Conformational differences between Fuc(α1-3)GlcNAc and its thioglycoside analogue

Article abstract of DOI:10.1016/S0008-6215(98)00066-4

NOE measurements and molecular mechanics calculations have been performed to study the conformational behaviour of Fuc(α1-3)GlcNAc and its thioglycoside analogue in solution. Experimental data show that, in contrast with the natural O-disaccharide, which

Full text of DOI:10.1016/S0008-6215(98)00066-4

Carbohydrate Research 308 (1998) 19±27
Conformational di€erences between
Fuc(ꢀ1±3)GlcNAc and its thioglycoside analogue
Begona Aguilera, Jesus Jimenez-Barbero *, Alfonso Fernandez-Mayoralas *
Â
Departamento de Quõmica Organica Biologica, Instituto de Quõmica Organica, CSIC, Juan de la Cierva
Â
Â
Â
3, 28006 Madrid, Spain
Â
Received 17 November 1997; accepted 5 March 1998
Abstract
NOE measurements and molecular mechanics calculations have been performed to study the
conformational behaviour of Fuc(ꢀ1±3)GlcNAc and its thioglycoside analogue in solution.
Experimental data show that, in contrast with the natural O-disaccharide, which is basically
monoconformational, the S-analogue shows two conformational families, namely syn and anti.
# 1998 Elsevier Science Ltd. All rights reserved
Keywords: NOE spectroscopy; Molecular mechanics calculations; Thiodisaccharides; Lewis antigens;
Conformational analysis
1. Introduction
resistant to glycosidase-catalysed hydrolysis, and
more stable at acidic pH than the corresponding O-
glycoside [4].
The sequence Fuc(ꢀ1±3)GlcNAc is a principal
constituent of Lex-bearing glycoconjugates, which
play an important role in numerous biological
phenomena. For instance, sialyl Lex tetra-
saccharide has been shown to be recognized by
E-selectin, a protein involved in the acute in¯am-
matory process [1], and is often expressed in tumor
cells and carcinomas [2]. Besides, we have reported
[3] that some oligosaccharides structurally related
to Lex are inhibitors of neural cell division. In
order to prepare oligosaccharides that are more
stable for in vivo experiments, we were interested
in the synthesis of Lex-analogues with a sulfur
atom linking the glucosaminyl and fucosyl moi-
eties. It is known that S-linked oligosaccharides are
An important aspect in understanding the mole-
cular bases of oligosaccharide-mediated biological
processes is the determination of the conforma-
tional ¯exibility around the interglycosidic bonds.
Therefore, it is important to determine whether the
conformational characteristics of the natural com-
pounds are re¯ected in the thio-oligosaccharide
analogues. We have carried out a comparative
conformational study of propyl O-(ꢀ-l-fucopyrano-
syl)-(1!3)-2-acetamido-2-deoxy-ꢁ-d-glucopyrano-
side (1) and its S-interglycosidic analogue 2, using
NMR spectroscopy and molecular mechanics
calculations. Thus, NOE measurements and mol-
ecular mechanics calculations [5,6] have been per-
formed to estimate the probability distribution of
conformers of both analogues in solution. Experi-
mental data show that, in contrast with the natural
* Corresponding authors. Fax: +34-1-5644853; e-mail:
iqojj01@pinar1.csic.es
0008-6215/98/$19.00 # 1998 Elsevier Science Ltd. All rights reserved
PII S0008-6215(98)00066-4

20
B. Aguilera et al./Carbohydrate Research 308 (1998) 19±27
O-analogue, which is basically monoconforma-
tional, the S-disaccharide presents two conforma-
tional families in solution, namely syn and anti.
Conformational analysis.ÐMolecular mechanics
calculations. The structures of 1 and 2 with the
atomic numbering is shown in Scheme 1. Torsional
angles around the glycosidic linkage are de®ned as
ꢂ H-10±C-10±O(S)±C-3 and C-10±O-10±O(S)±H-3.
