Carbohydrate-carbohydrate interactions in water with glycophanes as model systems

Article abstract of DOI:10.1021/jo9807823

The synthesis and conformational properties of glycophanes 2 and 3 (cyclodextrin-cyclophane hybrid receptors containing two maltose units linked by (4-hydroxymethyl) benzoic acid spacer) are described. The binding properties in water of these receptors with a series of 4-nitrophenyl glycosides with α- and β-configurations at the anomeric center have been studied using ¹H NMR spectroscopy and molecular mechanics calculations. A comparison of these properties with those of glycophane 1 (an α,α-trehalose containing glycophane) and α-cyclodextrin (αCD) using the same glycosides shows the existence of a stabilizing contribution to the free energy of binding in the case of of glycophanes but not in the case of the αCD system. This contribution is due to carbohydrate-carbohydrate interactions between both host and guest lipophilic sugar surfaces. Glycophanes 1, 2, and 3 show similar α/β selectivity on binding the ligands, despite the great flexibility of 3 related to 1 and 2. Parallels are drawn between the thermodynamic behavior of these model systems and that proposed for sugar- protein interactions.

Full text of DOI:10.1021/jo9807823

9212
J . Org. Chem. 1998, 63, 9212-9222
Ca r boh yd r a te-Ca r boh yd r a te In ter a ction s in Wa ter w ith
Glycop h a n es a s Mod el System s
J uan Carlos Morales, Da´cil Zurita, and Soledad Penade´s*
Grupo de Carbohidratos, Instituto de Investigaciones Qu´ımicas (CSIC-UNSE), Isla de la Cartuja,
Americo Vespucio s/ n, 41092 Sevilla, Spain
Received April 28, 1998
The synthesis and conformational properties of glycophanes 2 and 3 (cyclodextrin-cyclophane hybrid
receptors containing two maltose units linked by (4-hydroxymethyl) benzoic acid spacer) are
described. The binding properties in water of these receptors with a series of 4-nitrophenyl
1
glycosides with R- and â-configurations at the anomeric center have been studied using H NMR
spectroscopy and molecular mechanics calculations. A comparison of these properties with those
of glycophane 1 (an R,R-trehalose containing glycophane) and R-cyclodextrin (RCD) using the same
glycosides shows the existence of a stabilizing contribution to the free energy of binding in the
case of of glycophanes but not in the case of the RCD system. This contribution is due to
carbohydrate-carbohydrate interactions between both host and guest lipophilic sugar surfaces.
Glycophanes 1, 2, and 3 show similar R/â selectivity on binding the ligands, despite the great
flexibility of 3 related to 1 and 2. Parallels are drawn between the thermodynamic behavior of
these model systems and that proposed for sugar-protein interactions.
In tr od u ction
dence suggests that this interaction is an unfavorable
process.4 With the system formed by one of these
receptors, glycophane 1, and a series of 4-nitrophenyl
glycosides (PNPGly) (Chart 1), we have shown the
existence of CARB-CARB interactions in water between
lipophilic patches of both host and guest carbohydrate
moieties.5 The contribution of this interaction to the free
There exists strong evidence that, in addition to the
well-established carbohydrate-protein interactions,1 cells
use carbohydrate-carbohydrate (CARB-CARB) interac-
tions as molecular mechanisms for adhesion.2 The
molecules involved in this adhesion are glycolipids2 (Lex,
GM3, Gg3, etc.) and proteoglycans as a proteoglycan2b
isolated from a marine sponge. A characteristic feature
of this interaction is its extremely low affinity that has
to be compensated for by aggregation and subsequent
multivalency and its dependence on bivalent cations
(Ca2+, Mg2+, and Mn2+).2
Hakomori has proposed complementarity between
hydrophobic surfaces and subsequent Ca2+ chelation to
some oxygen atoms of the carbohydrate2a to reinforce the
interaction between glycolipids; however, structural data
about this interaction or absolute values of association
energies have not yet been determined.
Some years ago we began to use model systems to
understand how carbohydrates interact with each other
in water. We designed and synthesized a novel type of
receptors named glycophanes. These receptors are com-
posed of disaccharides and aromatic segments and may
be considered as cyclodextrin-cyclophane hybrids.3 The
question we addressed with these receptors was whether
it was possible to show CARB-CARB interactions in
water between the constituent sugars of these receptors
and simple saccharides, even though calorimetric evi-
energy of binding (∆G) was up to 1.8 kcal‚mol-1
.
In this paper we report a detailed account of the
synthesis and properties of two new glycophanes 2 and
3, and we present the similarities and differences in
saccharide binding among these receptors and glyco-
phane 1 toward 4-nitrophenyl glycosides (PNPGly).
One of the requirements of synthetic hosts is to be
relatively rigid so that loss of conformational entropy
upon binding of a guest will be minimized. On the
contrary, a more flexible receptor can allow better
complementarity between host and guest surfaces by an
“induced fit” mechanism. The different flexibilities of 2
and 3 and their reference partner the R-cyclodextrin
(RCD) may allow us to address this question. The
binding ability of RCD and glycophanes 1, 2, and 3 is
also compared in terms of differential contributions of
the CARB-CARB interaction to the free energy of
binding.
Resu lts a n d Discu ssion
Syn th esis a n d In ter con ver sion of Glycop h a n es 2
a n d 3. The first glycophane we reported, compound 1,
was designed on the basis of a conformationally restricted
and highly symmetric disaccharide, R,R-trehalose, and
an electron-rich aromatic segment, 2,7-dihydroxynaph-
* To whom correspondence should be addressed (Fax, Int. Code +
(95) 446-0565; E-mail, penades@cica.es).
(1) Weiss, W. I.; Drickamer, K. Annu. Rev. Biochem. 1996, 65, 441-
473 and references therein.
(2) (a) Hakomori, S. Pure Appl. Chem. 1991, 63, 473-482. (b)
Misevic, G. N.; Burger, M. M. J . Biol. Chem. 1993, 268, 4922-4929.
(c) Dammer, U.; Popescu, O.; Anselmetti, D.; Gu¨ntherodt, H.-J .;
Misevic, G. N. Science 1995, 267, 1173-1175. (d) Bovin, N. V. In
Glycosciences; Gabius, H.-J ., Gabius, S., Eds.; Chapman and Hall:
Weinheim, 1997; pp 277-289.
(4) (a) Gaffney, S. H.; Haslam, E.; Lilley, T. H.; Ward, T. R. J . Chem.
Soc., Faraday Trans. 1 1988, 84, 2545-2552. (b) Barone, G.; Cas-
tronuovo, G.; Elia, V.; Savino, V. J . Solution Chem. 1984, 13, 209-
219.
(5) J imene´z-Barbero, J .; J unquera, E.; Martin-Pastor, M.; Sharma,
S.; Vicent, C.; Penade´s, S. J . Am. Chem. Soc. 1995, 117, 11198-11204.
(3) Cotero´n, J . M.; Vicent, C.; Bosso, C.; Penade´s, S. J . Am. Chem.
Soc. 1993, 115, 10066-10076.
10.1021/jo9807823 CCC: $15.00 © 1998 American Chemical Society
Published on Web 11/17/1998

Carbohydrate-Carbohydrate Interactions
J . Org. Chem., Vol. 63, No. 25, 1998 9213
Ch a r t 1. Str u ctu r e of Glycop h a n es 1, 2, 3, a n d rCD a n d th e Axia l a n d Equ a tor ia l 4-Nitr op h en yl
Glycosid es
thalene3,5 (Chart 1). For the new glycophanes 2 and 3
we have used the disaccharide maltose [4-O-(R-D-glu-
copyranosyl)-D-glucopyranose] and (4-hydroxymethyl)-
benzoic acid as the aromatic segment (Chart 1). These
cyclodextrin-cyclophane hybrid receptors are the result
of the substitution of two glucose units (1 and 4) in
R-cyclodextrin by an aromatic ring. It was thought that
in the new glycophanes 2 and 3 the maltose moiety will
expose more surface to the solvent than in the case of
the torus-shaped cyclodextrins. The hydroxymethyl group
of the aromatic ring is linked to the anomeric position of
the maltose unit, while the carboxyl group is linked to
either position 4′ (glycophane 2) or position 6′ (glycophane
3) of the other maltose molecule. This confers on the
receptors different flexibilities and topologies, so that the
maltose moieties present different regions of their sur-
faces to interact with the ligands.
Glycophane 2 was synthesized starting from maltose,
and glycophane 3 was obtained by transacylation of 2 in
water in almost quantitative yield. A preliminary com-
munication on the synthesis of 2 has already been
reported.6 The synthetic route is summarized in Scheme
1, and the synthetic details are given in the Experimental
Section. For the synthesis of glycophane 3, a route
similar to that for 2 starting from compound 8 could be
envisaged (Scheme 1). However, the nearly quantitative
transacylation in water of the ester group from position
4′ to 6′ in 2 permitted ready access to glycophane 3.
In an aqueous solution, glycophane 2 undergoes a
spontaneous double transacylation by transposition of the
ester group from the 4′ to the 6′ position of the maltose
moiety to give the new isomeric glycophane 3. The
reaction was followed by HPLC in a reverse-phase
column, and after 2 days at 37 °C, the transesterification
was complete. During this process a new product, in
addition to 2 and 3, was detected, probably the asym-
metric glycophane that results from the transposition of
only one of the ester groups. The transacylation rate of
2 depends on temperature, solvent, and concentration.
At 50 °C, more than 80% of 3 was formed after 24 h. In
methanol or in a micellar aqueous solution of sodium
dodecyl sulfate, the rate is much slower, and in DMSO
the transposition is not observed at all. Furthermore,
as the concentration of 2 increases, the transposition rate
decreases dramatically. The same is observed when
glycophane 2 is dissolved in the presence of some
aromatic guests (picric acid; 2,3-disulfonic naphthalene;
or 1,3,5-trihydroxybenzene) that interact with the recep-
tor. The inclusion of these guests in the cavity of 2
probably changes its geometry so that the carbonyl group
is placed further away from the hydroxyl group at
position 6′. A similar effect could explain the decrease
of the transposition rate when the concentration of 2
increases. The amphiphilic nature of 2 makes possible
self-association by including an aromatic ring of one of
the molecules in a second molecule.
