Oxidative decarboxylation as a route to ketene acetals: Assignment of relative and absolute stereochemistry to the fungal metabolite benesudon by total synthesis

Article abstract of DOI:10.1021/ol702509a

(Chemical Equation Presented) The unusual ketene acetal benesudon, which is a bioactive fungal metabolite, was synthesized from D-glucose by a route involving radical cyclization to form the five-membered ring and oxidative decarboxylation to generate the

Full text of DOI:10.1021/ol702509a

ORGANIC
LETTERS
2007
Vol. 9, No. 25
5315-5317
Oxidative Decarboxylation as a Route to
Ketene Acetals: Assignment of Relative
and Absolute Stereochemistry to the
Fungal Metabolite Benesudon by Total
Synthesis
Derrick L. J. Clive* and Minaruzzaman
Chemistry Department, UniVersity of Alberta, Edmonton, Alberta, Canada T6G 2G2
derrick.cliVe@ualberta.ca
Received October 13, 2007
ABSTRACT
The unusual ketene acetal benesudon, which is a bioactive fungal metabolite, was synthesized from
D-glucose by a route involving radical
cyclization to form the five-membered ring and oxidative decarboxylation to generate the key central double bond. The originally suggested
stereochemistry for the quaternary center C(5) must be revised, as both possibilities were prepared and a comparison with an authentic
sample was made. The absolute configuration of benesudon is 4S,5R,6S.
Benesudon, a metabolite of the fungus Mollisia benesuada,
is a biologically active substance originally assigned structure
1 (absolute stereochemistry not implied), largely on the basis
of NMR measurements, with the relative stereochemistry
suggested by the observed nuclear Overhauser enhance-
ments.1 The structure is unusual, not only among natural
products but also in its own right, and a search of the
literature for related compounds having the substructure 2
retrieved only cyclogregatin (3), which is also a fungal
metabolite.2 Benesudon shows activity against the growth
of bacteria and fungi and is also cytotoxic.1 Because these
potentially significant properties are associated with a new
structural type, the compound merits attention as a synthetic
target. Although the molecule is small, it possesses a high
degree of complexity because it contains interrelated ketene
acetal, vinylogous ester, R-methylene carbonyl, and enol
ether subunitssby any measure, a striking level of closely
associated functional groups which would likely complicate
its synthesis.
There was no prior synthetic work in this area, and
therefore we first undertook a model study3 to explore a route
to the ketene acetal subunit. In that investigation, the core
structure 5 was eventually found to be accessible by way of
(1) Thines, E.; Arendholz, W.-R.; Anke, H.; Sterner, O. J. Antibiot. 1997,
50, 13-19.
(2) Anke, H.; Casser, I.; Schrage, M.; Steglich, W. J. Antibiot. 1988,
41, 1681-1684.
(3) Clive, D. L. J.; Minaruzzaman; Yang, H. Org. Lett. 2005, 7, 5581-
5583.
10.1021/ol702509a CCC: $37.00
© 2007 American Chemical Society
Published on Web 11/08/2007

