Glycosyl phosphonium halide as a reactive intermediate in highly α-selective glycosylation

Article abstract of DOI:10.1246/bcsj.78.910

Highly α-selective glycosylations of several glycosyl acceptors with 2,3,4,6-tetra-O-benzyl-D-glucopyranosyl bromide (4) proceeded smoothly in the presence of tri(1-pyrrolidino)phosphine oxide in CH₂Cl₂ at room temperature or in refl

Full text of DOI:10.1246/bcsj.78.910

910 Bull. Chem. Soc. Jpn., 78, 910–916 (2005)
ꢀ 2005 The Chemical Society of Japan
Glycosyl Phosphonium Halide as a Reactive Intermediate
in Highly ꢀ-Selective Glycosylation
ꢀ;1;2
Yohei Kobashi1 and Teruaki Mukaiyama
1Center for Basic Research, The Kitasato Institute (TCI), 6-15-5 Toshima, Kita-ku, Tokyo 114-0003
2The Kitasato Institute for Life Sciences, Kitasato University, 5-9-1 Shirokane, Minato-ku, Tokyo 108-8641
Received November 29, 2004; E-mail: mukaiyam@abeam.ocn.ne.jp
Highly ꢀ-selective glycosylations of several glycosyl acceptors with 2,3,4,6-tetra-O-benzyl-D-glucopyranosyl
bromide (4) proceeded smoothly in the presence of tri(1-pyrrolidino)phosphine oxide in CH2Cl2 at room temperature
or in refluxing CHCl3 via glycosyl phosphonium bromide intermediates and the corresponding disaccharides were af-
forded in good yields with high selectivities. Further, the one-pot glycosylation starting from glycosyl acetate using io-
dotrimethylsilane and phosphine oxide also proceeded efficiently in CH2Cl2 at room temperature to give the correspond-
ing disaccharides, although this one-pot synthesis of ꢀ-glycoside via in situ anomerization procedure is considered dif-
ficult because glycosyl halide reacts with trimethylsilyl acetate easily to regenerate glycosyl acetate, a starting glycosyl
donor.
To develop a new method for the stereoselective glycosyla-
I
BnO
Me
tion is one of the most important and fundamental topics in
carbohydrate chemistry.1 1,2-cis-Glycosylation is less success-
ful than 1,2-trans-glycosylation since the substituent at C-2
that does not participate in neighboring effect, for example
O-benzyl ether, must be chosen. Secondly, SN2 reaction at
anomeric center of an unstable ꢁ-glucosyl bromide is quite
difficult.2 Lemieux and co-workers showed that the reaction
of ꢀ-pyranosyl bromide with bromide ion produced highly
reactive ꢁ-pyranosyl bromide, which in turn reacted with an
acceptor much faster than the ꢀ-one, and consequently ꢀ-gly-
coside was formed predominantly.3 This method is generally
called in situ anomerization method and is employed frequent-
ly in the formation of ꢀ-glycosyl linkage4 when galactose and
fucose are employed as donors, although it is less effective in
the case when glucose is used. The above glycosyl bromide is
generally prepared from glycosyl acetate and trialkylsilyl bro-
mide, though its isolation is often difficult.5 To isolate glycosyl
bromide is necessary for the in situ anomerization method
because an undesired reverse reaction of forming glycosyl ace-
tate takes place in the subsequent glycosylation reaction when
the above mixture is used without purification.6 Thus, the one-
pot synthesis of ꢀ-glycoside starting from glycosyl acetate
involving in situ anomerization procedure has long been con-
sidered difficult.
Recently, successful ꢀ-stereoselective glycosylation of gly-
cosyl acceptors using glycosyl diphenylphosphinite 1, a glyco-
syl donor, and iodomethane, a promoter, was reported from
our laboratory (Scheme 1).7
It is considered that this glycosylation proceeds via glycosyl
phosphonium iodide intermediates formed by the reaction of
glycosyl diphenylphosphinite and iodomethane, as was ex-
plained in the previous experiment result that phosphine oxide
promoted the reaction of glycosyl iodide and 2 and the corre-
sponding disaccharide 3 was formed in ꢀ-selective manner
O
O
BnO
BnO
ROH
OPPh2
BnO
OR
BnO
β−
Intermediate
O
MeI
BnO
BnO
= 91/9 to 99/1
α/β
BnO
OPPh2
1
O
I
Me
OPPh2
α−
Intermediate
Scheme 1. Glycosylation of glycosyl acceptor with donor 1
and iodomethane.
Table 1. Effect of Additives on Glycosylation of Acceptor 2
with Glycosyl Iodide
Additive
BnO
HO
BnO
(1.2 equiv)
MS 5A
(3 g/mmol)
O
O
O
BnO
BnO
BnO
BnO
BnO
BnO
+
O
BnO
BnO
BnO
I
OMe
O
2
CH2Cl2,
rt, 5 h
BnO
BnO
Glycosyl Iodide
(1.2 equiv)
(1.0 equiv)
BnO
OMe
3
Entry
Additive
—
n-Bu4NI/(i-Pr)2NEt
Yield/% (ꢀ=ꢁ)a)
1
2
3
47 (71:3=28:7)
42 (92:2=7:8)
86 (94:2=5:8)
Ph2P(=O)Me
a) The ꢀ=ꢁ ratios were determined by HPLC analysis.
(Table 1).7 It is also noted that this system of using phosphine
oxide was more effective to promote the glycosylation than the
n-Bu4NI/(i-Pr)2NEt system of Lemieux’s in situ anomeriza-
tion process.