Con for m a tion a l Stu d ies of Glycop h a n es 2 a n d 3.
The solution conformation of the isomeric glycophanes 2
and 3 has been analyzed by NMR spectroscopy and
molecular mechanics and dynamics simulations, and the
results have already been published.7 The comparison
of the calculated minima for RCD, 2, and 3 given in
Figure 1 shows the structural differences between these
receptors. From these differences a differential binding
behavior toward the same ligands is to be expected.
The conformational behavior of 2 and 3 is rather
different. Glycophane 2 presents in solution only a
family of conformers with geometry similar to that of
RCD (Figure 1A,B). In the global minimum, the inter-
glycosidic angles φ (H1′-C1′-O4-C4) and æ (C1′-O4-
C4-H4) of the maltose units are -4° and -30°, respec-
tively, which are similar to those of the minimum B (φ/æ
(6) Morales, J . C.; Penade´s, S. Tetrahedron Lett. 1996, 37, 5011-
5014.
(7) Morales, J . C.; Penade´s, S. Angew. Chem., Int. Ed. Engl. 1998,
37, 654-657.

9214 J . Org. Chem., Vol. 63, No. 25, 1998
Sch em e 1a Syn th esis of Glycop h a n es 2 a n d 3
Morales et al.
a
Reagents and conditions: (a) NaOAc/Ac2O; (b) HBr 33%/AcOH/CH2Cl2, 25 °C, 45 min; (c) AllOH/Hg(OAc)2, 25 °C, 4 h; (d) NaOMe/
MeOH, 25 °C, 10-12 h, 44% overall; (e) p-MeOC6H4CH(OMe)2/p-TsOH/DMF, 60 °C, 2 h, 62%; (f) NaH/p-MeOC6H5CH2Cl (PMBCl)/DMF,
25 °C, 16 h, 89%; (g) NaCNBH3/CF3COOH/DMF, 25 °C, 16 h, 7: 74%, 8: 11%; (h) Ac2O/py, 25 °C, 15 h, 98%; (i) RhCl(PPh3)3/10%EtOH/
To, 85 °C, 16 h, then HgO/HgCl2/acetone/H2O, 100 °C, 2.5 h, 81%; (j) K2CO3/Cl3CCN/CH2Cl2, 25 °C, 15 h; (k) TMSOTf/HOCH2PhCOOMe/
CH2Cl2, -78 °C, 20 min, 80% (R/â 1.5:1); (l) MeONa/MeOH, 25 °C, 18 h, 13r: 56%, 13â: 39%; (m) KOHaq/MeOH/THF, 25 °C, 20 h, 100%;
(n) DCC/DMAP/DMAP‚HCl/CHCl3, reflux, 6 h, 29%; (o) CF3COOH 5%/CH2Cl2, 25 °C, 5 h, 95%; (p) H2O, 50 °C, 3 days, 98%.
F igu r e 1. Top view (a) and front view (b) of the calculated global minima for RCD (A); glycophane 2 (B), and glycophane 3 (C
and D).
-3°/-30°) found for maltose in solution.8 While glyco-
phane 2 is fairly rigid, glycophane 3 is rather flexible and
shows, in aqueous solution, an equilibrium between three
different conformers (the third conformer is not shown
in Figure 1).7 In the global minimum (Figure 1C), the
cavity is collapsed by a double sugar aromatic stacking
between the aromatic ring and the â-face of the glucose
(8) Ha, S. H.; Madsen, L. J .; Brady, J . W. Biopolymers 1988, 27,
1927-1952.

Carbohydrate-Carbohydrate Interactions
J . Org. Chem., Vol. 63, No. 25, 1998 9215
Ta ble 1. Associa tion Con sta n ts (Ka , M-1) a n d F r ee En er gy of Bin d in g (-∆G, k ca l‚m ol-1) of th e Com p lexes F or m ed
betw een r-Cyclod extr in or Glycop h a n es 1, 2, a n d 3 a n d 4-Nitr op h en yl Glycosid es or 1-(4-Nitr op h en yl)glycer ol (P NP G)
R-CD
-∆Gb
1
2
3
d
d
-∆Ga
-∆∆Gc
-∆Ga
-∆∆Gc
Ka
-∆G
-∆∆Gc
Ka
-∆G
-∆∆Gc
RGlc
2.9 ( 0.1
2.8 ( 0.1
3.0 ( 0.1
3.0 ( 0.1
2.9 ( 0.1
2.6 ( 0.1
2.9 ( 0.1
2.8 ( 0.1
2.8 ( 0.1
2.8 ( 0.1
2.9 ( 0.1
2.8 ( 0.01
2.7 ( 0.01
2.9 ( 0.01
2.9 ( 0.01
2.9 ( 0.01
2.7 ( 0.01
2.8 ( 0.01
2.8 ( 0.01
2.6 ( 0.01
2.8 ( 0.01
e
0
2.8 ( 0.1
2.6 ( 0.1
3.0 ( 0.1
3.0 ( 0.1
3.3 ( 0.1
2.7 ( 0.1
2.7 ( 0.1
2.6 ( 0.1
2.6 ( 0.1
2.8 ( 0.1
1.8 ( 0.1
1.0
0.8
1.2
1.2
1.5
0.9
0.9
0.8
0.8
1.0
78
37
100
92
101
32
36
55
60
64
23
2.6 ( 0.1
2.2 ( 0.1
2.8 ( 0.1
2.7 ( 0.1
2.8 ( 0.1
2.1 ( 0.1
2.2 ( 0.1
2.4 ( 0.1
2.5 ( 0.1
2.5 ( 0.1
1.9 ( 0.1
0.7
0.3
0.9
0.8
0.9
0.2
0.3
0.5
0.6
0.6
22
20
57
42
40
19
23
22
26
18
16
1.9 ( 0.1
1.8 ( 0.1
2.4 ( 0.1
2.2 ( 0.1
2.2 ( 0.1
1.8 ( 0.1
1.9 ( 0.1
1.9 ( 0.1
2.0 ( 0.1
1.7 ( 0.1
1.7 ( 0.1
0.2
0.1
0.7
0.5
0.5
0.1
0.2
0.2
0.3
0.0
RGal
-0.1
0.1
0.1
0
-0.3
0
-0.1
-0.1
-0.1
RMan
âLAra
RLFuc
âGlc
âGal
âMann
RLAra
âLFuc
PNPG
a
b
From NMR data in D2O at 303 K.5 From calorimetry data in H2O at 298 K.13 c ∆G(PNP-glycosides) - ∆G(PNPG) in kcal‚mol-1
.
From NMR data in D2O at 303 K; maximum percent of estimated error was 20%. e Not measured.
d
Ta ble 2. In d u ced Ch em ica l Sh ifts (ICS)a Obser ved in 1H NMR for Glycop h a n e 2 u p on th e F or m a tion of th e Com p lexes
w ith 4-Nitr op h en yl Glycosid es
axial glycosides
equatorial glycosides
2-RGlc
2-RGal
2-RMan
2-âLAra
2-RLFuc
2-âGlc
2-âGal
2-âMan
2-RLAra
2-âLFuc
Ha
-0.188
-0.080
-0.021
-0.027
-0.014
+0.040
-0.360
-0.361
-0.148
-0.039
-0.055
b
-0.211
-0.123
-0.048
-0.041
b
-0.287
-0.117
-0.057
-0.074
-0.056
+0.065
-0.299
-0.122
-0.056
-0.088
b
-0.127
-0.051
-0.017
-0.035
-0.025
+0.031
-0.091
-0.035
b
-0.024
b
-0.187
-0.073
-0.041
-0.052
b
Hb
H4′
H3′
H5′
H3
-0.161
-0.088
-0.046
-0.019
b
b
b
b
b
b
b
+0.079
+0.021
+0.068
+0.023
+0.042
a
b
For ICS: +, downfield shifts; -, upfield shifts. It could not be observed.
units. This interaction forces glycophane 3 to adopt a
folded geometry which places the maltose molecule in a
conformation (φ/ψ -55°/-60°, φa/ψa 69°/70°) different
from that found in glycophane 2 (φ/ψ -4°/-30°, φa/ψa
68°/-17°). The second minimum (Figure 1D) has inter-
glycosidic angles φ and ψ (11°/15°) similar to those found
for maltose in the solid state.9 The aromatic rings show
a face-to-edge π-π interaction, which closes the cavity.
The equilibrium between both conformers is influenced
by temperature and solvent7 as well as by the presence
of ligands in the aqueous solution (see below). The
different conformational presentation of the maltose
moiety in 2, 3, and RCD makes these molecules good
models to study the influence of oligosaccharide flex-
ibility7,10 and surface presentation11 in the molecular
recognition of carbohydrates.
Bin d in g Stu d ies. The new glycophanes 2 and 3 have
electron-deficient aromatic walls and, as is to be expected
from the nature of arene-arene interactions,12 will
interact favorably with the electron-deficient 4-nitrophen-
yl glycosides. The electron-rich benzyl â-D-glucopyrano-
side derivative interacts unfavorably with glycophanes
2 and 3.
To evaluate the contribution of the pyranose ring, the
Ka of the 1-(4-nitrophenyl) glycerol (PNPG), a lineal-chain
saccharide, was also determined as reference ligand. The
association constant of the other reference compound, the
4-nitrophenol, could not be determined in this case
because of solubility problems. Table 1 summarizes the
association constants and the corresponding free energies
of binding (∆G kcal‚mol-1) at 30 °C for all complexes. For
comparative reasons, the ∆G values previously deter-
mined for RCD5,13 and glycophane 15 complexes are also
given in Table 1.
The binding experiments were carried out at constant
concentration of the host and by the addition of increas-
ing concentrations of guests, as described in the Experi-
mental Section. Different induced chemical shift (ICS)
patterns for 2 and 3 were obtained upon addition of the
guests (Tables 2 and 3).
