Differential reactivity of α- and β-anomers of glycosyl acceptors in glycosylations. A remote consequence of the endo-anomeric effect?

Article abstract of DOI:10.1021/ol006039q

(equation presented) When phenyl tri-O-benzyl-1-thio-β-D-galactopyranosiduronic acid esters were coupled with a 1/1 mixture of a and β 2,3 di-O-protected D-galactopyranosiduronic acid esters, the β-anomer proved to be more reactive. Data from theoretical

Full text of DOI:10.1021/ol006039q

ORGANIC
LETTERS
2000
Vol. 2, No. 15
2275-2277
Differential Reactivity of r- and
â-Anomers of Glycosyl Acceptors in
Glycosylations. A Remote Consequence
of the endo-Anomeric Effect?
Didier Magaud, Rene´ Dolmazon, Daniel Anker, and Alain Doutheau*
Laboratoire de Chimie Organique, Institut National des Sciences Applique´es,
20 aVenue A. Einstein, 69621 Villeurbanne, France
Yves L. Dory and Pierre Deslongchamps
Laboratoire de Synthe`se Organique, De´partement de Chimie,
UniVersite´ de Sherbrooke, Sherbrooke (QC), Canada JIH 5N4
doutheau@insa-lyon.fr
Received May 10, 2000
ABSTRACT
When phenyl tri-O-benzyl-1-thio-â-D-galactopyranosiduronic acid esters were coupled with a 1/1 mixture of r and â 2,3 di-O-protected
D-galactopyranosiduronic acid esters, the â-anomer proved to be more reactive. Data from theoretical calculations suggested that the enhanced
reactivity of this anomer compared with the r one would be due to a stronger hydrogen bond of the C-4 OH with the ring oxygen.
We recently reported that direct coupling between two
D-galacturonic acid esters can be performed in good yields
when 1-thioglycosides are used as donors.1 For example, the
condensation, at - 60 °C, between 1a and 2â (Figure 1) in
the presence of N-iodosuccinimide and triflic acid used as
promoters afforded disaccharide 4a (Figure 2) in 88% yield.1b
In contrast, when 1a was reacted with disaccharide acceptor
4b we observed that the reaction became sluggish and had
to be conducted at room temperature, giving rise to the
expected trisaccharide in only 45% yield.1a We first envisaged
that the difference in reactivity between 2â and 4b could be
due to steric effects. Indeed, the presence of the R-oriented
bulky substituant at C-1′ in 4b would force the C-2′ and
* INSA. Phone (33) 4 72 43 82 21. Fax (33) 4 72 43 88 96.
(1) (a) Magaud, D.; Grandjean, C.; Doutheau, A.; Anker, D.; Shevchik,
V.; Cotte-Pattat, N.; Robert-Baudouy, J. Tetrahedron Lett. 1997, 38, 241-
244. (b) Magaud, D.; Grandjean, C.; Doutheau, A.; Anker, D.; Shevchik,
V.; Cotte-Pattat, N.; Robert-Baudouy, J. Carbohydr. Res. 1998, 314, 189-
199.
Figure 1. Monosaccharide reactants.
10.1021/ol006039q CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/21/2000

their favored 4C1 conformations, the C-4 hydroxyl group does
not present steric compression and remains accessible.
Therefore, we decided to examine the relative reactivity of
3R and 3â, the analogues of compounds 2 bearing methoxy
groups at C-1, C-2, and C-3. We reasoned that, in these
compounds, it would be rather unlikely that the steric
hindrance around the 4-OH group would depend on the
anomeric configuration.
The known acceptors6 3R and 3â were obtained in
respectively 64% and 33% yields from methyl 2,3-di-O-
methyl R- and â-D-galactopyranosides7 by selective oxida-
tion of the C-6 hydroxyl group (Pt/C, O2, NaHCO3, H2O,
95 °C) followed by treating the resulting sodium carboxylate
with methyl iodide in DMF. When a mixture of 3R (0.5
equiv) and 3â (0.5 equiv) was reacted with 1a (0.6 equiv),
we again observed that 3â was more rapidly consumed than
3R and we obtained a mixture of 4e and 4f (ratio 4e/4f )
4:18) in 83% yield based on donor 1a.
