Mechanism of the reductive cleavage reaction of permethylated methyl D-glycopyranosides

Article abstract of DOI:10.1016/S0008-6215(99)00157-3

The mechanism of the reductive cleavage reaction of permethylated methyl D-glycopyranosides was investigated by measuring the rate of reaction. Glycosides employed were of α-Glc, β-Glc, α-Man, α-Gal, and β-Gal. Seven silanes were used to explore the reactivities of the reducing agents as well as to examine the stereoelectronic effects of the agents. Trimethylsilyl trifluoromethanesulfonate was employed as catalyst. In general, the rates of β anomers were about twice as fast as those of the α anomers. The rates of anomerization were about five to ten times lower than those of reduction. A cyclic oxonium ion has been proposed as a sole intermediate for the reductive cleavage of the α-glycoside linkage, but the attack of the reducing agent on both cyclic and acyclic forms as well as on the substrate-Lewis acid complex seems to be involved for the β anomer. Copyright (C) 1999 Elsevier Science Ltd.

Full text of DOI:10.1016/S0008-6215(99)00157-3

www.elsevier.nl/locate/carres
Carbohydrate Research 320 (1999) 223–229
Mechanism of the reductive cleavage reaction of
permethylated methyl
D-glycopyranosides
Chang Kiu Lee *, Eun Ju Kim
Department of Chemistry, Kangwon National Uni6ersity, Chuncheon 200-701, South Korea
Received 24 February 1999; accepted 8 June 1999
Abstract
The mechanism of the reductive cleavage reaction of permethylated methyl
D
-glycopyranosides was investigated by
measuring the rate of reaction. Glycosides employed were of a-Glc, b-Glc, a-Man, a-Gal, and b-Gal. Seven silanes
were used to explore the reactivities of the reducing agents as well as to examine the stereoelectronic effects of the
agents. Trimethylsilyl trifluoromethanesulfonate was employed as catalyst. In general, the rates of b anomers were
about twice as fast as those of the a anomers. The rates of anomerization were about five to ten times lower than
those of reduction. A cyclic oxonium ion has been proposed as a sole intermediate for the reductive cleavage of the
a-glycoside linkage, but the attack of the reducing agent on both cyclic and acyclic forms as well as on the
substrate–Lewis acid complex seems to be involved for the b anomer. © 1999 Elsevier Science Ltd. All rights
reserved.
Keywords: Reductive cleavage of glycoside linkages; Anomerization; Cyclic oxonium ion; Acyclic oxonium ion
partially methylated and acylated 1,5- or 1,4-
1. Introduction
anhydroalditol products present in the final
reaction mixture. Any knowledge related to
the mechanism of the reductive cleavage step
should therefore be very useful for under-
standing the nature of the reaction, and
providing better reaction conditions that
could make the method more applicable to
various types of substrates.
Although a mechanistic aspect of the reac-
tion was not investigated in depth, a cyclic
oxonium ion was suggested to be formed dur-
ing the course of the reaction, based on the
stereochemistry of the 1-monodeuterio-1,5-an-
hydroglucitol produced upon reduction with
Et3SiD [2]. The stereochemistry of the deu-
terio-1,5-anhydroalditol products seems to
support this rationale, because the predomi-
nant (\90%) axial configuration of the deu-
terium atom in the products could arise from
The reductive cleavage method is one of the
important tools in determining the structure
of polysaccharides [1]. Various aspects of the
method have been studied extensively [2–13].
So far, the investigation has concentrated
mostly on the variation of polysaccharides
and catalysts while triethylsilane and
dichloromethane are employed as the reducing
agent and solvent, respectively.
The method consists of exhaustive methyla-
tion of the free OH groups in a polysaccha-
ride, reductively cleaving all glycosidic
linkages with Et3SiH in the presence of Lewis
acid catalyst, acylating the free OH group
generated by the cleavage, and identifying the
* Corresponding author. Fax: +82-361-253-7582.
