Completely β-selective glycosylation using 3,6- O-(o-xylylene)-bridged axial-rich glucosyl fluoride

Article abstract of DOI:10.1021/ja301480g

A completely β-selective glycosylation that does not rely on neighboring group participation is described. The novelty of this work is the design of the glycosyl donor locked into the axial-rich form by the o-xylylene bridge between the 3-O and 6-O of d-glucopyranose. The synthesized 2,4-di-O-benzyl-3,6-O-(o-xylyene)glucopyranosyl fluoride could efficiently react with various alcohols in a SnCl₂-AgB(C₆F ₅)₄ catalytic system. The mechanism composed of the glycosylation and isomerization cycles was revealed through comparative experiments using acidic and basic molecular sieves. The achieved perfect stereocontrol is attributed to the synergy of the axial-rich conformation and convergent isomerization caused by HB(C₆F₅)₄ generated in situ.

Full text of DOI:10.1021/ja301480g

Communication
pubs.acs.org/JACS
Completely β-Selective Glycosylation Using 3,6-O-(o-Xylylene)-
Bridged Axial-Rich Glucosyl Fluoride
Yasunori Okada, Noriaki Asakura,† Masafumi Bando, Yoshiki Ashikaga, and Hidetoshi Yamada*
School of Science and Technology, Kwansei Gakuin University, 2-1 Gakuen, Sanda 669-1337, Japan
S
* Supporting Information
O-(o-xylylene) bridge. We also describe the reaction mecha-
nism.
ABSTRACT: A completely β-selective glycosylation that
does not rely on neighboring group participation is
described. The novelty of this work is the design of the
glycosyl donor locked into the axial-rich form by the o-
xylylene bridge between the 3-O and 6-O of D-
glucopyranose. The synthesized 2,4-di-O-benzyl-3,6-O-(o-
xylyene)glucopyranosyl fluoride could efficiently react with
various alcohols in a SnCl2−AgB(C6F5)4 catalytic system.
The mechanism composed of the glycosylation and
isomerization cycles was revealed through comparative
experiments using acidic and basic molecular sieves. The
achieved perfect stereocontrol is attributed to the synergy
of the axial-rich conformation and convergent isomer-
ization caused by HB(C6F5)4 generated in situ.
Because simultaneous 3,6-bridge formation with inversion of
the pyranose into an axial-rich conformation is unattainable,10
we applied a temporary ring-opening strategy to synthesize 2
starting from known diol 3 (Scheme 1). The o-xylylene
Scheme 1. Synthesis of 2,4-O-Benzyl-3,6-O-
(o-xylylene)glucopyranosyl Fluoride (2)
a
omplete control of anomeric stereoselectivity has been
C
one of the most important needs to be achieved in
chemical glycosylation. Stereocontrol on the basis of neighbor-
ing group participation (NGP) by a 2-O-acyl group has often
fulfilled expectations to furnish 1,2-trans-O-glycosidic linkages.1
On the other hand, stereocontrol that does not rely on the
NGP strategy has also been developed, but the range of
reaction conditions allowing high selectivity is narrow.2,3 In
addition, despite the high selectivity, the production of a small
proportion of the undesired isomer eventually requires
separation processes to obtain the pure desired isomer. In
this regard, the development of a completely stereoselective
glycosylation methodology has been challenging.
