Regioselectivity of the glycosylation of N-dimethylmaleoyl-protected hexosamine acceptors. An experimental and DFT approach

Article abstract of DOI:10.1039/c1ob00021g

Both anomers of the methyl glycoside of 6-O-benzyl-N-dimethylmaleoyl-d- allosamine (6 and 7) are glycosylated exclusively on O3 when reacting with the trichloroacetimidate of peracetylated α-d-galactopyranose (5). This regioselectivity is expected for 6,

Full text of DOI:10.1039/c1ob00021g

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 3020
www.rsc.org/obc
PAPER
Regioselectivity of the glycosylation of N-dimethylmaleoyl-protected
hexosamine acceptors. An experimental and DFT approach†
Mar´ıa I. Colombo,*a Edmundo A. Ru´vedaa and Carlos A. Stortz*b
Received 5th January 2011, Accepted 9th February 2011
DOI: 10.1039/c1ob00021g
Both anomers of the methyl glycoside of 6-O-benzyl-N-dimethylmaleoyl-D-allosamine (6 and 7) are
glycosylated exclusively on O3 when reacting with the trichloroacetimidate of peracetylated
a-D-galactopyranose (5). This regioselectivity is expected for 6, the a-anomer, as a strong hydrogen
bond of its H(O)3 with the carbonyl group of the dimethylmaleoyl group occurs, as shown by NMR
temperature dependence. However, this hydrogen bond was not encountered experimentally for 7, the
b-anomer. A DFT study of the energies implied in an analog of the glycosylation reaction charged
intermediate has explained neatly this behavior, in terms of strong hydrogen bonds occurring at these
charged intermediates. This approach explains both the experimental regioselectivities found for 6 and
7, but furthermore the calculations have shown a marked agreement with the regioselectivities found for
other related compounds in the literature.
would allow glycosylations to be carried out without the need
Introduction
(or with minimal need) of protections. Several interesting studies
The increasing importance of the roles that oligosaccharides
directed to understand the factors that affect the regioselectivity
and glycoconjugates play in biological processes has led to a
of glycosylation reactions have been reported. In a series of
demand for reliable new procedures for their synthesis. A great
publications, Vasella and co-workers pointed out the importance
deal of attention has been devoted to glycosylation reactions and,
of hydrogen bonds in diol acceptors mainly in reactions with
diazirine glycosyl donors.5,6 The group of Fraser-Reid and Lo´pez
have shown that the nature (armed or disarmed) of the glycosyl
donor has a great influence on the regioselectivity of glycosylation
reactions. These observations were the basis for the development
of the new concept for “matching” donors with acceptors called
“reciprocal donor acceptor selectivity” (RDAS).7 More recently,
the group of Mart´ın-Lomas established, using experimental and
theoretical studies,8 that the regioselectivity for the glycosylation of
a given hydroxyl group can be considerably activated by increasing
its nucleophilicity, for the presence of groups able to form a
hydrogen bond with that hydroxyl group.
Given the biological significance of hexosamines,9 we have pre-
viously analyzed the regioselectivity of the 1,2-trans-diequatorial
diols of a- and b-anomers of 6-O-substituted N-dimethylmaleoyl-
protected D-glucosamine acceptors, using two disarmed donors
and also made theoretical calculations justifying the experimental
trends.10,11 More recently, we reported on the assessment of
the reactivity of isomeric N-dimethylmaleoyl (DMM) 4,6-O-
benzylidene protected D-glucosamine and D-allosamine acceptors
(1, 2, 3 and 4) through competition experiments with the disarmed
donor 5. The order of reactivity 3 ꢀ 1 > 4 > 2 suggested that
the strong hydrogen bond between the H(O)3 with a carbonyl of
the DMM, determined by 1H NMR and confirmed by modelling
experiments, activated O3 by increasing its nucleophilicity, and
compensating the steric hindrance expected for the most reactive
(axial) acceptor.12 In the formation of this critical hydrogen bond
consequently, the chemical synthesis of most glycosidic linkages
can now be readily achieved.1 However, the stereo- and regio-
chemical outcome of glycosylation reactions is often difficult to
predict. In fact, small changes in the structure of the glycosyl
donor and acceptor, such as protecting groups or stereochemistry,
often cause dramatic changes in the outcome of the reaction.2–4
The control of the regioselectivity in the synthesis of oligosac-
charides usually requires the use of protecting groups in order
to select reactions at a specific secondary hydroxyl group, as
they have often similar reactivities. This makes synthetic schemes
lengthy, laborious and time-consuming. However, sometimes the
reactivity of some of those hydroxyl groups is much larger than
that of their neighbors, and thus regioselective reactions can be
carried out in the presence of limiting amounts of glycosyl donor.