To rationalize the differential reactivity9 between the R-
and â-anomers in acceptors 2 or 3, we formulated the
following hypothesis: Due to the absence of an endo-
anomeric effect10 the basicity of the pyranosyl oxygen atom
in â-anomers would be greater than that in the R ones.11 As
a consequence, in the â-anomers the intramolecular hydrogen
bonding between the axial 4-OH group and its proximal
pyranosyl oxygen would be stronger than that in the
R-anomers and therefore the C-4 alcoholic function should
be more nucleophilic.12 This hypothesis is supported by the
results of theoretical ab initio calculations13 (Figure 3) carried
Figure 2. Disaccharide reactants and products.
C-3′ benzyloxy groups on the upper face and thus reduce
the accessibility of the C-4′ OH group. However, we also
noticed that a structural difference between the two acceptors
was the configuration of the anomeric carbons C-1 and C-1′,
respectively â in compound 2â and R in compound 4b. We
thus found it interesting to compare the relative reactivity
of the two anomeric acceptors 2R and 2â when simulta-
neously reacted with donor 1b.2
When coupling a mixture of 2R (0.5 equiv) and 2â (0.5
equiv) with donor 1b (0.6 equiv) under usual conditions,1b
monitoring of the reaction by TLC indicated that the former
compound reacted faster. After purification of the crude
product we obtained, in 67% yield based on donor, a mixture
of 4c and 4d in the ratio 4c/4d ) 4:1.3 As already envisaged
for acceptors 2â and 4b, the difference in reactivity between
2R and 2â would again result from a greater steric hindrance
around the 4-OH group in the former compound. However
molecular modeling4 data did not confirm this hypothesis
and failed to provide a rationale for the differential reactivity
between 2R and 2â based on steric considerations.5 Analysis
of conformational space showed that in both compounds in
(2) Compounds 1a-c and 2â were prepared according to ref 1b. The
acceptor 2R was obtained as follows: 1c was transformed into the
corresponding silyl ether (TBDMSOTf, 2,6-lutidine, CH2Cl2, 100%) which
was reacted with benzyl alcohol (NIS, TfOH, CH2Cl2/Et2O 1/2, 4 Å MS,
25 °C, 3 h) to give, in 75% yield, a mixture of anomeric benzyl glycosides
(R/â ) 1:1.5) which were separated by column chromatography. Cleavage
of the TBDMS group in the R-anomer (TFA, 10% H2O, 1 h 30 min, 93%)
gave 2R.
Figure 3. Ab initio 6.31G* geometries.
(3) This ratio was determined from the 1H NMR spectrum of the mixture
on the integration of the hydrogen signals at C-1′: 4c, 5.32 (d, 0.8 H, J1′-2′
) 2.9, H-1′), 4d, 5.20 (d, 0.2 H, J1′-2′ ) 3.3, H-1′). The major isomer in
the mixture was deduced from the integration of signals at 3.77 and 3.46
corresponding respectively to H-2 and H-3 in 4c (values to be compared
with 3.77 and 3.52 in 2â Versus 3.87 and 3.97 in 2R). The ratio 4c/4d was
confirmed by the ratio of recovered unreacted acceptors (2R/2â ) 80:20)
determined from the 1H NMR spectrum of the mixture (integration of H-2
and H-3 hydrogen signals).
(4) (a) Boswell, D. R.; Coxon, E. E.; Coxon, J. M. In AdVances in
Molecular Modeling; Liotta, D., Ed.; Jai Press Inc.: Greenwich, 1995; Vol.
