S. Rudrawar et al. / Tetrahedron Letters 54 (2013) 1198–1201
1199
AcO AcO
AcO AcO
CO2Me
OH
OH
H
H
H
O
O
CO2Me
Bu3SnAll, AIBN
AcO
AcO
toluene, 100 °C, 8 h
57% (Ref. 6)
AcHN
Br
AcHN
AcO AcO
AcO
AcO
H
O
CO2Me
4
6
NBS
AcO
66%
MeCN-H2O,
AcHN
80 °C, 30 min.
AcO
AcO AcO
AcO AcO
4
5
ratio 1:2)
95% (
/
OH
OH
O CO2Me
H
H
O
CO2Me
3
Bu3SnAll, AIBN
X
AcO
AcO
toluene, 100 °C, 8 h
AcHN
Br
AcHN
H
AcO
AcO
5
7
Scheme 1.
bromohydroxylation is governed by solvent composition, reaction
temperature and the stereoelectronic nature of the substituent at
C4.10 The two bromohydrins formed are, however, difficult to sepa-
rate on silica gel in a number of different solvent combinations, and
require several chromatographic purifications to isolate the pure
isomers. Given that the trans-2,3-diaxial bromohydrin 5 is always
formed in the bromohydroxylation of 3, we examined the reactivity
of 5 towards C3 allylation (Scheme 1). Reaction of 5 at 100 °C in tol-
uene with allyltributyltin using AIBN as the free-radical initiator
afforded a single C3-allylated product in 66% yield. Interestingly,
the main product was assigned as the C3 equatorially-allylated
derivative 6, rather than the C3 axially-substituted product 7, based
on the large J3,4 value (10.5 Hz) of the product, and comparison with
NMR data for 6 prepared previously from the C3 equatorial bro-
mohydrin.6,8 The yield of 6 was slightly improved when compared
to reaction with trans-2,3-diequatorial bromohydrin 4 (57%6) under
thesamereactionconditions, and importantly wasreproducibleona
large scale (up to 25 g, 44 mmol, of 5).
Scheme 2.
Table 1
Reaction conditions evaluated for C3 allylation of bromohydrins 4/5 using Bu3SnAll in
the presence of AIBNa
Precursor trans-2,3-diequatorial 4 and trans-2,3-diaxial 5 bro-
mohydrins each give rise to the same C3 equatorially-substituted
product 6, indicating that the stereoselectivity of the free-radical
reaction is evidently not influenced by the configuration of the
activating group (bromide) on the C3 position. Substituents adja-
cent to the radical centre in glycopyranosyl radicals influence the
direction of attack on that centre:11 where the radical centre is
flanked by equatorial substituents there is a marked preference
for an equatorial product; where the substituents are axial, the
preference is for axial attack.11 In the case of the radical species
8 formed from bromohydrins 4/5, the bulkiest substituents b to
the radical centre (4-OAc, 2-CO2Me) are equatorial. The formation
of the equatorially allylated product 6 is therefore consistent with
the reported11 trend.
Entry
Solvent and reaction conditionsb
Outcomec (yield %)
4/5
6
9
1
2
3
4
5
6
Toluene, 100 °C, 8 h
Benzene, 80 °C, 10 h
THF, 60 °C, 10 h
1,4-Dioxane, 100 °C, 8 h
MeCN, 80 °C, 10 h
—
—
5
10
15
18
68
79
8
20
40
26
22
8
80
35
35
40
1,2-Dichloroethane, 80 °C, 10 h
a
All reactions were carried out on 0.17 mmol (0.1 g) of a mixture of 4/5 mixture
(ratio = 1:2), except for entries 1 (1.75 mmol, 1 g) and 2 (0.35 mmol, 0.2 g).
Reaction composition: 4/5 (0.17 mmol), Bu3SnAll (0.84 mmol), AIBN
(0.017 mmol), anhydrous solvent (8 mL).
b
c
Isolated yields.
To investigate if the yield of 6 could be further improved, we
studied the effects of solvent on the outcome of the allylation
reaction. The use of toluene (chosen6 to achieve reaction at high
temperature) as a solvent for the free-radical allylation of
bromohydrin 4 always proceeded with the formation of de-
brominated Neu5Ac derivative 9 as a by-product. (e.g., 22% under
the optimised reaction conditions; Table 1, entry 1).
Presumably, the formed radical at the C3 position abstracts a
hydrogen atom from the solvent (toluene). Bromohydrin mixture
4/5 was therefore reacted with allyltributyltin in alternative sol-
vents (benzene, THF, 1,4-dioxane, acetonitrile, 1,2-dichloroethane)
(Scheme 2). The results of these reactions are summarised in
Table 1. Interestingly, replacing toluene with anhydrous benzene
led to an increase in reaction yield (79%, up from 68%), with the
de-brominated derivative 9 formed in only ꢀ8% yield (Table 1, en-
try 2). Reaction in THF, 1,4-dioxane, acetonitrile or 1,2-dichloroeth-
ane (DCE), however, resulted in less than satisfactory yields of 6
(Table 1, entries 3–6). While reaction in benzene gave the most
In light of this result, the reactivity of bromohydrin mixture 4/5
towards the allylation reaction was examined. Reaction of a mix-
ture of 4/5 (ratio 1:2) at 100 °C in toluene with allyltributyltin in
the presence of AIBN afforded the single C3 equatorially-allylated
product 6 in 68% yield (Scheme 2, Table 1, entry 1). Using the
bromohydrin mixture, the overall yield of C3 C-allyl derivative
6 from glycal 3 over two steps was 63%, a substantial increase
in yield over that previously achieved (22%6) via trans-2,3-diequ-
atorial bromohydrin 4.