Consequently, a better comprehension of the regioselectivity
aInstituto de Qu´ımica Rosario (CONICET-UNR), Facultad de Ciencias
Bioqu´ımicas y Farmace´uticas, Suipacha 531, 2000, Rosario, Argentina.
E-mail: colombo@iquir-conicet.gov.ar; Fax: 54 341 437 0477; Tel: 54 341
437 0477
bDepartamento de Qu´ımica Orga´nica-CIHIDECAR, Facultad de Ciencias
Exactas y Naturales, Universidad de Buenos Aires, Ciudad Universitaria,
1428, Buenos Aires, Argentina. E-mail: stortz@qo.fcen.uba.ar; Fax: 54 11
4576 3346; Tel: 54 11 4576 3346
† Electronic supplementary information (ESI) available: The NMR spectra
of compounds 6, 7, 8a and 10a, and the DFT-calculated Cartesian
coordinates of all the methylated intermediates under study. See DOI:
10.1039/c1ob00021g
3020 | Org. Biomol. Chem., 2011, 9, 3020–3025
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
the carbonyl of DMM plays a fundamental role, increasing the
benefits of using this protecting group.13
group with the C O group of the DMM moiety, by increasing
its nucleophilicity, compensated the expected steric hindrance.
Furthermore, this result is also in agreement with the report of
Maloisel and Vasella regarding the regioselective glycosidation at
O3 of the related N-phthaloylallosamine-derived acceptor (9) with
a trichloroacetimidate donor.16
Based on these results, we decided to study the differential
O3/O4 regioselectivity of the 1,2-cis-diols of the anomers of
methyl glycosides of N-DMM-D-allosamines 6 and 7, in an
attempt to advance our understanding of the electronic factors
governing the regioselectivity in the glycosylations of N-protected
hexosamine derived acceptors.
In the reaction mixture of the glycosidation of diol 7 only one
disaccharide was also detected (TLC, 1H NMR). This was shown
to be the 1→3-linked disaccharide 10a, isolated in 94% yield.
The regiochemistry of 10a was established by using the same
methodology described above for 8a (dH 4.98, H4 for the acetate
10b). The clear preference for the glycosidation at the axial O3
instead of occurring at the equatorial O4 of 7, showing both weak
hydrogen bonds, was surprising at first.
Furthermore, we have attempted to rationalize these observa-
tions using modelling tools which analyze the key step for the
glycosidation reaction. This analysis showed that the experimental
trends were matched by the theory, and it was also used with several
literature cases in order to show its ability to predict the trends in
other competitive reactions.
These results indicated that the predictive capability of the
regioselectivity outcome for acceptors exhibiting only weak hy-
drogen bonds is much more complex than for those, such as diol
6, showing a strong intramolecular hydrogen bond. Consequently,
in an attempt to rationalize these observations we turned our
attention to a more refined approach, focused on the mechanism
of the glycosidation reaction.
Results and discussion
The anomeric D-allosamine acceptors 6 and 7 were readily
obtained by reductive opening of the benzylidene protecting
groups of 3 and 4. With these diols in hand we first carried out an
analysis of their 1H NMR spectra in DMSO-d6 to establish if their
hydroxyl groups showed different abilities to form intramolecular
H-bonds in solution. The Dd/DT values obtained for 6 clearly
indicated a strong intramolecular H-bond for H(O)3, whereas a
weak one was detected for the H(O)4 (-3.5 and -7.1 ppb K-1,
respectively).5,6,14,15 For 7, the values obtained suggested at best
weak intramolecular H-bonds (H(O)3, -6.2 and H(O)4, -5.6 ppb
K-1, respectively). On the basis of these Dd/DT values and on the
reactivity observed with acceptors 1–4, we expected that acceptor
6 would yield mainly the 1→3-linked disaccharide when coupled
with the disarmed donor 5, whereas a mixture of 1→3- and 1→4-
linked disaccharides should be expected for acceptor 7.