In the case of glycophane 2 (Table 2), upfield shifts for
the aromatic protons (Ha and Hb) and of H3′, H5′, and
H4′ and downfield shifts for H3 were observed. This
pattern was the same in all cases, indicating a common
geometry for all complexes. These shifts are consistent
with a geometry where the aromatic ring of the ligand is
parallel to the aromatic walls of the host, placing the H3′
and H5′ protons in the shielding cone of the 4-nitrophenyl
ring and the H3 and, probably, H5 (its signal was
overlapped) outside the cone. The upfield shift observed
for the H4′ proton, which is outside of the cavity, can be
explained by changes in the position of the adjacent
carbonyl group upon complexation, as confirmed by
molecular mechanics simulations (see below).
Thus, the association constants, Ka in water between
2 and 3 and a series of 4-nitrophenyl glycosides with axial
R-D-gluco- (RGlc), R-D-galacto- (RGal), R-D-manno- (RMan),
â-L-arabino- (âLAra), and R-L-fucopyranoside (RLFuc) and
equatorial â-D-gluco- (âGlc), â-D-galacto- (âGal), â-D-
manno- (âMan), R-L-arabino- (RLAra), and â-L-fucopy-
ranoside (âLFuc) configuration at the anomeric center
(Chart 1) have been determined by 1H NMR spectroscopy.
For glycophane 3, upfield shifts for H1′, H4′, H5′, H6′R,
H6′S, and aromatic protons, and downfield shifts for H2,
H3, and H4, could be observed upon addition of the
(9) Gress, M. E.; J effrey, G. A. Acta Crystallogr. B 1977, 33, 2490-
2495.
(10) Peters, T. Curr. Opin. Struct. Biol. 1996, 6, 710-720 and
references therein.
(11) Carver, J . P. Pure Appl. Chem. 1993, 65, 763-770.
(12) Hunter, C. A. Angew. Chem., Int. Ed. Engl. 1993, 32, 1584-
1586 and references therein.
(13) J unquera, E.; Laynez, J .; Mene´ndez, M.; Sharma, S.; Penade´s,
S. J . Org. Chem. 1996, 61, 6790-6798.

9216 J . Org. Chem., Vol. 63, No. 25, 1998
Morales et al.
Ta ble 3. In d u ced Ch em ica l Sh ifts (ICS)a Obser ved in 1H NMR for Glycop h a n e 3 u p on th e F or m a tion of th e Com p lexes
w ith 4-Nitr op h en yl Glycosid es
axial glycosides
equatorial glycosides
3-RGlc
3-RGal
3-RMan
3-âLAra
3-RLFuc
3-âGlc
3-âGal
3-âMan
3-RLAra
3-âLFuc
Ha
Hb
-0.080
-0.023
-0.073
-0.107
-0.067
+0.110
c
-0.024
-0.026
+0.062
-0.085
-0.026
-0.082
-0.120
-0.084
+0.118
c
-0.017
-0.033
+0.059
-0.078
+0.008
-0.123
-0.269
-0.155
+0.162
+0.028
-0.089
-0.062
+0.086
-0.071
-0.018
-0.062
-0.091
-0.060
+0.102
+0.039
-0.033
-0.020
+0.059
-0.112
-0.038
-0.075
-0.069
-0.078
+0.145
+0.034
-0.002
-0.006
+0.077
-0.122
-0.048
-0.089
-0.096
-0.104
+0.129
+0.041
-0.052
-0.026
+0.026
-0.102
-0.038
-0.075
-0.075
-0.085
+0.122
c
-0.125
-0.040
-0.098
-0.125
-0.121
+0.159
c
-0.060
-0.018
-0.050
-0.064
-0.054
+0.084
+0.021
c
-0.071
-0.024
-0.071
-0.058
-0.061
+0.096
+0.020
-0.015
-0.040
+0.067
H1′
H6′R
H6′S
H2b
H3
H5′
H4′
H4b
c
c
-0.020
+0.074
-0.033
+0.090
-0.016
+0.043
a
b
For ICS: +, downfield shifts; -, upfield shifts. ICS caused by the conformational changes, from folded to unfolded forms. c It could
not be observed.
ligands (Table 3). In a first attempt it is difficult to
establish a geometry for the complexes of 3 that satisfies
all the observed ICS. However, it should be remembered
that the major conformer of 3 is in a folded conformation
(Figure 1C), which does not possess a cavity for binding,
so that the equilibrium has to be shifted to an open
conformation to accept the guest in a clear “induced-fit”
mechanism. The ICS observed for 3 can be explained as
the result of two effects: (a) the folded conformation being
shifted to an open one, as indicated by the downfield shift
of the H1, H2, H4, H6R, and H6S protons (H1, H6R, and
H6S are not shown in Table 3); and (b) the effect of
binding of the aromatic guest into the opened cavity,
shown by the upfield shift of the aromatic and the H5′,
H6′R, and H6′S protons and the downfield shift of the
H3 proton. Thus, the interaction of glycophane 3 with
the ligands implies both a strong conformational change
upon binding and a reduction of the conformational space
of 3, which may result in unfavorable entropy effects.
This fact may be the cause of the lower association
constants observed for glycophane 3 compared to those
observed for the more rigid receptors 1, 2, and RCD
(Table 1).
F igu r e 2. Stereoview of the calculated minima for the
complexes 2-RMan (a) and 2-âMan (b).
complexes obtained by entry of the ligands at the primary
hydroxyl side.
All minima found for the 2 complexes belong to one
major conformation in which the maltose moieties show
the same interglycosidic angles, φ and æ, as in the free
host. A stereoview of the global minima conformers for
the 2-RMan and 2-âMan complexes are given in Figure
2.
We have tried to assess the tridimensional structure
of the complexes by intermolecular NOEs. For the 2 and
3 complexes, monodimensional NOEs and ROESY ex-
periments were carried out. Intermolecular NOEs be-
tween the aromatic protons of the host and guests were
observed. However, the intermolecular NOEs observed
between the sugar residues were too small to draw an
unambiguous conclusion about the geometry of the
complexes.
The conformational change observed in the free host
upon complexation is only the relative orientation be-
tween the aromatic rings, which in the complexes maxi-
mizes the interactions with the 4-nitrophenyl group. This
change places the carbonyl group at position 4′, further
away from the OH6′, which would explain both the
upfield shift observed for the 4′ proton (Table 2) and the
decrease of the transacylation rate of 2 in the presence
of ligand. Besides the expected aromatic-aromatic
interactions, van der Waals contacts between lipophilic
patches of both host and guest carbohydrate moieties are
also present. This CARB-CARB interaction can account
for the additional stabilization found in these complexes
related to the 2-PNPG complex (-∆∆G in Table 1). The
calculated geometry for these complexes agrees reason-
ably with the experimental ICS observed in 2 upon
addition of the guests (Figure 2, Table 2).
Molecu la r Mod elin g. The interactions of 2 and 3
with 4-nitrophenyl R- and â-D-gluco- and -mannopy-
ranosides were examined using the MACROMODEL v4.5
package. The guests were docked into the glycophane
hosts as described in the Experimental Section. The
global minimum previously found for the free glycophane
2 (Figure 1B) was used as the starting geometry to dock
the different ligands. In the case of 3, the global
minimum is a folded conformation without a cavity.
Therefore, the nonfolded conformation (Figure 1D) was
taken as the starting geometry for the docking. The
structures of the free hosts present two different faces.
Calculations for the two modes of ligand entry into the
host cavity were performed. For glycophane 2, the
energies obtained for the two types of complexes were
similar. However, for 3, the complexes where the ligands
enter at the secondary hydroxyl side are more stable than
The docking of the 4-nitrophenyl glycosides into 3
results in two different structural families of complexes,
A and B, with similar steric energies due to the big cavity
and great flexibility of glycophane 3. Stereoviews of the

Carbohydrate-Carbohydrate Interactions
J . Org. Chem., Vol. 63, No. 25, 1998 9217
F igu r e 3. Stereoview of the calculated minima for the
complex 3-RMan: (a) minimum A; (b) minimum B.
F igu r e 4. Calculated minima for the complexes 1-RMan (a);
2-RMan (b); 3-RMan (c); and RCD-RMan (d) showing the
van der Waals radii.
minimum energy conformers of A and B structures are
given in Figure 3 for the complex formed between 3 and
the 4-nitrophenyl R-D-mannopyranoside. In both geom-
etries, the receptor shows a cavity to accept the ligand
so that its conformation is different from that of the free
receptor. In the calculated structure A (Figure 3a), the
sugar residue shows van der Waals contacts with the
sugar moiety of the receptor, and a hydrogen bond
between the OH3 of the manno residue and the OH4′ of
the receptor is also possible. The whole geometry re-
sembles the calculated geometry for the complexes with
compound 2. In the structure B (Figure 3b), one of the
maltose units changes the interglycosidic φ and ψ angles
to 0° and 3°, respectively. In the calculated geometry for
the 3-RMan complex (Figure 3b), contacts between the
R face of the mannosyl residue and the aromatic residue
of the receptor are favored. Sugar-arene stacking is one
of the most frequent binding motifs found in the crystal
structure of sugar-protein complexes.1 However, its
contribution to the energetic of binding has only been
determined in the case of a maltose-binding protein.14
The molecular mechanics calculations for the 3-PN-
PGly complexes are consistent with the idea of an
induced-fit mechanism for binding, as proposed on the
basis of the experimental ICS observed for glycophane 3
upon addition of the guests (see above and Table 3),
which implies two processes: (a) a conformational change
in the free host 3 (downfield shifts for the H1, H2, and
H4 protons) and, (b) the binding of the aromatic guests
into the cavity (upfield shift of the aromatic and H5′,
H6′R, and H6′S protons and downfield shift of the H3
and H5 protons).
F igu r e 5. Comparative graphic of the free energy of binding
(∆G/kcal‚mol-1) of RCD or glycophanes 1, 2, and 3 and axial
and equatorial PNPGly and PNPG.
models for studying the nature of the apolar binding
between neutral sugars in aqueous solution.