In order to maximize this interesting character of a glycosyl
phosphonium halide, we studied in detail glycosylations using
glycosyl halides by the promotion of phosphine oxide. Further,
Published on the web May 6, 2005; DOI 10.1246/bcsj.78.910

Y. Kobashi et al.
Bull. Chem. Soc. Jpn., 78, No. 5 (2005)
911
Table 3. Glycosylation of 2 with Glycosyl Bromide 4 in the
Presence of Triphenylphosphine Oxide in Various Solvent
the one-pot glycosylation of several acceptors using glycosyl
acetate and iodotrimethylsilane in the presence of phosphine
oxide was studied in order to establish a more useful glyco-
sylation method.
In this paper, we would like to report on highly ꢀ-selective
glycosylation of several acceptors with glycosyl bromide in the
presence of phosphine oxide as a promoter. Further, a one-pot
glycosylation via glycosyl phosphonium iodide by using gly-
cosyl acetate, iodotrimethylsilane, and triphenylphosphine ox-
ide is also described.
BnO
O
BnO
BnO
O
Ph3P=O (1.2 equiv)
MS5A (3 g/mmol)
BnO
HO
BnO
O
O
BnO
BnO
O
BnO
BnO
+
BnO
BnO
BnO
BnO
Solvent, rt,
BnO
OMe
Br
OMe
20 h
4 (1.2 equiv)
2 (1.0 equiv)
3
Entry
Solvent
Yield/%
ꢀ=ꢁa)
1
2
3
4
5
6
7
8
9
THF
i-Pr2O
DME
CH3CN
Toluene
EtOAc
Acetone
DMF
60
44
35
69
51
52
39
69
77
72
84
92.3/7.7
95.1/4.9
95.8/4.2
87.5/12.5
92.0/8.0
90.3/9.7
89.1/10.9
81.9/18.1
84.6/15.4
92.1/7.9
94.5/5.5
Results and Discussion
Glycosylation by Using Glycosyl Bromide by the Promo-
tion of Phosphine Oxide. 2,3,4,6-Tetra-O-benzyl-D-gluco-
pyranosyl bromide (4) was selected for a glycosyl donor since
the bromide 4 was more stable than the corresponding glycosyl
iodide. It is easily prepared by the reaction of glycosyl acetate
with bromotrimethylsilane in CHCl3, followed by evaporation
of solvent and TMSOAc.5a In the first place, glycosylations of
the acceptor 2 were tried with glycosyl bromide 4 in the pres-
ence of several promoters and molecular sieves 5A8 in di-
chloromethane at room temperature (Table 2). The glycosyla-
tion was promoted effectively by using phosphine oxides such
as diphenylmethylphosphine oxide, tributylphosphine oxide,
or triphenylphosphine oxide; the desired 3 was afforded in
good yields with high stereoselectivities in all cases (Table 2,
Entry 1–4). While yields and stereoselectivities of reactions
promoted by triphenylphosphine sulfide or triphenyl phosphate
were low (Table 2, Entry 5, 6), the corresponding disaccha-
rides were obtained in a highly ꢀ-selective manner when phos-
CH3NO2
CHCl3
CH2Cl2
10
11
a) The ꢀ=ꢁ ratios were determined by HPLC analysis.
phoric acid triamide derivatives were used (Table 2, Entries
7–10). Finally, it was found that tri(1-pyrrolidino)phosphine
oxide9 was the most effective promoter: it afforded 3 in 92%
yield (Table 2, Entry 10, ꢀ=ꢁ ¼ 94:3=5:7). The stereoselectiv-
ity of 3 increased as the amount of reagent increased (Table 2,
Entry 11, ꢀ=ꢁ ¼ 95:6=4:4), although the ꢀ-selectivity re-
mained low if the reaction was performed according to
Lemieux’s procedure (Table 2, Entry 12, ꢀ=ꢁ ¼ 89:5=10:5).
Next, the effect of solvents was examined (Table 3). In
ethers such as THF, i-Pr2O, or DME, the desired 3 was pro-
duced in moderate yields with high stereoselectivities
(Table 3, Entries 1–3). The reaction in dichloromethane sol-
vent afforded 3 in high yield with high stereoselectivity
(Table 3, Entry 11), while the stereoselectivities were low
when other typical organic solvents were used (Table 3, En-
tries 4–10).
Next, glycosylations of various acceptors with glycosyl bro-
mide 4 in the presence of tri(1-pyrrolidino)phosphine oxide
were tried (Table 4). Each reaction proceeded efficiently in di-
chloromethane at room temperature and afforded the corre-
sponding disaccharides in good to high yields with high ꢀ-se-
lectivities even when an acid-labile glycosyl fluoride10 was
used as an acceptor (Table 4, Condition a). The reactions were
highly accelerated in refluxing chloroform, but their stereose-
lectivities became slightly lower (Table 4, Condition b).
Thus, the glycosylation using glycosyl bromide could be
successfully carried out by the promotion of phosphine oxide
and various ꢀ-disaccharides were afforded in high yields with
high ꢀ-selectivities.
