3732
S. Tatsumi et al. / Tetrahedron Letters 54 (2013) 3731–3734
Table 2
Since we found that the preparation of sufficiently active TrClO4
was not always reproducible, we compared several sources of Tr+
and chose trityl N,N-bis(trifluoromethanesulfonic)imidate (TrN
Tf2), which was superior in terms of preparation, handling, and
reactivity to the others (TrClO4, TrOCOCF3, TrOTf, and TrPF6). First
of all, glycosylation of cholesterol 29 was chosen as a model reac-
tion and the reactions were carried out with donor 1 (3.0 equiv)
and cholesterol 2 (1.0 equiv) in the presence of TrNTf2 (6.0 equiv)
in pure or mixed solvents at 0 °C for optimization of the reaction
conditions (Table 1). CH2Cl2 was used as a solvent in every case
to dissolve highly lipophilic cholesterol 2, and toluene, MeCN, diox-
ane, and THF were tested as cosolvents. We found that the yield of
the desired product 3 was low when the reaction was conducted in
CH2Cl2 (entry 1) and the addition of nonpolar toluene did not im-
prove the result (entry 2). In contrast, the use of polar MeCN and
1,4-dioxane as cosolvents improved the yield (entries 1 vs 3 and
4) and 1,4-dioxane with a little higher Gutmann donor number
(DN)10 (14.8) than MeCN (14.1) gave a better result. However,
the use of THF as a cosolvent resulted in a poor yield of 3 (entry
5), which can probably be attributed to the deactivation of TrNTf2
by highly coordinating THF (DN = 20.0); the color of the reaction
mixture gradually changed from clear yellow to black.11 The mod-
est Lewis basicity of dioxane may be appropriate to promote this
reaction without deactivating TrNTf2. Thus, the desired product 3
was obtained in a modest yield via the novel Tr+-promoted glyco-
sylation. But the yield was not improved further by changing reac-
tion conditions. For example, reactions conducted at rt yielded
substantial byproducts. The main byproduct was 1-deoxy-2,3,4,6-
tetra-O-Bn-glucose, which was probably formed by the reduction
of the oxocarbenium ion generated from 1 with the BH3 group of
another molecule of 1. Lowering the temperature significantly slo-
wed down the reaction and the desired product 3 was not obtained
in sufficient yield even with extended reaction time (entry 6).
In order to improve the reaction, we next focused on the use of
trityl ethers as glycosyl acceptors. Thus, O-tritylcholesterol 512 was
used in the place of cholesterol 2 and detritylated in situ by TfOH.
Glycosylation using O-tritylcholesterol 5 as acceptor
OR
O
RO
RO
+
BH3
O
MeO
RO
P
OMe
TrO
1: R = Bn (α:β = 74:26)
5
4: R = Bz (β only)
(2.0 equiv)
(1.0 equiv)
OR
TfOH (0.57 equiv)
solvent, rt
O
RO
RO
O
RO
3: R = Bn
6: R = Bz
Entry Donor Solvent
Time (min) Yield (%)a :b)b
(a
1
2
3
4
5
1
1
1
4
4
4
1
CH2Cl2
Dioxane–CH2Cl2 (2:1, v/v)
Dioxane
CH2Cl2
Dioxane–CH2Cl2 (2:1, v/v)
dioxane
240
10
10
120
10
10
40 (54:46)
76 (77:23)
83 (83:17)
61 (b Only)
74 (b Only)
67 (b Only)
55 (74:26)
6
7c
Dioxane–CH2Cl2 (2:1, v/v)
30
a
b
c
Isolated yield.
Determined by 1H NMR.
2.0 equiv of TrNTf2 and 2.0 equiv of cholesterol were used as activator and
acceptor, respectively.
The resultant Tr+ would activate the glycosyl boranophosphotriest-
ers to promote the glycosylation of cholesterol generated from 5 in
situ. As reported in the literature,13 the use of O-tritylalcohols as
glycosyl acceptors in the place of alcohols is advantageous when
complex acceptors are used because protecting group-manipula-
tion is simplified and this would also allow us to circumvent the
use of moisture-sensitive TrNTf2.
Table 1
As shown in Table 2, per-O-Bn- and Bz-glycosyl boranophos-
photriesters (1 and 4) were allowed to react with O-tritylcholester-
ol 5 in the presence of TfOH in CH2Cl2, dioxane, or a 1:2 mixture of
these two solvents. Tf2NH was found to be ineffective probably be-
cause of the low solubility in CH2Cl2. Compared with the method
using TrNTf2 and cholesterol 2, the desired product 3 was obtained
in better yield (entries 2 vs 7). Per-O-Bn-donor 1 gave the product 3
Tr+-promoted glycosylation of cholesterol
2
with glycosyl boranophosphotriester
derivative 1
OBn
O
BnO
BnO
+
BH3
O
MeO
BnO
as an a,b-mixture, while the per-O-Bz counterpart 4 gave b-6 ste-
reoselectively, indicating that the reaction proceeded via the
P
OMe
HO
1
2
(3.0 equiv)
(1.0 equiv)
Table 3
OBn
Optimization of molar equivalent of trityl ether 5a
TrNTf2 (6.0 equiv)
solvent, 0 °C, MS3A or 4A
O
BnO
OBn
BnO
TrOChol 5
TfOH (0.57 equiv)
OBn
O
BnO
O
BnO
BnO
O
BnO
BnO
3
BH3
O
MeO
dioxane, rt, 10 min
BnO
P
OChol
BnO
Entry
Solvent
CH2Cl2
Time (min)
Yield (%)a :b)b
(a
OMe
3
1
1
2
3
40
30
30
40
60
21 (34:66)
22 (52:48)
35 (17:83)
60 (38:62)
22 (46:54)
33 (32:68)
(1.0 equiv)
Toluene–CH2Cl2 (2:1, v/v)
MeCN–CH2Cl2 (4:1, v/v)
Dioxane–CH2Cl2 (2:1, v/v)
THF–CH2Cl2 (4:3, v/v)
Entry
equiv of 5
Yield (%)b
a
:bc
4
5c
6d
1
2
3
1.2
1.5
2.0
34
45
83
80:20
87:13
83:17
Dioxane–CH2Cl2 (1:1, v/v)
120
a
b
c
Isolated yield.
Determined by 1H NMR.
Additional 8.0 equiv of TrNTf2 was added after 40 min.
Reaction was conducted at À20 °C.
a
b
c
Chol = cholesteryl.
Isolated yield.
Determined by 1H NMR.
d