As a recently developed methodology for stereoselective
glycosylation, the utilization of conformationally locked
glycosyl donors has been discussed.4 Bulky silyl protecting
groups have often been introduced onto the adjacent trans-diols
of carbohydrates in order to lock the conformation into axial-
rich forms, which are conformations possessing more axial
substituents (e.g., 1).5 In the case of O-glucosylation with
conformationally locked glycosyl donors, high but incomplete
β-selectivity has been observed.6,7 In the reactions, reductions
in the reactivity and migration of the trialkylsilyl groups due to
the intense steric hindrance of the bulky silyl groups have been
observed.6,8,9 To resolve these problems and extend this
methodology to stereocontrol based on conformational
restriction, sterically less hindered, more robust, and easily
removable locking methods are desired. We describe here a
completely β-selective glycosylation using the novel axial-rich
glycosyl donor 2, in which the conformation is locked by a 3,6-
a
Reagents and conditions: (a) α,α′-dibromo-o-xylene, NaH, DMF, rt,
10 min, then 100 °C, 1 h, 51%; (b) DDQ, CH2Cl2, phosphate buffer
(pH 7.41), 0 °C, 1.5 h, 85%; (c) OsO4, NMO, acetone/H2O (5/1 v/
v), rt, 2 h, then NaIO4, rt, 5 h; (d) DAST, THF, 0 °C to rt, 30 min,
76% for three steps; sonication in MeOH/n-hexane, α/β = 8/92.
Abbreviations: DAST, (diethylamino)sulfur trifluoride; DDQ, 2,3-
dichloro-5,6-dicyano-p-benzoquinone; DMF, N,N-dimethylforma-
mide; NMO, N-methylmorpholine N-oxide; PMB, p-methoxybenzyl;
THF, tetrahydrofuran.
bridge11 was introduced into 3 by using NaH and α,α′-
dibromo-o-xylene to produce 3,6-O-bridged 4. Treatment with
DDQ removed the PMB group to provide 5. Oxidative cleavage
Received: February 14, 2012
Published: April 4, 2012
© 2012 American Chemical Society
6940
dx.doi.org/10.1021/ja301480g | J. Am. Chem. Soc. 2012, 134, 6940−6943

Journal of the American Chemical Society
Communication
of the terminal double bond of 5 produced pyranose 6.
Fluorination at the anomeric position of 6 furnished the desired
product 2 as an air-stable white powder. H NMR analysis
Table 1. β-Glycosylation Reaction Using Glucosyl Fluoride 2
with Various Glycosyl Acceptors
1
revealed that the pyranose ring of β-2 is certainly locked into
12
1
the axial rich S3 conformation.
To develop an efficient glycosylation using 2, previously
reported methods for activation of glycosyl fluorides were
investigated.13−15 The preliminary screening revealed that
catalytic use of SnCl2−AgB(C6F5)4, as reported by Mukaiya-
ma,15 was promising. We thus optimized this reaction with 2
and cyclohexylmethanol as the glycosyl donor and acceptor,
respectively. Finally, we chose the following reaction conditions
for the completely β-selective glycosylations: 2 (1.0 equiv),
ROH (1.2 equiv), SnCl2 (0.2 equiv), AgB(C6F5)4 (0.2 equiv),
and 5A molecular sieves (MS) (3.0 g/mmol of 2) in
benzotrifluoride (BTF)16 at room temperature (rt). These
reaction conditions were applied in the following investigations
into the scope and clarification of the reaction mechanism.
Interestingly, the optimal reaction conditions included
isomerization of the α-anomeric isomer into the β-isomer (eq
1). At the beginning of the reaction, both the α- and β-isomers
a
The stereochemistries of β-7b−f were confirmed after removal of the
b
c
benzyl and o-xylylene groups. Isolated yields. 1.5 equiv of glycosyl
acceptor was used. The reaction temperature was 0 °C.
d
glycosylation reaction, we applied this method to the synthesis
of β,β-trehalose, which possesses a 1,1-β,β-glycosidic linkage.
Pyranose 6 successfully coupled with 2 (entry 5) to provide
only the β,β-trehalose derivative 7f, although glycosyl acceptor
6 was a mixture of anomers.
were detected, and then the α-isomer disappeared gradually.
This observation indicated that the β-isomer of 7a is
thermodynamically more stable than the α-isomer. Despite
the existence of the o-xylylene bridge over the β-face of the
pyranose, the thermodynamic stability was more affected by the
1,2-cis repulsion between the anomeric alkyloxy and axially
oriented 2-O-Bn groups. Density functional theory (DFT)
calculations17 indicated that β-7a is 21.0 kJ/mol more stable
than α-7a. The stereochemistry of β-7a was confirmed after
returning the pyranose ring to the 4C1 conformation by
hydrogenolytic cleavage of the benzyl and o-xylylene groups.