Figs. 1 and 2 show the adiabatic maps calculated
from the respective relaxed energy maps for MM3*
(MM2* for 2) and AMBER* force ®elds [10,11],
using E=80. For 1, both force ®elds predict two
low energy regions, corresponding to the so-called
syn and É-anti regions [12,13]. The central syn
region is fairly extended along É, including both
positive and negative É values. Indeed, MM3*
predicts the existence of two di€erent minima
within this region which have been dubbed syn-A
and syn-B conformers, respectively. The syn region
heavily dominated the probability distribution
since the central region is populated in more than
94%, independent of the force ®eld used. In addi-
tion, the region around conformer syn-A (global
minimum in both cases) with positive É angles is
populated between 63 (AMBER*) and 76%
(MM3*). The geometries of the minima are very
2. Results and discussion
Synthesis.ÐThe synthesis of 1 started from per-
acetylated ꢀ-glucosamine 3 following procedures
for related disaccharides [7] (Scheme 1). Glycosy-
lation of 3 with allyl alcohol using SnCl4 as acid
promotor gave 4, which after deacetylation and
subsequent benzylidenation provided alcohol 5.
Treatment of 5 with fucosyl bromide 6, in situ,
prepared from ethyl 2,3,4-tri-O-benzyl-1-thio-ꢁ-l-
fucopyranoside [8], in the presence of Bu4NBr
a€orded the fully-protected disaccharide 7 in 79%
yield. Catalytic hydrogenation of 7 provided the
desired disaccharide 1.
The synthesis of thio-disaccharide 2 was recently
reported [9]. Brie¯y, 2 was obtained by regioselec-
tive opening of the cyclic sulfamidate 8 with fucose
1-thiolate 9, followed by deprotection steps.
Scheme 1.

B. Aguilera et al./Carbohydrate Research 308 (1998) 19±27
21
similar for both force ®elds, with minor di€erences
around the glycosidic linkages (Table 1, Fig. 3),
and are in accordance with the exo-anomeric e€ect
[14]. The potential energy surfaces obtained are
also similar to that obtained previously for the
same disaccharide using the regular MM3 force
®eld [15]. With this force ®eld, the global minimum
conformation also corresponds to the syn-A con-
former with È/É values of 45/16. It is interesting to
mention that, with both MM3* and AMBER*
force ®elds, the obtained global minima for this
disaccharide also corresponds satisfactorily to the
conformation found for the Lex trisaccharide in
the solid state [16] (È/É 40/15 and 50/12, for the
two independent molecules which exist in the
crystal).
The situation is di€erent for 2. In this case, both
AMBER* and MM2* force ®elds predict again the
two low energy regions (Figs 1 and 2), similar to
those obtained above for the natural compound,
and again the geometries of the minima are closely
related (Table 2, Fig. 4). AMBER* and MM2*
predict, however, di€erent global minima, and
rather distinct population distributions. Using
MM2*, the most populated region is that corre-
sponding to the central minimum (syn, ꢂ, =55,
16), and the global minimum is the syn-A
Fig. 1. Comparison of the adiabatic maps calculated by using
AMBER* (a,c), MM3* (b) and MM2* (d) force ®elds for
compounds 1 (a,b) and 2 (c,d). Levels are drawn every 2 kJ/
mol.
Table 1
Ensemble average populations (calculated from the relative
steric energy values) of the low energy regions of the O-
disaccharide. È/É points are assigned to any of the basic
regions based on proximity
AMBER*
MM3*
min
ꢂ,
Pop. (%)
ꢂ,
Pop. (%)
syn-A
anti
syn-B
42, 20
34, 169
Converges to
syn-A
94.2
5.8
Ð
61, 43
35, 170
34, 49
76.4
0.3
23.3
Fig. 2. Comparison of the probability distribution maps cal-
culated by using AMBER* (a,c), MM3* (b) and MM2* (d)
force ®elds for compounds 1 (a,b) and 2(c,d). Levels are drawn
at 10, 1, and 0.1% probability. Relevant proton±proton short
distances are labelled and superimposed onto the maps. Con-
tour levels are drawn at 2.0 and 3.5 A.
Fig. 3. Stereoscopic view of the global minima of 1: (A) con-
former syn-A, AMBER*; (B) conformer syn-A, MM3*.

22
B. Aguilera et al./Carbohydrate Research 308 (1998) 19±27
conformer. The population of the anti region now
amounts to 14%. AMBER* locates the most
populated region (49%) around the global mini-
mum, in this case the so-called anti conformer [12].
Nevertheless, the addition of syn-A and syn-B
conformers is basically equal to those populating
the anti valley.
quently compared to those obtained from the
previously mentioned force ®eld calculations. An
e€ective average correlation time of 45 ps at 310 K
was estimated from the matching of the experi-
mental NOEs obtained for two intraresidue proton
pairs to those theoretically predicted. No over-
lapping between key protons was found, thus
facilitating the NOE analysis (Fig. 5).
