Synthesis of a BSA-Lex glycoconjugate and recognition of Le x analogues by the anti-Lex monoclonal antibody SH1: The identification of a non-cross reactive analogue

Article abstract of DOI:10.1016/j.bmc.2010.08.040

A Le^x trisaccharide functionalized with a cysteamine arm was prepared and this synthesis provided additional information on the reactivity of N-acetylglucosamine O-4 acceptors when they are glycosylated with trichloroacetimidate donors activated with excess BF₃·OEt ₂. In turn, this trisaccharide was conjugated to BSA lysine side chains through a squarate-mediated coupling. This BSA-Le^x glycoconjugate displayed 35 Le^x haptens per BSA molecule. The relative affinity of the anti-Le^x monoclonal antibody SH1 for the Le^x antigen and analogues of Le^x in which the d-glucosamine, l-fucose or d-galactose residues were replaced with d-glucose, l-rhamnose and d-glucose, respectively, was measured by competitive ELISA experiments. While all analogues were weaker inhibitors than the Le^x antigen, only the analogue of Le^x in which the galactose residue was replaced by a glucose unit showed no binding to the SH1 mAb. To confirm that the reduced or loss of recognition of the Le^x analogues by the anti-Le^x mAb SH1 did not result from different conformations adopted by the analogues when compared to the native Le^x antigen, we assessed the conformational behavior of all trisaccharides by a combination of stochastic searches and NMR experiments. Our results showed that, indeed, the analogues adopted the same stacked conformation as that identified for the Le^x antigen. The identification of a trisaccharide analogue that does not cross-react with Le^x but still retains the same conformation as Le^x constitutes the first step to the design of a safe anti-cancer vaccine based on the dimeric Le^x tumor associated carbohydrate antigen.

Full text of DOI:10.1016/j.bmc.2010.08.040

Bioorganic & Medicinal Chemistry 18 (2010) 7174–7185
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Synthesis of a BSA-Lex glycoconjugate and recognition of Lex analogues
by the anti-Lex monoclonal antibody SH1: The identification of a non-cross
reactive analogue
Jo-Wen Wang , Ari Asnani à, France-Isabelle Auzanneau
⇑
Department of Chemistry, University of Guelph, Guelph, Ontario, Canada N1G 2W1
a r t i c l e i n f o
a b s t r a c t
A Lex trisaccharide functionalized with a cysteamine arm was prepared and this synthesis provided addi-
tional information on the reactivity of N-acetylglucosamine O-4 acceptors when they are glycosylated
with trichloroacetimidate donors activated with excess BF3ꢀOEt2. In turn, this trisaccharide was conju-
gated to BSA lysine side chains through a squarate–mediated coupling. This BSA-Lex glycoconjugate dis-
played 35 Lex haptens per BSA molecule. The relative affinity of the anti-Lex monoclonal antibody SH1 for
Article history:
Received 8 July 2010
Revised 16 August 2010
Accepted 19 August 2010
Available online 22 August 2010
the Lex antigen and analogues of Lex in which the
replaced with -glucose, -rhamnose and -glucose, respectively, was measured by competitive ELISA
D-glucosamine, L-fucose or D-galactose residues were
Keywords:
Oligosaccharide synthesis
Lewis X
mAb Inhibitors
Conformational analysis
D
L
D
experiments. While all analogues were weaker inhibitors than the Lex antigen, only the analogue of
Lex in which the galactose residue was replaced by a glucose unit showed no binding to the SH1 mAb.
To confirm that the reduced or loss of recognition of the Lex analogues by the anti-Lex mAb SH1 did
not result from different conformations adopted by the analogues when compared to the native Lex anti-
gen, we assessed the conformational behavior of all trisaccharides by a combination of stochastic
searches and NMR experiments. Our results showed that, indeed, the analogues adopted the same
stacked conformation as that identified for the Lex antigen. The identification of a trisaccharide analogue
that does not cross-react with Lex but still retains the same conformation as Lex constitutes the first step
to the design of a safe anti-cancer vaccine based on the dimeric Lex tumor associated carbohydrate
antigen.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
attempting to use the Tumor Associated Carbohydrate Antigen
(TACA) dimLex as a target for anti-cancer vaccine development,
Aberrant glycosylation in human cancer was first indirectly sug-
gested in the mid-sixties when several papers reported the accu-
mulation of fucose-containing glycolipids in adenocarcinomas.1
One such glycoplipid displaying the dimeric Lex hexasaccharide
(dimLex, Chart 1) was reported to accumulate in colonic and liver
adenocarcinoma and it was suggested that additional fucosylation
of internal GlcNAc residues in polylactosamine chains was associ-
ated with the progression of colorectal cancer.2–4 However since it
was first characterized,5 the Lex antigenic determinant or X deter-
an important factor to consider is the possibility that such vaccine
would also elicit an immune response against the X determinant
that would eventually lead to the destruction of healthy cells. In
this context, our research program aims at discovering analogues
of the dimLex hexasaccharide that would retain internal epitopes
displayed on the surface of cancer cells by dimLex but no longer
possess epitopes associated with the Lex trisaccharide A0(B0)C0. In-
deed, the occurrence of such internal epitopes has been repeat-
edly2,4 established when monoclonal antibodies against dimLex
(FH4, SH2) were shown to recognize selectively the dimLex and
trimLex antigens but only weakly bind to the Lex antigen.2,4 In con-
trast anti-Lex monoclonal antibodies (e.g. FH3, SH1) were clearly
shown to also recognize the X determinant when it was displayed
at the non-reducing end of dimLex.2,4 In turn, these antibodies have
been extensively used to differentiate dimLex expressed by colo-
rectal tumor cells from the Lex antigen expressed on healthy tis-
sues.2,4 Thus, our goal is to discover dimLex analogues that retain
the internal epitopes recognized by anti-dimLex SH2-like antibod-
ies but no longer possess epitopes recognized by anti-Lex SH1-like
antibodies. We are investigating if hexasaccharides 2–4 (Chart 1) in
minant (b-D-Galp(1?4)[a-L-Fucp(1?3)]-D-GlcNAcp) displayed at
the non-reducing end of dimLex has also been found on numerous
normal cells and tissues such as kidney tubules, gastrointestinal
epithelial cells, and cells of the spleen and brain.6 Thus, when
⇑
Corresponding author.
E-mail address: fauzanne@uoguelph.ca (F.-I. Auzanneau).
Present address: Montreal Neurological Institute, 3801 University Street, Suite
207, Montreal, Canada H3A 2B4.
à
Present address: Jenderal Soedirman University, Purwokerto, 53122 Jawa Tengah,
Indonesia.
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.08.040

J.-W. Wang et al. / Bioorg. Med. Chem. 18 (2010) 7174–7185
7175
OH
HO
OH
HO
OH
OH
HO
O
OH
OH
O
O
O
O
C'
O
O
HO
O
A'
C
O
HO
O
O
A
OR
O
O
O
OH
OH
NHAc
OH
NHAc
NHAc
O
B'
O
O
OH
B
OH
OH
OH
HO
OH
OH
HO
HO
1 DimLex
10 R = (CH2)3S(CH2)2NH2
11 R = All
Chart 3.
