Low-barrier pathway for endo-cleavage induced anomerization of pyranosides with N-benzyl-2,3-trans-oxazolidinone groups

Article abstract of DOI:10.1002/ejoc.200801140

Pyranosides with N-benzyl-2,3-trans-oxazolidinone undergo anomerization from the β form to the α form even in the presence of a weak Lewis acid. Experimental evidence for endo-cleavage, the breaking of the bond between the pyran-oxygen and anomeric carbon

Full text of DOI:10.1002/ejoc.200801140

SHORT COMMUNICATION
DOI: 10.1002/ejoc.200801140
Low-Barrier Pathway for endo-Cleavage Induced Anomerization of
Pyranosides with N-Benzyl-2,3-trans-oxazolidinone Groups
Hiroko Satoh,*[a,b] Jürg Hutter,[c] Hans Peter Lüthi,[a] Shino Manabe,[d,e] Kazuyuki Ishii,[d,e]
and Yukishige Ito[d]
Keywords: Anomerization / Carbohydrates / Density functional calculations / Transition states / Protecting groups
Pyranosides with N-benzyl-2,3-trans-oxazolidinone undergo
anomerization from the β form to the α form even in the pres-
ence of a weak Lewis acid. Experimental evidence for endo-
cleavage, the breaking of the bond between the pyran-oxy-
gen and anomeric carbon atoms, in the anomerization reac-
tions was obtained. This unexpected phenomenon was in-
vestigated by quantum mechanical calculations, which found
clear differences in the transition states between anomerized
and non-anomerized substrates. The computations suggest
that BF3 induces endo-cleavage followed by rotation of the
C1–C2 bond to give the α form via lower-energy transition
states.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2009)
The second is endo-cleavage, where the bond between the
anomeric carbon and the pyranose ring oxygen atoms
breaks (Figure 1, b). For acetals, there have been many dis-
cussions about the presence of both exo- and endo-cleavage
pathways. However, in carbohydrate chemistry, the occur-
rence of endo-cleavage in pyranosides is not thought to be
very common. The possibility of endo-cleavage taking place
in pyranosides has been investigated ever since Post and
Karplus proposed a hypothesis of endo-type hydrolysis for
oligosaccharides based on molecular dynamics calculations
of N-acetylglucopyranoside with lysozyme, where the pyran
ring was fixed to a chair-form conformation.[2] Franck cap-
tured the cation generated through endo-cleavage by an in-
tramolecular aza-Diels–Alder reaction in alkyl β-THP ace-
tal methanolysis.[3]
Introduction
The mechanism of acetal hydrolysis with their stereoelec-
tronic aspects have attracted a great deal of attention in
both organic chemistry and biochemistry.[1] In carbohydrate
chemistry, since glycosides are asymmetric acetals, the
mechanism of glycosidic cleavage is also a fundamental is-
sue. Two possible pathways have been proposed for the
cleavage mechanism (Figure 1). The first is exocyclic cleav-
age, where the bond between the anomeric carbon and the
exocyclic oxygen atoms breaks to give the cyclic oxocarbon-
ium ion (Figure 1, a). The oxocarbonium ion reacts with
acceptors such as alcohols, phenols, and thiols to give gly-
cosides. It is known that typical glycosylation reactions are
based on this mechanism.
Fraser–Reid demonstrated the existence of linear ace-
tylium ions in acetic acid during the acetolysis of acyl-pro-
tected methyl glycosides in the presence of ferric chloride.[4]
Anslyn reported that only the cis-decalin-type of conforma-
tionally locked β-alkyl acetals undergo endocyclic cleavage
with a maximum yield of 30% in MeOH with a deuterium
scrambling test.[5] It was reported that under aprotic condi-
tions, endo-cleavage for typical pyranose sugars requires
strong Lewis acids, such as SnCl4.[6] Deslongchamps and
Dory performed a detailed investigation of the reaction
pathways of enzyme-catalyzed hydrolyses of α- and β-glyco-
sides.[7] Their RHF/6-31G(d,p) calculations indicated the
presence of a transition structure for the endo-cleavage of
tetrahydropyranyl acetals under protic conditions; the tran-
sition energy relative to the β form was ca. 15 kJmol–1 in
vacuo.
Figure 1. Exo- (a)- and endocyclic (b) cleavage of glycoside.