E-mail address: cklee410@cc.kangwon.ac.kr (C.K. Lee)
0008-6215/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(99)00157-3

224
C.K. Lee, E.J. Kim / Carbohydrate Research 320 (1999) 223–229
of acetals or ketals [17]. Needless to say, gly-
cosides are acetals and formation of an oxo-
nium ion in the course of both reactions may
imply mechanistic similarities. An A-2 mecha-
nism has been suggested for hydrolysis of
methyl (or phenyl)
According to this mechanism for the hydroly-
sis of -glucopyranosides, it is apparent that
D
-glycopyranosides [18].
D
the pyranoside ring does not open during the
course of reaction. It is also important to
point out that attack by a water molecule on
the oxonium ion takes place in the rate-deter-
mining step. However, this step is a fast one in
the hydrolysis of other acetals in general.
Kaczmarek et al. studied the acetolysis of
ethyl 2,3,4,6-tetra-O-acetyl-a and -b- -glu-
D
copyranosides in 1:1 (v/v) acetic anhydride–
acetic acid containing 1% of sulfuric acid [16].
They concluded that anomerization of the a
to b anomer proceeds 30 times, and anomer-
ization of the b to a anomer 300 times, as fast
as the acetolysis of the substrate to form the
anomeric glucose pentaacetates. Based on ki-
netic analysis, they ruled out the involvement
of a cyclic oxonium ion intermediate for the
process of anomerization, and instead pro-
posed an acyclic oxonium ion to be responsi-
ble for anomerization. On the other hand, the
former intermediate is the key intermediate
for acetolysis, which is an intermolecular
process.
Scheme 1.
the cyclic oxonium intermediate. A cyclic oxo-
nium ion has also been proposed as an inter-
mediate
for
the
anomerization
-glycosides, along with
of
permethylated methyl
D
It is reasonable to assume that an oxonium
ion, either cyclic or acyclic, may be involved
in the course of reductive cleavage because the
reducing agent itself is anionic by nature.
However, the most critical problem to be
solved is whether the rate-determining step is
the formation of an oxonium ion from the
substrate or the reduction of the ion by a
silane. This means that the factors which may
influence the observed rates should be: (1) the
structural characteristics of the substrates, es-
pecially the configuration of the anomeric car-
bons; (2) the nature of the reducing agents;
and (3) the nature of the catalysts.
an acyclic oxonium ion [14,15]. It has also
been suggested that a similar cyclic oxonium
ion is involved in the process of transglycosy-
lation and acetolysis [16].
We report here the results of our studies on
the rates of reaction for the reductive cleavage
of several permethylated methyl
nosides (Scheme 1).
D
-glycopyra-
The rate of disappearance of the substrate
should be the sum of the rate of the anomer-
ization and that of formation of the product.
Anomerization is known to be a reversible
process and the calculation of k2 would be
complicated if the product is formed from
both of the oxonium intermediates.
2. Results and discussion
The reductive cleavage of methyl glycosides
has some analogy with the hydrolysis reaction

C.K. Lee, E.J. Kim / Carbohydrate Research 320 (1999) 223–229
225
Table 1
The mole percent of substrate (1a), its anomer (1b) and 1,5-anhydroglucitol (5) present in the reaction mixture of 1a, triethylsilane,
and Me3SiOSO2CF3 in CH2Cl2 after 0.5, 1, and 24 h, as determined by GLC
0.5 h
1 h
24 h
Silane
1a
1b
5
1a
1b
5
a/b
1a
1b
5
4a
4b
4c
4d
4e
4f
71.3
96.7
75.7
86.9
72.3
80.8
63.2
2.6
1.4
3.8
4.6
2.3
1.6
1.6
26.1
1.9
20.5
8.5
25.4
17.6
35.2
57.0
93.6
62.0
76.8
54.8
75.5
43.6
2.0
2.2
5.2
5.2
3.3
1.7
1.5
41.0
4.2
32.8
18.0
41.9
22.8
54.9
28.5
42.5
11.9
14.8
16.6
44.4
29.1
1.2
34.4
9.8
26.5
0.0
41.7
3.8
2.2
2.2
1.8
3.9
0.0
1.4
1.0
96.6
63.4
88.4
69.6
100.0
56.9
95.2
4g
Table 2
The mole percent of substrate (1b), its anomer (1a) and 1,5-anhydroglucitol (5) present in the reaction mixture of 1b, triethylsilane,
and Me3SiOSO2CF3 in CH2Cl2 after 0.5, 1, and 24 h, as determined by GLC
0.5 h
1 h
24 h
Silane
1a
1b
5
1a
1b
5
a/b
1a
1b
5
4a
4b
4c
4d
4e
4f
9.8
23.8
18.1
27.0
16.0
6.0
46.8
73.6
40.1
38.0
23.5
63.0
44.5
43.4
2.6
41.8
35.0
60.5
31.0
34.3
12.4
37.1
23.0
27.0
14.1
5.6
28.1
57.6
24.0
27.6
10.7
62.0
27.8
59.5
5.3
53.0
45.4
75.2
32.4
43.0
0.44
0.64
0.96
1.00
1.32
0.09
0.98
1.4
11.8
0.5
16.0
0.0
6.9
7.3
0.1
2.9
0.2
3.2
0.1
7.2
0.0
98.5
85.2
99.3
80.7
100
85.9
92.7
4g
21.2
27.2
for the a anomer than those of the b anomer.