R4
OH
OH
OH
HO
OH
O
O
R5
HO
O
C'
O
O
A'
C
O
2. Results and discussion
2.1. Synthesis of Lex-BSA conjugate 23
O
O
A
O
O
OH
R1
OH
NHAc
O
B'
R2
O
R3
B
OH
OH
R3
The Lex-BSA glycoconjugate 23 was obtained through the squa-
rate–mediated7–10 coupling of trisaccharide 10 to the BSA lysine
residues (Chart 3). We describe here the preparation of trisaccha-
ride 10 from the allyl Lex trisaccharide 11 prepared in our labora-
tory by Asnani11 as well as through an alternate synthetic scheme
from the individual monosaccharide building blocks.
OH
R2
HO
2 R1 = R2 = R4 = OH, R3 = R5 = H
3 R1 = NHAc, R2 = R5 = H, R3 = R4 = OH
4 R1 = NHAc, R2 = R5 = OH, R3 = R4 = H
Chart 1.
2.1.1. Synthesis of trisaccharide 10: additional results on the
glycosylation at O-4 of N-acetylglucosamine glycosyl acceptors
There are numerous reports in the literature that deal with the
chemical12–17 or chemoenzymatic18 preparation of Lex analogues
as well as that of Lex intermediate building blocks to be further
converted into the Sialyl Lex tetrasaccharide. It has been estab-
lished early on by Jacquinet and Sinaÿ12 as well as by Hindsgaul
et al.13 that the glycosylation at O-4 of N-acetylglucosamine in
such syntheses was often difficult. Indeed, the majority of these
syntheses employed acceptors that carried an N-Phth, N-TCP,
N-Troc, azido or oxazolidinone at C-2, and only two reports16,17
in addition to these early syntheses12,13 use an N-acetylglucosa-
mine acceptor. The first one is that of Roy and co-workers16 that
describes the regioselective glycosylation of an acceptor free at
O-3 and O-4, and the second one is our recent synthesis17 of Lex
in which we have applied conditions (2 equiv BF3ꢀOEt2, 5 equiv do-
nor, rt or 40 °C) that we had found19–21 to efficiently promote glu-
cosylation at O-4 of N-acetylglucosamine. Indeed over the past few
years our studies have shown that the success of such glycosyla-
tion depended on the nature of the substituent already present
which one of the sugar units A0, B0 or C0 defining the X determinant
in dimLex (1) is replaced by another sugar residue, that is, N-acetyl-
glucosamine by glucose (2), fucose by rhamnose (3), and galactose
by glucose (4), would meet these requirements. Since our first goal
with these analogues is to abolish their cross-reactivity with the
Lex antigen we have first embarked on studying the cross-reactiv-
ity of the Lex-OMe trisaccharide 5 with the Lex analogue trisaccha-
rides 6–9 shown in Chart 2. In analogue 6 N-acetylglucosamine is
replaced by glucose; in analogue 7 fucose is replaced by rhamnose;
in analogue 8 galactose is replaced by glucose and in analogue 9 N-
acetylglucosamine and fucose are replaced with glucose and rham-
nose, respectively. To complete our study of the cross-reactivity of
these analogues with the natural Lex antigen towards binding with
the monoclonal antibody SH1,4 we report here the preparation of a
Lex-BSA glycoconjugate. In turn, the affinity of anti-Lex mAb SH1
for analogues 5–9 was studied by ELISA and a non-cross reactive
analogue was identified. Finally, the conformations adopted by
the Lex trisaccharide and the analogues were studied by a combi-
nation of NMR and stochastic searches.
HO
HO
OH
OH
OH
HO
HO
O
OH
O
O
O
OMe
O
O
O
OH
OH
OMe
R
O
OH
OH
O
NHAc
OH
O
HO
HO
5 R = NHAc (Lex)
6 R = OH
OH
7
OH
OH
O
OH
HO
HO
OH
O
O
HO
HO
O
O
O
OMe
O
OMe
O
OH
OH
NHAc
OH
O
O
OH
HO
OH
HO
OH
OH
8
9
Chart 2.

7176
J.-W. Wang et al. / Bioorg. Med. Chem. 18 (2010) 7174–7185
at O-3 in the acceptor20 but also and more surprisingly17 on the
structure of the aglycone carried by the acceptor. Since our re-
search required the preparation of additional cysteamine adduct
10 we investigated the O-4 galactosylation of N-glucosamine gly-
cosyl acceptors carrying an allyl aglycone at C-1. Thus, we report
here our results when galactosylating monosaccharide acceptor
14 and disaccharide acceptor 16 which were prepared from the
known22 intermediates 12 and 15, respectively (Scheme 1). Alcohol
12 was chloroacetylated at O-3 and the benzylidene ring in the
resulting intermediate 13 was reductively opened to give acceptor
14 while disaccharide acceptor 16 was obtained through the
reductive opening of the benzylidene ring in disaccharide 15.
Using the reactions conditions that we have previously estab-
lished17,19–21 (2 equiv of BF3ꢀOEt2, 40 °C) to glycosylate O-4 of N-
acetylglucosamine, we then attempted the coupling of monosac-
charide 14 with 5 equiv of the known23 trichloroacetimidate donor
17. Under these conditions, the reaction proceeded slowly and
unreacted acceptor was still present in the reaction mixture after
the usual reaction time of 1 h.17,20 After an extended reaction time,
the desired disaccharide 18 was isolated in a disappointing 43%
yield. This reaction was difficult to follow by normal phase TLC
and our attempts to use reverse phase HPLC to either follow the
glycosylation or purify the product were also unsuccessful.
In contrast galactosylation of disaccharide acceptor 16 with do-
nor 17 could easily be followed by RP HPLC and to limit the ex-
pected20 degradation of the product at 40 °C, the reaction was
carried out at room temperature. Two HPLC traces at 50 and
100 min of reaction are shown on Figure 1a and b, respectively.
As can be seen on Figure 1a most of the acceptor 16 had been con-
verted to the trisaccharide 19 after 50 min of reaction and the pres-
ence of two additional products 20 and 21 indicated that
degradation was already taking place. Indeed, at 100 min of reac-
tion most of the trisaccharide product 19 had disappeared and
two new products were now clearly present in the reaction mix-
ture. After additional 40 min at 40 °C, the reaction was quenched
and three products were separated by flash chromatography
(Scheme 2). The two compounds seen by RP HPLC were identified
as the known24 fucose hemiacetal 21 (15% yield) and the endo/exo
mixture of orthoester 20 (60% yield). The third product which
eluted with the solvent front in the RP HPLC was identified as
disaccharide 22 (25% yield). The major isomer present in com-
pound 20 (75%) was assumed to be the less sterically hindered
exo orthoester and presented in the 1H NMR spectrum a character-
istic signal at 1.74 ppm corresponding to the methyl orthoace-
tate.25 In addition, the galactosyl H-2 signal was found at
Figure 1. HPLC trace after (a) 50 min and (b) 100 min of reaction for the
glycosylation of disaccharide acceptor 16 with donor 17 (5 equiv) activated with
2 equiv of BF3ꢀOEt2 at rt.