[a] Laboratory of Physical Chemistry,
ETH Zürich, 8093 Zürich, Switzerland
[b] National Institute of Informatics,
Chiyoda-ku, Tokyo 101-8430, Japan
Fax: +81-3-3556-1916
E-mail: cheminfo@nii.ac.jp
[c] Institute of Physical Chemistry, Zürich University,
8057 Zürich, Switzerland
[d] RIKEN Advanced Science Institute,
Wako, Hirosawa, Saitama 351-0198, Japan
[e] PRESTO, Japan Science and Technology Agency (JST),
Honcho, Kawaguchi, Saitama 332-1102, Japan
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Some other experimental studies related to endo-cleavage
were reported by Crich and Manabe. They independently
found that pyranosides with N-benzyl-2,3-trans-oxazolidi-
Eur. J. Org. Chem. 2009, 1127–1131
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1127

H. Satoh, J. Hutter, H. P. Lüthi, S. Manabe, K. Ishii, Y. Ito
SHORT COMMUNICATION
none groups[8] are quite easily anomerized from the β- to
the α-direction, which made them the first cases of endo-
cleavage under mild acidic conditions in aprotic media.[9]
Olsson and Oscarson et al. very recently reported the ano-
merization of disaccharides with kinetic measurements.
However, they reported that evidence of endo-cleavage was
not obtained.[10]
We report here that we experimentally obtained evidence
for the endo-cleavage of the anomerization of pyranosides
with N-benzyl-2,3-trans-oxazolidinone groups. We detected
lower-energy transition states (TS) on the pathways for the
endo-cleavage induced anomerization under aprotic condi-
tions with a weak Lewis acid for pyranosides with oxazolid-
inone groups.
Figure 2. Proposed mechanism of anomerization of pyranosides
with 2,3-trans-oxazolidinone groups.
Results and Discussion
cal glucose 4C1 chair-form conformation. Calculations were
performed for two coordination modes of BF3 to the lone
pairs of the pyran oxygen atom. We describe here the results
of the coordination to the axially oriented lone pair in the
4C1 chair-form conformation, which showed the lowest-en-
ergy barriers of the two modes (Scheme 2).
To obtain evidence for the endo-cleavage of pyranoside
1, a reduction reaction under acidic conditions was carried
out (Scheme 1). After the reductive benzylidene cleavage re-
action of compound 1 with Et3SiH and BF3·OEt2 at 0 °C,
the α-thioglycoside 2b was obtained as a by-product in 14%
yield in addition to the β-glycoside 2a (72%). When the
reaction was carried out at room temperature, the yield of
the α-product 2b increased to 49%. Prolonged reaction
times increased the yield of the α-glycoside 2b and gave
alcohol 3 in 11% yield. It is assumed that the ring-opened
alcohol 3 was generated by the reduction of the ring-opened
cation by Et3SiH (Figure 2). It was previously reported that
the cation generated by endo-cleavage was captured in an
intramolecular Diels–Alder reaction in protic media.[3]
However, our examples highlight a new phenomenon in
that they indicate the existence of the endo-cleaved cation
in aprotic media by simply carrying out reduction reactions.
Furthermore, additional evidence for endo-cleavage was ob-
served in the course of the experimental studies.[11]
Scheme 2. Calculated model structures.
The relevant variables to model the reaction coordinate
Four structures (4a, 5a, 6a, 7a) were investigated by den- are the rotation of the C1–C2 bond, the O5–B distance, and
sity functional theory with the basis set B3LYP/6-31G(d,p) the C1–O5 distance. We chose the rotation of the C1–C2
by using Gaussian03.[12] BF3 was used as the Lewis acid. bond as the first target to investigate. We explored the po-
Initial geometries of 4a and 5a were constructed on the ba- tential energy surface by rotating the C1–C2 bond counter-
sis of NMR and X-ray data for the related pyranosides with clockwise from 0 to 180° in 5° steps, i.e. by decreasing the
2,3-trans-oxazolidinone groups, which indicate that the py- dihedral angle θ (H1–C1–C2–H2) from the β to the α direc-
ran ring takes a 4C1 chair-form conformation.[9b,11] Geome- tion (Figure 3) and allowing full relaxation of all other vari-
tries of 6a and 7a were constructed on the basis of the typi- ables.
Scheme 1. Experimental investigation of the anomerization process.
1128
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 1127–1131

endo-Cleavage Induced Anomerization of Pyranosides
which means that calculations using the vacuum condition
are adequate for discussing the differences in energies be-
tween oxazolidinones (4a and 5a) and the others (6a and
7a).
The other relevant variables, the O5–B and the C1–O5
distances as a function of the rotation about the C1–C2
bond are plotted as shown in Figure 5. The solid and dotted
lines represent the C1–O5 and O5–B distances, respectively.
The plots show that these two variables have a strong de-
pendence on the C1–C2 bond rotation. The results indicate
that the O5–B distance is important for the reaction.