There are proton±proton short distances which
give characteristic NOEs which may help in dedu-
cing the existence of the three di€erent regions of
the conformational map. Consequently, these
NOE intensities will be sensitive to their respective
populations. For both compounds, H-3±H-10
de®nes the central syn-region (both A and B con-
formers), while H-3±H-5 is sensitive to population
around region syn-A. On the other hand, H-2±H-10
and H-4±H-1 are exclusive for the anti conforma-
tion. Therefore, the existence of the di€erent con-
formational families in solution could be detected
by the presence of these NOEs. Fig. 2 shows the
relevant interresidue proton distances for 1 and 2,
superimposed on the probability distribution maps.
NMR results. Measurements of nuclear Over-
hauser enhancements [17] were made and subse-
The experimental NOEs, compared to those cal-
culated are collected in Tables 3 and 4. For 1, both
MM3* and AMBER* force ®elds produce a rea-
sonable agreement between expected and observed
NOEs. In particular, the matching produced by
AMBER* may be considered almost excellent,
within experimental error. On the other hand,
MM3* predicts a H-3±H-50 NOE higher than the
experimental one, which probably means that the
actual population around minimun syn-A is smal-
ler than that calculated, 76%, more in agreement
with the AMBER* value of around 60%. The H-
3±H-10 intensity is fairly well reproduced by both
force ®elds. On the other hand, AMBER* predicts
correctly the H-4±H-5 NOE. Taking into account
all these data, the experimental NOEs could be
explained by a population distribution similar
to that proposed by AMBER*, always with the syn
and anti regions populated in about 90±95 and
Table 3
Experimental and calculated (ꢀc, 45 ps) normalized NOESY
intensities (%) for the O-disaccharide at 310 K in D2O
solution, at 500 MHz
Proton pair
Mixing time 300 ms
Mixing time 450 ms
&
Exp. AMBER* MM3* Exp. AMBER* MM3*
H3±H10
H2±H10
H4±H10
H3±H50
H4±H50
3.8
0.9
0.4
0.2
0.5
4.0
0.1
0.2
0.1
0.4
2.6
0.4
0.2
1.5
0.7
5.3
1.3
0.6
0.5
1.0
6.0
0.2
0.4
0.1
0.8
3.9
0.8
0.4
2.3
1.1
Fig. 4. Stereoscopic view of the major minima of 2: (A) con-
former anti, AMBER*; (B) conformer syn-A, MM2*.
Table 4
Experimental and calculated (ꢀc, 45 ps) normalized NOESY
intensities (%) for the S-disaccharide at 310 K in D2O solution,
at 500 MHz
Table 2
Ensemble average populations (calculated from the relative
steric energy values) of the low energy regions of the S-
disaccharide. È/É points are assigned to any of the basic
regions based on proximity
Proton pair
Mixing time 300 ms
Mixing time 450 ms
&
Exp. AMBER* MM2* Exp. AMBER* MM2*
AMBER*
MM2*
H3±H10
H2±H10
H4±H10
H3±H50
H4±H50
2.4
0.8
1.0
0.6
0.8
1.5
1.1
2.7
0.2
2.1
0.4
0.9
0.4
3.2
0.9
1.6
0.8
0.5
2.3
1.6
4.0
0.4
3.2
0.9
1.4
0.9
min
ꢂ,
Pop. (%)
ꢂ,
Pop. (%)
syn-A
anti
syn-B
56, 27
47, 177
32, 25
43.9
49.4
6.7
55, 16
39, 170
27, 25
78.8
14.5
6.7

B. Aguilera et al./Carbohydrate Research 308 (1998) 19±27
23
1
Fig. 5. 1D DPFGNOE H NMR spectra of 2 at 500 MHz acquired with a mixing time of 450 ms.

24
B. Aguilera et al./Carbohydrate Research 308 (1998) 19±27
5±10%, respectively. Similar results have been
described for other disaccharides [13].
3. Experimental
In the case of 2, the experiments show a relative
decrease of the H-3±H-1 NOE along with the cor-
responding increase of the H-2±H-1 and H-4±H-1
NOEs, a clear indication of a ¯ow of population
from the syn to the anti region, in comparison with
the natural compound, 1. Nevertheless, from
inspection of Table 4, it is evident that the match
between experimental and theoretical results is
much better when the MM2* results are con-
sidered. In particular, AMBER* overestimates the
exclusive H-2±H-1 and H-4±H-1 NOEs, which
probably means that the actual population around
minimun anti is smaller than that calculated, 49%,
and closer to that calculated by MM2* (around
15%). The H-3±H-10 intensity is fairly well repro-
duced by both force ®elds, especially by MM2*.