4.05 ppm which was, as expected, considerably up field from the
signal that is typically found for this hydrogen if the galactosyl res-
idue is acetylated at C-2 (see for instance H-20 in disaccharide 22
found at ꢁ5.2 ppm). When the reaction was quenched after
40 min the desired trisaccharide 19 was obtained in 64% yield.
The protected trisaccharide 19 was then converted to the cyste-
amine derivative 10 in two steps: the UV-promoted addition of 2-
aminoethanethiol hydrochloride to allyl glycoside 19 followed by
the removal of all protecting groups under dissolving metal condi-
tions [Na/NH3(l)] (Scheme 3).17 The resulting trisaccharide 10 was
purified by gel permeation chromatography and was obtained pure
as the acetate salt in 73% yield after lyophilization from water. The
cysteamine adduct 10 was also prepared via the UV-promoted
addition of 2-aminoethanethiol hydrochloride8 to allyl glycoside
11.11
2.1.2. Squarate mediated coupling to BSA
Following the well established squarate mediated conjugation
method7–10 glycoconjugate 23 was prepared in two steps from
the cysteamine adduct 10 (Scheme 4). Trisaccharide 10 was dis-
solved in methanol, allowed to react with diethylsquarate at room
temperature and when TLC showed its complete conversion to the
less polar squarate derivative the solvent was evaporated. The
squarate adduct (51 equiv) was allowed to react at a 30 mM hap-
ten concentration with the lysine side chains of BSA (Sigma Cat.
No. A4378) in a pH 10 (0.1 M) carbonate buffer for 7 days at room
temperature. The resulting glycoconjugate was dialyzed against
water and the average level of Lex hapten incorporation per BSA
molecule was determined by the Dubois26 colorimetric assay to
be of 35 3. This loading corresponds to a 70% coupling yield
(based on the hapten) and a 59% occupancy level of the 59 lysine
residues present in the protein.27 This result is in excellent agree-
R3O
O
R2O
R1O
OAll
NHAc
12 R1 = H, R2, R3 = >CHPh
13 R1 = ClAc, R2, R3 = >CHPh
14 R1 = ClAc, R2 = H, R3 = Bn
ment with results obtained by Kovác and co-workers10 in their re-
ClAcCl, C5H5N
ˇ
NaCNBH3
HCl-Et2O
cent thorough study of this coupling method.
2.2. Competitive binding experiments with the Lex analogues
5–9
R2O
O
R1O
OAll
O
The relative affinity of the anti-Lex monoclonal antibody SH14
for the Lex antigen 515 and analogues (6–9)15,19 was measured by
competitive ELISA experiments using glycoconjugate 23 as the
immobilized ligand.28 Figure 2 shows the inhibition curves of
mAb SH1 for each of the soluble ligands 5–9 that were fitted to a
four parameters logistic sigmoidal equation.29 The concentration
of each inhibitor required for 50% inhibition (IC50) and the
NHAc
O
OBn
OBn
BnO
Et3SiH
TfOH
15 R1, R2 = >CHPh
16 R1 = H, R2 = Bn
Scheme 1.

J.-W. Wang et al. / Bioorg. Med. Chem. 18 (2010) 7174–7185
7177
OAc
O
AcO
AcO
OBn
O
OAll
NHAc
O
ClAcO
AcO
14
18
BF3 OEt2
.
OAc
O
AcO
AcO
AcO
OC(NH)CCl3
BF3 OEt2
17
OAc
O
AcO
AcO
.
OBn
O
OAll
NHAc
16
+
O
HO
AcO
OAc
O
AcO
OBn
22
O
O
AcO
OAc
O
AcO
AcO
OAll
O
AcO
OH
NHAc
O
OBn
19
O
O
OBn
O
+
OBn
BnO
+
OBn
BnO
O
21
O
OBn
OBn
BnO
20
Scheme 2.
ergy of binding to SH1 while replacing the fucose residue by a
rhamnose unit (entry 3) led to a change in free energy of binding
consistent with the loss of a key polar interaction to the SH1 com-
bining site.30 A cumulative effect was observed when both the N-
acetylglucosamine and fucose sugars were substituted by glucose
and rhamnose, respectively (entry 4) but cross-reactivity with
the native antigen was nonetheless still observed. In sharp con-
trast, analogue 8 in which the galactose residue is replaced by a
glucose unit, showed no binding to the SH1 mAb even at high con-
centration (entry 5). This result suggests that the non-reducing end
galactose residue is essential to conserve cross-reactivity with the
natural Lex antigen and that modification of this residue and par-
ticularly at O-4 may totally abolish this cross-reactivity. To confirm
that the loss of recognition of analogue 8 by the anti-Lex mAb SH1
did not result from different conformations adopted by analogue 8
when compared to the native Lex antigen we embarked on the con-
formational analysis of analogue 8 as well as that of compounds 5–
7 and 9.
19
HS(CH2)2NH3+,Cl- UV
OAc
O
AcO
AcO
OBn
O
O
O(CH2)3S(CH2)2NH2
NHAc
O
AcO
O
OBn
OBn
BnO
Na/NH3(l)
10
Scheme 3.
corresponding changes in free energy
D
(
D
G) of binding referenced
2.3. Conformational analysis on trisaccharide analogues 5–9
to the Lex soluble antigen 5 are given in Table 1. As can be seen
from these results all analogues 6–9 were weaker inhibitors than
the natural Lex antigen. Replacing the glucosamine unit by a glu-
cose residue (entry 2) resulted in only a small decrease in free en-
In a previous paper,31 we have described the results of biased sto-
chastic conformational searches on reducing analogues of trisaccha-
rides 5–8 using an early version of the Molecular Operating
HO
OH
O
O
OH
O
1) Diethylsquarate
O
BSA
O
HO
MeOH
O (CH2)3S(CH2)2NH
NH
O
10
OH
OH
2) BSA, pH 10
NHAc
O
OH
HO
35
23
Scheme 4.

7178
J.-W. Wang et al. / Bioorg. Med. Chem. 18 (2010) 7174–7185
our attention on the conformational features associated with these
linkages in the trisaccharides 5–9. The orientations adopted around
the glycosidic linkages are described below by the dihedral angles:
U
= O5–C1–O1–Cx and W = C1–O1–Cx–Cx+1 and the signs of the
torsions are in agreement with the recommendations of the
IUPAC-IUB Commission of Biochemical Nomenclature.42
2.3.1. Stochastic searches: conformational families of
trisaccharides 5–9
In these stochastic searches, minima in the potential energy
surface are randomly sampled through the rotation of all bonds
(including ring bonds) to random dihedral angles as well as
through a random 0.4 Å Cartesian perturbation of atom positions.
For each analogues 5–9, two searches of 150,000 iterations were
carried out to generate new conformers that were each minimized
using the AMBER94 force field39 with implicit solvation by the
GB/SA continuum model.34 In the first search a RMS tolerance of
0.01 Å on heavy atoms was used to determine whether generated
conformations were considered as new. In the second search a
RMS tolerance of 0.1 Å was applied on all atoms including all
hydrogens atoms to account for possible hydrogen bonding. For
each compound, the databases containing the conformations se-
lected by each search were combined and only those conforma-
tions found within 10 kcal molꢂ1 from the global minima were
retained. In turn we focused our attention on those conformations
found within 3 kcal molꢂ1 from the global minima since these con-
formations may be populated in solution. As shown in Table 2, the
searches yielded between 836 and 1865 conformations within
10 kcal molꢂ1 from the global minima but the numbers of confor-
mations found within 3 kcal molꢂ1 of the global minima were con-
siderably smaller from 36 conformations for analogue 7 to 100 for
analogue 6.