To investigate the scope and limitations of this reaction, the
optimal reaction conditions were applied to several alcohols
(Table 1). The perfect β-stereoselectivity observed in all cases
We next investigated the reaction mechanism and found that
this complete β-glycosylation is composed of glycosylation and
isomerization cycles (Scheme 2). The glycosylation cycle is
triggered by the generation of SnB(C6F5)4Cl (8) as the actual
catalyst.15 The catalyst activates glucosyl fluoride 2, generating
oxocarbenium ion 9 along with SnClF. Intermediate 9 couples
with an alcohol to provide glycoside 7, initially as an anomeric
mixture. This glycosylation is accompanied by the formation of
HB(C6F5)4 (10), which reacts with SnClF to regenerate 8. In
the isomerization cycle, 10 isomerizes α-7 into the
thermodynamically more stable β-isomer via oxocarbenium
ion 9 to afford β-7 only. The following paragraphs describe in
detail the investigations into the mechanism.
1
was noteworthy; the α-isomers were not detected by H NMR
analysis of the crude products. The yields of the glucosides
were good to excellent. With simple aliphatic alcohols (entries
1 and 2), the β-glycosylation proceeded quite smoothly,
providing the glycosides in >90% yield, even with the tertiary
alcohol. With partially protected sugars (entries 3 and 4), the
yields were ∼80%. As described above, convergent anomeric
isomerization was induced in the β-selectivity. Since the α-
methoxy groups of the glycosyl acceptors survived without
being isomerized, the isomerization rate for the axial-rich
glycosides would be faster than for the equatorial-rich
glycosides. This phenomenon could be explained by the
conformational-arming theory reported by Bols and co-
workers,9,18 which states that an axially oriented polar
substituent stabilizes a positive charge (oxocarbenium ion in
this case) more effectively than the corresponding equatorial
equivalent. To show the effectiveness of this new β-
The reaction between 9 and an alcohol provides glycoside 7
along with the protic acid 10, which has been reported by
Mukaiyama to be an effective catalyst for activating a glycosyl
fluoride.19 These facts prompted the following question: which
species is the real catalyst of this reaction: 8 or 10? To clarify
the actual catalytic species experimentally, we utilized the
properties of molecular sieves: acidic 5A MS20 and basic 4A and
3A MS.21,22
Table 2 summarizes the results of comparative experiments
using the acidic and basic MS. The two reagents 8 and 10 were
generated according to eqs 2 and 3,14,15 respectively. When 10
activated 2 (entries 1−4), differences appeared in the yield. An
approximately 80% yield of β-7a was produced without MS or
with 5A MS (entries 1 and 2), whereas no reaction occurred
with 4A and 3A MS (entries 3 and 4). In the former two
entries, 10 could act as the actual catalyst. In contrast, the basic
6941
dx.doi.org/10.1021/ja301480g | J. Am. Chem. Soc. 2012, 134, 6940−6943

Journal of the American Chemical Society
Communication
Scheme 2. Possible Catalytic Cycle
AgB(C6F5)4 + t‐BuBr
→ HB(C6F5)4 + H2CCMe2 + AgBr
(10)
(2)
(3)
SnCl2 + AgB(C6F5)4 → SnB(C6F5)4Cl + AgCl
(8)
When 8 was applied as the reagent (entries 5−8), 7a was
provided in >90% yield. Differences appeared in the anomeric
ratio. Under acidic reaction conditions (entries 5 and 6), both 8
and 10 could act as the actual catalyst, but 8 was likely to be
more efficient because the yield of 7a was 10−20% lower when
only 10 was applied (entries 1 and 2 vs 5 and 6). With the basic
MS (entries 7 and 8), the MS could counteract 10 generated by
the glycosylation reaction (see Scheme 2), and thus, only 8
remained as the actual catalyst. These results demonstrate that
8 is not affected by the basic MS and is the actual catalyst in
these reactions. Thus, there should be a pathway to revive 8 as
the final step of the glycosylation cycle.
