yonemitsu, O. Tetrahedron
Lett. 1990, 31, 6367. (b) Hikota, M.; Tone, H.; Horita, K.; Yonemitsu, O.
Tetrahedron 1990, 46, 4613.
(40) Ziegler, F. E. Chem. ReV. 1988, 88, 1423.
(41) Wipf, P. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Fleming, I., Paquette, L. A., Eds.; Pergamon Press: Oxford, 1991; Vol. 5,
Chapter 7.2.

9026 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
Bernardelli et al.
Scheme 6
Scheme 7
measurements. The strong (12%) enhancement between H2 and
the newly introduced methyl substituent proved particularly
convincing.
Following O-acetylation or benzoylation of this entity, the
long-serving silyl protecting group was removed, and the allylic
alcohols so formed were oxidized to ketones 37 and 38 with
the perruthenate reagent.45
Since little precedence was available, the decision was made
to briefly explore the stereochemistry of epoxidation of 37 to
assess the feasibility of regiocontrolled electrophilic attack at
C11. These functional group manipulations began by treating
37 with MCPBA, with resultant formation of 39 and 40 in a
2:1 ratio (Scheme 7). Thus, a modest preference for approach
of the peracid to the R-surface of the medium-ring double bond
was demonstrated.
Not entirely unexpected46 were our observations that the
enolate anions of these epoxy ketones are rather unreactive
chemical intermediates. Following the deprotonation of 39 under
kinetically controlled conditions with potassium hexamethyld-
isilazide in THF at -78 °C and the subsequent introduction of
excess methyl iodide, isomerically pure 41 was obtained in only
45% yield (or 53% based on recovered unreacted 39). Under
comparable conditions, 40 gave rise to a 7:1 mixture of 43 and
its R-methyl epimer (41%). The response of the enolate anion
of 39 toward the Davis oxaziridine47 was to deliver the
R-hydroxy ketone 42 with complete stereoselectivity and
somewhat more efficiently. Beyond this, intermediates 41-43
smoothly underwent 1,4-conjugate addition in the presence of
isopropylmagnesium chloride and the copper(I) bromide-
dimethyl sulfide complex to provide 44-46, respectively, in
good yield. A number of NOE studies supported the depicted
stereochemical assignments (see formulas).
thereby leading to the proper setting of the requisite R
configuration at C14.
Elaboration of Ring A. As a prelude to examining the
problem of setting the two anti-oxygen bridges present in the
target, we hoped to complete the proper functionalization of
ring A in an efficient manner. Two changes were instituted at
the outset. In an effort to deter premature deacylation at C3,
the acetate group was replaced with the more robust benzoate
alternative. In addition, the decision was made to probe the
reactivity of siloxy diene 47 as a way of facilitating alkylation
at C11 (Scheme 8). The matter of generating 47 was easily
resolved. To reach 48, 47 was directly exposed to chlorometh-
ylphenyl sulfide and silver tetrafluoroborate in CH2Cl2 as the
solvent. The alternate use of the more often utilized TiCl448 was
found to be destructive, and the milder Lewis acid ZnBr2 proved
inefficient. This reaction sequence furnished 48 in 55% yield
for the two steps. To reach primary carbinol 49 in an efficient
manner, recourse was made to 37% aqueous formaldehyde in
THF with ytterbium triflate as the promoter.49 This protocol
led to the isolation of 49 in 65% yield alongside 32% of
recovered enone 38. The subsequent O-silylation of 49 pro-
ceeded in reduced yield because of the sensitivity of this
compound to concurrent retroaldol cleavage.
Three important points had now been established: (a) R-attack
on the double bond positioned in the largest ring of these
heterotricyclic frameworks is not sterically impeded; (b) elec-
trophilic attack on the planarized enolate anion formed in either
series occurs preferentially from the â-face, with 39 exhibiting
particularly high stereoselectivity; (c) steric factors likewise
control the direction of entry of the isopropyl cuprate reagent,
While the copper-catalyzed addition of isopropylmagnesium
chloride to 48 could not be induced to proceed efficiently, it
was found that 50 gave rise reproducibly to 52 in 90% yield.
NOE measurements confirmed that the stereochemical course
of this process did not differ from earlier precedent. Therefore,
(45) Ley, S. V.; Norman, J.; Griffith, W. P.; Marsden, S. P. Synthesis
1994, 639.