Figure 2. Inhibition curves for mAb SH1 using glycoconjugate 23 as immobilized
antigen and analogues 5–9 as soluble competitive inhibitors: 5 (d), 6 (4), 7 (N), 8
(.), 9 ( ).
r
Table 1
These conformations clustered around well-defined values of
the glycosidic torsions that fell into 4 distinct families which are
represented by the overlaid lowest energy conformations found
for each family shown on Figure 3a (more detailed results of the
stochastic searches are given in the Supplementary data). As can
be seen on Figure 3a and in Table 3 (entry 1), the same global min-
imum (A) was found for all analogues. This conformation corre-
sponds to the global minimum identified by numerous modeling
studies of the Lex trisaccharide35–38 and which was experimentally
supported by NMR36,37 and X-ray crystallography.38 For all ana-
logues (5–9) this stacked conformation was in fact the most repre-
sented conformation in the final databases (Fig. 3b). Indeed, no
other conformation was found for analogue 9 while 98% of the con-
formations found for analogue 6 clustered around these glycosidic
torsions (Fig. 3b). The first and second local minima (Fig. 3a, B and
C) found for the Lex analogue 5 were also represented in the dat-
abases obtained for analogues 7 and 8 but were not found for ana-
logues 6 or 9. Finally, a third local minimum (Fig. 3a, D) was found
for analogues 6 and 8 which was not represented in the databases
of analogues 5, 7, and 9. The energy as well as glycosidic torsions
measured for the lowest energy representatives of each family
(B–D) and for each analogue (5–9) are given in the Supplementary
data. Shown Table 3 (entries 2–6) are the Boltzman-averaged
Inhibition data for MAb SH1 with analogues 5–9
Entry
Inhibitor
IC50a
(
l
M)
D
(D )
G)b (kcal molꢂ1
1
2
3
4
5
5
6
7
9
8
156
204
1068
2037
ꢃ5800
0
0.2 2%
1.1 4%
1.5 3%
c
—
a
Concentration of inhibitor required for 50% inhibition using solid phase glyco-
conjugate 23 as immobilized antigen.
b
Values determined from the expression
D(DG) = rt ln([I1]/[I2] where [I2] is the
IC50 measured for the reference inhibitor analogue 5 and [I1] the IC50 measured for
each analogue 6–9 with R = 1.98 cal Kꢂ1and T = 296 K.
c
No inhibition observed at 5800 lM.
Environment program suite (MOE2001)32 which offered a limited
choice of force fields. In that study the MMFF94 force field33 and
the Generalized Born/Surface Area continuum solvation model34
(GB/SA) were used to identify preferred conformations adopted by
the Lex trisaccharide. The global minima that were identified for
the reducing trisaccharide Lex and for the analogues in which the
glucosamine or galactosyl unit were replaced by glucose were in
agreement with the preferred conformation reported in the litera-
ture for the Lex antigen.35–38 In contrast, these preliminary results
suggested that the Lex analogue in which the fucose residue was re-
placed by a rhamnose unit was adopting a minimum energy
conformation drastically different from that adopted by the Lex
trisaccharide.31 Having now synthesized analogues 5–915,19 and re-
ceived a version of MOE (MOE2005) that implements the AMBER94
force field39 known to be efficient for carbohydrate modeling,40 we
performed new stochastic searches on analogues 5–9 and supported
these results by NMR experiments. Since the different orientations
around the glycosidic linkages have the most substantial impact
on the overall molecular shape of oligosaccharides,41 we focused
Table 2
Number of conformations found within 10 and 3 kcal molꢂ1 from
the global minima
Analogue
10 kcal molꢂ1
3 kcal molꢂ1
5
6
7
8
9
1318
1646
1524
836
56
100
36
61
52
1865

J.-W. Wang et al. / Bioorg. Med. Chem. 18 (2010) 7174–7185
7179
4 gives for each analogue (entries 2–6) the Boltzman-averaged in-
ter-proton distances: H-10 to H-3, H-100 to H-4, H-100 to H-6a and H-
6b, H-50 to H-200, that are typically short in the stacked minimum
energy conformation35–38 (entry 1) and have been supported
experimentally for the Lex antigen.36–38 As can be seen, these val-
ues are essentially identical to the average torsions and distances
measured in the global minimum (Tables 3 and 4, entry 1) common
to all analogues. Thus, it is clear that the stacked conformation (A)
previously identified as the rigid global minimum adopted by the
Lex antigen is also the preferred conformation adopted by all our
analogues 6–9. Indeed, although other conformations were fairly
well represented in the database obtained for analogue
8
(Fig. 3b) the Boltzman-averaged values of the glycosidic torsions
and distances given in Tables 3 and 4 (entry 5) show that the ener-
getically most favored conformation is the global minimum. Thus,
the results obtained in these stochastic searches confirm, as we
had previously found,31 that replacing the N-acetyl-
or D-galactose unit by a D
D
-glucosamine
-glucose residue in the Lex trisaccharide
does not lead to a change in the conformation. In contrast our
observation here that replacing the -fucose residue with an
L
L-
rhamnose unit did not either lead to a different conformation than
the Lex antigen contradicts our original findings using the MMFF94
force field.31 This difference is likely resulting from the different
parameterizations of the force fields involved. To confirm that in-
deed the analogues 6–9 adopt the same conformation than the
Lex antigen 5 the conformational behavior of compounds 5–9
was assessed by NMR experiments.
2.3.2. NMR measurements
The 1H chemical shift assignments for trisaccharides 5–9 mea-
sured at 300 K in D2O are reported in the Supplementary data.
The vicinal coupling constants measured for the three sugar units
supported an average 4C1 conformation for the galactose and N-
acetylglucosamine rings and a 1C4 conformation of the fucose
and rhamnose rings and have been reported.15,19 Inter-residue
NOE interactions were evaluated using 1D ROESY experiments at
300 K selectively irradiating H-10, H-50, and H-100 for each ana-
logues. Since we are particularly interested in the conformational
behavior of analogue 8 that is not recognized by the mAb SH1,
we show here in Figure 4a the 1H NMR spectrum and the 1H, 1H
ROESY spectra (mixing time 200 ms) upon selective irradiation of
protons H-10, H-50, and H-100 for this analogue. The 1H NMR and
1H, 1H ROESY spectra obtained for analogue 5–7 and 9 are given
in the Supplementary data. For all analogues, the expected intra-
residue cross-relaxation signals were evident: cross-relaxation sig-
nals were observed for the glycosidic linkages between H-10 and H-
3 as well as between H-100 and H-4; cross-relaxation signals were
observed from H-100 to H-6a and H-6b; and cross-relaxation signals
between H-50 and H-200 were also seen in all spectra. Although for
the rhamnosylated analogues 7 and 9, H-100 and H-50 gave overlap-
ping signals, the intensity of the cross-relaxation signal to H-200
when both H-100 and H-50 were irradiated together was clearly
indicative of a close proximity of H-50 and H-200. Cross-relaxation
buildup curves were obtained for those isolated signals over 9–
10 mixing times ranging from 20 to 400 ms. The normalized build-
up curves obtained for analogue 8 are given here Figure 4b, while
those obtained for analogue 5–7 and 9 are given in the Supplemen-
tary data. The normalized cross relaxation buildup curves (Fig. 3)
were fitted to a double exponential equation and the slopes at ini-
tial buildup were calculated and used to evaluate the inter-proton
distances using the isolated spin pair approximation.43 In all cases
the R2 values for the exponential fits were at least 0.99. The mea-
sured distances are given in Table 4, entries 7–11.