Figure 3. The dihedral angle and rotation.
The initial θ values and the C1–O5 bond lengths of the
β-formed structures 4a, 5a, 6a, and 7a were 174, 175, 177,
and 179°, and 0.145, 0.145, 0.144, and 0.143 nm, respec-
tively. When the dihedral angle θ is reduced, the C1–O5
bond length increases until the bond finally breaks at θ =
105, 95, 125, and 90°, for which the energies relative to the
corresponding
β form were 94.2, 90.5, 123.7, and
137.8 kJmol–1 for 4a, 5a, 6a, and 7a, respectively. After the
bond breaks, transition states (TS) were found between θ =
28–45° for all structures. The energy relative to the corre-
sponding β form and the values for θ of the TS for 4a, 5a,
6a, and 7a were 104.5 kJmol–1 at 44.2°, 83.6 kJmol–1 at
28.8°, 159.5 kJmol–1 at 35.6°, and 157.1 kJmol–1 at 37.3°,
respectively. Figure 4 shows the TS energies relative to the
β form as well as the relative energies of the α- and β forms
with their structures for 4a–7a. The results show that the
TS energies of the structures with 2,3-trans-oxazolidinone
groups, 4a and 5a, were clearly lower than the other struc-
tures without 2,3-trans-oxazolidinone groups, 6a and 7a. It
is noted that the oxazolidinone groups of the TS for 4a
Figure 5. Changes in the O5–B and C1–O5 distances during C1–
C2 rotation.
We therefore then focused on the O5–B distance and in-
and 5a obviously take a flat form rather than those of the vestigated the potential energy surface with BF3 approach-
corresponding β and α forms. The energy calculations that ing the oxygen atom in the pyran ring, without a solvent.
take the solvent CH2Cl2 into account with the IEF-PCM The initial O5–B and C1–O5 distances for optimized 4a,
(Integral Equation Formalism for the Polarizable Contin- 5a, 6a, and 7a were 0.26, 0.27, 0.25, and 0.25 nm, and 0.145,
uum Model) method of the SCRF (Self-Consistent Reac- 0.145, 0.144, and 0.143 nm, respectively. Starting from the
tion Field) theory stabilized the TS for all structures by al- equilibrium structures of 4a, 5a, 6a, and 7a, the O5–B dis-
most the same amount, shown as dotted lines in Figure 4. tance was reduced from 0.22 nm in steps of 0.01 nm. The
These results show the same tendency as those in vacuo, C1–O5 bond in the pyran ring became longer than 0.15 nm
Figure 4. Relative energies for the transition state and the α- and β forms for 4–7.
Eur. J. Org. Chem. 2009, 1127–1131
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1129

H. Satoh, J. Hutter, H. P. Lüthi, S. Manabe, K. Ishii, Y. Ito
SHORT COMMUNICATION
when the O5–B distance was 0.16, 0.16, 0.15, and 0.15 nm, site instead of the phenoxy group (8), since the p-meth-
for which the energies relative to the corresponding initial oxyphenyl group can be removed under oxidative condi-
β form were 19.4, 20.0, 34.7, and 36.3 kJmol–1 for 4a, 5a, tions and is commonly used in oligosaccharide synthesis.[13]
6a, and 7a, respectively (Figure 6). When the potential en-
ergy surface for the C1–C2 bond rotation was explored, the
C1–O5 bond became longer than 0.15 nm at θ = 135, 125,
125, and 130°, for which energies relative to the correspond-
ing β form were 35.9, 52.5, 54.1, and 47.2 kJmol–1, respec-
tively. These values are considerably larger than those ob-
tained in this investigation for the O5–B bond length. The
results show that the effect of BF3 is to extend the C1–O5
bond, which makes the rotation about the C1–C2 bond eas-
ier. Moreover, they indicate that pyranoses with the oxazoli-
dinone group (4a and 5a) can be endo-cleaved more easily
than compounds without the oxazolidinone ring (6a and
7a).
Figure 6. O5–B bond length (nm) and energy relative to the corre-
sponding initial β form (kJmol–1) at the state where the C1–O5
bond was elongated more than 0.15 nm.