Table 2. Glycosylations of Acceptor 2 with Glycosyl Bro-
mide 4 in the Presence of Various Reagents
BnO
O
BnO
BnO
Reagent
O
BnO
MS 5A
HO
BnO
O
(3 g/mL)
O
BnO
BnO
O
BnO
BnO
+
BnO
BnO
BnO
BnO
CH2Cl2, rt,
20 h
BnO
OMe
Br
OMe
4
2
3
Entry 1/equiv
Reagentb)
Yield/%
ꢀ=ꢁa)
1
2
3
4
5
6
7
8
1.0
1.2
1.2
1.2
1.2
1.2
1.0
1.2
Ph2P(=O)Me
Ph2P(=O)Me
Bu3P=O
Ph3P=O
Ph3P=S
(PhO)3P=O
(Me2N)3P=O
(Me2N)3P=O
72
86
86
84
10
6
93.8/6.2
94.0/6.0
93.1/6.9
94.5/5.5
79.7/20.3
75.4/24.6
94.3/5.7
93.8/6.2
74
89
O
N
P
O
9
1.2
1.2
1.5
1.2
64
92
95
84
93.0/7.0
94.3/5.7
95.6/4.4
89.5/10.5
3
N
P
O
10
11
12
Glycosylation via a Glycosyl Phosphonium Iodide Inter-
mediate Generated from Glycosyl Acetate. In order to es-
tablish a more useful glycosylation reaction, we tried a one-pot
in situ anomerization procedure starting from glycosyl acetate
in the co-existence of iodotrimethylsilane and phosphine oxide
even though the one-pot synthesis of ꢀ-glycoside via the in
situ anomerization procedure was not successful, as shown
in several reports.
3
c)
N
P
O
3
Et4NBr/(i-Pr)2NEt
a) The ꢀ=ꢁ ratios were determined by HPLC analysis. b) The
same equivalent to 4 was used. c) 3.0 equivalent of reagent
was used.

912 Bull. Chem. Soc. Jpn., 78, No. 5 (2005)
Glycosylation via Glycosyl Phosphonium Halide
Table 4. Glycosylation of Various Acceptors with Donor 4 by the Promotion of Tri(1-pyrrolidino)-
phosphine Oxide
N
P=O
3
BnO
BnO
O
O
O
(3.0 equiv)
BnO
BnO
BnO
BnO
+
HO
O
X
MS 5A (3g /mmol)
BnO
BnO
O
Br
X
Donor 4 (1.5 equiv)
Acceptor (1.0 equiv)
Disaccharide
Entry
Acceptor
Product
Conditiona)
Time/h
Yield/%
ꢀ=ꢁ
HO
O
BnO
BnO
BnO
1
2
3
a
b
a
b
a
b
a
b
a
b
20
3
95
94
91
83
83
66
88
87
88
76
95.6/4.4b)
94.7/5.3b)
97.2/2.8b)
95.6/4.4b)
98.3/1.7b)
95.8/4.2b)
95.0/5.0b)
91.5/8.5b)
95.5/4.5c)
93.3/6.7c)
OMe
2
BnO
O
BnO
HO
3
9
72
5
BnO
OMe
5
4
BnO
O
HO
5
10
11
12
96
5
BnO
BnO
OMe
6
6
HO
O
BnO
AcO
SEt
NPht
7
40
3
7
8
HO
O
BnO
BnO
F
9
24
3
BnO
8
10
a) Condition a: CH2Cl2, rt; b: CHCl3, reflux. b) The ꢀ=ꢁ ratios were determined by HPLC analysis. c)
The ꢀ=ꢁ ratios were determined by isolated yields of both isomers. d) 3.0 equiv of 4 was used.
BnO
BnO
BnO
O
O
BnO
BnO
O
BnO
BnO
BnO
BnO
O
O
BnO
BnO
O
BnO
BnO
BnO
BnO
BnO
O
O
O
BnO
BnO
O BnO
O
BnO
BnO
O
O
BnO
O
O
BnO
BnO
BnO
AcO
BnO BnO
SEt
F
BnO
BnO
OBn
OMe
BnO
OMe
BnO
BnO
PhthN
OMe
3
9
10
11
12
In the first place, glycosylations of acceptor 2 with glycosyl
iodide generated from glycosyl acetate 13 and iodotrimethyl-
silane in the presence of several promoters were tried in di-
chloromethane at room temperature (Table 5). When phos-
phoric acid triamide derivatives were used, the desired disac-
charides were obtained in moderate yields with high stereo-
selectivities (Table 5, Entries 1–3). When trialkylphosphine
oxides were used, the corresponding disaccharides were effec-
tively formed in a highly ꢀ-selective manner (Table 5, Entries
6–9), although yields and stereoselectivities of those promoted
by triphenylphosphine sulfide and phenyl diphenylphosphinite,
respectively, remained low (Table 5, Entries 4, 5). It is thus
noted that triphenylphosphine oxide excellently promoted the
glycosylation to afford the corresponding disaccharide in high
yield with high selectivity (Table 5, Entry 9). In the absence
of phosphine oxide, however, the stereoselectivity was not
controlled well (Table 5, Entry 10), and also 3 was afforded
in low yield with lower selectivity when bromotrimethylsilane,
a milder promoter than iodotrimethylsilane, was used (Table 5,
Entry 11). In the case when tetrabutylammonium iodide was
used as a promoter in the co-existence of N,N-diisopropyl-
ethylamine, a starting material, namely glycosyl acetate 13,
was recovered by the reverse reaction between glycosyl iodide
and trimethylsilyl acetate (Table 5, Entry 12). It is remarkable
to note that the glycosyl acetate formed by this reverse reaction
was not obtained if phosphine oxide was used as a promoter.