When the a-anomer of diol (6) was coupled with 1.1 equiv. of
donor 5, under activation with trimethylsilyl triflate (TMSOTf)
at -25 ◦C, the only compound, obtained in 87% yield, was
the 1→3-linked disaccharide 8a. The corresponding 1→4-linked
compound was neither detected by TLC nor by 1H NMR
spectroscopy. The assignment of the regioisomeric structure of
8a was readily accomplished by the presence of H1¢→C3 and
H3→C1¢ correlations in its HMBC spectrum and was confirmed
by acetylating the disaccharide and determining the chemical shift
of the newly downfield-shifted proton (dH 5.05, H4 for the acetate
8b). This result is in complete agreement with our observation that
in the a acceptor, the strong hydrogen bond of the axial H(O)3
As nowadays is generally accepted, we assumed that the
glycosidation reaction proceeds through the classical three (or
four) steps of an SN1-like mechanism.3,17 However, we were not
interested in the usually irreversible ionization step, the rate-
determining step in classical SN1 mechanistic discussions as this
step would be the same for every spot in the acceptor. Rather,
inspired by the report of Whitfield et al. we have paid specific
attention to the stabilities of the charged intermediates formed in
the second step, that is, in the nucleophilic attack of the acceptors
on the intermediate formed as a result of the leaving group
departure in the donor.2,18 We have also taken into consideration
in the present analysis that Cid et al. pointed out that the observed
weak hydrogen bonds of the acceptors might not explain the
stabilities of the transition states leading to the intermediates and,
finally, to the disaccharides.8 Recently, these authors, in their study
on the regioselective glycosylation of inositol-derived diols and
other isomeric diol acceptors, found, by DFT modelling of the
acceptors and the corresponding complexes prior to transition
states, that the weaker hydrogen bonds in the ground state of the
acceptors became much more important in these complexes, close
in geometry and energy to the transition states.8
Moreover, Whitfield et al. made a theoretical study directed
to rationalize the stereochemical outcome of glycosidation reac-
tions by analyzing the conformations of the electrophilic species
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3020–3025 | 3021
©

View Article Online
generated from the donor, and decided to replace the acceptor by a
simple methanol molecule in order to simplify the calculations.2,18
By joining the approaches of the groups of Whitfield and Mart´ın-
Lomas, and the accepted mechanism of the reaction3,17 we decided
to study the stability (energetics) of the charged intermediates
using a methyl group as the donor. Such a simple approach
assumes the attack of either of both hydroxyl groups to a methyl
carbocation, in an SN1-like fashion. The intermediate has a (+1)
charge, thus giving rise to very strong hydrogen bonds not present
in the free acceptor. The balance of hydrogen bonds should show
a better stabilization of the intermediate with the methoxyl group
on either oxygen, and this result could be then compared with
the experimental regioselectivities. This approach assumes that
the energy barriers from the intermediates to the disaccharides
are very similar, i.e. that the relative energies of the intermediates
and of their corresponding transition states are similar. One way
of testing this approach is to analyze the agreement between
the relative stabilities of the intermediates and the observed
experimental regioselectivities.
Fig. 2 Molecular representations of the most stable conformers of
the compound 7m methylated on O3 and O4, calculated at the
B3LYP/6-31+G** level. The H–O distances for the strong hydrogen bonds
are indicated.
compound. The experimental result is once more giving ground
to the theoretical approach: compound 7 shows also a marked
preference for O3 substitution.
In these examples the H(O)3 and the O C of the DMM group
are in the most favorable position to form a strong hydrogen
bond, as O3 is axial. It might thus be arguable if this agreement
between the experimental and theoretical data will hold for other
N-dimethylmaleoyl D-hexosamine acceptors. Thus, we decided
to apply the same approach to the D-glucosamine acceptors
11 and 12. For these acceptors carrying benzoyl and benzyl
groups (11a/11b and 12a/12b), we had already determined the
differential O3/O4 regioselectivity using two disarmed donors
and also made theoretical calculations11 on their simpler analogs
(11f/11m and 12f/12m) justifying the experimental trends.10
Furthermore, we have currently done the temperature dependence
NMR experiments in DMSO-d6 solution carried out on the
acceptors themselves. Their values (see the Experimental section)
indicated only weak intramolecular hydrogen bonds for their
hydroxyl groups.
Analogs of 6 and 7 carrying a methyl group on O6 instead of
a benzyl group (6m and 7m, respectively) were used as the basis
acceptors to which methyl groups were added on O3 and O4. These
were analyzed by DFT at the B3LYP/6-31+G(d,p) level. Several
different conformers were found for each adduct, depending on
the orientation of the exocyclic groups. For 6m, the two more stable
for the substitution on O3 and O4 are shown in Fig. 1.
Fig. 1 Molecular representations of the most stable conformers of
the compound 6m methylated on O3 and O4, calculated at the
B3LYP/6-31+G** level. The H–O distances for the strong hydrogen bonds
are indicated.
The structure methylated on O3 is more stable than the second
one by 8.64 kcal mol-1. In the former, a very strong hydrogen
bond is formed between the proton on O3 and the carbonyl of
the DMM protecting group. This is the same bond occurring in
the uncharged ground state acceptor, but much stronger because
of the charge. The proton is actually in-between both oxygen
atoms, and closer to the oxygen of the carbonyl group (Fig. 1).