We then investigated the possibilities of exo-cleavage with
BF3 approaching the oxygen atom (O1) for 4a, 6a, and 7a
or the sulfur atoms for 5a at the anomer site in the same Scheme 3. Experimental results based on the calculated model
structures.
manner. The initial O1–B(S–B for 5a) and C1–O1 (C1–S
for 5a) distances for the β-formed 4a, 5a, 6a, and 7a were
The pyranoside with trans-oxazolidinone 4 (α/β 18:82)
0.22, 0.33, 0.23, and 0.24 nm, and 0.140, 0.183, 0.140, and
was anomerized preferably in the α-direction [4a(β)/4b(α)
0.141 nm, respectively. The C1–O1 bond did not break even
36:64] in the presence of 2 equiv. BF3·OEt2 at –30 °C after
when the O1–B and S–B distances became 0.14 and
12 h. Similarly, the thiophenyl glycoside 5 gave 5a (β) and
0.17 nm, respectively, whereas the C1–O5 bond broke when
5b (α) in 16% and 63% yields, respectively, at –30 °C after
the O5–B distance became ca. 0.16 nm. The energies rela-
12 h. When the reaction was carried out at 0 °C, anomeriz-
tive to the corresponding initial β form with the O1–B and
ation equilibrium was almost reached within 60 min. On
S–B distances were 63.9, 121.9, 67.6, and 75.7 kJmol–1,
the other hand, for the typical pyranosides 6 and 8, the
respectively. These results indicate that BF3-induced endo-
corresponding α-anomers were not observed, and the start-
cleavage is much more dominant than exo-cleavage for these
ing materials were recovered in high yields.
pyranosides.
This series of computations shows that the C1–C2 bond
in the pyran ring with a 2,3-trans-oxazolidinone group will
rotate to give the α form via lower-energy transition states
Conclusions
following endo-cleavage induced by BF3. The calculations
We obtained experimental evidence that endo-cleavage
also show that pyranoses with oxazolidinones are anomer- takes place for pyranosides with 2,3-trans-oxazolidinone
ized as a result of their lower-energy barriers. The observed groups with a weak Lewis acid by isolation of the reductive
geometry change indicates that the oxazolidinone group product. On the basis of these results, we explored the path-
prefers to take a flat form but is forced to strain in the β way of endo-cleavage induced anomerization of glycosides
and α forms. As can be seen from the current calculations, using quantum mechanical computations and found that
this strained structure most likely makes endo-cleavage feas- transition states form as the C1–C2 bond rotates to trans-
ible for 4a and 5a.
form the β form to the α form. Our investigation showed
The tendency of structures 4a–7a to anomerize as pro- that pyranosides with N-benzyl-2,3-trans-oxazolidinone
posed by the calculations was verified experimentally groups clearly possess lower energies for both the transition
(Scheme 3). For the model structure 7a, we used a p-meth- state and the endo-cleavage than pyranosides without
oxylphenyl group as the protecting group at the anomeric oxazolidinone groups. The results are consistent with ex-
1130
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 1127–1131

endo-Cleavage Induced Anomerization of Pyranosides
R. Natarajan, J. Am. Chem. Soc. 1986, 108, 2425–2430; d) A. J.
Bennet, M. L. Sinnott, J. Am. Chem. Soc. 1986, 108, 7287–
7294; e) A. J. Kirby, P. G. Jones, J. Chem. Soc., Chem. Commun.
1986, 444–445; f) T. H. Fife, T. J. Przystas, J. Chem. Soc. Perkin
Trans. 2 1987, 143–150.
periments in which pyranosides with oxazolidinones were
anomerized by weak Lewis acids, whereas pyranosides
without oxazolidinones did not endo-cleave. The results
from experiments performed on compounds related to the
models are in good agreement with those from the compu-
tations. The calculated structural features indicate that the
strained oxazolidinone groups make endo-cleavage more
feasible. Several factors can influence the stabilization of
the TS energies of pyranosides with oxazolidinone groups,
such as fixing the conformation and/or electronic effects
from the oxazolidinones. Our calculations suggest that there
should be at least three relevant variables to model the reac-
tion coordinates. There are still further possibilities that
lower the reaction paths. Further experimental and theoret-
ical studies to determine the characteristics and precise re-
action paths of the pyranosides with oxazolidinones whilst
considering these variables will reveal the detailed nature of
the anomerization reactions.
[2] C. B. Post, M. Karplus, J. Am. Chem. Soc. 1986, 108, 1317–
1319.
[3] R. B. Gupta, R. W. Franck, J. Am. Chem. Soc. 1987, 109,
6554–6556.
[4] D. R. McPhail, J. R. Le, B. Fraser-Reid, J. Am. Chem. Soc.
1992, 114, 1905–1906.
[5] a) J. L. Liras, E. V. Anslyn, J. Am. Chem. Soc. 1994, 116, 2645–
2646; b) J. L. Liras, V. M. Lynch, E. V. Anslyn, J. Am. Chem.
Soc. 1997, 119, 8191–8200.