(46) For comparison, the enolate anion of 37 was found not to react
with the Eschenmoser salt, dry formaldehyde, or tosyl cyanide.
(47) (a) Davis, F. A.; Vishwakarma, L. C.; Billmers, J. G.; Finn, J. J.
Org. Chem. 1984, 49, 3241. (b) Davis, F. A.; Chattopadhyay, S.; Towson,
J. C.; Lal, S.; Reddy, T. J. Org. Chem. 1988, 53, 2087.
(48) Paterson, I. Tetrahedron 1988, 44, 4207.
(49) Kobayashi, S.; Hachiya, I. J. Org. Chem. 1994, 59, 3590.

Synthesis of Structure of Sclerophytins A and B
J. Am. Chem. Soc., Vol. 123, No. 37, 2001 9027
Scheme 8
Scheme 9
omeric mixture of the doubly bridged ether dihydro-7, whose
1H NMR spectrum showed the major product to be the same
substance produced earlier by Hochlowski and Faulkner24 from
the metabolite 6 of a Pacific soft coral. In a two-step sequence,
they generated dihydro-7 having the same characteristic peaks
δ 3.93 (br t, J ) 5 Hz, 1 H), 3.67 (s, 1 H), 1.27 (s, 3 H), 1.06
(s, 3 H), 0.95 (d, J ) 6.8 Hz, 3 H, 0.92 (d, J ) 6.8 Hz, 3 H),
and 0.78 (d, J ) 6.8 Hz, 3 H).
This linkup was perceived to be of direct relevance to our
working assumptions regarding the synthesis of 5 by a parallel
stratagem. Further, this validation constituted an endorsement
for the oxymercuration as a workable means for setting the pair
of ether bridges in an anti relationship.
Acquisition of 5. Following completion of the reconnaissance
phase just described, we returned to the oxymercuration of 53,
but presently with oxidative demercuration (NaBH4, O2)53 in
mind. To our delight, this protocol furnished the secondary
carbinol 57 (54%, R,â ) 3:7) as the major product in addition
to 18% of the tertiary alcohol 58 (Scheme 10). Following
acetylation of 57, it proved convenient to separate the R- and
â-epimers chromatographically. This development made feasible
the subsequent independent elaboration of both 5 and 60.
Introduction of the exocyclic double bond in both epimers of
57 was accomplished by fluoride ion-induced desilylation in
advance of dehydration via the o-nitrophenylselenocyanate54 and
Dibal-H reduction. In addition, the complementary R-isomer
60 was produced by oxidation of 5 to ketone 61 and reduction
of the latter intermediate. Interestingly, hydride attack occurred
from the â-direction to deliver 60 exclusively.
the stage was set for deoxygenative removal of the carbonyl
oxygen. This reductive maneuver was implemented in three
steps consisting of Luche reduction, formation of the thiocar-
bonyl imidazolide, and radical reduction.50 Subsequent removal
of the benzoate functionality provided 53 in 72% yield from
52.
Scrutiny of the Applicability of Oxymercuration Chem-
istry. Attainability of Prototype Sclerophytin 7. Having
removed the superfluous oxygen in ring A, we next set out to
conduct selected model experiments involving possible intramo-
lecular participation during the oxymercuration of 53. After
treatment with mercuric trifluoroacetate in DMF under ther-
modynamic conditions,51 the cyclized mercurials were subjected
directly to the reductive action of sodium borohydride. These
conditions led to the isolation of both 54 (57%) and 55 (23%)
(Scheme 9). The mercurinium ion H that materializes upon
complexation to the R-surface of the π-bond can consequently
be judged to be sufficiently unsymmetrical to position greater
positive charge at C7 than at C6. In this way, formation of a
pyran ring as in 54 can compete quite effectively with otherwise
kinetically favored five-ring closure to deliver 55.
The availability of 54 in this manner made possible its
conversion to the exo-methylene derivative 56 by oxidative
elimination of the o-nitrophenyl selenide derivative.52 Hydro-
genation of 56 under standard conditions furnished a diastere-
Neither targeted compound was identical to the natural
substance. While 5 is dextrorotatory, sclerophytin A is not.55
Beyond this, the significantly more polar nature of 4 commands
attention. This marked difference is also reflected in the derived
ketones 61 and 62. Quite unexpected at the time was the finding
that the previously unknown 62 (incorrect structure) was
(53) Harding, K. E.; Marman, T. H.; Nam, D.-h. Tetrahedron Lett. 1988,
29, 1627 and relevant references therein.