3. (b) Leach, A. R. In Molecular Modeling, Principles and Applications;
Longman, 1996.
out on 5-hydroxy-2-methoxytetrahydropyrans 5 used as
models of R and â galactopyranosides: the distance between
the two involved atoms would be respectively 2.37 Å in 5â
(6) Synthesis of 3R: Grishkovets, V. I.; Zemlyakov, A. E.; Chirva, V.
Y. Chem. Nat. Compd. (Engl. Transl.) 1983, 19, 522-524. Synthesis of
3â: Evtushenko, E. V.; Ovodov, Y. S. Chem. Nat. Compd. (Engl. Transl.)
1987, 23, 28-30.
(7) Williams, N. R.; Jeanloz, R. W. J. Org. Chem. 1964, 29, 3434-
3435.
(8) Column chromatography of the crude product gave pure samples of
4e and 4f and a mixture of these two compounds. In this mixture the ratio
4e/4f was determined from the integration of the signals of hydrogens H-1′
and H-5.
(9) This 4:1 ratio corresponds to a ratio of about 7 for the rate constants
for the glycosylation reactions with respectively â and R acceptors.
(5) The molecular dynamic trajectories were calculated using version
6.01 of SYBYL (SYBYL Molecular Modeling System, Tripos Associates
Inc., St. Louis, MO, 1993). The simulation of the two anomers was initiated
with zero kinetic energy using a time step of 0.5 fs, at 300 K. The lengths
of the trajectories are 1.000 ps.
2276
Org. Lett., Vol. 2, No. 15, 2000

and 2.43 Å in 5R in favor of a stronger hydrogen bonding
in the former compound. A similar distance difference was
observed when modeling hydroxonium species 6R and 6â
designed to mimic positively charged cationic intermediates
formed by nucleophilic attack of acceptors on activated
donors during glycosylation reactions.
â-anomers is far less pronounced than that with their
galacturonic analogues. The residual slightly enhanced
reactivity of the â-anomer could be due to the delocalization
of oxygen nonbonding electrons into the σ* orbital of the
C4-C5 bond (Figure 5) enhancing the electron density at
C-4 and consequently on the C-4 OH group.16
To try to confirm the above hypothesis, we compared the
reactivity of glucuronic acid ester derivatives 7 (Figure 4).
Figure 5. n f σ* delocalization.
Again this delocalization would be more important in the
â than in the R-anomer due to the greater basicity of the
ring oxygen in the former compound. Evidently the same
effect would also contribute to some extent to the enhance-
ment of the C-4 OH reactivity in acceptors 2â or 3â.
In conclusion, we report here that the reactivity of
galactopyranosiduronic acid esters possessing a C-4 free
hydroxyl depends significantly on the anomeric configura-
tion. To the best of our knowledge such clear-cut differences
of behavior between R- and â-anomers have been rarely
observed.17 Data from theoretical calculations suggested that
the C-4 alcohol would be more nucleophilic in the â- than
in the R-anomers because of a stronger hydrogen bonding
of the OH group, acting as H-donor, with the ring oxygen.
The greater reactivity of the â-anomers would also be due
to enhancement of the electron density on the C-4 OH group
in these compounds due to a more important n f σ*
delocalization. These two effects would result from the
greater basicity of pyranosyl oxygen atom in the â- than in
the R-anomers due to the absence of an endo-anomeric effect
in the former compounds.18 We will now examine if a similar
differential reactivity between anomers is also observed with
galactose or L-arabinopyranose derivatives.
Figure 4. Glucuronic acid reactants and products.