The conversion from 10 and SnClF to 8 (eq 4) is considered
to be possible on the basis of the acidities of HF (pKa = 3.2)
and 10. Although the pKa value of 10 is not described in the
−
literature, B(C6F5)4 is known as an activated anion whose
nucleophilicity and coordination ability have been reported to
be quite low.24 Accordingly, the acidity of 10 should be much
stronger than that of HF, shifting the equilibrium in eq 4 to the
right.
HB(C6F5)4 + SnCIF ⥂ SnB(C6F5)4Cl + HF
(10)
(8)
(4)
This consideration was supported by experiments. Although
SnF2 did not induce the glycosylation reaction (eq 5), the
simultaneous use of SnF2 and 10 gave glycoside 7 (eq 6). In
this experiment, 4A MS was applied to hinder the reaction
caused by 10 itself. This successful glycosylation revealed that
the transformation described in eq 7 could be possible,
supporting the production of 8 via eq 4.
SnF2 (0.2 equiv)
2 + C6H11CH2OH ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→ no reaction
BTF, rt, 44 h
(1.2 equiv)
(5)
Table 2. Effects of MS in the Glycosylation Reactions
SnF2 (0.3 equiv)
Catalyzed by SnB(C6F5)4Cl (8) or HB(C6F5)4 (10)
10 (0.2 equiv), 4A MS
reagent (0.2 equiv)
2 + C6H11CH2OH ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→
7a
BTF, rt, 16 h, 81%
(1.2 equiv)
2 + C6H11CH2OH ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯→ 7a
α/β=61/39
(6)
(7)
MS, BTF, rt
(1.2 equiv)
10 + SnF ⥂ SnB(C6F5)4F + HF
2
entry
reagent
MS
actual catalyst
yield (%)
α/β ratio
1
2
3
4
5
6
7
8
10
10
10
10
8
none
5A
10
78
80
>1/99
>1/99
−
In the isomerization cycle (Scheme 2), it was clear that 10
dominates the cycle. Thus, in Table 2, when 8 worked as the
actual catalyst (entries 7 and 8), low stereoselectivity was
observed. In contrast, when 10 was the actual catalyst (entries
1, 2, 5, and 6), complete β-selectivity was obtained. In addition,
isomerization of α-7a to β-7a using 10 was confirmed (eq 8),
10
a
4A
none
none
8 or 10
8 or 10
8
0
a
3A
0
−
none
5A
96
98
95
90
>1/99
>1/99
60/40
67/33
8
8
4A
8
3A
8
a
No reaction.
MS shown in the latter two entries deactivated the protic acid
10; therefore, no “actual catalyst” existed in the reaction
system.23 These results show that 10 can catalyze the
glycosylation reaction under acidic conditions but not under
basic conditions.
6942
dx.doi.org/10.1021/ja301480g | J. Am. Chem. Soc. 2012, 134, 6940−6943

Journal of the American Chemical Society
Communication
(6) Okada, Y.; Mukae, T.; Okajima, K.; Taira, M.; Fujita, M.; Yamada,
H. Org. Lett. 2007, 9, 1573.
(7) Okada, Y.; Nagata, O.; Taira, M.; Yamada, H. Org. Lett. 2007, 9,
which experiment also corroborated the generation of 9.
Despite the potential isomerization pathway via an endocyclic
cleavage,25 the isomerization involving exocyclic cleavage is
more probable, as the isomerization reaction in the presence of
Et3SiH provided 11 (eq 9).
