an oxygen atom. In addition, the elimination of Ph3PdS is not
as thermodynamically favorable as that of Ph3PdO, as the
phosphorus-sulfur double bond is not as strong as the
phosphorus-oxygen double bond (PdO bond energy ∼ 544
kJ/mol, whereas PdS ∼ 335 kJ/mol).24
complementary regioselectivity to standard esterification
conditions. The reason for the different regioselectivity is
perhaps due to the stereoelectronic differences between the
two positions. The 2-position is less hindered in comparison
to the 3-hydroxyl group and is acylated more favorably under
the mild conditions of the Mitsunobu reaction. The 3-position
is favored under normal esterification conditions, perhaps
due to the hydrogen bonding of the cis-3- and 4-hydroxyl
groups, increasing the nucleophilicity of the 3-position.
This exceptional selectivity disappears when using the
1-methoxy derivative 38. The direct esterification also favors
the 3-ester (45% 39, 12% 40, and 17% 41) but to a lesser
degree than the 1-deoxy compound 34. Under Mitsunobu
conditions, when using 2 equiv of reagents, the reaction leads
to almost equal amounts of the 3-ester 39 and 2-ester 40
and some unreacted starting material. The reaction can reach
completion by adding triethylamine as a coreactant giving
the 2- and 3-esters in a 1:1 ratio. When a larger excess of
reagents was used (4 equiv of each and stirring for 4 days),
the reaction gave about 5% diester 41, 35% 2-ester 40, and
60% 3-ester 39. The regioselectivity for the 2-ester decreases
probably due to the presence of the large excess of reagents.
In summary, we have carried out a series of esterification
reactions on chiral alcohols including vicinal diols derived
from monosaccharides under Mitsunobu conditions. The
Mitsunobu esterifications of the diols derived from D-glucose,
D-galactose, and D-mannose all proceeded with retention of
configuration. The glucose based diols 23, 15, and 26 and
1-deoxy galactose derivative 34 afforded 2-OBz with high
regioselectivity. The D-glucose derivatives regioselectively
afforded the 2-benzoate esters with excellent yields. How-
ever, D-mannose derivatives and 1-methoxy D-galactose
derivatives did not show good regioselectivity. The 1-deoxy
D-galactose 34 afforded excellent regioselectivity, and the
Mitsunobu esterification gave opposite regioselectivity com-
pared to that of the direct esterification reactions. The
stereochemistry of Mitsunobu reactions is very sensitive to
the environment, and hindered substrates tend to yield
products with retention of configuration. The interesting
regioselectivity and stereoselectivity of these Mitsunobu
reactions on diols derived from common sugars can be
applied to other chiral cyclic diols.
The regioselectivities of 15, 23, and 26 can be ration-
alized by stereoelectronic factors. In the diol 23, the 2-OH
is less hindered than the 3-OH since it is next to a primary
carbon. Therefore, the regioselectivity favors the formation
of the 2-ester. In compounds 15 and 26, the cis relationship
with the methoxyl group presumably promotes hydrogen
bonding and leads to the enhanced nucleophilicity of the
2-hydroxyl group. This makes the 2-hydroxyl group more
reactive, and under mild Mitsunobu conditions, the 2-ester
is the major product.
When the substrates are derivatives from D-mannose (29,
30), the Mitsunobu reaction results are quite different. Unlike
the D-glucose derivatives shown in Table 1, here the regiose-
lectivity and yields are low perhaps due to the 2-hydroxy
group being in the axial position and more sterically
hindered.25 For the 1-deoxy derivative 29, using the same
conditions utilized in the glucose series, only a very small
amount of the starting material was esterified. Two esteri-
fication products were obtained in equal ratio, affording no
regioselectivity, and the majority (>80%) of the reaction
mixture was unchanged starting material.
For the 1-methoxy mannoside 30, depending on the
reaction conditions, different esterification products with
retention of stereochemistry were obtained (Figure 2). Under
mild conditions, using 2 equiv of reagents, and heating at
∼45 °C for 24 h, the 2-ester was obtained in 38% yield along
with 62% unreacted starting materials. Under more harsh
conditions, using 3 equiv of each reagent and heating to about
80 °C for 3 days, the starting material was completely
converted to three products: the diester 31 in 40%, 2-ester
32 in 45%, and 15% 3-ester 33. The reaction is no longer
regioselective under these conditions; however, both condi-
tions lead to retention of configuration. This is confirmed
1
by analysis of the H NMR spectra. Additionally, simple
hydrolysis of 32 led to the starting material 30 as indicated
1
by H NMR spectra.
For D-galactose derivatives, an interesting regioselectivity
was observed. The direct esterification of 1-deoxy galactose
acetal 34 using benzoyl chloride26 led to the 3-ester 36 as
the major product in 85-95% yield along with a small
amount of diester 37. However, under Mitsunobu reaction
conditions using TPP and either DEAD or DMEAD,27 the
opposite regioselectivity was observed. The 2-ester 35 was
obtained as the major product in 90% yield along with about
5-10% of 36. Thus, the Mitsunobu reaction provides
Acknowledgment. This research is supported in part by
the National Science Foundation (Grant 518283), the Ameri-
can Heart Association (Grant 043285N), and Louisiana Board
of Regents Graduate Fellowship to K. Williams.
Supporting Information Available: The experimental
data for the preparation of 20, 22, 24, 25, 27, 35, 36, and
39-41 are provided. The 1H NMR and 13C NMR spectra of
these compounds and some starting materials are also
provided. This material is available free of charge via the
(24) Huheey, J. E.; Keiter, E. A.; Keiter, R. L. Inorganic chemistry,
ISBN 0-06-042995-X, 4th ed., page, A 31.
(25) Davis, A. P.; Dresen, S.; Lawless, L. J. Tetrahedron Lett. 1997,
38, 4305–4308.
(26) Kondo, Y. Carbohydr. Res. 1989, 193, 279–282.
(27) Sugimura, T.; Hagiya1, K. Chem. Lett. 2007, 36, 566–567.
OL801316F
4206
Org. Lett., Vol. 10, No. 19, 2008