The hydrogen bond occurring between H(O)4 and O6 in the O4-
methylated adduct is strong, but weaker than that occurring in the
O3-methylated adduct. This result matches the clear preference
for the glycosidation at O3 found experimentally for compound 6.
A similar analysis for compound 7m leads to the equivalent
conformers shown in Fig. 2. Although the uncharged compound
7 does not show experimentally a strong hydrogen bond to
the carbonyl group, its reaction intermediate seems to give
a very strong one, thus stabilizing again the compound with
O3 methylation by 6.93 kcal mol-1 against the O4-methylated
Different conformers of the O3- and O4-methylated adducts of
11f, 11m, 12f and 12m were analyzed by DFT. For the most stable
compounds, the results are shown in Table 1. An astonishing
match between the theoretical results and the experimental
regioselectivities is observed: 11a/11f show preference for O3
substitution and show a higher stability for the O3-methylated
intermediate, whereas 11b/11m show the opposite trend both
experimentally and theoretically. The other two cases follow the
same trend. Furthermore, the energy values are especially adjusted
to the reaction with the furanosyl donor 13: both methylated
compounds have about the same energy as 12f, for which parent
compound 12a no regioselectivity was found experimentally. The
other three compounds show the sign and magnitude of the energy
difference in agreement with the observed regioselectivity.10
We have previously assessed that the strong hydrogen bond be-
tween H(O)3 and the O C of DMM is responsible for the higher
3022 | Org. Biomol. Chem., 2011, 9, 3020–3025
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Table 1 Differences in energy for the O3- and O4-methylated adducts of
11m, 11f, 12m, and 12f, together with the experimental regioselectivities
of their parent compounds with the donors 5 and 13
has been shown experimentally to occur when comparing the
reactivities of 1 and 2.12
We are aware that in the current theoretical approach an
overestimation of the hydrogen bonds is feasible, as other factors
not considered (solvent, additives, counterion, etc.) are present.
However, we understand that this approach is useful to identify
the trends which might be useful to help in the prediction of
regioselectivity in glycosylation reactions.
Ratio O3/O4
Ratio O3/O4
EMeO3 - EMeO4/
Compound
with donor 5a
with donor 13a
kcal mol-1
11a/11f
11b/11m
12a/12f
12b/12m
2 : 1
1 : 1
1 : 13
0 : 1
1 : 0
3.2 : 1
1 : 1
-4.60
-1.85
-0.03
+2.15
1 : 2.9
a From ref. 10
Experimental
General
stability of O3-methylated adduct. This is true for both a-anomers
11f and 11m, for which a torsion of the DMM group is needed
for the hydrogen bond to occur, and where the a-methoxyl group
promotes this torsion. For the b-anomers 12f and 12m, the torsion
of the DMM group is not straightforward: it requires energy to
occur.11 For the main O3-methylated conformers the DMM group
leans in order to generate a hydrogen bond, but this is not as
favorable as occurred for a-anomers, thus destabilizing partially
these structures. The O4-methylated compounds gain ponderation
for these anomers, especially when a methyl group is present on
O6 (12m), as this oxygen atom has more electron availability for
a H(O)4 bond than that occurring on a formylated/benzoylated
derivative (12f).
A further example of the application of this approach can
be given using a non-nitrogenated acceptor like Vasella’s 14a
and 14b, 4,6-O-benzylidene-D-allose methyl glycosides having free
O2 (equatorial) and O3 (axial).6 The present approach (using a
ethylidene group instead of a benzylidene group) shows a higher
stability for the O3-methylated adduct of 4.39 kcal mol-1 (14a) and
2.25 kcal mol-1 (14b). The experimental reactivities against donor
15 favored O3 by 2.6 : 1 (14a) and 1.2 : 1 (14b), showing once again
that the experimental and theoretical trends match.
Melting points were determined on an Electrothermal 9100
apparatus and are uncorrected. The H and 13C NMR spectra
1
were recorded on Bruker Avance 300 spectrometer. For the
2D experiments, Bruker standard software was employed. High
resolution mass spectrometry (HRMS-ESI) was performed in
a Bruker microTOF-Q II instrument. Optical rotations were
measured with a Jasco DIP-1000 polarimeter. Column chro-
matography was performed on Silica Gel 60 H, slurry-packed,
run under low pressure of nitrogen and employing increasing
amounts of EtOAc in hexane as solvent. Analytical TLC was
carried out using Kieselgel GF254 (E. Merck) with a thickness
of 0.20 mm. The homogeneity of all compounds prior to the
high-resolution mass spectral determination was carefully verified
by TLC. Reactions were routinely run under a dry nitrogen
atmosphere with magnetic stirring. All chemicals were used as
purchased or purified according to standard procedures.