In conclusion, we have developed a completely β-selective
glycosylation that does not rely on the NGP strategy. To realize
this, we designed and synthesized the 3,6-O-(o-xylylene)-
bridged axial-rich glucosyl fluoride 2. The β-glycosylation
reaction using 2 and SnCl2−AgB(C6F5)4 catalyst was stably
applicable to several types of alcohols. In these reactions, the
actual catalyst might be SnB(C6F5)4Cl, generated in situ. The
perfect β-selectivity arises from isomerization of the α-anomeric
isomer into the β-isomer, which is catalyzed by HB(C6F5)4 that
is also generated in situ. This novel glycosylation offers a
fundamental concept for new trends in the design of chemical
glycosylation.
2755.
(8) (a) Broddefalk, J.; Bergquist, K.-E.; Kihlberg, J. Tetrahedron 1998,
54, 12047. (b) Yamada, H. Trends Glycosci. Glycotechnol. 2011, 23, 122.
(9) On the other hand, Bols reported enhancement of the reactivity
on axial-rich glycosyl donors induced by silylations. See: (a) Pedersen,
C. M.; Nordstrøm, L. U.; Bols, M. J. Am. Chem. Soc. 2007, 129, 9222.
(b) Pedersen, C. M.; Marinescu, L. G.; Bols, M. Chem. Commun. 2008,
2465.
(10) Yamada, H.; Nagao, K.; Dokei, K.; Kasai, Y.; Michihata, N. J.
Am. Chem. Soc. 2008, 130, 7566.
(11) (a) Poss, A. J.; Smyth, M. S. Synth. Commun. 1989, 19, 3363.
(b) Balbuena, P.; Rubio, E. M.; Mellet, C. O.; Fernandez, J. M. G.
́
Chem. Commun. 2006, 2610. (c) Imamura, A.; Lowary, T. L. Org. Lett.
2010, 12, 3686.
(12) For the conformational descriptor, see: Dixon, H. B. F. Pure
Appl. Chem. 1981, 53, 1901.
(13) (a) Mukaiyama, T.; Murai, Y.; Shoda, S. Chem. Lett. 1981, 431.
(b) Takahashi, Y.; Ogawa, T. Carbohydr. Res. 1987, 164, 277.
(c) Matsumoto, T.; Maeta, H.; Suzuki, K.; Tsuchihashi, G. Tetrahedron
Lett. 1988, 29, 3567. (d) Suzuki, K.; Maeta, H.; Matsumoto, T.;
Tsuchihashi, G. Tetrahedron Lett. 1988, 29, 3571.
(14) Jona, H.; Takeuchi, K.; Mukaiyama, T. Chem. Lett. 2000, 1278.
(15) Mukaiyama, T.; Maeshima, H.; Jona, H. Chem. Lett. 2001, 388.
(16) Ogawa, A.; Curran, D. P. J. Org. Chem. 1997, 62, 450.
(17) DFT geometry optimizations were performed with the B3LYP
hybrid functional and the 6-31G(d,p) basis set using the Gaussian 03
program: Frisch, M. J.; et al. Gaussian 03, revision E.01; Gaussian, Inc.;
Wallingford, CT, 2004.
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental procedures and characterization data for all
reactions and products, including H NMR and 13C NMR
1
spectra, the results of DFT calculations, and complete ref 17 (as
SI ref 3). This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
hidetosh@kwansei.ac.jp
■
(18) (a) Jensen, H. H.; Bols, M. Acc. Chem. Res. 2006, 39, 259.
(b) McDonnell, C.; Lopez, O.; Murphy, P.; Fernandez Bolanos, J. G.;
́
́
̃
Present Address
Hazell, R.; Bols, M. J. Am. Chem. Soc. 2004, 126, 12374.
(19) Jona, H.; Mandai, H.; Chavasiri, W.; Takeuchi, K.; Mukaiyama,
T. Bull. Chem. Soc. Jpn. 2002, 75, 291.
†Department of Chemistry, Graduate School of Science, Osaka
University, Toyonaka, Osaka 560-0043, Japan.
(20) Asakura, N.; Hirokane, T.; Hoshida, H.; Yamada, H. Tetrahedron
Lett. 2011, 52, 534.
Notes
The authors declare no competing financial interest.
(21) Kreuzer, M.; Thiem, J. Carbohydr. Res. 1986, 149, 347.