Figure 3. (a) Overlaid global (A) and local (B–C) minima identified in the stochastic
searches for compounds 5–9. (b) Representation (%) of each conformation A–D in
the databases (3 kcal molꢂ1) calculated for analogue 5–9.
Table 3
Calculated torsion anglesa
Entry
U1
w1
U2
w2
Global minimumb
1
A
ꢂ72 4c
148 3c
ꢂ72 5c
ꢂ107 3c
Boltzman averaged
2
3
4
5
6
5 (Lex)
ꢂ71
ꢂ70
ꢂ75
ꢂ75
ꢂ63
148
150
145
145
155
ꢂ69
ꢂ70
ꢂ74
ꢂ77
ꢂ72
ꢂ111
ꢂ110
ꢂ108
ꢂ110
ꢂ109
6
7
8
9
a
U1, w1
=
a-
L
-Fuc/Rha-(1?3)-GlcNAc/Glc; U2, w2 = b-
D-Gal/Glc-(1?4)-GlcNAc/
Glc.
b
Torsions calculated for global minimum A averaged on all analogues.
Standard deviation amongst the analogues.
c
d
Boltzman-averaged torsions on the 3 kcal molꢂ1 databases for each analogue.
values for the U1, w1 and U2, w2 angles measured respectively for
the -Fuc/Rha-(1?3)-GlcNAc/Glc and b- -Gal/Glc-(1?4)-Glc-
NAc/Glc glycosidic bonds for the five analogues. In addition, Table
As can be seen these experimental distances are in very good
agreement with the results obtained for all analogues in the
stochastic searches. It is therefore clear that all analogues adopt
a-
L
D

7180
J.-W. Wang et al. / Bioorg. Med. Chem. 18 (2010) 7174–7185
Table 4
Calculated and measured inter-proton distances
Entry
H-10/H-3 (Å)
H-100/H-4 (Å)
2.4
H-100/H-6a (Å)
2.4
H-100/H-6b (Å)
3.8
H-50/H-200 (Å)
2.7
Lowest energy confomersa
1
A
2.5
Boltzman averageb
2
3
4
5
6
5
6
7
8
9
2.6
2.6
2.5
2.6
2.8
2.4
2.4
2.3
2.4
2.3
2.9
3.1
2.6
2.8
3.0
2.9
3.1
3.3
3.5
3.1
2.6
2.6
2.9
3.0
2.8
NMR measurements
7
8
5
6
7
8
9
NDc
2.6d
ND
ND
2.5
2.4
2.5
2.7
2.6
2.7
2.7
2.7
2.7
NDc
2.3
NDc
2.7d
2.4
2.7
9
10
11
2.8d
2.7
2.4
NDc
NDc
2.8d
a
b
c
Distances calculated for global minimum A averaged on all analogues.
Boltzman-averaged distances on the 3 kcal molꢂ1 databases for each analogue.
Signal overlap did not allow distance calculation.
d
Signals were corrected as they overlapped with cross relaxation signals generated from known short inter-proton distances.
a similar conformation in solution which corresponds to the
stacked conformation already known for the Lex antigen.35–38
3. Conclusions
From a synthetic standpoint, the results presented here for the
preparation of trisaccharide 10 support once again that the out-
come of galactosylations at O-4 of N-acetylglucosamine acceptors
depend on the structure of the acceptor. Together with our pub-
lished results17,20,21 this study indicates that glycosylations of
monosaccharide acceptors using trichloroacetimidate donors acti-
vated with 2 equiv of BF3ꢀOEt2 are successful (ꢁ70%) when the
acceptor carries a methyl or 6-chlorohexyl aglycone. In contrast
glycosylation of monosaccharides carrying an allyl and 6-azi-
dohexyl aglycone are less successful giving yields of 30–40%. In
addition, while such activation conditions have proven useful to
achieve glycosylation at O-4 of N-acetylglucosamine in good yields,
it is important to recognize that they may lead to the degradation
of the desired product when the acceptor carries an a-linked sugar
unit at O-3.20 Indeed, the results presented here show that the
galactosylation of the allyl glycoside disaccharide acceptor 16
was faster than the glucosylation of the analogous methyl glyco-
side disaccharide acceptor20 but that, concurrently, the degrada-
tion of the acceptor and/or trisaccharide also occurred at a faster
rate.
We have also prepared a BSA-Lex glycoconjugate and used it as
the immobilized ligand to measure the relative affinity of the anti-
Lex mAb SH18 for the Lex antigen 515 and analogues (6–9)15,19 by
competitive ELISA. Our results showed that all analogues were
weaker inhibitors of the anti-Lex mAb SH1 than the natural anti-
gen. To evaluate whether conformational differences between the
Lex analogues and Lex could induce these drops in affinity to the
SH1 mAb, the conformational behavior of analogues 6–9 was
compared to that of the natural Lex trisaccharide 5 by a combina-
tion of stochastic searches and NMR experiments. Our results indi-
cate that all analogues adopt as a preferred conformation the
stacked conformation that has been identified for the Lex trisaccha-
ride.35–38 Therefore, it is clear that the lower binding of analogues
6–9 to mAb SH1 when compared to the natural antigen Lex does
not result from changes in conformational behavior of the ana-
logues. While replacing the glucosamine unit by a glucose residue
or the fucose residue by a rhamnose unit led to analogues having
lower affinity for the SH1 combining site, some cross-reactivity
with the native antigen Lex was nonetheless still retained.
Figure 4. (a) 1H NMR spectrum (600 MHz, 300 K) of trisaccharide 8 and 1H, 1H
ROESY spectra (mixing time 200 ms) upon selective excitation of proton H-1, H-50,
and H-100. (b) 1H, 1H ROESY cross-relaxation buildup curves for trisaccharide 8.

J.-W. Wang et al. / Bioorg. Med. Chem. 18 (2010) 7174–7185
7181
Therefore, it is likely that vaccine candidates 2 or 3 that display
analogues of Lex at their reducing end in which the glucosamine
and fucose residues are replaced by a glucose and a rhamnose unit,
respectively, may still induce the production of antibodies cross-
reacting with the Lex antigen.