(54) Grieco, P. A.; Gilman, S.; Nishizawa, M. J. Org. Chem. 1976, 41,
1485.
(55) The optical rotation of natural sclerophytin A has not previously
been reported: [R]D -6.9 (c 0.087, CHCl3).
(50) Barton, D. H. R.; McCombie, S. W. J. Chem. Soc., Perkin Trans.
1 1975, 1574.
(51) Broka, C. A.; Lin, Y.-T. J. Org. Chem. 1988, 53, 5876.
(52) Grieco, P. A.; Takigawa, T.; Schillinger, W. J. J. Org. Chem. 1980,
45, 2247.

9028 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
Bernardelli et al.
Scheme 10
Scheme 11
appreciably more polar than either 5 or 60. We now recognize
this to be because the structural assignment to natural 4 was
misformulated by Alam and co-workers. More recently, the
rightmost ether bridge in natural 4 (and by extrapolation to 62)
has been shown to be constituted instead of two tertiary
hydroxyl-bearing centers.56
sufficiently unreliable in delivering 64 so as to warrant the
dismissal of this alternative from further consideration.
Copper-catalyzed addition of isopropylmagnesium chloride
to 64 proceeded smoothly, giving ketone 65 after fluoride ion-
induced cleavage of the intermediate silyl enol ether. The
stereochemical assignment to 65, unequivocally demonstrated
by a series of NOE experiments (see I), was consistent only
with â-entry of the cuprate as before.
Alternative Synthesis of 60. Concurrent with the effort
described above, an alternate and somewhat shorter route to 60
was explored. This alternative began with enantiopure ketone
37 and was designed with early migration of the C12 carbonyl
to C11, concurrent with incorporation of the C14 isopropyl
group from the more open â-face. In line with these objectives,
the potassium enolate of 37 was generated and oxidized with
the Davis oxaziridine47 to produce R-ketol 63, which was
acetylated in turn to give 64 (Scheme 11).
Attempts to circumvent the modest yield of this oxidative
protocol prompted the examination of a Mn(III)-mediated
process intended to lead directly to 64.57 Although roughly
comparable efficiencies were realized during small-scale runs
involving Mn(OAc)3, recourse to larger quantities proved
The acetylated nature of the hydroxyl groups in 65 presented
a target of opportunity for removal of the superfluous C12
carbonyl oxygen. This goal could not be reached by possible
desulfurization of a dithioacetal because of the observed high
sensitivity of 65 to Lewis acids. Alternatively, its sequential
(56) (a) Friedrich, D.; Doskotch, R. W.; Paquette, L. A. Org. Lett. 2000,
2, 1879. (b) Gallou, F.; MacMillan, D. W. C.; Overman, L. E.; Paquette,
L. A.; Pennington, L. D.; Yang, J. Org. Lett. 2001, 3, 135.
(57) (a) Dunlap, N. K.; Sabol, M. R.; Watt, D. S. Tetrahedron Lett. 1984,
25, 5839. (b) Jeganathan, A.; Richardson, S. K.; Watt, D. S. Synth. Commun.
1989, 19, 1091.

Synthesis of Structure of Sclerophytins A and B
J. Am. Chem. Soc., Vol. 123, No. 37, 2001 9029
Scheme 12
graphic means. For this reason, the 72/73 mixture was directly
oxidized with TPAP and NMO. The tertiary alcohol 73 was
expectedly inert to these conditions and was now easily retrieved
as a single diastereomer. The desired less polar ketone 61 forms
without difficulty, and its availability by means of this protocol
allowed for the implementation of lithium aluminum hydride
reduction to furnish (-)-60. In line with precedent, the exclusive
product was deduced to be the R-carbinol resulting from
sterically controlled approach to the carbonyl group in a nonchair
conformation of the C-ring.
The linear sequence beginning from 37 consisted of 13 steps.
The key reactions involved carbonyl transposition in tandem
with Barton deoxygenation, as complemented by intramolecular
oxymercuration.
NMR Spectroscopic Analysis. The 600 MHz 1H NMR
spectrum of ketone 61 in CDCl3 reveals the presence of two
nonequivalent exocyclic methylene protons, two methine protons
geminal to oxygen, three additional methines to higher field,
two methyl groups attached to fully substituted carbons, and
two methyls that form part of an isopropyl group (Figure 1).