Next, the effect of the molar ratio of 13, iodotrimethylsilane
and triphenylphosphine oxide was studied (Table 6). The de-
sired disaccharide was obtained in 86% yield with high ꢀ-se-
lectivity (ꢀ=ꢁ ¼ 96:1=3:9) when 1.5 equivalents each of 13
and iodotrimethylsilane and 3.0 equivalents of triphenylphos-
phine oxide were used (Table 6, Entry 3). When an excess
amount of triphenylphosphine oxide was used, the glycosyl
phosphonium iodide intermediate was stabilized by coordina-
tion of the phosphine oxide and the yield of the disaccharide
decreased (Table 6, Entry 4). On the other hand, the glycosy-
lation was accelerated and 3 was obtained in 95% yield when
double amounts of 13, iodotrimethylsilane and triphenylphos-

Y. Kobashi et al.
Bull. Chem. Soc. Jpn., 78, No. 5 (2005)
913
Table 5. Glycosylations of Acceptor 2, with Glycosyl Ace-
tate 13 via Glycosyl Phosphonium Iodide Using Various
Reagents
of triphenylphosphine oxide were needed for completion of the
reactions (Table 7, Entries 2–5, 8, 9).
Thus, the glycosylations using glycosyl phosphonium iodide
generated from the corresponding glycosyl acetate, i.e., an
easily accessible and manageable glycosyl donor, proceeded
successfully in CH2Cl2 at room temperature. In this reaction,
the glycosyl acetate regenerated by the reaction of glycosyl
halide and trimethylsilyl acetate formed by the one-pot in situ
anomerization procedure was not obtained. The result causes
the present one-pot glycosylation to proceed effectively.
Reagent
(3.0 equiv)
BnO
O
BnO
BnO
TMSI
HO
BnO
BnO
O
O
BnO
(1.5 equiv)
MS 5A
(3.0 g/mmol)
BnO
BnO
O
BnO
BnO
OMe
O
BnO
BnO
2
OAc
BnO
BnO
OMe
CH2Cl2, rt,
7 h
13
CH2Cl2, 0 °C,
30 min
3
(1.5 equiv)
Entry
Reagent
Yield/%
ꢀ=ꢁa)
Mechanism
N
P
3
O
1
2
3
61
69
61
94.1/5.9
94.3/5.7
95.9/4.1
The glycosylation of 2 with 13 using Ph3P=O–TMSI adduct
was tried in order to show the reason why the reverse reaction
did not take place in this one-pot in situ anomerization proce-
dure.11 Interestingly, the desired 3 was obtained in 54% yield
along with the recovery of 13 (Table 8, Entry 1). On the other
hand, the disaccharide 3 was not obtained in the case when n-
Bu4NI/i-Pr2NEt system was employed in place of Ph3P=O in
the above-mentioned reaction (Table 8, Entry 2). This result
indicates that Ph3P=O–TMSI adduct is reactive enough to pro-
mote the formation of glycosyl phosphonium iodide from the
glycosyl acetate, although the acetate is formed by the reverse
reaction in general.
The glycosyl phosphonium iodide was not detected by
1H NMR study of the reaction mixture of glycosyl iodide
and triphenylphosphine oxide in CDCl3. This indicates that
very small amounts of glycosyl phosphonium iodide existed
in equilibrium with a mixture of glycosyl iodide and phosphine
oxide. The proposed reaction mechanism is illustrated below
(Scheme 2). The reactive intermediate, glycosyl phosphonium
iodide 20, is formed from the corresponding glycosyl iodide 19
and phosphine oxide. A glycosyl acceptor in turn reacts dom-
inantly with highly-reactive ꢀ-20 to afford ꢁ-21 while less-re-
active ꢁ-20 epimerizes to the above more reactive ꢀ-20, which
also affords ꢁ-21. There is an alternative pathway to convert
glycosyl acetate 18 directly into glycosyl phosphonium iodide
20 by the promotion of Ph3P=O–TMSI adduct, though the
glycosyl iodide 19 is generally converted to 18 by the reverse
reaction. In addition, hydrogen iodide, a co-product, is rapidly
neutralized by phosphine oxide to form a Ph3P=O–HI adduct
(22);12 therefore, the present glycosylation proceeds smoothly
under almost neutral condition.
(Me2N)3P=O
O
N
P O
3
4
5
6
7
Ph3P=S
58
56
72
73
71.2/22.8
95.2/4.1
95.0/5.0
95.4/4.6
Ph2P(=O)OPh
n-Bu3P=O
Ph2P(=O)Me
MeO
P
O
3
8
69
94.8/5.2
9
Ph3P=O
—
Ph3P=Ob)
n-Bu4NI/(i-Pr)2NEt
89
59
8
96.1/3.9
68.4/31.6
93.1/6.9
91.6/8.4
10
11
12
19
a) The ꢀ=ꢁ ratios were determined by HPLC analysis. b)
TMSBr was used instead of TMSI.
Table 6. Optimization of the Molar Ratios of 13, TMSI, and
Ph3P=O
BnO
Ph3P=O
O
BnO
BnO
HO
O
TMSI
MS 5A
(3.0 g/mmol)
BnO
O
BnO
BnO
BnO
O
BnO
BnO
BnO
OMe
O
BnO
BnO
2
OAc
BnO
BnO
OMe
CH2Cl2, rt,
7 h
13
CH2Cl2, 0 °C,
30 min
3
(1.5 equiv)
Entry TMSI/equiv Ph3P=O/equiv Yield/%
ꢀ=ꢁa)
1
2
3
4
1.5
1.5
1.5
1.5
3.0
3.0
0.1
1.5
3.0
6.0
3.0
6.0
51
63
86
73
82
95
86.9/13.1
95.0/5.0
96.1/3.9
96.0/4.0
94.6/5.4
94.5/5.5
5b)
6b)
Conclusion
The glycosylations of various acceptors with glycosyl hal-
ides such as glycosyl bromide and glycosyl iodide are effec-
tively promoted by phosphine oxide to afford the correspond-
ing ꢀ-disaccharides in good yields with high selectivities. In
the case of one-pot glycosylation using glycosyl acetate and
iodotrimethylsilane, the glycosyl acetate is converted directly
into glycosyl phosphonium iodide when Ph3P=O–TMSI
adduct, a promoter to activate glycosyl acetate, is generated,
and consequently, the present one-pot glycosylation proceeded
effectively. This method is considered practical in the synthe-
ses of complex oligosaccharides because the glycosyl acetate,
the glycosyl donor, is easy to access and handle, and the reac-
tion always gives ꢀ-disaccharide with quite high selectivity.