The free energies of binding (-∆G, Table 1) have been
1
obtained by H NMR spectroscopy, and additionally, for
the RCD system, accurate values of enthalpy and entropy
could be determined by titration calorimetry.13 Although
glycophanes 2 and 3 have little structural homology with
glycophane 1, they clearly show common tendencies in
their energetic behavior, as the results in Table 1 and
Figure 5 reveal: (a) The free energies of binding for the
axial glycosides are usually higher than those for the
equatorial glycosides. (b) In the equatorial complexes,
∆G values do not change with the stereochemistry of the
ligands, while in the axial glycosides, these values
increase, depending on the stereochemistry, in the order
RGal < RGlc < âLAra < RMan < RLFuc. This influence
is not observed either in the axial or in the equatorial
complexes with RCD. (c) The free energy of binding
decreases in the order 1 > 2 > 3 for all complexes. This
decrease is proportional to the increase in conformational
flexibility on going from 1 to 3. (d) In all complexes, a
Although the calculated geometry might be different
from the real situation in solution, the results presented
here agree reasonably with the experimental data ob-
served for these interactions.
Com p a r ison of th e Bin d in g P r op er ties of 1, 2, 3,
a n d rCD. The results presented here, together with our
previous results, constitute a useful data set on the
interaction in water of four different model systems
(Figure 4). These systems, constituted by either R-cy-
clodextrin or glycophanes 1, 2, and 3 as receptors and a
series of 4-nitrophenyl glycosides of varied stereochem-
istry as ligands, have been proven to be appropriate
(14) Spurlino, J . C.; Rodseth, L. E.; Quiocho, F. A. J . Mol. Biol. 1992,
226, 15-22.

9218 J . Org. Chem., Vol. 63, No. 25, 1998
Morales et al.
stabilizing contribution to the binding (-∆∆G, Table 1)
was observed due to the presence of the pyranose ring.
This stabilization could be evaluated by comparing the
∆G values to those obtained for the complexes that are
formed between the PNPG and the glycophanes. The
highest contribution was observed for the 1-RLFuc
complex with 1.5 kcal‚mol-1 stabilization related to the
glycerol derivative. The lowest contribution corresponds
to the complexes with the most flexible glycophane 3, in
which a small stabilization was found only for the
3-RMan, 3-âLAra, and 3-RLFuc complexes. This con-
tribution was not found in the RCD system. We have
attributed this stabilization to CARB-CARB interactions
between both host and guest pyranose moieties, present
in the glycophane systems but absent in the cyclodextrin
system (Figure 4). This interaction may also be the
origin of the binding selectivity observed in glycophane
systems related to the cyclodextrin system.
assumed that the net observed thermodynamic properties
are controlled by the behavior of water surrounding both
host and guest complementary surfaces.17 This can be
the case in our systems. The common tendencies ob-
served and the similar thermodynamics in the interaction
of glycophanes 1, 2, and 3 with 4-nitrophenyl glycosides
might indicate that the small differences in the energetics
of binding observed in these systems can only be due to
differential dehydration process related to the ability of
sugars to structure water molecules in a stereospecific
way. However, to confirm this hypothesis, accurate
values of enthalpy and entropy will be necessary.
In conclusion, we have shown with our glycophane
systems that CARB-CARB interactions in water be-
tween small sugars (our hosts and guests may be
considered as small glycoconjugates) are possible and
contribute to the affinity of binding. The aromatic rings
are necessary to bring both host and guest carbohydrate
moieties together. Clearly, the presence of carbohydrate
residues increases the affinity. Divalent cations were not
present, and nonpolar interactions between sugar sur-
faces seem to be involved in stabilizing binding. The
interaction shows R/â selectivity similar to that of
lectins.15a,d,e Flexibility does not seem to oppose selectiv-
ity in glycophanes, and despite the great flexibility of 3,
which results in lower association constants, the R/â
selectivity observed in this receptor is of the same order
of magnitude as that in the more rigid receptors 1 and
2.
Establishing the structural origin of affinity and
selectivity as well as the connection between structure
and thermodynamics in carbohydrate interactions re-
mains a major challenge. The results presented here
show that our small systems may be considered good
models to mimic carbohydrate interactions in water. They
reproduce the nonpolar interactions between lipophilic
surfaces found in sugars binding to proteins. They show
some selectivity as a function of the stereochemical
diversity of these molecules, and additionally, their
thermodynamic behavior resembles that of biological
systems.
The differences among ∆G values for all the systems
are too small to draw conclusions about the differential
thermodynamic behavior of glycophanes related to the
RCD system (Figure 5). These differences should be more
evident in the enthalpy and entropy terms. Unfortu-
nately, accurate enthalpy and entropy values could be
determined only for the RCD system.13
For glycophanes 1 and 3, we have determined, by van’t
Hoff analysis, the enthalpies for the 1-RGlc, 1-âGlc,
3-RGlc, and 3-RMan complexes. 1H NMR spectra were
recorded at 283, 293, 303, and 313 K, which resulted in
∆H ) -6.6 kcal‚mol-1 and T∆S ) -4.2 kcal‚mol-1 for
3-RGlc complex and ∆H ) -4.1 kcal‚mol-1 and T∆S )
-1.7 kcal‚mol-1 for the 3-RMan complex at 298 K. For
the 1-RGlc and 1-âGlc complexes, these values are ∆H
) -4.8 and -4.1 kcal‚mol-1 and T∆S ) -1.9 and -1.4
kcal‚mol-1, respectively, as previously determined.5 The
values found by calorimetry for the RCD complexes13 at
this temperature are ∆H ) -5.4 kcal‚mol-1 and T∆S )
-2.6 for RCD-RGlc, ∆H ) -4.5 kcal‚mol-1 and T∆S )
-1.8 kcal‚mol-1 for the RCD-âGlc, and ∆H ) -5.8
kcal‚mol-1 and T∆S ) -2.9 kcal‚mol-1 for the RCD-
RMan complexes. These results indicate that all com-
plexes are exothermic and enthalpy driven near room
temperature with an unfavorable entropy term. The
contribution provided by the CARB-CARB interaction
to the enthalpy of binding seems to be the only difference
between glycophanes and cyclodextrin.
The proposal that nonpolar interactions are involved
in CARB-CARB recognition may be quite unexpected for
structures as polar as carbohydrates. However, as
proposed by Hakomori,2a “complementarity of two inter-
acting carbohydrates could be based on hydrophobic
interaction between the respective hydrophobic sur-
faces...”.
These thermodynamic patterns are similar to those
found for monosaccharides binding to lectins;15 the bulk
of the binding energy is enthalpic with a small entropy
term which opposes binding. Toone15c has called atten-
tion to the similarity of this energetic pattern with that
of the binding, in aqueous solution, of apolar ligands to
their corresponding receptors.16 Since the only similarity
among these systems is the solvent water, it has been
Exp er im en ta l Section
Gen er a l Tech n iqu es. Chemicals, including 4-nitrophenyl
glycosides, were obtained from commercial sources (Sigma,
Aldrich, or Fluka). Solvents were dried as recommended in
the literature. Thin-layer chromatography (TLC) was per-
formed on aluminum sheets coated with silica gel 60 F254
(Merck). Column chromatography was carried out on silica
gel 60 (Merck, 230-400 mesh). HPLC was performed on a
C18 reverse-phase column using UV detection. NMR spectra
were recorded on 300 or 500 MHz spectrometers with the
solvent reference as internal standards. Deuterated solvents
used were CD3OD (99.8% purity, Scharlau), D2O (99.8% purity,
Merck), and DMSO (99.96% purity, Aldrich). The steady-state
NOEs were obtained at 500 MHz through the interleaved
differential technique using a saturation delay of 5-10 s. The
2D rotating-frame NOE (ROESY, CAMELSPIN) experiments
(15) (a) Chervenak, M. C., Toone, E. J . Biochemistry 1995, 34, 5685-
5695. (b) Toone, E. J . Curr. Opin. Struct. Biol. 1994, 4, 719-728 and
references therein. (c) Williams, B. A.; Chervenak, M. C.; Toone, E. J .
J . Biol. Chem. 1992, 267, 22907-22911. (d) Schwarz, F. P.; Puri, K.;
Suriola, A. J . Biol. Chem. 1991, 266, 24344-24350. (e) Farina, R. D.;
Wilkins, R. G. Biochim. Biophys. Acta 1980, 631, 428-438.
(16) For some examples see: (a) Smithrud, D. B.; Wyman, T. B.;
Diederich, F. J . Am. Chem. Soc. 1991, 113, 5420-5426. (b) Diederich,
F.; Smithrud, D. B.; Sanford, E. M.; Wyman, T. B.; Ferguson, S. B.;
Carcanagne, D. R.; Chao, I.; Houk, K. N. Acta Chem. Scand. 1992, 46,
205-215. (c) Stauffer, D. A.; Barrans R. E. J .; Dougherty D. A. J . Org.
Chem. 1990, 55, 2762-2767. (d) Rekharsky, M. V.; Goldberg, R. N.;
Schwarz, F. P.; Tewari, Y. B.; Ross, P. D.; Yamashoji, Y.; Inoue, Y. J .
Am. Chem. Soc. 1995, 117, 8830-8840.
(17) Lemieux, R. U. Acc. Chem. Res. 1995, 29, 373-380.

Carbohydrate-Carbohydrate Interactions
J . Org. Chem., Vol. 63, No. 25, 1998 9219
were performed in the phase-sensitive mode at 500 MHz. The
spin-locking period consists of a train of 30° pulses (2.5 µs)
separated by delays of 50 µs, and total mixing time was set to
350 ms. The radio frequency carrier was set at δ ) 6 ppm to
minimize Hartmann-Hahn effects.
kJ ‚mol-1 over the global minimum was chosen. Finally, 25
conformers for 2 and 20 conformers for 3 were selected. The
global minimum was extensively minimized, and then its
conformational stability was checked using 500 ps of molecular
dynamic simulations (MD) by using the MM3* force field. The
Shake option21 to fix C-H bonds was employed during the
simulation, and the temperature was kept fixed at 300 K by
coupling to a temperature bath.22 Trajectory frames were
saved every 1 ps of simulation. Additionally, 900 ps of MD
simulations were performed at 500 K to check the stability of
the “folded” and “nonfolded” minima of 3.