Methyl
6-O-benzyl-2-deoxy-2-dimethylmaleimido-a-D-allo-
pyranoside (6). To a solution of the 4,6-O-benzylidene acetal
3 (630 mg, 1.61 mmol) and BH3·N(CH3)3 (230 mg, 3.15 mmol)
in CH3CN (16.5 ml) in an ice-water bath was added dropwise
BF3·OEt (0.407 ml, 3.73 mmol). After 1 h at this temperature, the
solution was stirred at room temperature for an additional 0.5 h
(TLC), and NaHCO3 (243 mg) was then added, and the solution
was evaporated to dryness. The crude product in Cl2CH2 was
washed with brine, dried (Na2SO4), and evaporated. The residue
was chromatographed to yield 6 (426 mg, 67% yield), as a foamy
solid [a]2D7 +125.5 (c 0.3, CHCl3); Rf 0.18 (1 : 1 hexane–EtOAc).
1H NMR (300 MHz; CDCl3; Me4Si): d 7.40–7.26 (5H, m, ArH),
5.77 (1H, d, J = 1.6 Hz, H(O)3), 4.69 (1H, d, J1,2 = 3.6 Hz, H1),
These results show that the current approach is useful to
compare the relative reactivities of the hydroxyl groups present in
diols but we have wondered if it can be used to compare the relative
reactivities of two hydroxyl groups present in different compounds.
As explained above, the experimental relative reactivities of the
monosaccharide derivatives 1 and 2 carrying a single hydroxyl
group each are known (1 ꢀ 2). Using analogs of 1 and 2 where the
phenyl group was replaced by a methyl group (1e and 2e), we have
analyzed both the ground states (reactants) and the O-methylated
derivatives, in order to determine the energy difference for the
reaction:
4.64 (2H, s, CH2Ph), 4.29 (1H, dd with appearance of t, J2,3
3.4 Hz, H2), 4.26–4.22 (1H, m, H3), 4.08 (1H, ddd, J5,6a = 2.4,
J5,6b = 4.6, J5,4 = 9.9 Hz, H5), 3.85 (1H, dd, J6a,6b = 10.7 Hz, H6a),
=
3.79 (1H, dd, H6b), 3.74 (1H, ddd with appearance of dt, J4,3
=
2.8, J4,OH = 10.4 Hz, H4), 3.34 (3H, s, OCH3), 2.76 (1H, d, H(O)4),
2.00 (6H, s, CCH3 ¥ 2). 13C NMR (75 MHz; CDCl3; Me4Si): d
172.50 (CO ¥ 2), 138.30 (C-Ar), 137.77 (C ¥ 2) 128.31–127.50
(C-Ar), 98.39 (C1), 73.49 (CH2Ph), 69.38 (C6), 68.49 (C3), 68.10
(C4), 67.53 (C5), 55.86 (OCH3), 54.57 (C2), 8.93 (CCH3 ¥ 2).
ESI-HRMS: calcd for [C20H25NO7 + Na]+ 414.1523. Found, m/z
414.1515.
where the lower-energy conformers of Sug–OH and its methoxy-
lated counterpart were used for the calculation. For 1e, the
calculation leads to a DE of -107.6 kcal mol-1, whereas for 2e the
DE is -99.7 kcal mol-1. This large difference is clearly indicating
that the O3 in 1e has higher reactivity towards donors, as it
Methyl
6-O-benzyl-2-deoxy-2-dimethylmaleimido-b-D-allo-
pyranoside (7). Following the same procedure described for
the preparation of 6, starting with 4 (348 mg, 0.89 mmol), 7
was obtained (250 mg, 72% yield), as a foam; [a]2D9 -42.5 (c 0.6,
CHCl3); Rf 0.15 (1 : 1 hexane–EtOAc). 1H NMR (300 MHz;
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3020–3025 | 3023
©

View Article Online
CDCl3; Me4Si): d 7.38–7.27 (5H, m, ArH), 5.39 (1H, d, J1,2
=
ESI-HRMS: calcd for [C36H45NO17 + Na]+: 786.25797. Found,
m/z: 786.25748.