Acceptors 7R and 7â were obtained by following the pro-
cedure used for the preparation of compounds 3, from known
methyl 2,3-di-O-methyl R- and â-D-glucopyranosides.14
When a mixture of acceptors 7R (0.5 equiv) and 7â (0.5
equiv) was reacted with donor 1b (0.6 equiv), we obtained
a mixture of disaccharides 8R and 8â (in 62% yield based
on donor 1b and with a 8â/8R ratio of 1.2/1); some unreacted
acceptors 7 were also recovered as a 1.1/1 mixture of 7R
and 7â.15
Thus, as expected, with glucuronic acceptors 7, in which
the equatorial orientation of the C-4 OH group prevents the
formation of an internal hydrogen bond with the pyranosyl
oxygen atom, the difference in reactivity between the R- and
Acknowledgment. Some of us (D.M.,D.A., A.D.) thank
the European Community (Contract AIR2-CT94-1345) for
financial support and undergraduate students for technical
assistance.
(10) (a) Deslongchamps, P. In Stereoelectronic Effects in Organic
Chemistry; Baldwin, J. E., Ed.; Pergamon Press: Oxford, England, 1983.
(b) Juaristi, E.; Cuevas, G Tetrahedron 1992, 48, 5019-5087. (c) Kirby,
A. J.; Williams, N. H. In The Anomeric Effect and Associated Stereoelec-
tronic Effects; Thatcher, G. R., Ed.; ACS Symp. Series 539; American
Chemical Society: Washington, DC, 1993.
(11) For the influence of the anomeric configuration of 1,5-dithiogluco-
pyranosides on the ring sulfur nucleophilicity, see: Yuasa, H.; Kamata,
Y.; Hashimoto, H. Angew. Chem., Int. Ed. Engl. 1997, 36, 868-870.
(12) For hydrogen bonding between C-4 axial groups and ring oxygen
in six-membered rings, see: (a) Alonso, J. L.; Wilson, E. B. J. Am. Chem.
Soc. 1980, 102, 1248-1251. (b) Kwon, O.; Danishefsky, S. J. J. Am. Chem.
Soc. 1998, 120, 1588-1599.
Supporting Information Available: Typical experimen-
tal procedure for competitive glycosylation reactions. Full
1
characterization for compound 2R. H NMR spectra for
compounds 3R and 3â . 1H NMR and 13C NMR spectra and
optical rotations for compounds 7R and 7â. This material is
available free of charge via the Internet at http://pubs.acs.org.
(13) The calculations were carried out at the RHF 6.31G* level of theory
by means of GAMESS; see: Schmidt, M. W.; Baldridge, K. K.; Boatz, J.
A.; Elbert, S. T.; Gordon, M. S.; Jensen, J. H.; Koseki, S.;. Matsunaga, N.;
Nguyen, K. A.; Su, S. J.; Windus, T. L.; Dupuis, M.; Montgomery, J. A. J.
Comput. Chem. 1993, 14, 1347-1363.
OL006039Q
(16) Fan, Y.-H.; Haseltine, J. Tetrahedron Lett. 1996, 37, 9279-9282.
(17) For recent examples see: (a) Zhu, X. X.; Cai, M. S.; Zhou, R. L.
Carbohydr. Res. 1997, 303, 261-266. (b) Rochepeau-Jobron, L.; Jacquinet,
J.-C. Carbohydr. Res. 1997, 305, 181-191.
(18) Even if it turned out that the R-anomeric configuration of C-1′ in
4b was unfavorable for the glycosylation, it seems that this factor could
not account alone for the poor reactivity of this acceptor when coupled
with 1a. The difficulty of undergoing coupling in this case could be also
due, in part, to a steric mismatch between donor and acceptor.
(14) Nicoll-Griffith, D. A.; Weiler, L. Tetrahedron 1991, 47, 2733-
2750.
(15) In this case, due to the weaker reactivity of acceptors 7 compared
with acceptors 3, the glycosylation had to be carried out at -30 °C.
Disaccharides 8R and 8â and recovered acceptors 7R and 7â were separated
by column chromatography of the crude product. Structural assignments
for compounds 8 were based on the signals for H-2 (3.30, dd, J2-3 ) 9.6,
J2-1 ) 3.4) in 8R and for H-1 (4.18, d, J1-2 ) 7.5) for 8â.
Org. Lett., Vol. 2, No. 15, 2000
2277