The docking studies were performed from the global minima
found in the previous analysis for the isolated glycophanes 2
and 3. The “nonfolded” conformation of 3 was taken as
starting geometry for 3. The 4-nitrophenyl glycosides were
minimized as previously described.5 The protocol followed was
as follows: First, every substrate was docked manually into
the cavity and the complex was extensively minimized. Since
the cavities of the hosts present two different faces, calculation
for the two modes of substrate entry into the cavity was
performed. Second, 100 MC steps with random translation
and rotation of the guest around the cavity of the host were
carried out. Limits between 0 and 2 Å and between 0° and
15° were used for rotation and translation, respectively. In
each step, 5000 iterative minimizations using the PR conjugate
gradient algorithms were performed. An energy window of
50 kJ ‚mol-1 was used as the criterion to accept a given
conformation. In all cases, those complexes with a steric
energy difference of 12 kJ ‚mol-1 over the corresponding global
minimum were not considered for subsequent analysis.
Allyl 4-O-r-D-Glu cop yr a n osyl-â-D-glu cop yr a n ose (4).
Maltose was converted into the octa-O-acetyl-â-D-maltoside as
described.23 The octa-O-acetyl-â-D-maltoside (50 g, 74 mmol)
in CH2Cl2 (300 mL) was treated with a solution of 33%
hydrogen bromide in AcOH (100 mL) and stirred at rt for 45
min. The mixture was then diluted with CH2Cl2 (500 mL)
before being washed successively with ice (ca. 500 g), a
saturated solution of NaHCO3 (300 mL), and water (300 mL).
The organic extracts were dried (Na2SO4) and concentrated
to yield a white solid (50 g, 97%), which was identified as
hepta-O-acetyl-R-D-maltosyl bromide.24 This compound (12.38
g, 17.5 mmol) was dissolved in allylic alcohol (40 mL), mercuric
acetate (4.78 g, 0.85 equiv) was then added, and the reaction
mixture was stirred at rt for 4 h. The solvents were evapo-
rated in vacuo, and the residue was dissolved in CH2Cl2 (ca.
150 mL) and washed with a 10% aqueous solution of KI (5 ×
50 mL). The organic phase was dried (Na2SO4) and concen-
trated, and the residue was purified by column chromatogra-
phy (AcOEt/hexane 1:2 to AcOEt) to obtain a white solid, which
could be crystallized in isopropyl ether/AcOEt (7:1) to yield
the allyl hepta-O-acetyl-â-D-maltoside. The product was sus-
pended in MeOH, and NaOMe/MeOH (1 M, 1 mL) was added.
The mixture was stirred overnight, and then the solution was
neutralized with Amberlite-H+ IR-120, filtered, and concen-
trated to dryness, to obtain the allyl-â-D-maltoside (4).25
Yield: 44% starting from maltose.
Bin d in g Stu d ies. Titration experiments were performed
using NMR spectroscopy at a constant temperature of 303 K.
Chemical shifts in D2O were referenced to external C6D6 (7.15
ppm) in a coaxial tube. Solutions of glycophanes 2 and 3 (0.2-
0.9 mM) were freshly prepared for every new experiment.
Solutions for the guests (10-100 mM) were prepared from the
host solution in order to have a constant concentration of the
host during the titration. Binding constants were determined
by adding in portions, via microsyringe, a solution of guest to
a solution of host. The 1H NMR spectrum of each solution
was recorded, and the chemical shifts of different protons
obtained at 10-12 different host:guest concentration ratios
were used in an iterative least-squares-fitting procedure,
assuming formation of a 1:1 complex.18 Each titration experi-
ment was performed two or three times, and the association
constants given in Table 1 are the weighted averages of
different protons when the observed ICS was at least 0.08 ppm.
Saturation reached 30-60%, depending on association con-
stants. The maximum percent of estimated error was 20%.
Com p u ta tion P r oced u r e. All calculations were per-
formed in a Silicon Graphics workstation with the MM3* force
field as integrated in MACROMODEL v4.5. The initial
structures for glycophanes 2 and 3 were obtained from an
MM3* minimization using the MACROMODEL program and
the GB/SA solvation model described by Still and co-workers.19
The starting geometries for 2 and 3 were built from glucose
units in their stable 4C1 chair conformation, and the inter-
glycosidic torsional angles of the maltose unit, φ and ψ, defined
as H1′C1′O4C4 and C1′O4C4H4, were fixed at -3° and -30°,
one of the typical minima found for maltose (minimum B). The
glycosidic angles of reducing glucose, φa and ψa defined as
H1C1O1C(CH2) and C1O1C(CH2)C(Ph), were established at
φa ) -60° following the exo-anomeric effect, and ψa ) -180°
to allow closure of the macrocycle. The conformation of the
hydroxymethyl groups for glycophane 2 was set to trans-
gauche (tg). This position is the least favored in glucose
derivatives;20 therefore, the hydroxymethyl groups will proceed
to gg or gt during the minimization. In the case of glycophane
3, the hydroxymethyl group of the reducing glucose was set
to tg as for 2, and the hydroxymethyl group linked to the
aromatic ring was set to gt. This is the preferred orientation
in solution for 3 as indicated by the coupling constants J 5′,6′S
) 10.2 Hz, J 5′,6′R < 1 Hz, and J 6′S,6′R ) 11.8 Hz.
A Monte Carlo (MC) approach was used because of the
elevated number of torsional angles in both hosts. One of the
bonds between the carbonyl group and oxygen (O4′ for 2, or
O6′ for 3) was chosen as the ring closure bond. The GB/SA
solvation model was used to simulate the water during the
conformational search. A global search using 1000 MC steps
was performed for each glycophane. A number of torsions
between 2 and 13 were randomly selected for modification at
each MC step. After each MC step, the resultant geometry
was minimized using 5000 gradient conjugate steps. Struc-
tures were tested for duplication according to a least-squares
criterion (0.25 Å) and the Numbering System Rotation imple-
mented in MACROMODEL for highly symmetric molecules.
An energy window of 50 kJ ‚mol-1 was used as the criterion to
accept a given conformation. The total number of conformers
provided in the search was 310 for glycophane 2 and 222 for
glycophane 3. A criterion of 12 kJ ‚mol-1 over the global
minimum was established before comparing the structures
obtained in order to select the final set of conformers. Because
of the higher flexibility of glycophane 3, a criterion of 25
Allyl 4-O-(4,6-O-4-Meth oxyben zylid en e-r-D-glu cop yr a -
n osyl)-â-D-glu cop yr a n osid e (5). To a solution of allyl-â-D-
maltoside (4) (12.44 g, 32.49 mmol) in dry DMF (80 mL) were
added 4-methoxybenzaldehyde dimethyl acetal (11.84 g, 64.98
mmol) and p-toluenesulfonic acid (1.54 g, 8.12 mmol). The
reaction mixture was heated at 60 °C in a rotary evaporator
under vacuum for 2 h. Triethylamine (25 mL) was then added,
solvent was evaporated in vacuo, and the residue was purified
by column chromatography (CH2Cl2/MeOH/NEt3, 14:1:0:1) to
(21) van Gunsteren, J . W. F.; Berendsen, H. J . Mol. Phys. 1977, 34,
1311-1327.
(22) Berendsen, H. J .; Postma, J . P. M.; van Gunsteren, W. F.; De
Nola, A.; Hack, J . R. J . Chem. Phys. 1984, 81, 3684-3690.
(23) Wolfrom, M. L.; Thompson, A. In Methods in Carbohydrate
Chemistry; Whistler, R. L., Wolfram, M. L., Eds.; Academic Press: New
York, London, 1962; Vol. 1, p 334.
(24) Brauns, D. H. J . Am. Chem. Soc. 1929, 51, 1820-1831.
(25) Takahashi, Y.; Ogawa, T. Carbohydr. Res. 1987, 164, 277-296.
(18) We thank Prof. C. S. Wilcox (Pittsburgh University) and Prof.
C. A. Hunter (Sheffield University) for kindly providing the fitting
programs.
(19) Still, W. C.; Tempczy, K. A.; Hawley, R. C.; Hendrickson, T. J .
Am. Chem. Soc. 1990, 112, 6127-6129.
(20) Bock, K.; Duus, J . P. J . Carbohydr. Chem. 1994, 13, 513-543.

9220 J . Org. Chem., Vol. 63, No. 25, 1998
Morales et al.
obtain 5 (10.00 g, 62%) as a syrup: 1H NMR (300 MHz, CDCl3)
δ ) 7.41 (d, 2H, J ) 8.7 Hz), 6.88 (d, 2H, J ) 8.7 Hz), 6.05-
5.84 (m, 1H), 5.46 (s, 1H), 5.32 (m, 1H, J ) 17.2, 1.6 Hz), 5.21
(m, 1H, J ) 10.2, 1.2 Hz), 5.13 (d, 1H, J ) 3.6 Hz), 4.40-3.29
(m, 15H), 3.79 (s, 3H), 3.07-2.98 (m, 1H; OH), 2.55 (broad s,
1H), 1.80 (broad s, 3H); 13C NMR (CDCl3) δ ) 161.9, 135.7,
131.47, 128.82, 117.46, 114.34, 103.42, 103.31, 103.02, 82.46,
81.60, 77.79, 76.54, 74.70, 74.65, 72.12, 71.11, 69.77, 65.01,
62.30, 55.69, 50.27, 50.11, 49.85. Anal. C23H32O12: calcd C
55.19, H 6.44; found C 54.93, H 6.26.