Furthermore, the reduction takes place at a
far greater rate than the anomerization, re-
gardless of the reducing agent. These results
are consistent with the process proceeding via
a cyclic oxonium ion intermediate. The silane,
which is present in five-fold excess, should be
a better nucleophile than methoxytrimethyl-
silane. Consequently, the rate of formation of
the product is much greater than that of
anomerization.
On the other hand, the b substrate forms a
substantial amount of a anomer, in addition
to 5, in 1 h. It seems to favor the acyclic
oxonium pathway because the intermediate
can readily recyclize intramolecularly to the a
anomer. Then the question remains as to
whether the cyclic oxonium ion is the sole
intermediate for reduction of both a and b
anomers. This may be true if a weak reducing
agent, such as triethoxysilane (4b) is em-
ployed, which gives only 2.6 and 5.3% of 5 in
0.5 and 1 h, respectively (Table 2). However, it
Our investigation of the rate of anomeriza-
tion of permethylated methyl -glycosides in-
D
dicates that the rate of anomerization of b to
a is about four times faster than that of a to
b for methyl -glucopyranosides [14]. Further-
D
more, the former process seems to take place
via an acyclic oxonium ion, whereas the latter
favors a cyclic oxonium ion [15]. In the pres-
ence of a reducing agent both ions may com-
pete with reduction and anomerization. The
cyclic oxonium ion, if formed, should compete
with methoxytrimethylsilane and the reducing
agent. Since the reducing agent is used at a
five or ten times molar excess, it is probable
that the rate of reduction is much higher than
that of anomerization.
Tables 1 and 2 show clearly that both
anomerization and reduction are much slower

226
C.K. Lee, E.J. Kim / Carbohydrate Research 320 (1999) 223–229
when Et3SiD was used for the reductive cleav-
age of 1b [2]. Also, this is why the reduction
takes place much faster than the anomeriza-
tion. The acyclic oxonium ion II may undergo
mostly intramolecular recyclization to form
the a anomer.
should be pointed out that the rate of forma-
tion of the product from the b anomer is
about twice as high as that from the a
anomer. If the cyclic oxonium ion is the sole
intermediate, the rates should be very similar
for both the a and b anomers.
On the other hand, in the case of the a
anomer, the complexation to O-1 and O-2 is
plausible (such as III) and the resulting anti
arrangement of the lone-pair orbital in the
O-5 and 1-CꢀO bond may readily lead to
formation of a cyclic oxonium ion (IV). Once
it is formed, a silane can readily attack from
the axial direction to produce the alditol 5
(and 5Da when Et3SiD is used). An SN2 type
of reaction of III with Et3SiD should give 5Db
as the major product, and this was not ob-
served. For such a displacement to occur, a
silane has to approach from the direction
parallel to the one of the lone-pair orbitals of
the ring oxygen atom, as shown in VI. This
seems unfavorable because of the repulsive
nature of the interaction between the lone-pair
electrons and the partially negatively charged
hydrogen of the reducing agent.
