Rozhkov and Larock
Great improvements were subsequently achieved by
using a 4:1 1,4-dioxane/water mixture as the solvent at
higher temperatures. Increasing the reaction tempera-
ture to 80 and 100 °C improved the yield of the desired
product 2 to 44% (entry 3) and 64% (entry 4), respectively.
Besides the desired product, significant amounts of Heck-
type products 12 and 13 were detected among the
inseparable mixture of side products. A further increase
in the water concentration resulted in a decrease in the
yield of 9, perhaps due to partial hydrolysis of the
starting material 11 (entry 5). The use of polar solvents,
such as DMF, DMA, and acetonitrile, apparently resulted
in rapid hydrolysis of the starting material 11. Further
(entry 13). Surprisingly, the use of 95% pure trans,trans-
2,4-hexadiene gave a 20:1 ratio of isomers 25a and 25b
in a 70% yield (entry 13). This result will be discussed
further in our later discussion of the reaction mechanism.
Remarkably, in all of our previous palladium annulation
chemistry, relatively hindered 1,3-dienes, like those
employed in entries 2, 3, 9-11, and 13, were completely
unreactive and only 1,3-dienes bearing monosubstituted
terminal double bonds have given satisfactory results.
The exclusive generation of E-stereochemistry in the
newly formed double bond in products 17, 18, 20, and
21 is consistent with the intermediacy of a syn-π-
allylpalladium intermediate in these reactions. The
annulation of methyl trans,trans-2,4-hexadienoate failed,
presumably because of the low affinity of the cationic
arylpalladium intermediate for the electron-deficient
double bond. Relatively sterically hindered 2,3-diphenyl-
1,3-butadiene gave only a 10% yield of the desired
product 26 (entry 14), while the even more hindered
dienes 1,4-diphenyl-1,3-butadiene and 2,5-dimethyl-2,4-
hexadiene were completely unreactive. The attempted
annulations of 2,3-dimethoxy-1,3-butadiene, 1-methoxy-
1,3-cyclohexadiene, 1,4-pentadiene, cycloheptatriene, and
cyclohexene did not afford any recognizable products.
1
8
optimizations, which utilized Pd(OAc)
dppp, dppb, BINAP, and PPh as the phosphine ligand;
and AgOAc, Ag PO and Ag O as the silver salt, only
2
as the catalyst;
3
3
4
2
resulted in a lower yield of the annulated product 9. We
have thus used the following “optimal” procedure for all
subsequent annulations: the iodoacetoxycoumarin (0.25
mmol), Pd(dba)
2
(5 mol %, 0.0125 mmol), dppe (5 mol %,
CO (0.5 mmol), 1,3-diene (1.0 mmol),
0
.0125 mmol), Ag
2
3
and 5 mL of a 4:1 1,4-dioxane/water mixture were stirred
at 100 °C for 24 h.
Next, the scope and limitations of this annulation have
been studied by using various 1,3-dienes and representa-
tive examples are shown in Table 2. An increase in the
ring size of the cyclic 1,3-diene leads to a significantly
lower yield of annulation product (entries 1-3). Cyclo-
pentadiene failed to give any annulation products, pre-
sumably due to rapid dimerization or some other side
reaction. Most terminal 1,3-dienes have given the ex-
pected annulation products 16-23 in 61 to 83% yields
with excellent regioselectivity (entries 4-11). Running
the reaction of 2,3-dimethyl-1,3-butadiene on a 2.0-mmol
scale with only 10 mL of the 4:1 1,4-dioxane/water
mixture resulted in an even higher 91% yield (entry 11),
indicating the utility of this procedure for practical
applications. The higher yield in the larger scale reaction
is presumably due to an increase in the concentration of
the reagents by a factor of 4 that facilitates coordination
of the 1,3-diene to the arylpalladium intermediate. The
regioselectivity in these experiments can be explained by
the greater affinity of the arylpalladium intermediate for
the less hindered terminal double bond over an internal
double bond. The annulation of isoprene gave a 3:2
mixture of regioisomers 24a and 24b in a 73% yield
In an effort to broaden the scope of this reaction,
similar reactions have been performed on coumarins 27,
35, and 37. All 1,3-dienes investigated have reacted with
coumarin 27 to give the expected annulation products
28-34 in high yields (entries 15-21). Coumarin 35 gave
the expected product 36 in a good yield, even when using
relatively hindered 2,3-dimethyl-1,3-butadiene (entry 22).
Annulation of the coumarin 37 under our “optimal”
reaction conditions resulted in hydrolysis of the acetyl
group. The facile hydrolysis is consistent with the higher
1
9
acidity of 4-hydroxycoumarin than 7-hydroxycoumarin.
The same reaction without the addition of water gave a
48% yield of dihydrofurocoumarin 38 (entry 23). In this
experiment, the acetyl group is quite possibly still being
hydrolyzed by trace amounts of water present in com-
mercial 1,4-dioxane.
Chemical modification of the prepared dihydrofuro-
coumarins enhances the utility of our approach for the
synthesis of various potentially biologically interesting
products. For example, the hydroxymercuration/demer-
20
21
curation and dihydroxylation of coumarin 23 gave the
corresponding alcohol 39 and diol 40, close analogues of
columbianetin (2) and prandiol (7), respectively, in high
(entry 12). The annulation of isoprene by o-iodophenol
is mostly governed by steric factors, favoring addition to
the less hindered double bond and affording a 7:1 ratio
of the corresponding annulation products.13 The poor
regioselectivity in entry 12 presumably results from the
higher reactivity of the presumed cationic arylpalladium
intermediate toward the more electron-rich disubstituted
double bond leading to a competition between steric and
electronic factors, which produces the two different
products.
22
yields (Scheme 1). The dehydrogenation of coumarin 9
afforded benzofurocoumarin 41 in a 90% yield. According
to previous reports benzofurocoumarins, have a high
potential for the treatment of psoriasis and related
22,23
diseases.
A proposed mechanism for this annulation process is
shown in Scheme 2. Initial oxidative addition of the
(18) Maitlis, P. M.; Espinet, P.; Russell, M. J . H. Compr. Organomet.
2
,4-Hexadiene (a 3:2 mixture of trans,trans and cis,-
Chem. 1982, 6, 385.
(19) Georgievskii, V. P. Khim. Prir. Soedin. 1985, 6, 770; Chem.
Abstr. 1986, 104: 144967.
1
7
trans stereoisomers), which has generally been unre-
active and afforded dismal yields in most of our previous
palladium annulation chemistry, gave a 3:2 ratio of trans-
and cis-stereoisomers 25a and 25b in a 60% overall yield
(20) Noland, W. E. Organic Syntheses; J ohn Wiley and Sons: New
York, 1988; Collect. Vol. VI, p 766.
(21) Panek, J . S.; Cirillo, P. F. J . Am. Chem. Soc. 1990, 112, 4873.
(
22) Via, L. D.; Gia, O.; Magno, S. M.; Santana, L.; Teijeira, M.;
Uriarte, E. J . Med. Chem. 1999, 42, 4405.
(23) Vedaldi, D.; Caffieri, S.; Miolo, G.; Guiotto, A.; Dall’Acqua, F.
J . Photochem. Photobiol. B 1992, 14, 81.
(
17) The ratio of trans,trans to cis,trans isomers was determined
by gas chromatography.
6
316 J . Org. Chem., Vol. 68, No. 16, 2003