though that would have involved attack by the less hindered
primary hydroxyl group.
of excellent diastereoselectivity (dr 9:1). Thus, both diaster-
eomers 7 and 8 are available in high diastereoselectivity.
Because the base equilibration and diastereoselective ketone
can be performed in one pot, any mixture of 7 and 8 can
now be transformed into diastereomer 8 of high selectivity,
requiring just four steps from starting alkyne 3.
As depicted in Scheme 1, diastereomer 8 conceptually can
diverge to the cycloisomer 11 (eq 3, path a) or the oxidatively
cyclized lactone 12 (eq 3, path b) by simple manipulation
Epoxidation of dihydropyran 5b proceeded diastereose-
lectively (4:1) to give 6 as the major diastereomer. Its lability
led to its in situ reaction with allenylmagnesium bromide to
give the trans hydroxy alkynes 7 and 8 in a 4:1 ratio in a
78% overall yield (eq 2).9 Stereoelectronic effects account
of the ligands on the Ru catalyst.1 Thus, subjecting terminal
alkyne 8 to the conditions of eq 1 provides the dihydropyran
11 in 70% yield with 6% of lactone 123b as a minor side
product. On the other hand, increasing the amount of the
N-hydroxysuccinimide and using a more electron-rich phos-
phine, p-anisylphosphine, provides the lactone 12 as the
major product in 60% yield with only 8% of the dihydro-
pyran 11.
The dihydropyran 11 corresponds to the B and C rings of
the ladder toxin yessotoxin (1). As previously, stereoelec-
tronic considerations predict that epoxidation of 11 would
give the â-epoxide10 predominantly, which indicates that an
iterative approach would necessitate a sequence as shown
in Scheme 2. This sequence has the advantage that neither
the stereochemistry of the epoxidation nor of the epoxide
ring opening is relevant because all stereoisomers funnel to
the same requisite diastereomer. Scheme 3 illustrates that
next iterative cycle. Epoxidation followed directly by addition
of allenylmagnesium bromide gave a mixture of diastereo-
mers that were not characterized.11 Subjecting the mixture
to PCC oxidation gave a good yield of ketone 14 predomi-
nantly as one diastereomer (dr 11:1). Base equilibration
followed by reduction either in two separate operations or
for the diastereoselectivity. Thus, this route provides good
access to diastereomer 7. For synthetic purposes, good access
to diastereomer 8 was also desirable. Direct epoxidation
methods to favor the opposite (i.e., R) epoxide diastereomer
proved fruitless. On the other hand, a very simple solution
evolved as shown in Scheme 2. The ketone 9 undergoes
Scheme 2. Diastereoselective Synthesis of 8
(8) Peng, Z.-H.; Li, Y.-L.; Wu, W.-L.; Liu, C.-X.; Wu, Y.-L. J. Chem.
Soc., Perkin Trans. 1 1996, 1057.
(9) (a) The structure of 7 and 8 was confirmed by conversion into the
corresponding acetates, which were easily separated by flash chromatog-
raphy. For detailed experimental procedure, see Supporting Information.
(b) The complete trans stereoselectivity in the propargylation step has been
established; see ref 1a.
(10) A similar facial selectivity in the epoxidation of trans-fused bicyclic
dihydropyrans has been reported; see: Rainier, J. D.; Allwein, S. P.
Tetrahedron Lett. 1998, 39, 9601.
facile base equilibration that strongly favors epimer 10 (dr
12:1). Simple LAH reduction then provides diastereomer 8
(11) The desired diastereomer 15 was formed as a minor diastereomer.
4312
Org. Lett., Vol. 6, No. 23, 2004