Allyl 2,3,6-Tr i-O-4-m eth oxyben zyl-4-O-(2,3-di-O-4-m eth -
oxyb en zyl-4,6-O-4-m et h oxyb en zylid en e-r-D-glu cop yr a -
n osyl)-â-D-glu cop yr a n osid e (6). NaH (0.165 g, 6.89 mmol)
was added to a solution of 5 (0.230 g, 0.459 mmol) in DMF (6
mL), and the mixture was stirred at 0 °C for 20 min. Then
4-methoxybenzyl chloride (0.622 mL, 4.59 mmol) was added
dropwise, and stirring was continued at room temperature for
16 h. MeOH was added and the mixture poured into ice. The
aqueous phase was extracted with Et2O (5 × 25 mL), and the
combined organic extracts were dried (Na2SO4) and concen-
trated. The residue was purified by column chromatography
(AcOEt/hexane, 1:4) to obtain 6 (0.450 g, 89%) as a syrup: 1H
NMR (300 MHz, CDCl3) δ ) 7.43 (d, 2H, J ) 8.7 Hz), 7.28 (d,
3H, J ) 9.2 Hz), 7.20 (d, 3H, J ) 9.0 Hz), 7.09 (d, 4H, J ) 8.7
Hz), 6.90 (d, 2H, J ) 8.8 Hz), 6.81 (d, 6H, J ) 8.5 Hz), 6.80 (d,
2H, J ) 8.7 Hz), 6.71 (d, 2H, J ) 8.7 Hz), 6.18-5.76 (m, 1H),
5.63 (d, 1H, J ) 3.9 Hz), 5.49 (s, 1H), 5.34 (m, 1H, J ) 17.3,
1.7 Hz), 5.22 (m, 1H, J ) 10.3, 1.5 Hz), 4.85 (d, 2H, J ) 10.9
Hz), 4.80 (d, 2H, J ) 11.5 Hz), 4.68-4.38 (m, 10H), 4.19-3.84
(m, 5H), 3.82 (s, 3H), 3.79 (s, 3H), 3.78 (s, 6H), 3.74 (s, 3H),
3.72 (s, 3H), 3.69-3.43 (m, 6H); 13C NMR (CDCl3) δ ) 159.88,
159.18, 159.08, 159.00, 158.77, 134.10, 130.88, 130.29, 130.10,
129.97, 129.79, 129.47, 129.43, 129.18, 128.21, 127.28, 117.08,
113.68, 113.60, 113.44, 102.55, 101.03, 97.09, 84.54, 82.16,
81.90, 78.42, 78.25, 74.78, 74.21, 73.35, 73.02, 71.89, 70.07,
81.33, 78.89, 77.14, 47.85, 74.40, 74.29, 74.00, 73.26, 72.81,
72.67, 72.05, 71.54, 69.83, 68.21, 61.37, 54.88, 54.81. Anal.
C63H74O17: calcd C 68.58, H 6.76; found C 68.25, H 6.93.
Allyl 2,3,6-Tr i-O-4-m eth oxyben zyl-4-O-(4-O-acetyl-2,3,6-
tr i-O-4-m eth oxyben zyl-r-D-glu cop yr a n osyl)-â-D-glu cop y-
r a n osid e (9). A solution of 7 (18.7 g, 16.95 mmol) in 2:1
pyridine/acetic anhydride was stirred at rt for 15 h. The
mixture was concentrated and evaporated with toluene (3 ×
100 mL) to obtain 9 (19.00 g, 98%) as a syrup: 1H NMR (300
MHz, CDCl3) δ ) 7.28-7.07 (m, 12H), 6.92-6.69 (m, 12H),
6.07-5.91 (m, 1H), 5.57 (d, 1H, J ) 3.7 Hz; H1′), 5.34 (m, 1H,
J ) 17.2, 1.6 Hz), 5.22 (m, 1H, J ) 10.4, 1.6 Hz), 5.03 (t, 1H,
J ) 10.1 Hz; H4′), 4.85 (d, 2H, J ) 10.4 Hz), 4.75-4.38 (m,
13H), 4.26 (d, 1H, J ) 12.5 Hz), 4.15 (m, 1H), 3.97 (t, 1H, J )
10.0 Hz), 3.94-3.71 (m, 3H), 3.78 (s, 9H), 3.77 (s, 6H), 3.72 (s,
3H), 3.57-3.47 (m, 3H), 3.35-3.21 (m, 2H), 1.84 (s, 3H; OAc);
13C NMR (CDCl3) δ ) 169.31, 159.05, 158.92, 158.65, 133.97,
130.73, 130.57, 130.33, 130.29, 129.86, 129.79, 129.66, 129.26,
129.08, 128.97, 128.83, 128.18, 128.03, 125.11, 116.98, 113.50,
113.45, 102.42, 96.39, 84.20, 81.66, 78.66, 78.55, 74.42, 74.12,
73.37, 73.04, 72.94, 72.86, 72.81, 70.55, 69.96, 69.27, 69 03,
68.32, 55.00, 54.93, 53.31, 20.73. Anal.
68.16, H 6.69; found C 68.40, H 6.84.
C65H76O18: calcd C
2,3,6-Tr i-O-4-m eth oxyben zyl-4-O-(4-O-a cetyl-2,3,6-tr i-
O-4-m eth oxyben zyl-r-D-glu cop yr a n osyl)-â-D-glu cop yr a n -
ose (10). To a solution of 9 (953 mg, 0.832 mmol) and 1,4-
diazabicyclo[2.2.2]octane (47 mg, 0.416 mmol) in toluene (4
mL), were added tris(triphenylphosphine)rhodium(I) chloride
(308 mg, 0.333 mmol) and 10% aqueous EtOH (35 mL), and
the mixture was stirred for 16 h at 85 °C. Then it was
concentrated, CH2Cl2 was added, and the mixture was washed
with water. The aqueous phase was extracted with CH2Cl2,
and the organic combined extracts were concentrated and
dissolved in acetone/water (10:1) (27 mL). Mercury(II) oxide
(450 mg, 2.08 mmol) and a solution of mercury(II) chloride (450
mg, 1.66 mmol) in acetone/water (10:1) (10 mL) were added,
and the mixture was stirred at 100 °C for 2 h and 20 min.
The mixture was cooled at rt and filtered through Celite; the
bed was washed with CH2Cl2, and the filtrate and washings
were concentrated. The residue was diluted with CH2Cl2,
washed with 10% IK, dried (Na2SO4), concentrated, and
purified by column chromatography (AcOEt/hexane, 1:2 to 1:1)
to obtain 10 (745 mg, 81%) as a syrup: 1H NMR (300 MHz,
CDCl3) δ ) 7.27-7.09 (m, 24H), 6.87-6.72 (m, 24H), 5.55 (d,
1H, J ) 3.7 Hz; H1′R or â), 5.53 (d, 1H, J ) 3.6 Hz; H1′R or â),
5.17 (m, 1H), 5.05 (t, 1H, J ) 10.0 Hz; H4′R or â), 5.02 (t, 1H,
J ) 10.0 Hz; H4′R or â), 4.85-4.27 (m, 26H), 4.03-3.31 (m,
22H), 3.79 (s, 12H), 3.79 (s, 9H), 3.78 (s, 3H), 3.77 (s, 3H), 3.76
(s, 3H), 3.74 (s, 3H), 3.74 (s, 3H), 1.88 (s, 3H; OAc), 1.87 (s,
3H; OAc); 13C NMR (CDCl3) δ ) 169.77, 169.26, 159.04, 158.87,
158.81, 158.53, 138.52, 130.77, 130.55, 130.46, 130.41, 130.21,
129.89, 129.68, 129.49, 129.40, 129.22, 129.11, 129.02, 128.97,
128.90, 128.24, 128.06, 124.02, 123.51, 123.46, 122.98, 113.51,
113.41, 113.34, 113.01, 112.48, 111.97, 111.93, 111.45, 111.42,
97.12, 96.52, 96.32, 90.39, 83.87, 82.43, 80.69, 79.34, 78.52,
78.35, 74.32, 73.85, 73.56, 73.29,73.18, 73.04, 72.83, 72.74,
72.59, 72.17, 70.46, 69.28, 69.14, 68.87, 68.19, 54.89, 54.80,
46.59, 20.61, 12.52. Anal. C62H72O18: calcd C 67.37, H 6.56;
found C 67.28, H 6.41.
68.84, 68.58, 63.15, 55.13, 55.04. Anal.
68.71, H 6.58; found C 69.00, H 6.68.
C63H72O17 calcd C
Allyl 2,3,6-Tr i-O-4-m et h oxyb en zyl-4-O-(2,3,6-t r i-O-4-
m et h oxyb en zyl-r-D-glu cop yr a n osyl)-â-D-glu cop yr a n o-
sid e (7) a n d Allyl 2,3,6-Tr i-O-4-m eth oxyben zyl-4-O-(2,3,4-
t r i-O-4-m et h oxyb en zyl-r-D-glu cop yr a n osyl)-â-D-glu co-
p yr a n osid e (8). To a suspension of sodium cyanoborohydride
(25.00 g, 397.8 mmol) and 3 Å powdered molecular sieves in
DMF (130 mL) was added a solution of 6 (25.12 g, 22.81 mmol)
in DMF (160 mL) at 0 °C. A solution of trifluoroacetic acid
(17.45 mL, 228 mmol) in DMF (90 mL) was then slowly added
for 40 min. The mixture was stirred at rt for 16 h, and solvent
was evaporated. The residue was extracted with Et2O (9 ×
150 mL), and the combined organic extracts were dried
(Na2SO4) and concentrated. The residue was purified by
column chromatography (toluene/AcOEt, 9:2) to obtain 7 (18.7
g, 74%) and 8 (2.79 g, 11%), both as syrups. Com p ou n d 7:
1H NMR (300 MHz, CDCl3) δ ) 7.29-7.10 (m, 12H), 6.89-
6.72 (m, 12H), 6.04-5.93 (m, 1H), 5.63 (d, 1H, J ) 3.6 Hz),
5.36 (m, 1H, J ) 17.2, 1.6 Hz), 5.23 (m, 1H, J ) 10.4, 1.6 Hz),
4.91-4.33 (m, 15H), 4.22-4.15 (m, 1H), 4.01 (t, 1H, J ) 8.9
Hz), 3.80 (s, 6H), 3.79 (s, 9H), 3.75 (s, 3H), 3.74-3.39 (m, 10H);
13C NMR (CDCl3) δ ) 159.03, 158.91, 158.64, 137.55, 134.02,
130.81, 130.38, 130.31, 129.92, 129.61, 129.21, 129.04, 128.99,
128.80, 128.14, 128.00, 125.09, 116.89, 113.65, 113.50, 102.42,
96.28, 84.32, 81.61, 80.78, 78.45, 74.61, 74.41, 74.02, 73.19,
72.99, 72.75, 72.45, 71.22, 70.58, 69.89, 69.33, 68.79, 54.94,
21.19. Anal. C63H74O17: calcd C 68.58, H 0.76; found C 68.83,
2,3,6-Tr i-O-4-m eth oxyben zyl-4-O-(4-O-a cetyl-2,3,6-tr i-
O-4-m et h oxyb en zyl-r-D-glu cop yr a n osyl)-r,â-D-glu cop y-
r a n osyl Tr ich lor oa cetim id a te (11). A mixture of compound
10 (2.49 g, 2.35 mmol), trichloroacetonitrile (1.88 mL, 18.82
mmol), and flame-dried potassium carbonate (650 mg, 470
mmol) in CH2Cl2 (12 mL) was stirred at rt for 15 h. The
mixture was filtered through Celite, and the filtrate was
evaporated in vacuo to obtain 11 (R/â 1:2, 2.85 g, quantitative)
as a syrup. It was used without further purification: 1H NMR
(300 MHz, CDCl3) δ ) 8.71 (s, 1H; CdNH â), 8.62 (s, 1H; Cd
NH R), 7.28-7.10 (m, 24H), 6.85-6.72 (m, 24H), 6.51 (d, 1H,
J ) 3.7 Hz; H1R), 5.89 (d, 1H, J ) 6.8 Hz; H1â).