The combination of 1H, 1H COSY, and NOE difference (DPFGE
method) studies provided convincing evidence for adoption of
the indicated conformation. Recourse to HMBC and HSQV
analyses at 150 MHz allowed for clear definition of all relevant
carbon chemical shifts and specific C/H interactions. (Supporting
Information).
Entirely parallel analyses of the epimeric alcohols 5 and 60
showed these reduced compounds to possess comparable three-
dimensional characteristics. The coupling of H-6 to H-5R and
H-5â in 5 (J ) 10.7, 3.9 Hz) defined this isomer to be the
â-carbinol. The R-orientation featured in 60 is met with
significantly reduced spin-spin interactions (J ) 3.4, 2.5 Hz)
(Supporting Information). The spectrum of neither epimer
compares at all favorably with that recorded for authentic
scleophytin A. Particularly diagnostic of the substantive dis-
similarities of the three isomers is the upfield segment of their
13C NMR spectra (Figure 2).
Mechanistic and Biogenetic Considerations. A structural
feature shared in common by cladiellins such as 6, lithophynins
A-D,22c,58,59 calicophirins A and B,60,61 and ophirin60-63 is a
trans double bond between C6 and C7. Conformational analysis
of a prototype system such as 6 clearly indicates the “hydrogen
down” conformer to be thermodynamically preferred relative
to the “hydrogen up” arrangement 74 (Scheme 13). The
nonbonded steric compression of the vinylic methyl against the
C2-C9 oxygen bridge is the prime destabilizing factor.
Preferential protonation of 6 would give rise to carbocation 75,
simple collapse of which with loss of a proton would deliver 7,
as observed experimentally.
treatment with sodium borohydride, conversion to the xanthate,
and heating of the latter with tri-n-butyltin hydride50 led
uneventfully to 67. The major carbinol formed midway in this
sequence was judged to have its hydroxyl oriented to the R-face,
since the geminal H11 proton exhibits two large couplings (J10,11
) 9.9 Hz, J11,12 ) 9.4 Hz) as a necessary consequence of dual
anti arrangements involving the vicinal hydrogens H10 and H12.
Progress toward the target continued with the low-temperature
Dibal-H reduction of 67, thereby producing the diol 68.
Controlled oxidation of the secondary hydroxyl in 68 with tetra-
n-propylammonium perruthenate (TPAP),45 in conjunction with
NMO as the co-oxidant, furnished the crystalline ketone 69 in
83% yield. At this stage, it was not possible to confirm
unequivocally by spectroscopic means whether epimerization
may have taken place adventitiously at C10 under the basic
conditions of the Wittig olefination with methylenetriph-
enylphosphorane required for conversion to 70. However, by
pressing forward until ultimate arrival at 60, it was subsequently
established that any concern regarding possible erosion of
stereochemical integrity at this center during this particular
olefination step was unwarranted.
The oxymercuration of dienol 70 proceeded chemoselectively
to the trisubstituted double bond, obviously as a direct conse-
quence of assistance provided by the C-ring hydroxyl substitu-
ent25 (Scheme 12). Treatment of the intermediate organomer-
curials with sodium borohydride gave rise to the doubly bridged
ethers 56 and 71 in near-comparable amounts. As before, the
lack of preference for intramolecular hydroxyl capture at the
tertiary olefinic center over the secondary one is believed to
stem from geometrical constraints inherent to this particular
structural framework. Notwithstanding, the high efficiency of
this process (96% combined yield) boded well for its possible
deployment in the preparation of 60.
In a marine environment, epoxidation of the C6/C7 double
bond could transpire.64 Attack must necessarily be relegated to
the exterior of the trans-cyclodecene ring for the usual reasons.
For 6, this process would give rise to 76 and conceivably set
(58) Ochi, M.; Futatsugi, K.; Kotsuki, H.; Ishii, M.; Shibata, K. Chem.
Lett. 1987, 2207.
(59) Ochi, M.; Futatsugi, K.; Kume, Y.; Kotsuki, H.; Asao, K.; Shibata,
K. Chem. Lett. 1988, 1661.
(60) Ochi, M.; Yamada, K.; Shirase, H.; Kotsuki, H.; Shibata, K.
Heterocycles 1991, 32, 19.
(61) Seo, Y.; Rho, J.-R.; Cho, K. W.; Shin, J. J. Nat. Prod. 1997, 60,
171.