8.7 Hz, H1), 4.62 (2H, s, CH2Ph), 4.15 (1H, br t, H3), 4.02 (2H,
dd, J2,3 = 2.4 Hz, H2), 3.96 (1H, dd, J5,6 = 5.0, J5,4 = 9.9 Hz, H5),
3.79 (1H, dd, H6), 3.78 (1H, br s, H(O)3), 3.77–3.68 (1H, m, H4),
3.43 (3H, s, OCH3), 3.02 (d, 1H, J = 6.5 Hz, H(O)4), 1.98 (6H, s,
CH3 ¥ 2). 13C NMR (75 MHz; CDCl3; Me4Si): d 172.50 (CO),
137.76 (C-Ar), 137.55 (C ¥ 2), 128.48–127.76 (C-Ar), 96.87 (C1),
73.72 (CH2Ph), 72.43 (C5), 70.93 (C3), 70.76 (C6), 69.87 (C4),
56.78 (OCH3), 55.62 (C2), 8.86 (CCH3, ¥ 2). ESI-HRMS: calcd
for [C20H25NO7 + Na]+ 414.1523. Found, m/z 414.1520.
Methyl 2,3,4,6-tetra-O-acetyl-b-D-galactopyranosyl-(1→3)-6-
O-benzyl-2-deoxy-2-dimethylmaleimido-b-D-allopyranoside (10a).
93.5%; as a foamy solid: [a]3D2 -66.1 (c 0.34, CHCl3); Rf.0.15 (1 : 1
hexane–EtOAc). 1H NMR (300 MHz; CDCl3; Me4Si): d 7.40–7.27
(5H, m, ArH), 5.66 (1H, d, J = 8.7 Hz, H1), 5.30 (1H, d, J4¢,3¢
=
3.2 Hz, H4¢), 5.09 (1H, dd, J2¢,1¢ = 7.8, J2¢,3¢ = 10.4 Hz, H2¢), 4.94 (1H,
dd, H3¢), 4.60 (1H, d, J = 11.6 Hz, CH2Ph), 4.52 (1H, d, CH2Ph),
4.51 (1H, d, H1¢), 4.08 (1H, dd with appearance of br t, H3), 3.90
(1H, dd, J6¢a,5 = 7.4, J6¢a,6¢b = 10.5 Hz, H6¢a), 3.86–3.68 (6H, m, H2,
H4, H5, H6a, H6b, and H5¢), 3.54 (1H, dd, J6¢b,5¢ = 6.4 Hz, H6¢b),
3.47 (3H, s, OCH3), 3.01 (1H, d, J = 2.8 Hz, H(O)4), 2.13 (3H, s,
COCH3), 2.02 (3H, s, COCH3), 1.95 (6H, s, CCH3 ¥ 2), 1.93 (6H, s,
COCH3). 13C NMR (75 MHz; CDCl3; Me4Si): d 172.29–169.45
(CO), 137.94 (C), 137.11 (C-Ar), 135.97 (C), 128.58–127.99 (C-
Ar), 102.01 (C1¢), 97.96 (C1), 77.47 (C3), 74.00 (CH2Ph), 71.57
(C6), 71.46 (C5), 70.89 (C4), 70.64 (C3¢), 69.93 (C5¢), 69.24 (C2¢),
66.75 (C4¢), 60.85 (C6¢), 56.91 (OCH3), 55.47 (C2), 20.77–20.54
(COCH3 ¥ 4), 8.67 (CCH3), 8.48 (CCH3). ESI-HRMS: calcd for
[C34H43NO16 + Na]+: 744.24741. Found, m/z: 744.24604.
Temperature dependence of the chemical shift of the hydroxyl
1
groups of 6, 7, 11 and 12 in H NMR. 1H NMR (300 MHz)
spectra were recorded for solutions of 6, 7, 11a, 11b, 12a and
12b in DMSO-d6 (internal standard, for the 1H residual DMSO).
Assignments of proton resonances were based on two-dimensional
1H–1H correlation experiments. Four spectra were recorded at
different temperatures in the 298–350 K range. The Dd/DT (ppb
K-1) were obtained from a linear fit.
The Dd/DT values obtained for 11a: H(O)3 (-5.9 ppb K-1) and
H(O)4 (-5.5 ppb K-1); for 11b: H(O)3 (-5.9 ppb K-1) and H(O)4
(-6.0 ppb K-1); for 12a: H(O)3 (-5.4 ppb K-1) and H(O)4 (-5.6 ppb
K-1) and for 12b: H(O)3 (-5.5 ppb K-1) and H(O)4 (-6.0 ppb K-1).
Acetate 10b. [a]3D2 -7.29 (c 0.38, CHCl3); Rf 0.20 (1 : 1 hexane–
EtOAc). 1H NMR: d 4.98 (1H, dd, J4,3 = 2.7, J4,5 = 10.1 Hz, H-4).