1
H 6.65. Com p ou n d 8: H NMR (300 Mz, CDCl3) δ ) 7.27-
7.21 (m, 8H), 7.16-7.1 (m, 4H), 6.89-6.81 (m, 10H), 6.78-
6.75 (m, 2H), 6.05-5.94 (m, 1H), 5.59 (d, 1H, J ) 3.7 Hz; H1′),
5.40 (m, 1H, J ) 17.2, 1.6 Hz), 5.35 (m, 1H, J ) 10.4, 1.3 Hz),
4.90-4.46 (m, 13H), 4.45 (m, 1H), 4.17 (m, 1H), 4.02 (t, 1H, J
) 9.3 Hz), 3.88 (t, 1H, J ) 9.3 Hz), 3.83-3.65 (m, 4H), 3.81 (s,
3H), 3.81 (s, 3H), 3.80 (s, 3H), 3.79 (s, 3H), 3.78 (s, 3H), 3.76
(s, 3H), 3.76-3.42 (m, 5H), 3.41 (dd, 1H, J ) 9.6, 3.7 Hz; H2′),
1.83 (m, 1H; OH6′); 13C NMR (CDCl3) δ ) 159.04, 158.95,
158.85, 158.57, 133.91, 130.72, 130.66, 130.23, 129.90, 129.82,
129.59, 129.38, 129.09, 128.93, 128.74, 128.07, 127.94, 125.03,
116.86, 113.55, 113.46, 113.39, 102.32, 96.07, 84.28, 81.58,
4-(Meth oxyca r bon yl)ben zyl 2,3,6-Tr i-O-4-m eth oxyben -
zyl-4-O-(4-O-a cetyl-2,3,6-tr i-O-4-m eth oxyben zyl-r-D-glu -
cop yr a n osyl)-r,â-D-glu cop yr a n osid e (12). A mixture of 11
(6.04 g, 4.83 mmol), methyl 4-(hydroxymethyl)benzoate (2.00

Carbohydrate-Carbohydrate Interactions
J . Org. Chem., Vol. 63, No. 25, 1998 9221
g, 12.08 mmol) and 3 Å molecular sieves in dry CH2Cl2 (120
mL) was stirred at rt for 90 min. Then the temperature was
reduced to -78 °C, a solution of trimethylsilyl triflate (0.1 M,
2.4 mL) in dry CH2Cl2 was slowly added, and the reaction
mixture was stirred for 20 min. NEt3(5 mL) was added, and
the mixture was allowed to rise to rt, filtered through Celite,
and washed with water and diluted NaCl solution. The
aqueous phase was extracted with CH2Cl2, and the combined
organic extracts were dried (Na2SO4) and concentrated. The
residue was purified by column chromatography (hexane/
acetone, 2:1) to give 12 (R/â 1.5:1, 4.82 g, 80%,). The R/â
mixture of compound 12 could be separated by column chro-
matography (CH2Cl2/acetone, 30:1), but it was more convenient
to separate them after deacetylation of position 4′ yielding
compounds 13R and 13â. Com p ou n d 12r: 1H NMR (300
MHz, CDCl3) δ ) 8.11 (d, 2H, J ) 8.3 Hz), 7.55 (d, 2H, J ) 8.3
Hz), 7.30-7.18 (m, 12H), 6.93-6.78 (m, 12H), 5.68 (d, 1H, J
) 3.7 Hz; H1′), 5.12 (t, 1H, J ) 9.8 Hz; H4′), 5.02-4.29 (m,
15H), 4.18-3.30 (m, 11H), 3.98 (s, 3H), 3.84 (s, 3H), 3.83 (s,
9H), 3.82 (s, 3H), 3.78 (s, 3H), 1.92 (s, 3H; OAc); 13C NMR
(CDC3) δ ) 169.30, 166.65, 159.14, 158.98, 158.89, 158.62,
142.35, 130.86, 130.53, 130.03, 129.84, 129.80, 129.71, 129.48,
129.46, 129.32, 129.11, 128.82, 128.32, 127.67, 113.57, 113.48,
113.44, 113.38, 96.50, 95.55, 81.33, 79.62, 78.59, 74.44, 73.85,
72.91, 72.83, 72.72, 70.52, 69.81, 69.17, 68.59, 68.24, 54.99,
54.91, 51.91, 20.74. Com p ou n d 12â: 1H NMR (300Mz, CDCl3)
δ ) 8.08 (d, 2H, J ) 8.2 Hz), 7.52 (d, 2H, J ) 8.2 Hz), 7.32-
7.14 (m, 12H), 6.91-6.76 (m, 12H), 5.65 (d, 1H, J ) 3.4 Hz;
H1′), 5.10-4.36 (m, 16H), 4.14-3.38 (m, 11H), 3.97 (s, 3H),
3.82 (s, 15H), 3.77 (s, 3H), 1.92 (s, 3H; OAc); 13C NMR (CDCl3)
δ ) 169.36, 166.68, 159.07, 158.95, 158.67, 142.69, 130.64,
130.54, 130.20, 130.15, 129.84, 129.75, 129.56, 129.34, 129.30,
129.13, 128.99, 128.85, 128.35, 128.19, 127.70, 127.22, 113.57,
113.51, 102.46, 96.49, 84.20, 81.67, 78.67, 78.52, 74.54, 74.47,
74.26, 73.49, 72.95, 72.88, 70.50, 70.15, 69.28, 8.91, 68.27,
60.18, 55.03, 54.95, 51.92, 30.70, 20.77, 14.02.
4-(Meth oxyca r bon yl)ben zyl 2,3,6-Tr i-O-4-m eth oxyben -
zyl-4-O-(2,3,6-tr i-O-4-m eth oxyben zyl-r-D-glu copyr an osyl)-
r-D-glu cop yr a n osid e (13r) a n d 4-(Meth oxyca r bon yl)ben -
zyl 2,3,6-Tr i-O-4-m eth oxyben zyl-4-O-(2,3,6-tr i-O-4-m eth -
oxyben zyl-r-D-glu copyr an osyl)-â-D-glu copyr an oside (13â).
To a solution of 12 (4.76 g, 3.797 mmol) as an R/â mixture in
CH2Cl2/MeOH (1:1) (116 mL) was added a solution of NaOMe
in MeOH (1M, 2.9 mL). The mixture was stirred at rt for 18
h, neutralized with Amberlite-H+ IR-120, filtered, and con-
centrated. The crude was purified by column chromatography
(CH2Cl2/acetone, 23:1) to obtain 2.54 g of 13R (56%) and 1.78
g of 13â (39%). Com p ou n d 13r: 1H NMR (300 MHz, CDCl3)
δ ) 8.05 (d, 2H, J ) 8.3 Hz), 7.49 (d, 2H, J ) 8.3 Hz), 7.30-
7.13 (m, 12H), 6.91-6.74 (m, 12H), 5.67 (d, 1H, J ) 3.6 Hz;
H1′), 5.00-4.31 (m, 15H), 4.17-3.42 (m, 12H), 3.95 (s, 3H;
COOCH3), 3.81 (s, 3H), 3.80 (s, 12H), 3.75 (s, 3H); 13C NMR
(CDCl3) δ ) 166.51, 159.08, 158.91, 158.82, 158.57, 142.35,
130.88, 130.78, 130.06, 129.82, 129.38, 129.30, 129.16, 128.94,
128.78, 128.18, 127.52, 113.54, 113.51, 113.44, 113.40, 96.28,
95.44, 81.39, 80.70, 79.65, 78.41, 74.56, 73.57, 72.88, 72.58,
72.51, 72.41, 72.22, 71.16, 70.50, 69.80, 69.28, 68.46, 66.12,
at rt for 20 h and neutralized with 2 N HCl solution. NEt3
(0.2 mL) was added to keep the pH slightly basic. The phases
were separated, the aqueous phase was extracted with CH2Cl2
(3 × 100 mL), and the combined organic extracts were dried
(Na2SO4) and concentrated to yield 14r (0.71 g, quantitative)
as a syrup. This compound was used without further purifica-
tion: 1H NMR (300 Mz, CDCl3) δ ) 8.13 (d, 2H, J ) 7.9 Hz),
7.48 (d, 2H, J ) 8.0 Hz), 7.31-7.12 (m, 12H), 6.91-6.74 (m,
12H), 5.68 (d, 1H, J ) 3.6 Hz; H1′), 5.01-4.32 (m, 15H), 4.19-
3.15 (m, 12H), 3.80 (s, 9H), 3.79 (s, 3H), 3.77 (s, 3H), 3.75 (s,
3H); 13C NMR (CDCl3) δ ) 170.67, 159.11, 158.99, 158.95,
158.85, 158.59, 141.25, 132.09, 130.97, 130.78, 130.13, 129.85,
129.81, 129.42, 129.28, 129.25, 129.02, 128.84, 128.24, 127.50,
113.61, 113.57, 113.50, 113.43, 96.28, 95.03, 81.46, 80.78,
79.63, 78.46, 74.66, 73.60, 72.95, 72.44, 72.18, 71.14, 70.55,
69.78, 69.26, 54.94, 44.90.