Our observation can be explained by sug-
gesting two reaction pathways, as shown in
Scheme 2. In the case of the b anomer, the
reduction may take place without the forma-
tion of the cyclic oxonium ion. The trime-
thylsilyl cation derived from Me3SiOSO2CF3
may coordinate to both O-5 and O-1 atoms of
the b anomer to form a complex such as I.
This kind of coordination is plausible if one
looks at the spatial arrangement of the lone-
pair orbitals in the two oxygen atoms. Then
the complex I may be directly attacked by a
silane to form an anhydroalditol product 5
and methoxytrimethylsilane, or the ring may
open to form an acyclic oxonium ion II. The
direct attack may be considered as an SN2
process, which can be illustrated as V. This
may be the reason that the a-deuterio-1,5-an-
hydroglucitol 5Da formed almost exclusively
When the substrate was mixed with
Me3SiOSO2CF3 in the absence of the reducing
agent, rapid equilibrium was reached within 1
h [15]. The a/b ratio was approximately four
for permethylated methyl
D
-glucopyranoside
and five for the galactopyranoside, regardless
of the anomeric configuration of the starting
material. The presence of a silane significantly
decreases the rate of anomerization of 1a to
1b, as shown in Table 1. This is not surprising
because the cyclic oxonium ion intermediate
IV should react with the silane (which should
be present in at least five-fold excess) at a
higher rate than recombination with
methoxytrimethylsilane to form 1b. The
anomerization of 1b to 1a is also retarded by
the presence of a silane, the ratio of a/b
ranging from 0.09 to 1.32. However, the retar-
Scheme 2.

C.K. Lee, E.J. Kim / Carbohydrate Research 320 (1999) 223–229
227
Table 3
of two hydrogen atoms may increase the re-
ducing power of 4g.
Rates of formation of 1,5-anhydroalditol by reductive cleav-
age of permethylated methyl
silane and Me3SiOSO2CF3 (1:5:5 by molar equiv) at 25 °C
D-glycopyranosides with triethyl-
The rates of the formation of the 1,5-anhy-
droalditols (5–7) are listed in Table 3. The
highest rate is observed with the b-galactoside
(3b) and the a-mannoside shows the lowest
rate among the five glycosides examined. The
rate enhancement in 3b may be due to the
complexation of Me3Si+ with the oxygen
atom at C-4 and the ring oxygen atom, as
shown in Scheme 3. Such complexation may
increase the partial charge at C-1, which
should be more susceptible to the attack by a
silane.
Glycoside
k (min−1
)
1a
1b
2a
3a
3b
9.6 (90.5)×10−3
3.7 (90.5)×10−2
6.0 (90.5)×10−3
2.2 (90.5)×10−2
1.2 (90.5)×10−1
A similar kind of complex formation is also
possible with the a-mannoside (2a). Once the
complex is formed, the lone-pair orbital of the
ring oxygen atom, which should be used for
complexation with Me3Si+ and the oxygen
atom of 2-CꢀOCH3, is no longer available to
push out the a-methoxy group at C-1 to form
a cyclic oxonium ion (Scheme 4). This may be
the reason for the lowest rate of the reductive
cleavage of 2a.
We also examined the effect of the molar
ratio of the substrate, triethylsilane, and
Me3SiOSO2CF3, and have listed the results in
Table 4. Surprisingly, the rate does not vary
much as might be expected due to the ratio
changes. The highest rate was observed when
the ratio was 1:5:10. This may constitute addi-
tional evidence that the formation of the cyclic
oxonium ion is rate determining in the case of
the a anomer.