Therefore, the experimental NOEs could be
explained by a population distribution similar to
that proposed by MM2*, with the syn and anti
regions populated in about 80±85 and 15±20%,
respectively. A clear increase in the population of
the anti region is evident when the interglycosidic
oxygen is substituted by sulfur. Similar results have
been recently described for other thio-oligo-
saccharides [18±20].
In conclusion, both compounds present a con-
formational equilibrium in solution between two
major regions. In the case of the natural com-
pound, 1, one of these is heavily populated (ca.
95%). Nevertheless, the glycosidic linkages still
present certain ¯exibility. In the case of the thio
analogue 2, the syn region is still dominant,
although the anti conformer is clearly present (ca.
15±20%). This increase of ¯exibility could prob-
ably be due to the fact that the C±S bond is longer
than the C±O analogue, and fewer steric con¯icts
occur for the anti conformation. It has also to be
considered that the C±S±C bond angle is ca. 98ꢀ
compared to ca. 117ꢀ for C±O±C. According to the
observed populations, the change in relative energy
is smaller than 1 kcal/mol. From the point of view
of the molecular recognition of these sugars by a
natural receptor, it seems that the global three-
dimensional shape of both compounds in solution
is fairly similar, and therefore the thio dis-
accharide could be chosen as a mimic of the nat-
ural compound, especially for studying interactions
with enzymes that may destroy O-glycosidic
linkages, but that cannot a€ect the S-glycosidic
linkages.
General Methods.ÐMelting points are uncor-
rected. TLC was performed using TLC plates
GF254 with detection by charring with 5% H2SO4
in EtOH. Column chromatography was performed
on silica gel (230±400 mesh). The eluent used is
indicated and solvent ratios refer to volume. Sol-
vents were distilled over drying agents: dimethyl-
formamide (DMF), BaO; dichloromethane
(DCM), CaH2; acetonitrile (MeCN), CaH2. 1H
NMR spectra were registered at 500 or 300 Mz. 13C
NMR spectra were obtained at 125 or 50 MHz.
Allyl 2-acetamido-3,4,6-tri-O-acetyl-2-deoxy-b-d-
glucopyranoside (4).ÐTo a solution of 2-acet-
amido-1,3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-glucopyran-
ose (3) (11.7 g, 30 mmol) and AllOH (6.1 mL,
90 mmol) in MeCN (100 mL) under argon, was
added 4 A molecular sieves (5 g). After 10 min,
SnCl4 (4.2 mL, 36 mmol) was added dropwise, and
ꢀ
the mixture was stirred at 70 C for 8 h. After
cooling, the mixture was neutralized with triethy-
lamine, ®ltered through a pad of celite and con-
centrated. The residue was puri®ed by column
chromatography (1:2 hexane±EtOAc) to give
2-acetamido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-gluco-
pyranoside (1.39 g, 12%) and then 4 (5.6 g, ꢀ49%) as
ꢀ
a white solid, mp 163±165 C (lit 164±167 C [21]).
[ꢀ]D 13.8ꢀ (c 1, CHCl3) (lit. 14.4ꢀ [21]). H
1
NMR (CDCl3, 300 MHz): ꢃ 5.92±5.79 (m, 1 H, ±
CH ˆ CH2), 5.62 (d, 1 H, JNH,2 8.8 Hz, NH), 5.30
(dd, 1 H, J3,4 9.4 Hz, J2,3 10.6 Hz, H-3), 5.25±5.17
(m, 2 H, ±CH ˆ CH2), 5.07 (t, 1 H, J3,4&J4,5
9.5 Hz, H-4), 4.71 (d, 1 H, J1,2 8.3 Hz, H-1), 4.37±
4.30 (m, 1 H, OCH2±CH ˆ ), 4.26 (dd, 1 H, J5,6a
4.7 Hz, J6a,6b 12.2 Hz, H-6a), 4.13 (dd, 1 H, J5,6b
2.4 Hz, J6a,6b 12.2 Hz, H-6b), 3.88 (dt, 1 H,
J2,NH&J1,2 8.6 Hz, J2,3 10.6 Hz, H-2), 3.72±3.67
(m, 1 H, H-5), 2.08 (s, 3 H, Me), 2.03 (s, 3 H, Me),
2.02 (s, 3 H, Me), 1.95 (s, 3 H, Me). 13C NMR
(CDCl3, 50 MHz): ꢃ 170.75 (CO), 172.62 (CO),
170.23 (CO), 169.33 (CO), 133.57 (±CH ˆ CH2),
117.59 (±CH ˆ CH2), 99.65 (C-1), 72.41, 71.72,
69.86 (OCH2±CH ˆ ), 68.78, 62.19 (C-6), 54.63 (C-
2), 23.19 (Me), 20.64 (Me), 20.60 (Me), 20.53 (Me).