Syn th esis of Glycop h a n e 15. A mixture of dicyclohexyl-
carbodiimide (0.628 g, 3.048 mmol), 4-(dimethylamino)pyridine
(0.447 g, 3.658 mmol), and 4-(dimethylamino)pyridine hydro-
chloride (0.483 g, 3.048 mmol) in ethanol-free CHCl3 (97 mL)
was heated to reflux. Then a solution of compound 14R (1.46
g, 1.219 mmol) in CHCl3 (20 mL) was infused via syringe pump
over 90 min (the needle was inserted through the top of the
condenser and positioned such that refluxing chloroform
washed the drops of the solution). When the addition was
completed, the syringe was rinsed with 2 mL of chloroform,
and this material was delivered to the reaction vessel by
syringe pump over 30 min. The reaction mixture was refluxed
for 3 h and 30 min more, cooled to rt, concentrated, and
purified by column chromatography twice: hexane/acetone, 10:
7, and CH2Cl2/acetone, 28:1, affording pure 15 as a white solid
(425 mg, 29%): 1H NMR (500Mz, CDCl3) δ ) 8.02 (d, 2H, J )
8.3 Hz), 7.49 (d, 2H, J ) 8.3 Hz), 7.34 (d, 4H, J ) 8.8 Hz),
7.25 (d, 4H, J ) 8.5 Hz), 7.23 (d, 4H, J ) 8.8 Hz), 7.12 (d, 4H,
J ) 8.8 Hz), 6.99 (d, 4H, J ) 8.5 Hz), 6.98 (d, 4H, J ) 8.5 Hz),
6.83 (d, 4H, J ) 8.3 Hz), 6.81 (d, 4H, J ) 8.6 Hz), 6.79 (d, 4H,
J ) 8.8 Hz), 6.78 (d, 4H, J ) 8.5 Hz), 6.57 (d, 4H, J ) 8.8 Hz),
6.56 (d, 4H, J ) 8.8 Hz), 5.26 (d, 2H, J ) 10.3 Hz; CH2Ph),
5.18 (t, 2H, J ) 10.3 Hz; H4′), 5.00 (d, 2H, J ) 3.3 Hz; H1′),
4.82 (d, 2H, J ) 3.7 Hz; H1), 4.79-4.72 (m, 8H; CH2Ph), 4.62
(d, 2H, J ) 11.5 Hz; CH2Ph), 4.56-4.46 (m, 8H; CH2Ph), 4.34
(s, 4H; CH2Ph), 4.23-4.17 (AB system, 4H, J ) 11.5 Hz;
CH2Ph), 4.12-4.10 (m, 2H; H6), 4.06-4.00 (m, 6H; H3 + H3′
+ H5′), 3.96 (t, 2H, J ) 9.0 Hz; H4), 3.79 (s, 6H), 3.79 (s, 6H),
3.77 (s, 6H), 3.76 (s, 6H), 3.70 (broad d, 2H; H5), 3.65 (s, 6H),
3.64 (s, 6H), 3.59 (dd, 2H, J ) 9.5, 3.64 Hz; H2), 3.54 (dd, 2H,
J ) 9.8, 3.2 Hz; H2′), 3.39 (broad d, 2H; H6S), 3.37-3.31 (m,
4H; H6′R + H6′S); 13C NMR (CDCl3) δ ) 164.84, 159.03,
158.86, 158.78, 158.68, 158.60, 143.46, 131.58, 130.44, 130.11,
130.02, 129.64, 129.53, 129.42, 129.29, 129.09, 129.02, 128.80,
125.89, 113.51, 113.39, 113.23, 99.14, 96.84, 80.89, 79.34,
78.81, 77.19, 75.83, 74.23, 72.92, 72.62, 72.48, 71.98, 70.95,
70.63, 69.37, 68.91, 68.41, 68.35, 54.90, 0.54.74, 29.05; FABMS
m/z ) 2419.6 [M + guanidinium]+, C136H148O36 + guanidiniun
ion+ calcd m/z ) 2419.6. Anal. C136H148O36: calcd C 69.25, H
6.32; found C 68.98, H 6.21.
Syn th esis of Glycop h a n e 2. To a solution of 15 (100 mg,
0.042 mmol) in CH2Cl2 (5 mL) was added a solution of
trifluoroacetic 5% in CH2Cl2 (5 mL). The mixture was stirred
at rt for 5 h, concentrated, washed with Et2O (3 × 10 mL),
and dried under vacuum for 20 h. It was then purified in
water solution by desalting through a C18 Sep-pak column.
Lyophilization of the filtrate affords 2 as a white solid (37 mg,
95%): 1H NMR (500 MHz, D2O, 30 °C) δ (to C6D6) ) 8.12 (d,
4H, J ) 8.43 Hz), 7.63 (d, 4H, J ) 8.0 Hz), 5.25 (d, 2H, J )
3.7 Hz; H1′), 5.14 (d, 2H, J ) 3.3 Hz; H1), 5.12 (t, 2H, J ) 9.9
Hz; H4′), 4.87-4.78 (AB system, 4H; CH2), 4.16 (t, 2H, J )
9.5 Hz; H3′), 4.14 (m, 2H; H5′), 4.07 (t, 2H, J ) 9.5 Hz; H3),
3.78 (dd, 2H, J ) 9.5, 3.66 Hz; H2′), 3.78 (m, 2H; H5), 3.76
(m, 2H; H6), 3.71-3.61 (m, 8H; H2 + H6′R + H4 + H6′S),
3.56 (m, 2H; H6); 13C NMR (CD3OD) δ ) 167.17, 145.57,
130.95, 130.25, 128.13, 103.94, 100.30, 83.86, 74.77, 74.40,
74.03, 72.91, 70.33, 64.30, 62.48, 61.82; HRFABMS: m/z )
939.2754 [M + Na]+, C40H52O24Na calcd m/z ) 939.2746.
54.85, 54.80, 53.22, 51.72, 30.45. Anal.
C69H78O19: calcd C
68.41, H 6.49; found C 68.22, H 6.31. Com p ou n d 13â: 1H
NMR (300 MHz, CDCl3) δ ) 8.03 (d, 2H, J ) 8.4 Hz), 7.46 (d,
2H, J ) 8.4 Hz), 7.30-7.10 (m, 12H), 6.90-6.72 (m, 12H), 5.61
(d, 1H, J ) 3.6 Hz; H1′), 5.05-4.34 (m, 15H), 4.15-3.40 (m,
12H), 3.94 (s, 3H; COOCH3), 3.80 (s, 6H), 3.79 (s, 6H), 3.78 (s,
3H), 3.75 (s, 3H); 13C NMR (CDCl3) δ ) 166.68, 158.99, 158.90,
158.64, 142.71, 130.81, 130.70, 130.19, 129.90, 129.47, 129.21,
129.02, 128.95, 128.13, 127.19, 113.63, 113.53, 113.49, 113.45,
102.41, 96.36, 84.29, 81.59, 80.77, 78.43, 74.61, 74.50, 74.08,
73.27, 72.97, 72.73, 72.48, 71.17, 70.61, 70.04, 69.30, 68.72,
54.94, 54.88, 53.26, 51.79. Anal. C69H78O19: calcd C 68.41, H
6.49; found C 68.29, H 6.31.
4-(Ca r boxy)ben zyl 2,3,6-Tr i-O-4-m eth oxyben zyl-4-O-
(2,3,6-t r i-O-4-m et h oxyb en zyl-r-D-glu cop yr a n osyl)-r-D-
glu cop yr a n osid e (14r). To a solution of 13r (0.72 g, 0.596
mmol) in THF (22 mL) were added MeOH (22 mL) and
aqueous KOH solution (6 N, 22 mL). The mixture was stirred

9222 J . Org. Chem., Vol. 63, No. 25, 1998
Morales et al.
Syn th esis of Glycop h a n e 3. A solution of 2 (50 mg, 0.054
mmol) in MilliQ quality water (20 mL) was stirred at 50 °C in
an orbitalic shaker for 3 days. The mixture was lyophilized
to obtain 3 as a white solid (49 mg, 98%): 1H NMR (500 MHz,
D2O, 30 °C) δ (to C6D6) ) 8.05 (d, 4H, J ) 8.5 Hz), 7.50 (d, 4H,
J ) 8.5 Hz), 5.46 (d, 2H, J ) 3.4 Hz; H1′), 4.96 (m, 2H, J )
11.8, 1 Hz; H6′), 4.65 (d, 2H, J ) 3.9 Hz; H1), 4.33 (m, 2H, J
) 11.8, 10.2 Hz; H6′S), 3.95 (t, 2H, J ) 9.3 Hz; H3), 3.93 (m,
2H; H5′), 3.68 (t, 2H, J ) 10.2 Hz; H3′), 3.63 (dd, 2H, J )
10.2, 3.4 Hz; H2′), 3.61-3.54 (m, 6H; H5 + H6 + H6S), 3.45
(t, 2H, J ) 9.3 Hz; H4), 3.40 (t, 2H, J ) 9.2 Hz; H4′), 3.18 (dd,
2H, J ) 10.2, 3.4 Hz; H2); 13C NMR (D2O) δ ) 170.29, 145.78,
132.91, 132.75, 131.58, 131.39, 101.88, 99.82, 79.45, 76.07,
75.88, 74.73, 73.97, 73.66, 73.09, 71.60, 67.68, 63.20; HR-
FABMS: m/z ) 939.2778 [M + Na]+, C40H52O24Na calcd m/z
) 939.2746.
Ack n ow led gm en t. We thank Dr. J . J ime´nez-Bar-
bero and co-workers for their NMR and molecular
mechanics calculations help and Prof. M. Mart´ın-Lomas
for his continuous support. This work was suported by
the Spanish DGICYT (PB93-0127). J .C.M. and D.Z.
thank the Ministerio de Educacio´n y Ciencia for finan-
cial support.
J O9807823