Scheme 3.
dation is much less significant than with the
case of 1a to 1b.
tert-Butyldimethylsilane (4e) seems to be
the most effective among the reducing agents
employed in the present investigation. This
may be an indication that the silane itself, not
the free hydride ion, approaches the anomeric
carbon in the rate-determining step. The
stereoelectronic factor seems to play an im-
portant role. Triethoxysilane (4b), triphenylsi-
lane (4d), and dimethylphenylsilane (4f) are
less effective than the trialkylsilanes 1a, 1c,
and 1e. Diphenylsilane (4g) seems to be as
effective as triethylsilane (4a) in spite of the
presence of two phenyl groups. The presence
Table 4
Rates of formation of 1,5-anhydroglucitol by reductive cleav-
age of permethylated methyl a-D-glucopyranoside (1a) under
different ratios of substrate, triethylsilane, and catalyst at
25 °C
a
Ratio
k (s−1
)
1:5:5
9.0 (90.5)×10−3
1.0 (90.5)×10−2
9.3 (90.5)×10−3
7.4 (90.5)×10−3
7.5 (90.5)×10−3
8.8 (90.5)×10−3
1.6 (90.5)×10−2
1.4 (90.5)×10−2
2.8 (90.5)×10−3
1:10:20
1:10:5
1:10:2
1:10:1
1:20:10
1:5:10
1:2:10
1:1:10
a 1a:Et3SiH:Me3SiOSO2CF3.
Scheme 4.

228
C.K. Lee, E.J. Kim / Carbohydrate Research 320 (1999) 223–229
Method B (use of NMR). Stock solutions of
3. Experimental
a mixture of a glycoside and a reducing agent
were prepared in a 1 mL volumetric flask by
dissolving 36.5 mg of the substrate and appro-
priate amount of Et3SiH in CD2Cl2 so that the
concentrations of substrate and reducing
agent were about 0.146 and 0.730 M, respec-
tively. A soln of Me3SiOSO2CF3 was prepared
by dissolving 0.200 mL in CD2Cl2 in a 2 mL
volumetric flask so that the concentration was
0.517 M. The soln of glycoside (0.2 mL) was
taken with a gas-tight syringe (0.25 mL) and
placed in an NMR tube (5 mm diameter) and
the soln of Me3SiOSO2CF3 (0.29 mL) was
introduced by a gas-tight syringe (0.50 mL).
The molar ratio of the glycoside, the silane,
Starting materials.—Methyl a- and b- -glu-
D
copyranosides, methyl a- -mannopyranoside,
D
and methyl a- and b-
were all commercial products. Permethylated
methyl -glycopyranosides (1a, 1b, 2a, 3a, 3b)
D-galactopyranosides
D
were prepared as described previously [19].
Triethylsilane, (C2H5)3SiH (4a), triethoxysi-
lane, (C2H5O)3SiH (4b), triisopropylsilane, (i-
C3H7)3SiH (4c), triphenylsilane, (C2H5)3SiH
(4d), tert-butyldimethylsilane, t-C4H9(CH3)2-
SiH (4e), dimethylphenylsilane, C6H5(CH3)2-
SiH (4f), and diphenylsilane (C6H5)2SiH2 (4g)
were commercial products and were used
without purification. Trimethylsilyl trifluo-
romethanesulfonate (Me3SiOSO2CF3) was
also a commercial product and was distilled
prior to use. CH2Cl2 was dried over CaH2 and
distilled prior to use.
1
and Me3SiOTf was 1:5:5. H NMR spectra of
the soln were obtained at predetermined inter-
vals using a kinetics program. The spectra
were retrieved and the peaks corresponding to
H-1 at around l 4.2 for b, l 4.8 for a, l 4.0
for the product, and CH2Cl2 at around l 5.2
were integrated. The pseudo-first-order rate
constants (kobs) were calculated from the slope
of a plot of time versus ln[(A0−Aꢀ)/(At −
Aꢀ)] for a period of 1 h [20] and the results
are listed in Table 3. ‘A’ values were calcu-
lated from the ratio of the integrations of the
anomeric proton and CH2Cl2 that was present
in the solvent. The value of A0 was the one
obtained from the first spectrum of each run
and the value of Aꢀ was the one at which the
concentration of the product reached a maxi-
mum. The first spectrum was usually obtained
within 2 min and the progress of the reaction
during the period was less than 1 and 3% for
the a and b anomers, respectively.