ESI-HRMS: calcd for [C36H45NO17 + Na]+: 786.25797. Found,
m/z: 786.25489.
General procedure for glycosylation and acetylation reactions
A suspension of the acceptor 6 or 7 (0.1 mmol), donor 5
˚
(0.11 mmol), activated 4 A molecular sieves (106 mg) in anhydrous
CH2Cl2 (4.9 ml) and CH3CN (133 ml) were stirred at room
temperature. After 40 min, the mixture was cooled to -25 ◦C,
TMSOTf (0.21 mmol) was slowly added and the stirring continued
for 30 min. The mixture was then neutralized by addition of solid
NaHCO3 (188 mg) and filtered through a silica gel pad with
copious washings with EtOAc. The filtrate was dried (Na2SO4)
and evaporated. The residue was chromatographed to give the
products. The acetylations were carried out under standard
conditions: pyridine, DMAP, Ac2O, room temperature, overnight.
Computational determinations
Quantum mechanical calculations were performed using Jaguar
6.0,19 using default minimization methods and termination con-
ditions. All the DFT calculations were made at the B3LYP/6-
31+G(d,p) level, starting with different geometries around the
exocyclic moieties, before and after the introduction of the methyl
group. For the latter group, it has been shown that the group leads
to a lower energy isomer when it is gauche (rather than anti) to
the hydrogen attached to the carbon (dihedral angle Hn–Cn–On–
Me ª 60◦). For the determination of the DE of methylation of
compounds 1e and 2e, the lower-energy conformers for the non-
methylated and methylated were used for the calculation, as well
as a methyl carbocation optimized separately.
Methyl 2,3,4,6-tetra-O-acetyl-b-D-galactopyranosyl-(1→3)-6-
O-benzyl-2-deoxy-2-dimethylmaleimido-a-D-allopyranoside (8a).
86.5%; as a foamy solid: [a]3D1 + 26.8 (c 0.4, CHCl3); Rf 0.09 (1 : 1
1
hexane–EtOAc). H NMR (300 MHz; CDCl3; Me4Si): d 7.36–
7.26 (5H, m, ArH), 5.31 (1H, d, J4¢,3¢ = 3.1 Hz, H4¢), 5.12 (1H,
dd, J2¢,1¢ = 8.0, J2¢,3¢ = 10.3 Hz, H2¢), 5.02 (1H, d, J1,2 = 4.7 Hz,
H1), 4.94 (1H, dd, H3¢), 4.63 (1H, dd with appearance of t, J3,2
ª J3,4 = 4.3 Hz, H3), 4.61 (1H, d, J2¢,1¢ = 8.0 Hz, H1¢), 4.60 (1H,
d, J = 12.0 Hz, CH2Ph), 4.54 (1H, d, CH2Ph), 4.34 (1H, dd with
Conclusions
We have herein presented experimental evidences of the influence
of a strong hydrogen bond in the regioselectivity of the a-anomer
of the N-DMM-D-allosamine diol acceptor (6) when coupled with
the disarmed donor 5. Owing to the difficulties for predicting the
regioselectivity of the glycosylation reaction of the b-anomer 7
(showing only weak hydrogen bonds) with the same donor, we
have calculated by DFT the relative stabilities of the charged
intermediates using a rather simple model, which assumes that
these stabilities might account for the observed regioselectivity.
On the basis of these calculations we have rationalized the
experimental results. Furthermore, by applying the same approach
we were able to rationalize the more complex regioselectivity
outcome observed previously with the N-DMM-D-glucosamine
acceptors 11a,b and 12a,b. The theoretical data have agreed with
experimental results even better than we expected. Our intention to
appearance of t, H2), 4.07–4.00 (1H, m, H5), 3.95 (1H, dd, J6¢a,5¢
=
7.6, J6¢a = 10.7 Hz, H6¢a), 3.90–3.73 (4H, m, H4, H5¢, H6¢b,
,
6¢b
H(O)4), 3.70 (2H, d, J6,5 = 4.1 Hz, H6), 3.37 (3H, s, OCH3), 2.12
(3H, s, COCH3), 1.99 (3H, s, COCH3), 1.95 (3H, s, COCH3), 1.94
(6H, s, CCH3 ¥ 2), 1.92 (3H, s, COCH3). 13C NMR (75 MHz;
CDCl3; Me4Si): d 172.17–169.15 (CO), 137.72 (C-Ar), 137.18 (C ¥
2), 128.46–127.71 (C-Ar), 100.49 (C1¢), 96.85 (C1), 73.64 (CH2Ph),
73.10 (C3), 70.89 (C3¢), 70.57 (C6), 70.37 (C5¢), 69.76 (C5), 69.28
(C2¢), 66.79 (C4¢), 66.66 (C4), 60.86 (C6¢), 55.57 (OCH3), 51.60
(C2), 20.69–20.54 (COCH3 ¥ 4), 8.75 (CCH3 ¥ 2). ESI-HRMS:
calcd for [C34H43NO16 + H]+: 722.26546. Found, m/z: 722.26684.