a) The ꢀ=ꢁ ratios were determined by HPLC analysis. b) 3.0
equiv of 13 was used.
phine oxide were used (Table 6, Entry 6).
Next, glycosylations of the acceptors with glucosyl acetate
13 or galactosyl acetate 14 using iodotrimethylsilane and tri-
phenylphosphine oxide were tried (Table 7). Each reaction
proceeded efficiently in dichloromethane at room temperature
and afforded the corresponding disaccharides in high yields
with high ꢀ-selectivities, as expected. In the case of using less
reactive acceptors such as 5 and 6, however, 3.0 equivalents
each of the donor and iodotrimethylsilane and 6.0 equivalents

914 Bull. Chem. Soc. Jpn., 78, No. 5 (2005)
Glycosylation via Glycosyl Phosphonium Halide
Table 7. Glycosylation of Various Acceptors with Donor 13 or 14 via Glycosyl Phosphonium Iodide Using
TMSI and Ph3P=O
R1
Ph3P=O
TMSI
MS 5A
OBn
O
O
O
HO
R2
BnO
X
O
(3.0 g/mmol)
Acceptor
OAc
O
BnO
X
13: R1 = H, R2 = OBn
14: R1 = OBn, R2 = H
CH2Cl2, rt,
CH2Cl2, 0 °C
30 min
Disaccharide
Entry
1
Donor
Acceptor
Product
Conditiona)
a
Time/h
7
Yield/%
89
ꢀ=ꢁb)
HO
O
BnO
BnO
BnO
13
3
96.1/3.9
OMe
2
BnO
O
BnO
HO
2
3
4
5
6
a
b
a
b
a
24
24
24
34
21
72
91
52
80
85
98.3/1.7
97.4/2.6
98.7/1.3
98.9/1.1
94.9/5.1
9
BnO
OMe
OMe
5
BnO
O
HO
BnO
10
11
BnO
6
7
HO
O
BnO
AcO
SEt
NPht
7
8
9
14
2
5
6
15
16
17
a
a
b
7
34
34
93
94
91
96.6/3.4
93.1/6.9
96.3/3.7
a) Condition a: 13 or 14 (1.5 equiv), Ph3P=O (3.0 equiv), TMSI (1.5 equiv); b: 13 or 14 (3.0 equiv),
Ph3P=O (6.0 equiv), TMSI (3.0 equiv). b) The ꢀ=ꢁ ratios were determined by HPLC analysis.
BnO
BnO
O
BnO
BnO
O
BnO
BnO
O
BnO
BnO
O
BnO
BnO
BnO
BnO
O
O
O
BnO
BnO
O BnO
O
BnO
BnO
O
BnO
BnO
O
O
BnO
AcO
BnO BnO
SEt
BnO
OBn
OMe
BnO
BnO
OMe
OMe
PhthN
3
9
10
11
BnO
OBn
O
BnO
BnO
OBn
O
OBn
BnO
BnO
BnO
O BnO
OBn
O
O
O
O
BnO
O
O
BnO
BnO
BnO
BnO BnO
BnO
OMe
OBn
BnO
OMe
BnO
OMe
15
16
17
Table 8. Glycosylation of 2 with 13 Promoted by Ph3P=O–
TMSI Adducta)
Experimental
General. All melting points were measured on a Yanaco MP-
S3 micro melting point apparatus. Infrared spectra were recorded
BnO
O
BnO
BnO
HO
O
O
BnO
1
BnO
BnO
Reagent,
TMSI,
MS 5A
on a Perkin Elmer Spectrum One infrared spectrometer. H NMR
BnO
BnO
O
BnO
BnO
OMe
O
spectra were recorded on a JEOL JNM-EX270L (270 MHz) or
JEOL JNM-LA500 (500 MHz) spectrometer; chemical shifts (ꢂ)
are reported in parts per million relative to tetramethylsilane.
Splitting patterns are designated as s, singlet; d, doublet; t, triplet;
q, quartet; m, multiplet; br, broad. 13C NMR spectra were recorded
on a JEOL JNM EX270L (68 MHz) or a JEOL JNM-LA500 (125
MHz) spectrometer with complete proton decoupling. Chemical
shifts are reported in parts per million relative to tetramethyl-
silane, with the solvent resonance as the internal standard (CDCl3;
ꢂ 77.0 ppm). 31P NMR spectra were recorded on a JEOL JNM
EX270L (108 MHz) spectrometer with complete proton decou-
pling. Chemical shifts are reported in parts per million relative
BnO
BnO
2
OAc
BnO
BnO
OMe
CH2Cl2, rt,
7 h
CH2Cl2, 0 °C,
30 min
13
3
Entry
Reagent
Yield/%
ꢀ=ꢁb)
1
2
n-Bu4NI/i-Pr2NEt
Ph3P=O
N.R.
