2(3H)-Furanones Via Electrophilic Cyclization
the 3-alkynoate ester 14 bearing â-ketoester functionality
proceeded smoothly and resulted in an 80% yield of the
spirocyclic product 15 (Table 1, entry 13). Although the oxygens
of both of the carbonyl groups could potentially attack the
presumed iodonium intermediate (see the later mechanistic
discussion), only the ester oxygen reacted to give the lactone
15 as the exclusive product. Previous work on the electrophilic
cyclization of acetylenic ketones and aldehydes has been
reported by others and our group.13,18 This chemoselectivity may
be explained by the closer proximity of the ester oxygen to the
hypothetical iodonium intermediate.
When we introduced an alkyl group instead of an aryl group
into the 4-position of the 3-alkynoate ester, the reaction was
somewhat faster (Table 1, entries 14-20). The cyclization of
3-alkynoate ester 16 proceeded cleanly in 1 h using I2 and
afforded an 86% yield of the iodolactone 17 (Table 1, entry
14). The yield for the analogous ICl cyclization was significantly
lower (Table 1, entry 15). However, none of the corresponding
chlorolactone was observed. For the 3-alkynoate ester 18 bearing
a cyclohexane moiety, the reaction proceeded more slowly than
that of 16, when employing I2 (Table 1, entry 16). However,
we still obtained a 78% yield of the corresponding iodolactone
19. Here ICl cyclization of 18 gave exclusively iodolactone 19
in 1 h, but in only a 60% yield (Table 1, entry 17). When we
employed I2 in the cyclization of the 2,2-diallyl-3-alkynoate ester
20, the reaction proceeded smoothly affording an 84% yield of
iodolactone 21 (Table 1, entry 18). Thus, neighboring carbon-
carbon double bonds are not problem. However, the terminal
double bonds of 20 were vulnerable to ICl and the reaction with
this reagent was messy. In the case of the cyclopentane-
containing ester 22, the yield of the corresponding iodolactone
23 was almost quantitative when using I2 (Table 1, entry 19).
The ICl cyclization of 22 gave exclusively iodolactone 23 in
only 0.5 h in an 86% yield (Table 1, entry 20).
We next considered the possibility of preparing 2-furanones
by the cyclization of 3-alkynoate esters bearing one or no
R-substituents. Since mono-R-substituted alkynoate esters are
not very stable and can be easily isomerized to the corresponding
allenic esters in the presence of a base,28 we chose to explore
instead the cyclization of the allenic ester 24, rather than a
monosubstituted 3-alkynoate ester. This cyclization resulted in
a 96% yield of the lactone 25, bearing the more stable
conjugated 2-furanone ring (Table 1, entry 21). Similar allenic
ester and acid electrophilic cyclizations with different electro-
philes have been explored by others under different reaction
conditions.29 Thus, we chose not to examine any additional
allenic esters.
When we explored the R-unsubstituted 3-alkynoate ester 26,
extensive isomerization occurred, affording the conjugated
2-furanone 27 in a 62% yield (Table 1, entry 22). Substrate 28
with a terminal alkyne failed to give the desired cyclization
product 29 even after 2 days (Table 1, entry 23). Some starting
material reacted with I2 to form a product resulting from addition
of the I2 across the acetylene moiety. Introducing two methyl
groups in the 2 position did nothing to improve the situation
(Table 1, entry 24). Introduction of a trimethylsilyl group on
the acetylene failed to afford any of the desired lactone using
either I2 or ICl (Table 1, entry 25). This ester only gave products
of addition of the I2 across the carbon-carbon triple bond, which
decomposed back to the starting material 32 upon column
chromatography. The R-unsubstituted substrate 34 also failed
to provide any iodolactone; only starting material was recovered
after aqueous work up and neither the desired conjugated
cyclization product 35 nor the corresponding 2(3H)-furanone
was observed (Table 1, entry 26). Comparing entries 1 and 26,
the Thorpe-Ingold effect of the gem-dimethyl group clearly
benefits the cyclization.30 When we attempted to cyclize the
R-hydroxy substrate 36, where the oxygen from the OH group
and the oxygen of the carbonyl group can both serve as
nucleophiles, we obtained an unidentified compound, which was
not the desired lactone (Table 1, entries 27 and 28). Even after
we protected the OH group with a TBS group, the desired I2
cyclization product was not formed (Table 1, entry 29).