Acetate 8b. [a]3D2 + 46.4 (c 0.44, CHCl3); Rf 0.13 (1 : 1 hexane–
EtOAc). 1H NMR: d 5.05 (1H, dd, J4,3 = 3.1, J4,5 = 10.5 Hz, H4).
3024 | Org. Biomol. Chem., 2011, 9, 3020–3025
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
identify trends was fulfilled with this approach, which in turn can
be used hopefully by synthetic chemists to predict regioselectivity
outcomes, and thus to improve the efficiency of glycosylation
reactions.
6 P. R. Mudasani, E. Bozo´, B. Bernet and A. Vasella, Helv. Chim. Acta,
1994, 77, 257–290; P. R. Mudasani, B. Bernet and A. A. Vasella, Helv.
Chim. Acta, 1994, 77, 334–350.
7 B. Fraser-Reid, J. C. Lo´pez, A. M. Go´mez and C. Uriel, Eur. J. Org.
Chem., 2004, 1387–1395; C. Uriel, A. M. Go´mez, J. C. Lo´pez and B.
Fraser-Reid, Eur. J. Org. Chem., 2009, 403–411.
8 M. B. Cid, F. Alfonso, I. Alonso and M. Mart´ın-Lomas, Org. Biomol.
Chem., 2009, 7, 1471–1481.
Acknowledgements
9 E. S. H. El Ashry and M. R. E. Aly, Pure Appl. Chem., 2007, 79, 2229–
2242; D. B. Werz, R. Ranzinger, S. Herget, A. Adibekian, C.-W. von
der Lieth and P. H. Seeberger, ACS Chem. Biol., 2007, 2, 685–691.
10 M. L. Bohn, M. I. Colombo, P. L. Pisano, C. A. Stortz and E. A.
Ru´veda, Carbohydr. Res., 2007, 342, 2522–2536.
11 M. L. Bohn, M. I. Colombo, E. A. Ru´veda and C. A. Stortz, Org.
Biomol. Chem., 2008, 6, 554–561.
We thank Dr Manuel Gonza´lez-Sierra for NMR spectral de-
terminations and interpretation and to Dr G. Labadie and Dr
C. Delpiccolo for HRMS determinations. Financial support
from UNR, UBA, ANPCyT and CONICET are also acknowl-
edged. M.I.C., C.A.S and E.A.R. are Research Members of the
National Research Council of Argentina (CONICET).
12 M. I. Colombo, C. A. Stortz and E. A. Ru´veda, Carbohydr. Res.,
DOI: 10.1016/j.carres.2011.01.017.
13 M. R. E. Aly, J. Castro-Palomino, E. I. Ibrahim, E. S. H. El Ashry and
R. R. Schmidt, Eur. J. Org. Chem., 1998, 2305–2316.
14 J. Kroon, L. M. J. Kroon-Batenburg, B. R. Leeflang and J. F. G.
Vliegenthart, J. Mol. Struct., 1994, 322, 27–31.
15 R. J. Abraham, J. J. Byrne, L. Griffiths and R. Koniotu, Magn. Reson.
Chem., 2005, 43, 611–624.
16 J.-L. Maloisel and A. Vasella, Helv. Chim. Acta, 1992, 75, 1491–1514.
17 For a detailed description of the accepted mechanism of glycosylation
and of the characteristics of the charged intermediate, see ref. 3.
18 T. Nukada, A. Be´rces and D. M. Whitfield, Carbohydr. Res., 2002, 337,
765–774.
Notes and references
1 X. Zhu and R. R. Schmidt, Angew. Chem., Int. Ed., 2009, 48, 1900–
1934.
2 D. M. Whitfield, Adv. Carbohydr. Chem. Biochem., 2009, 62, 83–159.
3 L. K. Mydock and A. V. Demchenko, Org. Biomol. Chem., 2010, 8,
497–510.
4 D. Crich, Acc. Chem. Res., 2010, 43, 1144–1153.
5 P. Ulhmann and A. Vasella, Helv. Chim. Acta, 1992, 75, 1979–
1994; E. Bozo´ and A. Vasella, Helv. Chim. Acta, 1994, 77, 745–
753.
19 Jaguar 6.0, release 107, Schrodinger, LLC, Portland, OR, 2005.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3020–3025 | 3025
©