followed by successive acrylate ester formation under the
standard protocol (acryloyl chloride, Et3N, DMAP) afforded
the dienic-ester. Subjecting the acrylate ester to Grubbs’
carbene catalyst 108 readily allowed for the formation of
lactenone 11 via a ring-closing olefin metathesis9 with a
combined yield of 67% over two steps from 9.
Scheme 4a
As shown in Scheme 3, an ensuing stereoselective epoxi-
dation of the corresponding lactenone intermediate 11 with
basic hydroperoxide provided the epoxy-lactone followed by
a subsequent regioselective reduction of the oxirane by means
of the in situ generated PhSeH10 afforded intermediate 12
and set the stage for the tandem nucleophilic addition-
oxonium cation generation-diastereoselective reduction
sequence in anticipation of providing the vital â-C-glycoside
component. Thus, nucleophilic addition of allylmagnesium
bromide to â-hydroxy lactone 12 furnished the lactol, which
was immediately transformed into the oxocarbenium cation
in the presence of TFA and consequently reduced with Et3-
SiH to furnish the cis-2,6-disubstituted tetrahydropyran 13
with a 76% yield from 12.11 To avoid removal of the TBS
groups during the reduction, TFA was employed instead of
BF3•OEt2 as used in Kishi’s conventional procedure. Unex-
pectedly, the C13 hydroxyl group was concomitantly protected
as a TES ether under these conditions.
Selective deprotection of the TES group was unsuccessful
upon the treatment of 13 with PPTS in methanol, and other
selective reagents such as Pd/C12 and DDQ13 also failed to
provide the free hydroxy-pyran. In lieu of selective depro-
tection, global desilylation of 13 was carried out with TBAF,
and the corresponding 1,2-diol was reprotected as the
acetonide 14 in 81% yield over two steps. Oxidation of the
free hydroxyl group resident in 14 was accomplished with
PCC, and ketone 15 was subsequently transformed via the
methylene Wittig reagent to the cis-2,6-disubstituted 4-exo-
methylene tetrahydropyran unit 16, which possessed all the
functionality required for the synthesis of 2. The relative
configuration of the â-C-glycoside subunit 16 was confirmed
by NOE enhancements as shown in Scheme 3.
a (a) triethyl phosphonoacetate (2.0 equiv), LiHMDS (2.0 equiv),
THF, 0 °C, 15 min, 98%. (b) pyridine-HF (10 equiv), THF, rt, 4 h,
92%. (c) Dess-Martin periodinane (1.5 equiv), CH2Cl2, rt, 1 h,
61%. (d) vinylMgBr (2.5 equiv), THF, -78 °C, 2 h, 49%. (e)
TBSCl (2.0 equiv), imidazole (3.0 equiv), DMF, 20 h, 85%. (f)
NaOH (1.0 M, 2.5 equiv), EtOH, 0 °C to rt, 20 h, 95%. LiHMDS
) lithium salt of 1,1,1,3,3,3-hexamethyldisilazane, Dess-Martin
periodinane ) 1,1,1-tris(acetyloxy)-1,1-dihydro-1,2-benziodoxo-3-
(1H)-one, TBSCl ) tert-butyldimethylsilyl chloride.
for the formation of the desired allylic alcohol 19 as an
inconsequential racemic mixture at C7 with a moderate yield.
Finally, protection of the allylic alcohol as a TBS ether
(TBSCl, imid.) and hydrolysis (NaOH, EtOH) of the ester
group provided the conjugated acid 4, which positioned us
for the convergence of the two synthetic intermediates 3 and
4, followed by the completion of the targeted compound 2.
To commence the convergence of 3 and 6, selective access
to the stereogenic secondary alcohol of 3 was essential.
Along this line, the acetonide protecting group was cleaved
under acidic conditions (TFA), and the primary hydroxyl
group was selectively reprotected as a TBS ether under
standard conditions in a quantitative yield.The first attempted
coupling 3 and 4 utilizing DCC was unsuccessful, giving
the desired ester 20 in low yield along with inseparable
impurities. Fortunately, esterification proceeded smoothly
under Yamaguchi conditions15 utilizing two equivalents of
4, affording the hexaeneic ester intermediate 20 as a
diastereomeric 3:1 mixture. Apparently, the two C7 epimers
of acid 4 displayed distinct kinetic reactivity toward the
enantiopure alcohol 3. Ring-closing olefin metathesis was
then attempted on bis-TBS-protected hexaene 20 employing
a variety of reaction conditions. In general, Grubbs’ first-
generation catalyst led to recovery of the starting material,
while the second-generation catalyst resulted in decomposed
starting material even at room temperature. At this point,
we conjectured that the TBS protecting group on allylic
alcohol was impeding the ring closure.16 Unfortunately,
cleavage of the silyl ethers was problematic, because of
sensitive functional groups present in 20 giving rise to
With the C8-C20 intermediate in hand, attention was
focused on the synthesis of the triene C1-C9 subunit.
Accordingly, the known R,â-unsaturated aldehyde 1714 was
treated with the corresponding Horner-Emmons reagent,
which afforded the expected (E,Z)-conjugated ester. Subse-
quent HF-pyridine-mediated desilylation of the TBS group
furnished the free hydroxyl intermediate 18 in a 90% yield
over two steps. Dess-Martin oxidation of the corresponding
homoallylic alcohol afforded the labile aldehyde with a
modest 61% yield. Because of the highly acidic nature of
the R-proton, addition of the vinyl Grignard reagent allowed
(8) Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett. 1999, 1,
953.
(9) Ghosh, A. K.; Cappiello, J.; Shin, D. Tetrahedron Lett. 1998, 39,
4651.
(10) Miyashita, M.; Suzuki, T.; Hoshino, M.; Yoshikoshi, A. Tetrahedron
1997, 53, 12469.
(11) Kraus, G. A.; Molina, M. T.; Walling, J. A. J. Chem. Soc., Chem.
Commun. 1986, 1568.
(12) Rotulo-Sims, D.; Prunet, J. Org. Lett. 2002, 4, 4701.
(13) Tanemura, K.; Suzuki, T.; Horaguchi, T. J. Chem. Soc., Perkin
Trans. 1 1992, 2997.
(15) Inanaga, J.; Hirata, K.; Saeki, H.; Katsuki, T.; Yamaguchi, M. Bull.
Chem. Soc. Jpn. 1979, 52, 1989.
(16) Hoye reported ring closure via olefin metathesis at the same C8-
C9 alkene with a compound that is somewhat structurally related to 20.
Similar obstacles with respect to ring-closing olefin metathesis have been
described, see: Matsuya, Y.; Kawaguchi, T.; Nemoto, H. Org. Lett. 2003,
5, 2939.
(14) McLaughlin, M. J.; Hsung, R. P. J. Org. Chem. 2001, 66, 1049.
Org. Lett., Vol. 7, No. 12, 2005
2323