Allyl 2-acetamido-4,6-O-benzylidene-2-deoxy-b-
d-glucopyranoside (5).ÐCompound
4
(1.43 g,
3.71 mmol) was treated with a solution of NaOMe
in MeOH (0.1M, 140 mL) at r.t. for 1 h. After this
time, the reaction mixture was neutralized with
Amberlita IR-120 (H+), and ®ltered. The solvent
was evaporated, and the crude was suspended in

B. Aguilera et al./Carbohydrate Research 308 (1998) 19±27
25
0
0
MeCN (23 mL) and treated with PhCH(OMe)2
.
(2.77 mL, 11.5 mmol) and p-TsOH H2O (35.3 mg,
H-50, OCH2±CH ˆ ), 3.93 (dd, 1 H, J3 ,4 2.6 Hz,
J2 ,3 10.0 Hz, H-30), 3.75 (t, 1 H, J5,6a&J6a,6e
10.0 Hz, H-6a), 3.36±3.59 (m, 4 H, H-2, H-4, H-5,
H-40), 1.6 (s, 3 H, Me), 0.8 (d, 3 H, JMe,5 6.3 Hz,
0
0
0.18 mmol) at r.t. for 2 h. Then, the mixture was
diluted with dichloromethane (50 mL), neutralized
with triethylamine and concentrated. The residue
was puri®ed by column chromatography
(40:1!20:1 DCM±MeOH) to give 5 (1.17 g, 91%)
0
Me-Fuc). 13C NMR (CDCl3, 50 MHz): ꢃ 170.47
(CO), 138.61 (C-ipso), 138.54 (C-ipso), 137.28 (C-
ipso), 133.71 (±CH ˆ CH2), 128.98, 128.60, 128.38,
128.22, 128.16, 127.86, 127.71, 127.64, 127.28,
126.19 (Ar), 117.57 (±CH ˆ CH2), 101.58 (CHPh),
99.91 (C-1), 98.33 (C10), 80.81, 79.82, 77.64, 76.37,
74.88, 74.16, 72.51, 70.26, 68.81, 66.86, 66.19, 58.27
(C-2), 23.21 (Me), 16.27 (Me-Fuc). Anal. Calcd for
C45H51NO10: C, 70.57; H, 6.71; N, 1.83. Found: C,
70.27; H, 6.80; N, 1.92.
Propyl O-(a-l-fucopyranosyl)-(1!3)-2-acetamido-
2-deoxy-b-d-glucopyranoside (1).ÐA solution of 7
(90 mg, 0.117 mmol) in MeOH (25 mL) and EtOAc
(5 mL) was added Pd/C (250 mg) and the mixture
was stirred under H2 atmosphere for 4.5 h. After
this time, the reaction mixture was ®ltered through
a pad of celite, and concentrated. The residue was
puri®ed by column chromatography (4:1 DCM±
ꢀ
as a white solid, mp 279±282 C (dec) (lit, 279±
ꢀ
1
281 C (dec) [22]). H NMR (CDCl3, 300 MHz): ꢃ
7.35±7.52 (m, 5 H, Ar), 5.85±5.92 (m, 1 H, OCH2±
CH ˆ ), 5.71 (d, 1 H, JNH,2 5.7 Hz, NH), 5.57 (s, 1
H, CHPh), 5.24±5.36 (m, 2 H, ±CH ˆ CH2), 4.78
(d, 1 H, J1,2 8.3 Hz, H-1), 4.33±4.41 (m, 2 H,
OCH2±CH ˆ , H-6e), 4.20 (t, 1 H, J2,3&J3,4 9.5 Hz,
H-3), 4.07±4.15 (m, 1 H, OCH2±CH ˆ ), 3.81 (t, 1
H, J5,6a&J6a,6b 9.8 Hz, H-6a), 3.59 (t, 1 H,
J3,4&J4,5 9.4 Hz, H-4), 3.44±3.54 (m, 2 H, H-2, H-
5), 2.06 (s, 3 H, Me). 13C NMR (CDCl3±CD3OD
7:1, 50 MHz): ꢃ 172.33 (CO), 136.92 (C-ipso),
133.48 (±CH ˆ CH2), 129.04, 128.09, 126.11 (Ar),
117.34 (±CH ˆ CH2), 101.65 (CHPh), 100.20 (C-1),
81.49, 70.03, 68.47, 66.09, 57.26 (C-2), 22.74 (Me).