54c)
—
94.9/5.1
a) Reaction condition: 13 (1.5 equiv), Reagent (3.0 equiv),
TMSI (1.5 equiv), MS 5A (3.0 g/mL). b) The ꢀ=ꢁ ratios were
determined by HPLC analysis. c) 13 was recovered in 46%
yield.

Y. Kobashi et al.
Bull. Chem. Soc. Jpn., 78, No. 5 (2005)
915
I−
I−
Ph3P=O
ROH
O
O
O
OPPh3
Ph3P OTMS
OR
TMSI
I
β−20
(more reactive)
19
22
α−
21
O
TMSOAc
(favor)
OAc
18
Ph3P=O−TMSI
I−
I−
(slow)
O
O
ROH
OR
Ph3P OTMS
OPPh3
22
β−21
(disfavor)
α−
20
(less reactive)
Scheme 2. Reaction mechanism of glycosylation with glycosyl acetate by the promotion of TMSI and Ph3P=O via glycosyl phos-
phonium iodide.
to trimethylphosphite in CDCl3 as the external standard (ꢂ 140.0
ppm). High-resolution mass spectra were recorded on a Micro-
mass Q-TOF2 instrument [ESI positive, 0.01 M (1 M = 1
mol dmꢁ3) AcONH4 in H2O/MeCN]. High-performance liquid
chromatography (HPLC) was carried out using a Hitachi LC-
Organizer, L-4000 UV Detector, L-6200 Intelligent Pump, and
D-2500 Chromato-Integrator. Optical rotations were recorded on
a Jasco-P-1020 polarimeter. Analytical TLC was done on precoat-
ed (0.25 mm) silica gel 60 F254 plates (E. Merck). Thin-layer chro-
matography was performed on Wakogel B-5F. Column chroma-
tography was performed on Silica gel 60 (Merck).
cosylation reaction was confirmed by monitoring TLC, diluted
with EtOAc, filtered through Celite, and evaporated. The resulting
residue was purified by preparative TLC (silica gel) to give the
corresponding disaccharide. The ꢀ to ꢁ ratios were determined
by HPLC analysis.
Methyl
2,3,4-Tri-O-benzyl-6-O-(2,3,4,6-tetra-O-benzyl-D-
glucopyranosyl)-ꢀ-D-glucopyranoside (3): The ꢀ=ꢁ ratio was
determined by HPLC analysis13 (Shodex SIL-5B, ꢃ4:6 ꢃ 250
mm2; n-hexane/ethyl acetate = 4/1; 1.0 mL/min; 254 nm; ꢀ-iso-
mer, 14.3 min; ꢁ-isomer, 12.5 min).
Methyl
2,4,6-Tri-O-benzyl-3-O-(2,3,4,6-tetra-O-benzyl-D-
All reactions were carried out under argon atmosphere in dried
glassware, unless otherwise noted. All reagents were purchased
from Tokyo Kasei Kogyo, Kokusan Chemical, Kanto Chemical,
Aldrich, or Fluka and were used without further purification, un-
less otherwise noted. CH2Cl2 was distilled from P2O5 and then
from CaH2 and was stored over molecular sieves 4A. Toluene
and CH3CN were distilled from P2O5 and stored over molecular
sieves 4A. DMF was distilled from CaH2 under reduced pressure
(pre-dried over P2O5) and stored over molecular sieves 4A. Dehy-
drated DME, IPE, CHCl3, THF, and Et2O were purchased from
Kanto Chemical. Dehydrated EtOAc and acetone were purchased
from Kokusan Chemical. Powdered and pre-dried (at 200 ꢂC/133
Pa, 6 h) molecular sieves 5A (MS 5A) was used in glycosylation
reactions.
Glycosylation Using Glycosyl Bromide and Phosphine Ox-
ide (General Procedure). To a stirred suspension of MS 5A
(240 mg), 4 (0.12 mmol), and glycosyl acceptor (0.08 mmol) in
CH2Cl2 (0.8 mL) was added tri(1-pyrrolidino)phosphine oxide
(0.24 mmol) at room temperature. After the completion of glyco-
sylation reaction was confirmed by monitoring TLC, the reaction
mixture was diluted with EtOAc and filtered through Celite. After
having been dried over MgSO4, filtered, and evaporated, the re-
sulting residue was purified by preparative TLC (silica gel) to give
the corresponding disaccharide. The ꢀ to ꢁ ratios for 3, 9, 10, and
11 were determined by HPLC analysis and the ratio for 16 was de-
termined by isolation of both isomers according to published con-
ditions.
glucopyranosyl)-ꢀ-D-glucopyranoside (9): The ꢀ=ꢁ ratio was
determined by HPLC analysis13 (Shodex SIL-5B, ꢃ4:6 ꢃ 250
mm2; n-hexane/ethyl acetate = 4/1; 1.0 mL/min; 254 nm; ꢀ-iso-
mer, 10.4 min; ꢁ-isomer, 11.9 min).
Methyl
2,3,6-Tri-O-benzyl-4-O-(2,3,4,6-tetra-O-benzyl-D-
glucopyranosyl)-ꢀ-D-glucopyranoside (10):
The ꢀ=ꢁ ratio
was determined by HPLC analysis13 (YMC J’sphere ODS M80,
ꢃ4:6 ꢃ 250 mm2; MeOH=H2O ¼ 4=1; 1.0 mL/min; 254 nm; ꢀ-
isomer, 14.6 min; ꢁ-isomer, 16.1 min).