Analytical methods.—1H NMR spectra
were recorded on a Varian 500 VXR-FT
NMR spectrometer in CD2Cl2 containing
Me4Si as the internal standard. A Hewlett–
Packard 5890 Plus gas–liquid chromatograph
equipped with a capillary column (HP-5, 25
m, 0.53 mm×1.0 mm) and a flame-ionization
detector was used for analysis. The column
was held at 120 °C for 2 min after injection
then programmed to 250 °C at 6 °C/min. Ni-
trogen was used as the carrier gas at a flow
rate of 1 mL/min.
Measurement of rates
Method A (use of GLC). A stock soln of
permethylated
methyl
D-glycopyranoside
(1.8×10−1 M) in CH2Cl2 was prepared. The
reaction mixture was prepared by mixing the
soln (2.6 mL) with a silane (10 equiv),
Me3SiOSO2CF3 (0.84 mL, 10 equiv) and do-
cosane (47 mg). The mixture was divided into
12 V vials (0.2 mL each) and stirred at room
temperature. At the predetermined time inter-
val, the reaction was quenched by adding a
soln (0.5 mL) of satd NaHCO3. The organic
layer was carefully separated with a syringe
and stored in a freezer (−10 °C) prior to
conducting the GLC. The mole percentages of
each component present after 30 min, 1 h, and
24 h are listed in Tables 1 and 2 for 1a and 1b,
respectively.
Acknowledgements
We thank Dr Gary Kwong of 3M for help
in preparing the manuscript. This work was
supported by the Research Center for New
Bio-Materials in Agriculture and by Korea
Science and Engineering Foundation.
References
[1] D. Rolf, G.R. Gray, J. Am. Chem. Soc., 104 (1982)
3539–3541.

C.K. Lee, E.J. Kim / Carbohydrate Research 320 (1999) 223–229
229
[2] D. Rolf, J.A. Benneck, G.R. Gray, J. Carbohydr. Chem.,
2 (1983) 373–383.
[3] J.A. Benneck, D. Rolf, G.R. Gray, J. Carbohydr. Chem.,
[11] D. Rolf, G.R. Gray, Carbohydr. Res., 131 (1984) 17–28.
[12] D. Rolf, G.R. Gray, Carbohydr. Res., 152 (1986) 343–
349.
2 (1983) 385–393.
[4] J.U. Bowie, G.R. Gray, Carbohydr. Res., 129 (1984)
[13] S.A. Vodonik, G.R. Gray, Carbohydr. Res., 172 (1989)
255–266.
87–97.
[14] C.K. Lee, E.J. Kim, I.-S.H. Lee, Carbohydr. Res., 240
(1993) 197–206.
[5] J.U. Bowie, P.V. Trescony, G.R. Gray, Carbohydr. Res.,
125 (1984) 301–307.
[6] J.-G. Jun, G.R. Gray, Carbohydr. Res., 163 (1987) 247–
[15] C.K. Lee, E.J. Kim, I.-S.H. Lee, Carbohydr. Res., 309
(1998) 243–250.
261.
[16] J. Kaczmarek, M. Preyss, H. Lonnberg, J. Szafranek,
Carbohydr. Res., 279 (1995) 107–116.
[7] C.K. Lee, G.R. Gray, J. Am. Chem. Soc., 110 (1988)
1292–1293.
[17] E.H. Cordes, H.G. Bull, Chem. Re6., 74 (1974) 581–503.
[18] P. Collins, R. Ferrier, Monosaccharides, Wiley, New
York, 1995, pp. 71–73.
[8] C.K. Lee, G.R. Gray, Carbohydr. Res., 269 (1995) 163–
179.
[9] J.A. Benneck, M.J. Rice, G.R. Gray, Carbohydr. Res.,
157 (1986) 125–137.
[10] D. Rolf, J.A. Benneck, G.R. Gray, Carbohydr. Res., 137
(1985) 183–196.
[19] J. Ciucanu, F. Kerek, Carbohydr. Res., 131 (1984) 209–
217.
[20] B.K. Carpenter, Determination of Organic Reaction
Mechanisms, Wiley, New York, 1984, p. 36.
.