To this end, we investigated the possibility of the oxidative
demercuration53 of the same pair of structurally complex
mercurials. The isomeric alcohols 72 and 73 were identified
spectroscopically as having been formed in this reaction but
proved to be exceedingly difficult to separate by chromato-
(62) Fusetani, N.; Nagata, H.; Hirota, H.; Tsuyuki, T. Tetrahedron Lett.
1989, 30, 7079.
(63) Kashman, Y. Tetrahedron Lett. 1980, 21, 879.
(64) Solenopodins A-C are naturally occurring trans epoxy-substituted
cladiellins: Bloor, S. J.; Schmitz, F. J.; Hossain, M. B.; van der Helm, D.
J. Org. Chem. 1992, 57, 1205.

9030 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
Bernardelli et al.
Figure 1. 1H NMR assignments (δ values, 600 MHz, CDCl3, J in Hz) and conformational features of ketone 61. The values in square brackets are
hidden patterns that were revealed by NOE and TOCSY methods. The numbering scheme is adapted from that reported by Sharma and Alam (ref
18).
Figure 2. Comparison of the high-field sectors of the 13C NMR spectra of alcohols 5 and 60 with that of authentic sclerophytin A (150 MHz,
CDCl3).
the stage for transannularly assisted cleavage of the oxirane ring.
Proper alignment of the nucleophilic hydroxyl requires some
degree of conformational rotation about the C3-C4 (and other)
bonds. The net consequences of this sequence of events are the
installation of two ether oxygen bridges having a syn relationship
and the presence of a â-hydroxyl at C6.
However, no substance of type 5 has yet been isolated and
correctly identified.
Accordingly, no evidence exists to support the epoxidative
pathway. The possibility does exist that a dihydroxylative option
transforms 6 into sclerophytin A. In this case, exo attack would
directly set the diol configuration at C6 and C7.
bond. Oxymercuration likewise proceeds from exterior attack
on this π-surface with subsequent intramolecular backside attack.
Adherence to this mechanistic paradigm serves to invert the
absolute configuration at C6. With these facts now deduced, it
remains to determine why Nature does not opt to proceed via
76 to 5.
Access to Authentic Sclerophytins A and B. A detailed
reexamination of sclerophytin B by NMR56a has resulted in
revised structural assignments for sclerophytins A and B. In
light of this critical study, the structural motif resident in 52
was considered to be well suited to a targeted synthesis of both
84 and 85. Therefore, this functionalized cycloalkene was
subjected to dihydroxylation. Under conditions involving cata-
lytic quantities of OsO4 with NMO as the stoichiometric reagent,
The synthetic strategy developed in the present investigation
involves intermediates characterized by a cis C6/C7 double

Synthesis of Structure of Sclerophytins A and B
J. Am. Chem. Soc., Vol. 123, No. 37, 2001 9031
Scheme 13
Scheme 14
overoxidation occurred, and the products proved difficult to
separate. Attempts to control the stereochemical outcome with
the AD-mix formulations65 led to anticipated results. While the
â-reactant gave rise predominantly to 78, the R-form led to no
visible reaction. In the final analysis, recourse was made to the
use of a molar equivalent of OsO4, since it furnished diols 77
and 78 efficiently in a 1:1.5 ratio (Scheme 14). The reversed
bias observed above during the epoxidation of related systems
is noted. The ease of chromatographic separation of 77 from
78 and their concomitant production in reasonable amounts
ultimately made possible the acquisition of all four possible
sclerophytin-like triols. The important distinction between 77
and 78 rests on NOESY measurements performed on both
diastereomers. In particular, the strong correlation noted for 78,
but absent in the spectra of 77, defined rather unequivocally
the configuration at C-7.
The oxidation of these diols to their respective R-hydroxy
ketones 79 and 80 was satisfactorily accomplished in DMSO
with o-iodoxybenzoic acid (IBX), a reagent purported to lack
the capability of engaging in cleavage of the glycolic C-C
bond.66 This reaction serves two purposes. First, it sets the stage
for intentional introduction of the hydroxyl configuration at C-6
in the final synthetic targets. Second, one gains the tactical
advantage of “protecting” the C-6 hydroxyl, since this func-
tionality, if left unmasked, is not compatible with the ensuing
dehydration in ring A.67
Following the desilylation of 79 and 80, the resulting diols
were carried directly into the Grieco process.68 While both
substrates formed the o-nitrophenylselenide at roughly compa-
rable rates, the relative ease of selenoxide elimination proved
to be markedly distinctive. When the C-7 hydroxyl configuration
was â as in 80, several hours at room temperature were sufficient
to introduce the exocyclic double bond in 82 (63%). Placement
of the remote C-7 hydroxyl on the R-surface eventuated instead
in the formation of an appreciably more stable selenoxide
intermediate. Overnight heating at 50 °C furnished 81 with
enhanced efficiency (98%).