Ethyl 3-O-Acetyl-4-O-benzyl-6-O-(2,3,4,6-tetra-O-benzyl-D-
glucopyranosyl)-2-deoxy-2-phthalimido-1-thio-ꢁ-D-glucopyra-
noside (11): The ꢀ=ꢁ ratio was determined by HPLC analysis13
(Shodex SIL-5B, ꢃ4:6 ꢃ 250 mm2; n-hexane/ethyl acetate = 4/1;
1.0 mL/min; 254 nm; ꢀ-isomer, 21.8 min; ꢁ-isomer, 19.4 min).
2,3,4-Tri-O-benzyl-6-O-(2,3,4,6-tetra-O-benzyl-ꢀ-D-gluco-
pyranosyl)-ꢁ-D-glucopyranosyl fluoride (12): The ꢀ=ꢁ ratio
was determined by isolated yields of both isomers.14
Methyl
2,3,4-Tri-O-benzyl-6-O-(2,3,4,6-tetra-O-benzyl-D-
galactopyranosyl)-ꢀ-D-glucopyranoside (15): The ꢀ=ꢁ ratio
was determined by HPLC analysis13 (Shodex SIL-5B, ꢃ4:6 ꢃ
250 mm2; n-hexane/ethyl acetate = 4/1; 1.0 mL/min; 254 nm;
ꢀ-isomer, 11.0 min; ꢁ-isomer, 12.8 min).
Methyl
2,4,6-Tri-O-benzyl-3-O-(2,3,4,6-tetra-O-benzyl-D-
galactopyranosyl)-ꢀ-D-glucopyranoside (16): The ꢀ=ꢁ ratio
was determined by HPLC analysis15 (GL Sciences Intersil SIL,
ꢃ4:6 ꢃ 250 mm2; n-hexane/ethyl acetate = 4/1; 1.0 mL/min;
254 nm; ꢀ-isomer, 8.3 min; ꢁ-isomer, 9.6 min).
20
Glycosylation Using Glycosyl Acetate, Iodotrimethylsilane,
and Phosphine Oxide (General Procedure). To a stirred sus-
pension of MS 5A (240 mg) and glycosyl acetate (0.12 mmol)
in CH2Cl2 (1.2 mL) was added iodotrimethylsilane (0.12 mmol)
16ꢁ: Colorless oil; ½ꢀꢄD þ53:3 (c 1.00, CHCl3); IR (Neat)
2907, 2867, 1606, 1586, 1496, 1454, 1363, 1208, 1156, 1133,
1096, 1055, 1028, 912, 735, 697 cmꢁ1 1H NMR (500 MHz,
;
CDCl3) ꢂ 3.23 (s, 3H), 3.50–3.76 (m, 6H), 3.95–3.97 (m, 1H),
4.01 (dd, J ¼ 10:2, 2.6 Hz, 1H), 4.07 (dd, J ¼ 10:2, 3.7 Hz,
1H), 4.24 (dd, J ¼ 9:5, 8.2 Hz, 1H), 4.33 (d, J ¼ 11:3 Hz, 1H),
4.35 (d, J ¼ 11:6 Hz, 1H), 4.39 (d, J ¼ 11:6 Hz, 1H), 4.43 (d, J ¼
ꢂ
at 0 C. After stirring for 30 min, Ph3P=O (0.24 mmol) and gly-
cosyl acceptor (0.08 mmol) were added in turn. The reaction mix-
ture was stirred at room temperature until the completion of gly-

916 Bull. Chem. Soc. Jpn., 78, No. 5 (2005)
Glycosylation via Glycosyl Phosphonium Halide
11:9 Hz, 1H), 4.44–4.51 (m, 1H), 4.48 (d, J ¼ 4:0 Hz, 1H, H-1),
4.50 (d, J ¼ 4:6 Hz, 1H), 4.57, (d, J ¼ 11:6 Hz, 1H), 4.58 (d, J ¼
11:9 Hz, 1H), 4.64, (d, J ¼ 11:9 Hz, 1H), 4.68–4.76 (m, 4H), 4.90,
(d, J ¼ 11:6 Hz, 1H), 4.94 (d, J ¼ 11:2 Hz, 1H), 5.59 (d, J ¼ 3:7
Hz, 1H, H-10), 6.98–7.01 (m, 2H), 7.11–7.33 (m, 33H); 13C NMR
(68 MHz, CDCl3) ꢂ 54.92, 68.50, 68.85, 69.02, 69.49, 72.61,
73.11, 73.34, 73.49, 73.70, 73.79, 74.85, 75.20, 76.03, 76.29,
78.57, 79.12, 79.29, 97.76 (C-1), 97.85 (C-10), 127.14, 127.26,
127.33, 127.44, 127.48, 127.66, 127.89, 127.91, 128.07, 128.12,
128.17, 128.22, 128.24, 128.26, 128.33, 128.36, 128.39, 137.92,
138.29, 138.34, 138.38, 138.43, 138.72, 138.82; HRMS m=z calcd
for C62H70NO11 [M þ NH4]þ 1004.4943, found 1004.4948.
etc. (2000), Part 1. h) K. J. Jensen, J. Chem. Soc., Perkin Trans.
1, 2002, 2219. i) A. V. Demchenko, Synlett, 2003, 1225.
2
T. K. Lindhorst, ‘‘Essentials of Carbohydrate Chemistry
and Biochemistry,’’ WILEY-VCH (2003).