a modified version of the classical Kochi oxidative decar-
boxylation4 (Scheme 1). With this method available for
Scheme 2. Preparation of Key Glucal
Scheme 1. Generation of Ketene Acetal System
introducing the characteristic C(3a)-C(7a) double bondsat
least into a simple modelswe began a synthesis of benesu-
don itself. During that work, isolation of the marine fungal
metabolite aigialone (6) was reported.5 The similarity of its
structure, which was deduced by spectroscopic means and
X-ray analysis, to that proposed for benesudon prompted a
reconsideration5 of the original1 NOE data. This led to the
suggestion5 that the observed NOEs for benesudon might
also be compatible with the relative stereochemistry shown
in 7 (absolute stereochemistry not implied). The evidence
in favor of revising the original assignment was certainly
suggestive but not compelling and, as we were already far
advanced in our route to 1, we decided to complete that
synthesis. However, examination of the NMR spectra of
synthetic 1 showed that structural revision was indeed
required, and so we turned our attention to the proposed
alternative 7, expecting (in the event, wrongly) that the
reactions used to make 1 would be equally applicable to 7.
Tosylate 8, made from D-glucose,6 was homologated6a with
represents the six-membered ring segment and is properly
functionalized for attachment of the five-membered ring.
Hydrolysis of the acetates liberated the trans-diol 16
(Scheme 3). On the basis of experience in the synthesis of
1, we masked the secondary alcohol as a tert-butyldimeth-
ylsilyl ether and the tertiary hydroxyl as a triethylsilyl ether.
However, the tert-butyldimethylsilyl group proved too robust,
and its removal in the last step of the route to 7 could not be
achieved. Consequently, both hydroxyls were protected with
Et3SiOSO2CF3. The triethylsilyl group withstood all subse-
quent reactions, and its use allowed both hydroxyls to be
protected in a single step. Attempts to introduce a cyano
group at C(2) in 17, by reaction with NBS and MeOH,
followed by replacement of the resulting anomeric methoxy
groupsa method we had used in making 110 swere not
successful, despite extensive efforts. It appears that the
orientation of the C(5) oxygen in 17 greatly decreases the
lability of leaving groups later installed at C(2).11 Conse-
quently, we treated glucal 17 with PCC12 to obtain the
â-siloxy lactone 18 (66%). Little, if any, elimination occurred
during this step, and the lactone was a stable, easily handled
compound. Conversion to the enol triflate 19 was readily
achieved with (Me3Si)2NK and 2-[N,N-bis(trifluoromethyl-
sulfonylamino]pyridine,13 and carbonylation in the presence
of MeOH then afforded the ester 20. The fact that 18 can be
deprotonated en route to the enol triflate 19 without
the organocuprate derived from n-C6H13MgBr (Scheme 2).
Swern oxidation then afforded ketone 10, and reaction with
MeLi gave mainly (24:1) the equatorial alcohol 11 (80%),7,8
The stereochemistry9 of this step (10f11) depends on the
reagent and temperature; with MeMgI in Et2O at -78 °C,
the corresponding axial alcohol is the major product, and
this pathway was used in our synthesis of 1.7,8 Hydrogenoly-
sis of the benzyl groups and acetylation afforded the
tetraacetates 13, and the anomeric acetoxy group was
replaced by bromine (13f14). Treatment with Zn then
generated glucal 15, which is a key intermediate, as it
(4) Kochi, J. K.; Bacha, J. D. J. Org. Chem. 1968, 33, 2746-2754.
(5) Vongvilai, P.; Isaka, M.; Kittakoop, P.; Srikitikulchai, P.; Kongsaeree,
P.; Thebtaranonth, Y. J. Nat. Prod. 2004, 67, 457-460.
(6) (a) Toshima, H.; Sato, H.; Ichihara, A. Tetrahedron 1999, 55, 2581-
2590. (b) Davis, N. J.; Flitsch, S. L. J. Chem. Soc., Perkin Trans. 1 1994,
359-368.
(7) Sato, K.; Kubo, K.; Hong, N.; Kodama, H. Bull. Chem. Soc. Jpn.
1982, 55, 938-942.
(8) Miljkovic, M.; Gligorijevic, M.; Satoh, T.; Miljkovic, D. J. Org.
Chem. 1974, 39, 1379-1384.
(10) In the sequence leading to 1, the reaction with NBS in MeOH places
a methoxy group at C(2) and a bromine at C(3); the anomeric methoxy
group was then replaced by CN, using Me3SiCN in the presence of BF3‚
OEt2. Base treatment (DBU) served to generate the C(2)-C(3) double bond
and the C(2) CN was then hydrolyzed to CO2H, which was esterified.
(11) Jensen, H. H.; Bols, M. Acc. Chem. Res. 2006, 39, 259-265.
(12) Rollin, P.; Sinay¨, P. Carbohydr. Res. 1981, 98, 139-142.
(13) Comins, D. L.; Dehghani, A. Tetrahedron Lett. 1992, 33, 6299-
6302.
(9) The correctness of the present stereochemical assignment is based
on the X-ray structure of an intermediate in the synthesis of 1, where MeMgI
had been used to generate the tertiary alcohol. The material was isomeric
with that obtained from the use of MeLi.
5316
Org. Lett., Vol. 9, No. 25, 2007