Surprisingly, however, the TMS-protected compound 40 af-
forded the desired 2-furanone 41 in a 66% yield (Table 1, entry
30). However, the ICl cyclization of 40 did not give the desired
product (Table 1, entry 31). We also examined substrate 42,
which has an sp2 carbon center in the R position. No reaction
took place using I2. With ICl, we obtained a mixture of the
desired cyclization product and an acetylene addition product,
which were hard to separate (Table 1, entry 33). When we
introduced a carbonyl group into the R position (44), we only
observed acetylene addition products when employing either I2
or ICl (Table 1, entries 34 and 35).
In conclusion, the substitution and the hybridization of the R
position of the 3-alkynoate ester are crucial for the electrophilic
cyclization to take place successfully. If there is an sp2 carbon
center present on the R carbon, cyclization is difficult, apparently
due to the wider angle between the alkyne and the ester groups.
The nucleophilic oxygen of the ester group is apparently simply
too far away for the iodonium intermediate to undergo cycliza-
tion. For similar reasons, the 3-alkynoate ester bearing a
cyclopropane ring in the R position also failed to cyclize when
allowed to react with I2 and only the product of addition to the
alkyne was observed (Table 1, entry 36). The addition product
was not stable and quickly reverted back to starting material
upon aqueous work up. Due to the strained, rather reactive
cyclopropane ring system, a ring opened product was observed
in the crude 1H NMR spectrum of the corresponding ICl reaction
(Table 1, entry 37).
In an attempt to overcome some of the limitations encountered
in the electrophilic cyclization of the 3-alkynoate esters, we have
examined the cyclization of the corresponding alkynoic acids.
We hydrolyzed our previous best substrate 22 to the corre-
sponding 3-alkynoic acid 48. This 3-alkynoic acid reacted with
both I2 and ICl in the presence of 3 equiv of NaHCO3 in CH3-
CN in slightly better yields than the corresponding ester (Table
2, entries 1 and 2), suggesting that an acid group or rather the
corresponding carboxylate is a better nucleophile than the ester
in these cyclization reactions. PhSeCl was also a good electro-
phile and gave the desired product 49 in an 85% yield (Table
2, entry 3). Next, we transformed an unsuccessful 3-alkynoate
ester 46 to the corresponding acid 50 and conducted the same
cyclization reaction. As we desired, the acid 50 gave the
cyclization product 47 using either I2 or ICl (Table 2, entries 4
and 5). Under the basic conditions, the acidic hydrogen is
removed and the anionic carboxylate becomes a better nucleo-
(27) (a) Rossi, R.; Bellina, F.; Bechini, C.; Mannina, L.; Vergamini, P.
Tetrahedron 1998, 54, 135. (b) Chui, C. C.; Jordan, F. J. J. Org. Chem.
1994, 59, 5763.
(28) Sleeman, M. J.; Meehan, G. V. Tetrahedron 1989, 30, 3345.
(29) (a) Fu, C.; Ma, S. Eur. J. Org. Chem. 2005, 3942. (b) Ma, S.; Wu,
S. Tetrahedron Lett. 2001, 42, 4075. (c) Ma, S.; Pan, F.; Hao, X.; Huang,
X. Synlett 2004, 85.
(30) Jung, M. E.; Piizzi, G. Chem. ReV. 2005, 105, 1735.
J. Org. Chem, Vol. 73, No. 7, 2008 2665