R. U. Lemieux, K. B. Hendriks, R. V. Stick, and K. James,
J. Am. Chem. Soc., 97, 4056 (1975).
a) S. N. Lam and J. Gervay, Carbohydr. Res., 337, 1953
3
4
(2002). b) A. K. Sarkar, A. K. Ray, and N. Roy, Carbohydr.
Res., 190, 181 (1989). c) A. Chernyak, S. Oscarson, and D. Turek,
Carbohydr. Res., 329, 309 (2000).
5
a) J. W. Gillard and M. Israel, Tetrahedron Lett., 22, 513
(1981). b) T. Ishikawa and H. G. Fletcher, Jr., J. Org. Chem.,
34, 536 (1969).
Methyl
2,3,6-Tri-O-benzyl-4-O-(2,3,4,6-tetra-O-benzyl-D-
galactopyranosyl)-ꢀ-D-glucopyranoside (17): The ꢀ=ꢁ ratio
was determined by HPLC analysis13 (DAICEL Chiralcel OD-H,
ꢃ4:6 ꢃ 250 mm2; n-hexane/2-propanol = 10/1; 1.0 mL/min;
254 nm; ꢀ-isomer, 13.4 min; ꢁ-isomer, 25.5 min).
6
7
M. Hadd and J. Gervay, Carbohydr. Res., 320, 61 (1999).
T. Mukaiyama, Y. Kobashi, and T. Shintou, Chem. Lett.,
32, 900 (2003).
8
Molecular sieves 5A improved the yields of glycosylations
Mechanistic Study of the Glycosylation with Glycosyl Ace-
tate Promoted by Ph3P=O–TMSI Adduct: To a stirred sus-
pension of MS 5A (240 mg), Ph3P=O (0.67 g, 0.24 mmol), and
iodotrimethylsilane (17.1 mL, 0.12 mmol) in CH2Cl2 (1.2 mL)
was added 13 (0.070 g, 0.12 mmol) at 0 ꢂC. After stirring for
30 min, 2 (0.037 g, 0.08 mmol) was added and the resulting mix-
ture was stirred for 7 h at room temperature, diluted with EtOAc,
filtered through Celite and evaporated. The resulting residue was
purified by preparative TLC (silica gel) to give 3 (0.042 g,
0.043 mmol, 54%, ꢀ=ꢁ ¼ 94:9=5:1).
using glycosyl phosphinite and iodomethane: see Ref. 7. Recent
examples of glycosylation using molecular sieves: a) G. H. Posner
and D. S. Bull, Tetrahedron Lett., 37, 6279 (1996). b) K. Toshima,
K. Kasumi, and S. Matsumura, Synlett, 1998, 643. c) H. Jona, K.
Takeuchi, and T. Mukaiyama, Chem. Lett., 2000, 1278. d) K.
Toshima, H. Nagai, K. Kasumi, K. Kawahara, and S. Matsumura,
Tetrahedron, 60, 5331 (2004).
9
It is known to be a polar aprotic solvent with high electron-
donating power: Y. Ozari and J. Jagur-Grodzinski, J. Chem. Soc.,
Chem. Commun., 1974, 295.
10 Reviews on glycosyl fluoride: a) M. Shimizu, H. Togo, and
M. Yokoyama, Synthesis, 1998, 799. b) M. Yokoyama, Carbo-
hydr. Res., 327, 5 (2000). c) K. Toshima, Carbohydr. Res., 327,
15 (2000). d) T. Mukaiyama and H. Jona, Proc. Jpn. Acad., Ser.
B, 78, 73 (2002).
11 The confirmation of Ph3P=O–TMSI adduct was supposed
based on the measurement of its 31P NMR in CD2Cl2 (ꢂ 41.9 ppm,
cf. 31P shift of Ph3P=O was ꢂ 29.0 ppm).
This study was supported in part by a Grant of the 21st Cen-
tury COE Program from the Ministry of Education, Culture,
Sports, Science, and Technology (MEXT). We thank Mr. Akira
Higuchi, Taisho Pharmaceutical Company Ltd., for his kind
help concerning High Resolution Mass Spectrometry analyses.
References
12 L. D. Quin, ‘‘A Guide to Organophosphorous Chemistry,’’
John Wiley and Sons (2000), p. 103.
1
Reviews on glycosylations: a) H. Paulsen, Angew. Chem.,
Int. Ed. Engl., 21, 155 (1982). b) R. R. Schmidt, Angew. Chem.,
Int. Ed. Engl., 25, 212 (1986). c) H. Kuntz, Angew. Chem., Int.
13 H. Jona, H. Mandai, W. Chavasiri, K. Takeuchi, and T.
Mukaiyama, Bull. Chem. Soc. Jpn., 75, 291 (2002).
14 H. Chiba, S. Funasaka, and T. Mukaiyama, Bull. Chem.
Soc. Jpn., 76, 1629 (2003).
15 Analytical date of 16ꢀ has been reported from our group:
T. Mukaiyama, K. Takeuchi, H. Jona, H. Maeshima, and T.
Saitoh, Helv. Chim. Acta, 83, 1901 (2000).
Ed. Engl., 26, 294 (1987). d) P. Sinay, Pure Appl. Chem., 63,
¨
519 (1991). e) K. Suzuki and T. Nagasawa, J. Synth. Org. Chem.,
Jpn., 50, 378 (1992). f) K. Toshima and K. Tatsuta, Chem. Rev.,
93, 1503 (1993). g) B. Ernst, G. W. Hart, and P. Sinay, ‘‘Carbo-
¨
hydrate in Chemistry and Biology,’’ WILEY-VCH, Weinheim