In sharp contrast, analogue 8 in which the galactose residue is
replaced by a glucose unit, showed no binding to the SH1 mAb
even at high concentration. The reason for this observed loss of
binding is unknown at this time. Indeed, it has been well-recog-
nized in the past two decades that binding affinities between pro-
(internal standard, for 1H residual CH3OD d 3.30; for 13C: CD3OD d
49.0) or D2O (external standard 3-(trimethylsilyl)-propionic acid-
d4, sodium salt (TSP) for 1H d 0.00, for 13C d 0.00). 1H NMR and 13
C
NMR chemical shifts are reported in parts per million (ppm).
Coupling constants (J) are reported in hertz (Hz). Chemical shifts
and coupling constants were obtained from a first-order analysis
of one-dimensional spectra. Assignments of proton and carbon res-
onances were based on two dimensional 1H–1H COSY and 13C–1H
HSQC correlation experiments. Multiplicities are abbreviated as fol-
lows: singlet (s), doublet (d), triplet (t), quartet (q) and multiplet
(m). Analytical thin-layer chromatography (TLC) was performed
teins and carbohydrates resulted from
a combination of
factors.30,44,45 These include the formation of direct or water-med-
iated hydrogen bonds between the ligand and amino acid residues
in the combining site, the favorable interactions between non-po-
lar amino acid residues and hydrophobic patches displayed by the
ligand, as well as the favorable displacement of high energy water
molecules from the combining site. Thus binding affinities result
from a combination of enthalpic, entropic and solvation effects
and are often the result of favorable enthalpic contributions that
are offset by unfavorable entropic contributions.30,44,45 It is
interesting to point out that replacing the galactose residue by a
glucose residue in analogue 8 brought a polar hydroxyl group
using silica gel 60 F254 precoated plates (250 lm), were visualized
under UV and/or charring with 10% H2SO4 in EtOH. Flash chroma-
tography was performed using silica gel 60 (230–400 mesh) from
EM science. All reactions were carried out under nitrogen atmo-
sphere with anhyd freshly distilled solvents unless otherwise
noted. Solvents were distilled and dried according to standard
procedures,47 and organic solutions were dried over Na2SO4 and
concentrated under reduced pressure. Molecular sieves were acti-
vated by flame drying under reduced pressure. All reagents were
purchased from commercial suppliers and used without further
purification unless otherwise noted. Reverse-phase HPLC purifica-
within the
a-face of what used to be the exposed hydrophobic
tion were carried out on Prep Nova PakÒ HR C18, 6
lm 60 Å
patch of galactose44 (see conformation A, Fig. 3a). Thus we may
speculate that the presence of this hydroxyl group disturbs favor-
able hydrophobic interactions between the ligand and the combin-
ing site of mAb SH1 and leads to this loss of binding. However, one
may also speculate that either a difference in solvation of analogue
8 versus Lex, or the loss of a strong H bond involving the axial 400-
OH group in Lex contribute to the lack of recognition of analogue 8
by mAb SH1. Thus, only additional studies with 400-deoxy, 400-flu-
oro, or 400-methyloxy Lex analogue inhibitors will provide a better
understanding of these phenomena30 and these analogues are cur-
rently being synthesized in our group. Nevertheless based on the
loss of recognition of analogue 8 by mAb SH1, we can postulate
that a vaccine candidate displaying an analogue of dimeric Lewis
X such as hexasaccharide 4 in which the non-reducing end galacto-
syl residue is replaced by a glucose unit would no longer trigger the
production of antibodies that cross-react with Lex. Most impor-
tantly since the galactose residue and particularly the 400-OH are
not likely involved in the internal epitopes displayed by the TACA
dimLex, we may expect that this vaccine candidate will trigger the
production of antibodies that are cross-reactive with the natural
TACA dimLex. We are currently preparing a glycoconjugate that
displays the hexasaccharide 4 conjugated to an immunostimulant
protein. This glycoconjugate will be used to immunize mice and
the polyclonal serum will be tested for its cross-reactivity with
the dimLex TACA. Since it is known that polymeric Lex antigens
can trigger the production of antibodies against internal epi-
topes,2,4,46 we expect that our vaccine candidate displaying hexa-
(8 ꢄ 100 mm) column using mixtures of acetonitrile and water as
eluent. Purifications by gel permeation chromatography were car-
ried out on a Biogel P2 column (95 cm ꢄ 1.5 cm) eluted with water
unless otherwise stated. UV irradiation was accomplished with an
Ace Glass Incorporated Hanovia photochemical (medium pressure)
quartz mercury-vapor lamp (approx. 40–48% of the total energy
radiated is in the UV portion of the spectrum) unless otherwise
noted. One time crystallized BSA (fatty acid-free, P97% by agarose
gel electrophoresis) was purchased as lyophilized powder from
Sigma; catalogue number: A4378. For Dubois tests, absorbance
values were read at 490 nm employing a Biomate 5 Thermo Elec-
tron Corporation plate reader. Optical rotations were measured
on a Rudolph Research Autopol III automatic polarimeter at 298 K
and are given in deg mL dmꢂ1 gꢂ1. High-resolution electrospray
ionization mass spectra (HRESIMS) were recorded by the analytical
services of the McMaster Regional Center for Mass Spectrometry,
Hamilton, Ontario.
4.1.2. 3-(2-Aminoethylthio)propyl 2-acetamido-2-deoxy-3-O-
(
a
-
L
-fucopyranosyl)-4-O-(b-
glucopyranoside (10)
4.1.2.1. From protected trisaccharide 19. MeOH (900
D-galactopyranosyl)-b-D-
l
L) and
2-aminoethanethiol hydrochloride (110 mg, 0.97 mmol, 49 equiv)
were added to a solution of the allyl glycoside 19 (22 mg,
0.020 mmol) dissolved in a minimum amount of CH2Cl2 (250 lL)
and placed into a quartz tube. The mixture was stirred until full
dissolution and irradiated by UV light (Rayonet photochemical
reactor) for 50 min. The solution was then diluted with CH2Cl2
(50 mL) and washed with 1 M NaOH (5 ꢄ 50 mL). The aqueous lay-
ers were re-extracted with CH2Cl2 (50 mL) and the organic layers
were combined, dried and concentrated to give the crude cyste-
amine Lex adduct. A solution of this adduct in a mixture of THF
(1.5 mL) and EtOH (1.5 mL) was added at ꢂ78 °C to liquid NH3
(ꢁ15 mL), and Na (70 mg, 3.0 mmol, 152 equiv) was added to the
resulting mixture. After 50 min at ꢂ78 °C, the reaction was
quenched with MeOH (2 mL), the ammonia allowed to evaporate
at rt and the remaining solvent evaporated. The resulting solid
was dissolved in water (5 mL) and neutralized with acetic acid
saccharide
4 will trigger the production of antibodies that
recognize the TACA dimLex while they do not bind to the Lex trisac-
charide. The results described in this paper place us one step closer
to the design of a safe anti-cancer vaccine based on the TACA dim-
Lex though a considerable amount of research remains to be car-
ried out.