The final transformations, removal of the benzoate function-
ality and reduction of the ketone carbonyl, were expected to be
accomplishable in a single laboratory operation (Scheme 15).
The use of diisobutylaluminum hydride for this purpose resulted
in the conversion of 81 to the cis-R-triol 83, and of 82 to the
(65) (a) Crispino, G. A.; Jeong, K.-S.; Kolb, H. C.; Wang, Z.-M.; Xu,
D.; Sharpless, K. B. J. Org. Chem. 1993, 58, 3785. (b) Sharpless, K. B.;
Amberg, W.; Bennani, Y. L.; Crispino, G. A.; Hartung, J.; Jeong, K.-S.;
Kwong, H.-L.; Morikawa, K.; Wang, Z.-M.; Xu, D.; Zhang, X.-L. J. Org
Chem. 1992, 57, 2768.
(68) Grieco, P. A.; Gilman, S.; Nishizawa, M. J. Org. Chem. 1976, 41,
1485.
(66) Frigerio, M.; Santagostino, M. Tetrahedron Lett. 1994, 35, 8019.
(67) Gallou, F. unpublished findings.
(69) Results in square brackets obtained from NOED and sel-TOCSY
experiments.

9032 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
Bernardelli et al.
Scheme 15
dynamic control. Indeed, these conditions resulted in the
introduction of a â-hydroxyl at C-6 in both stereoisomeric series,
with 81 giving rise to authentic sclerophytin A (84).
The polarities of the four end products hold interest. In the
7R-OH subset, sclerophytin A (84) is more polar than epimer
83. This trend is reversed in the 7â-OH series, with cis-isomer
87 being more polar than its trans counterpart 86. Thus, the
6â-OH (2°) compounds are the more polar irrespective of 7-OH
(3°) configuration.
The synthetic 84 generated in this investigation was spec-
troscopically identical to the original isolate in every detail. In
addition, its optical rotation, [R]D -2.7 (c 0.11, CHCl3) was
20
weakly negative in a manner similar to that recorded for the
authentic material, [R]D20 -6.9 (c 0.087, CHCl3). This evidence,
together with the ease with which 84 can be acetylated to give
sclerophytin B (85), reduces the known existence of any
octocoral-derived 2,11-cyclized cembranoids that contain two
ether bridges to a very small number. Furthermore, 83, 86, and
87 are not identical to any other reported sclerophytin or
analogue thereof. In the least, they have not yet been uncovered.
Acknowledgment. This work was made possible by funding
from the Eli Lilly Company and the National Institutes of
Health. O.M. is grateful to CONICET, Republic of Argentina,
for a postdoctoral fellowship. T.L. is the recipient of a Ciba-
Geigy-Jubilaeumsstiftung-Stipendium. We thank Professor K.
Euler (University of Houston) for samples of authentic sclero-
phytins A and B, Professor Larry Overman (UC Irvine) for
helpful exchange of information and spectra, Professor John
Faulkner for copies of the NMR spectra of 7, and Dr. Kurt
Loening for assistance with nomenclature.
trans-triol 86. Consequently, this reagent attacks from the
â-direction irrespective of the hydroxyl configuration preexisting
at C-7. This is not so with lithium aluminum hydride. In this
case, 81 is again transformed into 83 while 82 leads to 87.
However, the yields are less satisfactory and studies involving
this reagent were discontinued, although the consequences of
kinetic control when 81 is involved were already quite apparent.
For this reason, we turned to sodium in ethanol, a dissolving
metal process recognized for its ability to proceed via thermo-
Supporting Information Available: Experimental proce-
dures and spectral characterization of all intermediates and
particularly synthetic and natural sclerophytins A and B, together
with detailed structural mapping of ketone 61 and alcohols 5
and 60 (PDF). This material is available free of charge via the
Internet at http://pubs.acs.org.
JA011285Y