Allyl O-(2,3,4-tri-O-benzyl-a-l-fucopyranosyl)-
(1!3)-2-acetamido-4,6-O-benzylidene-2-deoxy-b-
d-glucopyranoside (7).ÐEthyl 2,3,4-tri-O-benzyl-1-
thio-ꢁ-l-fucopyranoside [8] (430 mg, 0.9 mmol)
was dissolved in dichoromethane (6.5 mL) under
ꢀ
MeOH) to give 1 (46 mg, 96%), mp 228±231 C
(dec). [ꢀ]D 112.6ꢀ (c 0.68, MeOH). H NMR
1
(D2O, 500 MHz): ꢃ 4.98 (d, 1 H, J1 ,2 4.1 Hz, H-10),
4.53 (d, 1 H, J1,2 8.7 Hz, H-1), 4.33 (m, 1 H, H-50),
3.93 (dd, 1 H, J5,6a 2.3 Hz, J6a,6b 12.3 Hz, H-6a),
3.86±3.78 (m, 4 H, OCH, H-30, H-40, H-2), 3.76
(dd, 1 H, J5,6b 5.6 Hz, J6a,6b 12.3 Hz, H-6b), 3.70
0
0
ꢀ
argon. At 0 C, bromine (47.3 mL, 0.91 mmol) was
added and the mixture was stirred for 30 min. The
mixture was concentrated in vacuo and residual
bromine was removed by evaporation with toluene
(3Â10 mL), to give crude bromide 6. Acceptor 2
(209 mg, 0.6 mmol) was dissolved in DMF (5.5 mL)
and 4 A molecular sieves and Bu4NBr (193.4 mg,
0.6 mmol) were added. The mixture was kept under
an argon atmosphere and bromide 6 (0.9 mmol)
dissolved in dichloromethane (1 mL) was added.
After stirring for 18 h, the reaction was quenched
with ethanol. After 30 min, the mixture was diluted
with dichloromethane and ®ltered through a pad of
celite. The ®ltrate was washed with saturated aqu-
eous NaHCO3, brine, dried (Na2SO4) and con-
centrated in vacuo. The residue was puri®ed by
column chromatography (2:1 hexane±EtOAc) to
give 7 (361 mg, 79%) as a white solid, mp 82±85 ꢀC,
[ꢀ]D 84.5ꢀ (c 1, CHCl3). 1H NMR (CDCl3,
300 MHz): ꢃ 7.17±7.47 (m, 15 H, Ar), 5.75±5.86 (m,
1 H, ±CH ˆ CH2), 5.63 (d, 1 H, JNH,2 7.1 Hz, NH),
5.49 (s, 1 H, CHPh), 5.14±5.28 (m, 2 H, ±
(dd, 1 H, J1 ,2 4.1 Hz, J2 ,3 10.5 Hz, H-20), 3.64 (t, 1
H, J2,3&J3,4 8.5 Hz, H-3), 3.57 (m, 1 H, OCH),
3.52 (dd, 1 H, J3,4 8.5 Hz, J4,5 8.9 Hz, H-4), 3.45
(m, 1 H, H-5), 1.81 (s, 3 H, Me), 1.33 (m, 2 H,
0
0
0
0
0
CH2), 0.94 (d, 1H, JMe,5 6.7 Hz, Me-Fuc), 0.65 (t, J
7.3 Hz, Me). 13C NMR (D2O, 125 MHz): ꢃ 175.56
(CO), 101.85 (C-1), 100.88 (C-10), 81.43, 76.86,
73.31 (OCH2), 72.83, 70.54, 69.62, 68.98, 67.89,
61.80 (C-6), 56.3 (C-2), 23.17 (Me), 23.06 (CH2),
16.16 (Me-Fuc), 10.56 (Me). Anal. Calcd for
C17H31NO10: C, 49.87; H, 7.63; N, 3.42. Found: C,
49.54; H, 7.51; N, 3.40.