[6] a) Y. Guindon, P. C. Anderson, Tetrahedron Lett. 1987, 28,
2485–2488; b) R. Olsson, U. Berg, T. Frejd, Tetrahedron 1998,
54, 3935–3954; c) C. O’Brien, M. Poláková, N. Pitt, M. Tosin,
P. V. Murphy, Chem. Eur. J. 2007, 13, 902–909 and references
cited therein.
[7] P. Deslongchamps, S. Li, Y. L. Dory, Org. Lett. 2004, 6, 505–
508.
[8] a) K. Benakli, C. Zha, R. J. Kerns, J. Am. Chem. Soc. 2001,
123, 9461–9462; b) R. J. Kerns, C. Zha, K. Benakli, Y.-Z. Li-
ang, Tetrahedron Lett. 2003, 44, 8069–8072; c) P. Wei, R. J.
Kerns, Tetrahedron Lett. 2005, 46, 6901–6905; d) M. Boysen,
E. Gemma, M. Lahmann, S. Oscarson, Chem. Commun. 2005,
3044–3046; e) P. Wei, R. J. Kerns, J. Org. Chem. 2005, 70, 4195–
4198; f) M. L. Bohn, M. I. Colombo, C. A. Stortz, E. A.
Rúveda, Carbohydr. Res. 2006, 341, 1096–1104; g) Y. Geng, L.-
H. Zhang, X.-S. Ye, Chem. Commun. 2008, 596–599; h) Y.
Geng, L.-H. Zhang, X.-S. Ye, Tetrahedron 2008, 64, 4949–4958.
[9] a) D. Crich, A. U. Vinod, J. Org. Chem. 2005, 70, 1291–1296;
b) S. Manabe, K. Ishii, Y. Ito, J. Am. Chem. Soc. 2006, 128,
10666–10667.
Experimental Section
To a solution of compound 5a (66.9 mg, 0.128 mmol) in CH2Cl2
(1.7 mL) was added BF3·OEt2 (32 µL, 0.259 mmol) at 0 °C. After
the mixture was stirred at 0 °C for 1 h, saturated NaHCO3 was
added. The aqueous layer was extracted with CHCl3. The com-
bined layers were washed with brine and dried with Na2SO4. After
filtration, the solvent was removed. The residue was purified by
preparative TLC (CHCl3/EtOAc, 20:1) to give 5a (50.6 mg, 76%)
and 5b (3.8 mg, 6%).
[10] J. D. M. Olsson, L. Eriksson, M. Lahmann, S. Oscarson, J.
Org. Chem. 2008, 73, 7181–7188.
[11] S. Manabe, K. Ishii, D. Hashizume, H. Koshino, Y. Ito, Acta
Crystallogr., Sect. E. 2008, 64, o1868-o1868.
Supporting Information (see also the footnote on the first page of
this article): Details of computations and experiments are provided.
[12] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. A. Montgomery Jr, T. Vreven,
K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tom-
asi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega,
G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota,
R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratch-
ian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gom-
perts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C.
Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A.
Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dap-
prich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick,
A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q.
Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov,
G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin,
D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayak-
kara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen,
M. W. Wong, C. Gonzalez, and J. A. Pople, Gaussian 03, Re-
vision C.02, Gaussian, Inc., Wallingford CT, 2004.
[13] Y. Matsuzaki, Y. Nakahara, Y. Ito, T. Ogawa, Tetrahedron Lett.
1993, 34, 1061–1064.
Acknowledgments
Calculations were executed with Obelix of ETH Zürich, a cluster
computer of the Competence Center for Computational Chemistry
(C4). The HA8000 cluster computer (t2k) at Tokyo University and
the RIKEN super combined cluster (RSCC) system were also used
in part. S. M. (RIKEN) thanks Ms. Akemi Takahashi (RIKEN)
for her technical assistance and Ms. Kaori Otsuki and Mr. Masaya
Usui at the Research Resource Center of RIKEN’s Brain Science
Center for performing the high-resolution MS measurements. S. M.
also expresses thanks for the Grant-in-Aid for Scientific Research
(C) (Grant No. 19590032) from the Japan Society for the Pro-
motion of Science and an Incentive Research Grant from RIKEN.
We thank Professor David M. Birney (Texas Tech University) for
valuable discussions.
[1] a) H. Cordes, H. G. Bull, Chem. Rev. 1974, 74, 581–603; b)
A. J. Kirby, Acc. Chem. Res. 1984, 17, 305–311; c) T. H. Fife,
Received: November 17, 2008
Published Online: January 29, 2009
Eur. J. Org. Chem. 2009, 1127–1131
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1131