(Bu3Sn)2O. Accordingly, this reagent16 was used in the
present case, and the resulting unstable carboxylic acid was
subjected to the oxidative decarboxylation procedure (23f24)
developed in the model study. Methylation of the ketene
acetal 24 at C(2) always led to extensive bismethylation.
Phenylselenation (24f25) was likewise troublesome at first,
but we eventually found a reliable procedure that gave the
required selenide in 39% yield, as well as substantial starting
material (66% corrected yield of 25) and the corresponding
bisphenylselenide (19%), which was reconverted into 24
(82%) by reaction with Ph3P in water-CH2Cl2. Once the
PhSe group was in place, methylation became straightforward
(25f26, 91%), and there remained only selenoxide frag-
mentation and deprotection of the silylated hydroxyls.
Formation of the exocyclic double bond17 (26f27) went in
high yield (93%), but many procedures had to be tried for
deprotection to 7, with the best (53%) being the use of a
controlled amount of HF-pyridine.
Scheme 3. Formation of ent-Benesudon
Although the reported NOEs were observed with 7, the
1H and 13C NMR spectra showed small but disconcerting
differences from the published1 values. Fortunately, the
original material1 had been preserved at a low temperature,
and we were able to rerun the spectra. These new measure-
ments show that synthetic 7 and natural benesudon have
identical NMR spectra. The natural compound has [R]D )
-120.5 (c ) 0.1, CHCl3), while 7 has [R]D ) +124.2 (c )
0.11, CHCl3); accordingly, the absolute configuration of
benesudon is 4S,5R,6S, and 7 (as depicted) is actually ent-
benesudon.
The dramatic influence on chemical behavior exerted by
the stereochemistry at C(5) was a challenging surprise, and
attachment of the ester group (cf. 17f20) required a different
approach in the series leading to 7 from that used in the
route to 1.10 Also, 7 showed a greater sensitivity to fluoride
ion than 1. The present synthesis tests in a natural product
context our oxidative decarboxylation route to the ketene
acetal substructure, and the method now provides opportuni-
ties to explore structure-activity relationships for this
unusual compound class.
â-elimination is noteworthy, as preservation of such a
â-oxygen substituent on a δ-lactone is unusual in the
presence of a strong base.14 Bromoetherification of 20, using
a large excess (180-fold) of propargyl alcohol and NBS, took
the route as far as 21. From this point, radical cyclization
by the triethylborane method15 at room temperature generated
the desired bicyclic framework. The standard method of slow
addition of a stannane to a hot solution of the bromides did
not work well, but the borane-air method was sufficiently
effective (51% yield) for our purpose. The C(3) carbonyl
was then introduced by ozonolysis, so as to set the stage for
introduction of the critical C(3a)-C(7a) double bond. In our
model study,3 ester hydrolysis could be achieved only with
Acknowledgment. We thank Crompton Co. and NSERC
for financial support, and Dr. H. Yang for assistance. We
are extremely grateful to Professors O. Sterner (Lund
University, Sweden) and H. Anke (University of Kaiser-
slautern, Germany) for their sample of natural benesudon.
Supporting Information Available: Full experimental
details and copies of NMR spectra for all new compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OL702509A
(14) Mckay, C.; Simpson, T. J.; Willis, C. L.; Forrest, A. K.; O’Hanlon,
P. J. Chem. Commun. 2000, 1109-1110.
(15) Yamazaki, O.; Yamaguchi, K.; Yokoyama, M.; Togo, H. J. Org.
Chem. 2000, 65, 5440-5442.
(16) Salomon, C. J.; Mata, E. G.; Mascaretti, O. A. J. Org. Chem. 1994,
59, 7259-7266.
(17) We did not examine alternatives to the selenium-based route for
attachment of the exocyclic methylene.
Org. Lett., Vol. 9, No. 25, 2007
5317