4. Conformational calculations: molecular
mechanics
Glycosidic torsion angles are de®ned as ꢂ H-10±
C-10±O(S)±C-3 and C-10±O(S)±C-3±H-3 for both
analogues. This convention is related to that which
employs heavy atoms (O-5 for È and C-4 for É) by
adding 120ꢀ. The reducing end was considered to
be ꢁ-O-methylated for the sake of simplicity.
CH ˆ CH2), 5.05 (d, 1 H, J1 ,2 3.4 Hz, H-10), 4.55±
4.93 (m, 7 H, 3 CH2Ph, H-1), 4.25±4.35 (m, 3 H,
H-6e, H3, OCH2±CH ˆ ), 4.01±4.08 (m, 3 H, H-20,
0
0

26
B. Aguilera et al./Carbohydrate Research 308 (1998) 19±27
Relaxed (ꢂ, ) potential energy maps were calcu-
lated using MM3* (MM2* for the S-analogue due
to the lack of parameters of MM3* for sulfur) and
AMBER* force ®elds as implemented in MACRO-
MODEL 4.5 [23], with dielectric constant E=80.
These MACROMODEL force ®elds di€er from
the regular force ®elds in the treatment of the elec-
trostatics since they employ charge±charge instead
of dipole±dipole interactions. The AMBER* force
®eld uses Homans parameters for carbohydrates
[24].
uate the in¯uence of three spin e€ects. At short
mixing times, multispin e€ects are fairly negligible
but NOEs are small. At longer mixing times, the
multispin e€ect becomes relevant but NOEs are
stronger and easy to measure [17]. The spectra were
6
simulated from the average distances hr ikl calcu-
lated from the relaxed energy maps at 310 K. Iso-
1
tropic motion and external relaxation of 0.1 s
were assumed. A ꢄc of 45 ps was used to obtained
the best match between experimental and calcu-
lated NOEs for two given intraresidue proton pairs
(H-10±H-20 and H-50±H-30+H-40).
All calculations were made for the ꢁ-O-methyl
anomer of GlcNAc. Four initial geometries were
considered: cc, cr, rr and rc, obtained by combin-
ing the positions r (reverse clockwise) and c
(clockwise) for the orientation of the secondary
hydroxyl groups of both pyranoid moieties. The
®rst character corresponds to the non-reducing
fucose moiety, and the second one, to the GlcNAc.
The gauche±trans rotamer [25] of the GlcNAc
moiety was considered for the ! torsion angle (O-
5±C-5±C-6±O-6), as +60ꢀ. Four relaxed energy
maps were obtained following a similar protocol to
that described previously [13]. A grid search of 18ꢀ
was employed for the generation of the potential
energy surfaces. Adiabatic surfaces [10,11] were
built, and the probability distributions calculated
for each ꢂ, point according to a Boltzmann
function.
All the NOE calculations were automatically
performed by a home-made program, available
from the authors upon request [13].
Acknowledgements
The authors thank DGICYT (Grant PB96-0833)
and the Mizutani Glycoscience Foundation for
®nancial support. B.A. is grateful to the Ministerio
de Educacion y Cultura for a fellowship.
References
[1] L.A. Lasky, Science, 258 (1992) 964±969.
[2] K. Fukushima, M. Hirota, P.I. Terasaki, A.
Wakisaka, H. Togashi, D. Chia, N. Suyama, Y.
Fukushi, E. Nudelman, and S. Hakomori, Cancer
Res., 44 (1984) 5279±5285.
5. NMR experiments
[3] F.F. Santos-Benito, A. Fernandez-Mayoralas, M.
Martõn-Lomas, and M. Nieto-Sampedro, J. Exp.
Med., 176 (1992) 915±918; J.M. Coteron, S.
Kamaljit, J.L. Asensio, M. Domõnguez-Dalda, A.
Fernandez-Mayoralas, J. Jimenez-Barbero, M.
Martõn-Lomas, J. Abad-Rodrõguez, and M. Nieto-
Sampedro, J. Org. Chem., 60 (1995) 1502±1519; M.
Nieto-Sampedro, C. Bailon, A. Fernandez-Mayor-
alas, M. Martõn-Lomas, B. Mellstrom, and J.R.
Naranjo, J. Neuropath. Exp. Neur., 55 (1996) 169±
177.