4. Experimental procedures
4.1. Synthesis of Lex-BSA conjugate
(180 lL, 3.0 mmol, 1 equiv to Na). After freeze-drying the crude
4.1.1. Materials and methods
trisaccharide was passed twice through a Biogel P2 column
eluted with 0.05 M ammonium acetate. The fractions containing
trisaccharide 10 were combined, freeze-dried, concentrated from
MeOH (4 ꢄ 10 mL), and freeze dried again from distilled water
1H NMR (300.13, 400.13 and 600.13 MHz) and 13C NMR (75.5,
100.6 and 150.9 MHz) spectra were recorded in CDCl3 (internal
standard, for 1H residual CHCl3 d 7.24; for 13C: CDCl3 d 77.0), CD3OD

7182
J.-W. Wang et al. / Bioorg. Med. Chem. 18 (2010) 7174–7185
(2 ꢄ 10 mL) to remove the ammonium acetate. The pure acetate
salt of cysteamine Lex 10 (9.4 mg, 73% over two steps) was ob-
tained as a white amorphous solid.
rt, filtered through CeliteÒ and the solids were rinsed with MeOH
(100 mL). The combined filtrate and washings were concentrated
and flash chromatography (15:1, CHCl3–MeOH) gave the pure chlo-
roacetate acceptor 14 (151 mg, 56%) as an amorphous solid.
4.1.2.2. From LexOAll trisaccharide 11. An aq solution of cyste-
[a]
D = ꢂ12.3 (c 0.3, CHCl3). 1H NMR (400 MHz, CDCl3, 295 K) d
amine hydrochloride (7.5 M, 75
l
L, 57 equiv) was added to a solu-
L)
7.37–7.21 (m, 5H, Ar); 5.88–5.76 (m, 2H, –CH@CH2, NH);
5.28–5.12 (m, 2H, –CH@CH2); 4.62–4.51 (m, 3H, H-1, CH2Ph);
4.32–4.26 (m, 1H, OCHH allyl); 4.13–4.01 (m, 3H, OCHH allyl,
O-chloroacetate); 3.95–3.86 (m, 1H, H-2); 3.83–3.71 (m, 3H, H-4,
H-6a, H-6b); 3.58–3.51 (m, 1H, H-5); 3.24 (s, 1H, OH); 1.95 (s, 3H,
N-acetate). 13C NMR (100 MHz, CDCl3, 295 K) d 133.5 (–CH@CH2);
128.5, 128.0, 127.8 (Ar); 117.7 (–CH@CH2); 99.8 (C-1); 76.6 (C-3);
73.8 (OCH2Ph); 73.6 (C-5); 70.6 (C-4); 70.2 (C-6); 69.9 (OCH2 allyl);
54.1 (C-2); 40.9 (CH3 chloroacetate); 23.4 (CH3 acetate). HRESIMS
calcd for C20H26ClNO7 [M+H]+: 428.1476, found: 428.1481.
tion of allyl glycoside 11 (6.1 mg, 0.011 mmol)11 in water (125
l
and placed in a quartz tube. The mixture was irradiated under UV
light for 90 min, diluted with aq ammonium acetate (0.25 M,
250 lL) and submitted to gel permeation chromatography on a
Biogel P2 column eluted with 0.25 M ammonium acetate. The
fractions containing trisaccharide 10 were combined and after
freeze-drying the crude adduct was dissolved in distilled water
and de-ionized with Dowex HOꢂ resin. After filtration and freeze-
drying, compound 10 (4 mg, 58%) was obtained pure as a white
amorphous powder.
4.1.5. Allyl 2-acetamido-3-O-(2,3,4-tri-O-benzyl-a-L-
4.1.2.3. Analytical data for 10. [
a
]
D = +31.2 (c 0.5, MeOH). 1H
fucopyranosyl)-6-O-benzyl-2-deoxy-b-D-glucopyranoside (16)
NMR (400 MHz, CD3OD, 295 K) d 5.03 (d, 1H, J = 3.9 Hz, C-10);
4.86–4.79 (m, 1H, H-50); 4.47–4.40 (m, 2H, H-100, H-1); 4.01–3.82
(m, 7H, H-2, H-4, H-6, H-30, H-400, OCHH); 3.82–3.67 (m, 4H, H-40,
H-300, H-600); 3.66–3.55 (m, 2H, H-20, OCHH); 3.53–3.39 (m, 4H, H-
3, H-5, H-200, H-500); 3.06 (m, 2H, SCH2CH2N); 2.76 (m, 2H,
SCH2CH2N); 2.63 (m, 2H, OCH2CH2CH2S); 1.97 (s, 3H, N-acetate);
1.83 (m, 2H, OCH2CH2CH2S); 1.18 (d, 3H, J = 6.5 Hz, H-60). 13C
NMR (100 MHz, CD3OD, 295 K): 173.7 (C@O); 103.8 (C-100); 102.5
(C-1); 100.2 (C-10); 77.4 (C-5); 76.7 (C-500); 76.5 (C-40); 75.0 (C-
4); 74.9 (C-3); 73.7 (C-40); 72.7 (C-200); 71.2 (C-30); 69.9 (C-200);
69.9 (C-300); 68.6 (OCH2); 67.7 (C-50); 62.8 (C-600); 61.3 (C-6); 57.4
(C-2) 39.7 (SCH2CH2N); 30.4 (OCH2CH2CH2S); 29.7 (SCH2CH2N);
28.8 (OCH2CH2CH2N); 23.2 (CH3 acetate); 16.7 (C-60). HRESIMS
calcd for C25H46N2O15S [M+H]+: 647.2673, found: 647.2697.
A solution of the known22 benzylidene disaccharide 15 (128 mg,
0.17 mmol) in anhyd CH2Cl2 (2.5 mL) containing powdered acti-
vated 4 Å MS (625 mg) was stirred under N2 for 1 h then cooled
to ꢂ78 °C. Et3SiH (80
lL, 0.50 mmol, 3.0 equiv) and TfOH (38 lL,
0.57 mmol, 3.4 equiv) were added to the reaction mixture and
the solution was stirred for 1 h. The reaction was quenched with
Et3N (333 lL, 2.4 mmol, 14 equiv) and MeOH (333 lL) and the
reaction mixture diluted with CH2Cl2 (50 mL) was filtered through
CeliteÒ. The filtrate was washed with satd aq NaHCO3 (50 mL),
dried, concentrated, and flash chromatography (1:1, EtOAc–hex-
anes) of the residue gave the pure N-acetate acceptor 16 (99 mg,
77%) as a colorless glass. [
a
]
D = ꢂ31.6 (c 1.0, CHCl3). 1H NMR
(400 MHz, CDCl3, 298 K) d 7.35–7.14 (m, 20H, Ar); 5.86–5.74 (m,
1H, –CH@CH2); 5.41 (d, 1H, J = 6.7 Hz, NH); 5.22–5.06 (m, 2H, –
CH@CH2); 4.92–4.85 (m, 2H, H-10, CHHPh); 4.84–4.66 (m, 4H, H-
1, CHHPh, CH2Ph); 4.62–4.49 (m, 4H, 2 ꢄ CHHPh, CH2Ph); 4.27
(dd, 1H, J = 13.0, 5.2 Hz, OCHH allyl); 4.09–3.97 (m, 4H, H-20, H-
50, OH, OCHH allyl); 3.90–3.72 (m, 3H, H-3, H-30, H-6a); 3.66–
3.58 (m, 2H, H-40, H-6b); 3.46–3.24 (m, 3H, H-2, H-4, H-5); 1.52
(s, 3H, N-acetate); 1.08 (d, 3H, J = 6.4 Hz, H-60). 13C NMR
(100 MHz, CDCl3, 298 K): 169.2 (C@O); 138.2 (Ar quaternary);
134.2 (–CH@CH2); 128.4, 128.2, 128.1, 127.9, 127.7, 127.6 (Ar);
117.3 (–CH@CH2); 99.2 (C-1, C-10); 83.9 (C-3); 79.2 (C-30); 77.2
(C-40); 76.0 (C-20); 75.0 (OCH2Ph); 74.9 (C-5); 74.1, 73.5, 72.8
(OCH2Ph); 70.4 (C-4); 69.8 (OCH2 allyl); 69.6 (C-6); 68.1 (C-50);
56.2 (C-2); 22.1 (CH3 acetate); 16.6 (C-60). HRESIMS calcd for
4.1.3. Allyl 2-acetamido-4,6-O-benzylidene-3-O-chloroacetyl-2-
deoxy-b-D-glucopyranoside (13)
A solution of the known22 benzylidene alcohol 12 (400 mg,
1.15 mmol) in anhyd CH2Cl2 (20 mL) and pyridine (1.6 mL) was
cooled to 0 °C under N2 and chloroacetyl chloride (194 lL, 2.1 equiv)
was added drop wise over 10 min. The reaction mixture was allowed
to slowly warm up to rt and stirred overnight. Solvents were evapo-
rated and residual pyridine was removed by co-concentration with
toluene (3 ꢄ 20 mL). The resulting amorphous solid was washed
with ethanol (2 ꢄ 20 mL) and the pure chloroacetate 13 (411 mg,
84%) was obtained as an amorphous white powder after drying un-
der high vacuum. [
a
]
D = ꢂ4.1 (c 1.0, DMF). 1H NMR (400 MHz, CDCl3,
C
45H53NO10 [M+H]+: 768.3748, found: 768.3777.