(22) (a) Mizuno, M.; Kobayashi, K.; Nakajima, H.; Koya, M.; Inazu,
T. Synth. Commun. 2002, 32, 1665. (b) Kartha, K. P. R.;
Mukhopadhyay, B.; Field, R. A. Carbohydr. Res. 2004, 339, 729.
(23) Similar deactivation has been observed in the perchloric acid-
mediated glycosylation. See: Evans, D. A.; Kaldor, S. W.; Jones, T. K.;
Clardy, J.; Stout, T. J. J. Am. Chem. Soc. 1990, 112, 7001.
(24) Piers, W. E.; Chivers, T. Chem. Soc. Rev. 1997, 26, 345.
(25) (a) Manabe, S.; Ishii, K.; Hashizume, D.; Koshino, H.; Ito, Y.
Chem.Eur. J. 2009, 15, 6894. (b) Satoh, H.; Manabe, S.; Ito, Y.;
ACKNOWLEDGMENTS
Kakenhi and a SUNBOR grant partially supported this work.
■
REFERENCES
■
(1) (a) Koenigs, W.; Knorr, E. Ber. Dtsch. Chem. Ges. 1901, 34, 957.
(b) Zhu, X.; Schmidt, R. R. Angew. Chem., Int. Ed. 2009, 48, 1900.
(c) Demchenko, A. V. Synlett 2003, 1225. (d) Jensen, K. J. J. Chem.
Soc., Perkin Trans. 1 2002, 2219. (e) Toshima, K.; Tatsuta, K. Chem.
Rev. 1993, 93, 1503. (f) Banoub, J.; Bundle, D. R. Can. J. Chem. 1979,
57, 2091.
Luthi, H. P.; Laino, T.; Hutter, J. J. Am. Chem. Soc. 2011, 133, 5610.
̈
(c) Manabe, S.; Ishii, K.; Satoh, H.; Ito, Y. Tetrahedron 2011, 67, 9966.
(2) Hashimoto, S.; Hayashi, M.; Noyori, R. Tetrahedron Lett. 1984,
25, 1379.
(3) (a) Schmidt, R. R.; Michel, J. Angew. Chem., Int. Ed. 1980, 19,
731. (b) Ito, Y.; Ogawa, T. Tetrahedron Lett. 1987, 28, 4701.
(c) Schmidt, R. R.; Behrendt, M.; Toepfer, A. Synlett 1990, 694.
(d) Vankar, Y. D.; Vankar, P. S.; Behrendt, M.; Schmidt, R. R.
Tetrahedron 1991, 47, 9985. (e) Hashihayata, T.; Mandai, H.;
Mukaiyama, T. Bull. Chem. Soc. Jpn. 2004, 77, 169. (f) Hashimoto,
S.; Umeo, K.; Sano, A.; Watanabe, N.; Nakajima, M.; Ikegami, S.
Tetrahedron Lett. 1995, 36, 2251.
(4) (a) Hosoya, T.; Ohashi, Y.; Matsumoto, T.; Suzuki, K.
Tetrahedron Lett. 1996, 37, 663. (b) Tamura, S.; Abe, H.; Matsuda,
A.; Shuto, S. Angew. Chem., Int. Ed. 2003, 42, 1021. (c) Ikeda, T.;
Yamada, H. Carbohydr. Res. 2000, 329, 889.
(5) (a) Yamada, H.; Nakatani, M.; Ikeda, T.; Marumoto, Y.
Tetrahedron Lett. 1999, 40, 5573. (b) Yamada, H.; Tanigakiuchi, K.;
Nagao, K.; Okajima, K.; Mukae, T. Tetrahedron Lett. 2004, 45, 5615.
(c) Yamada, H.; Tanigakiuchi, K.; Nagao, K.; Okajima, K.; Mukae, T.
Tetrahedron Lett. 2004, 45, 9207.
6943
dx.doi.org/10.1021/ja301480g | J. Am. Chem. Soc. 2012, 134, 6940−6943