[4] J. Defaye and J. Gelas, in Atta-ur-Rahman (Ed.),
Studies in Natural Products Chemistry, Vol. 8,
Elsevier, Amsterdam, 1991, pp 315±357; H. Dri-
guez, Top. Curr. Chem., 187 (1997) 85±116.
[5] T. Peters and B.M. Pinto, Curr. Opin. Struct. Biol.,
6 (1996) 710±716.
NMR spectra of both derivatives were recorded
at 310 K in D2O, using a Varian Unity 500 MHz
spectrometer. Proton chemical shifts were refer-
enced to residual HDO at ꢃ 4.64 ppm. COSY and
TOCSY spectra permitted the assignment of the
proton spin systems. Selective 1D double pulse
®eld gradient spin echo experiments [26] were
recorded for both sugars using four di€erent mix-
ing times, namely 150, 300, 450, and 600 ms. Only
the data at 300 and 450 ms are shown.
6. NOE calculations
NOESY spectra were simulated according to a
complete relaxation matrix approach, following
the protocol previously described [13], using four
di€erent mixing times (between 150 and 600 ms).
Di€erent mixing times were used in order to eval-
[6] A. Imberty, Curr. Opin. Struct. Biol., 7 (1997) 617±
623.
[7] S.S. Rana, J.J. Barlow, and K.L. Matta, Carbo-
hydr. Res., 91 (1981) 149±157; B.M. Heskamp,
G.H. Veeneman, G.A. van der Marel, C.A.A. van

B. Aguilera et al./Carbohydrate Research 308 (1998) 19±27
27
Boeckel, and J.H. van Boom, Recl. Trav. Chim.
Pays-Bas, 114 (1995) 398±402.
[17] D. Neuhaus and M.P. Williamson, The Nuclear
Overhauser E€ect in Structural and Conformational
Analysis, VCH, New York, 1989.
[18] K. Bock, J.O. Duus, and S. Refn, Carbohydr. Res.,
253 (1994) 51±67.
[19] A. Geyer, G. Hummel, T. Eisele, S. Reinhardt, and
R.R. Schmidt, Chem. Eur. J., 2 (1996) 981±988.
[20] U. Nilsson, R. Johansson, and G. Magnusson,
Chem. Eur. J., 2 (1996) 295±302.
[21] M. Kiso and L. Anderson, Carbohydr. Res., 72
(1979) C12±C14.
[8] H. Lohn, Carbohydr. Res., 139 (1985) 105±113.
[9] B. Aguilera, A. Fernandez-Mayoralas, and C. Jar-
amillo, Tetrahedron, 53 (1997) 5863±5876.
[10] A. French and J.W. Brady (Eds.), Computer Mod-
eling of Carbohydrate Molecules, ACS Symp. Ser.
430 (1990).
[11] M. Martõn-Pastor, J.F. Espinosa, J.L. Asensio, and
J. Jimenez-Barbero, Carbohydr. Res., 298 (1997)
15±49.
[12] J. Dabrowski, T. Kozar, H. Grosskurth, and N. E
Nifant'ev, J. Am. Chem. Soc., 117 (1995) 5534±5539.
[13] J.L. Asensio and J. Jimenez-Barbero, Biopolymers,
35 (1995) 55±73.
[14] H. Thogersen, R.U. Lemieux, K. Bock, and B.
Meyer, Can. J. Chem., 60 (1982) 44±57.
[15] A. Imberty, E. Mikros, J. Koca, R. Mollicone, R.
Oriol, and S. Perez, Glycoconjugate J., 12 (1995)
331±349.
[22] M.A.E. Shaban, V.N. Reinhold, and R.W. Jean-
loz, Carbohydr. Res., 55 (1977) 213±233.
[23] F. Mohamadi, N.G.J. Richards, W.C. Guida, R.
LisKamp, C. Cau®eld, G. Chang, T. Hendrickson,
and W.C. Still, J. Comput. Chem., 11 (1990) 440±
467.
[24] S.W. Homans, Biochemistry, 29 (1990) 9110±9118.
[25] K. Bock and J. Duus, J. Carbohydr. Chem., 13
(1993) 513±543.
[16] S. Perez, N. Mouhous-Riou, N.E. Nifant'ev, Y.E
Tsvetkov, B. Bachet, and A. Imberty, Glycobiol-
ogy, 6 (1996) 537±542.
[26] K. Stott, J. Stonehouse, J. Keeler, T.L. Hwang,
and A.J. Shaka, J. Am. Chem. Soc., 117 (1995)
4199±4200.