295 K) d7.39–7.26 (m, 5H, Ar); 5.84–5.73 (m, 1H, –CH@CH2); 5.46 (s,
1H, CHPh); 5.42 (d, 1H, J = 9.1 Hz, NH); 5.31 (t, 1H, J = 9.9 Hz, H-3);
5.26–5.13 (m, 2H, –CH@CH2); 4.61 (d, 1H, J = 8.4 Hz, H-1); 4.33–
4.26 (m, 2H, H-6a, OCHH allyl); 4.06–3.95 (m, 4H, H-2, OCHH allyl,
O-chloroacetate); 3.76 (t, 1H, J = 10.3 Hz, H-6b); 3.67 (t, 1H,
J = 9.4 Hz, H-4); 3.48 (m, 1H, H-5); 1.95 (s, 3H, N-acetate). 13C NMR
(100 MHz, CDCl3, 295 K) d 133.3 (–CH@CH2); 131.9, 129.1, 128.2,
126.1 (Ar); 117.9 (–CH@CH2); 101.8 (CHPh); 100.4 (C-1); 78.4 (C-
4); 73.4 (C-3); 70.1 (OCH2 allyl); 68.6 (C-6); 66.3 (C-5); 54.9 (C-2);
40.7 (O-chloroacetate); 23.4 (N-acetate). HRESI calcd for
4.1.6. Allyl 2-acetamido-4-O-(2,3,4,6-tetra-O-acetyl-b-
galactopyranosyl)-6-O-benzyl-3-O-chloroacetyl-2-deoxy-b-D-
glucopyranoside (18)
The chloroacetate acceptor 14 (27 mg, 0.063 mmol) and the
known23 trichloroacetimidate 17 (164 mg, 0.33 mmol, 5.3 equiv)
were dissolved in anhyd CH2Cl2 (2 mL) and the solution was
D
-
warmed up to 40 °C under N2. BF3ꢀOEt2 (17
2.1 equiv) was added and the mixture was stirred for 1 h 40 min
at 40 °C. The reaction was quenched with Et3N (21 L, 0.15 mmol,
lL, 0.14 mmol,
l
C
20H24ClNO7 [M+H]+: 426.1320, found: 426.1325.
2.4 equiv), the mixture diluted in CH2Cl2 (30 mL) and washed with
satd aq NaHCO3 (30 mL). The aq phase was re-extracted with
CH2Cl2 (20 mL) and the combined organic layers were dried and
concentrated. After two successive flash chromatography columns
(6:4 EtOAc–hexanes followed by 50:1 CHCl3–MeOH) the pure
disaccharide 18 (20.9 mg, 43%) was isolated as an amorphous glass.
4.1.4. Allyl 2-acetamido-6-O-benzyl-3-O-chloroacetyl-2-deoxy-
b-D
-glucopyranoside (14)
A solution of benzylidene chloroacetate 13 (270 mg, 0.63 mmol)
and NaCNBH3 (420 mg, 6.9 mmol, 10.1 equiv) in anhyd THF (12 mL)
containing activated 3 Å MS (770 mg) and methyl orange (5 mg)
was cooled to 0 °C and stirred under N2 for 30 min. A 1 M solution
of HCl in Et2O (6.3 mL, 0.63 mmol, 1 equiv) was slowly added over
10 min to the reaction mixture, the solution turned dark pink and
H2 (g) was generated. The reaction mixture was stirred for 1 h at
[
a
]
D = ꢂ5.5 (c 1.0, CHCl3). 1H NMR (400 MHz, CDCl3, 295 K) d 7.40–
7.29 (m, 5H, Ar); 5.89–5.77 (m, 1H, –CH@CH2); 5.57 (d, 1H,
J = 9.2 Hz, NH); 5.27–5.08 (m, 4H, H-3, H-40, –CH@CH2); 4.97 (dd,
1H, J = 10.4, 8.0 Hz, H-20); 4.81 (dd, 1H, J = 10.3, 3.5 Hz, H-30);
4.73 (d, 1H, J = 12.1 Hz, CHHPh); 4.51 (d, 1H, J = 7.8 Hz, H-1); 4.45

J.-W. Wang et al. / Bioorg. Med. Chem. 18 (2010) 7174–7185
7183
(d, 1H, J = 12.1 Hz, CHHPh); 4.39 (d, 1H, J = 8.0 Hz, H-10); 4.31 (m,
1H, OCHH allyl); 4.09–3.99 (m, 6H, H-2, H-6a0, H6b0, OCHH allyl,
O-chloroacetate); 3.94 (t, 1H, J = 8.1 Hz, H-4); 3.70 (m, 2H, H-6a,
H-6b); 3.63 (m, 1H, H-50); 3.45 (m, 1H, H-5); 2.13–1.92 (5 s,
5 ꢄ 3H, O-acetates and N-acetate). 13C NMR (75 MHz, CDCl3,
300 K): 171.0, 170.9, 170.7, 169.7, 168.0 (C@O); 138.4 (Ar quater-
nary); 133.6 (–CH@CH2); 128.6, 128.1, 128.0 (Ar); 117.6
(–CH@CH2); 100.2 (C-10); 99.8 (C-1); 74.5 (C-5); 74.4 (C-3); 74.2
(C-4); 73.7 (OCH2Ph); 70.8 (C-30); 70.7 (C-50); 69.6 (OCH2 allyl);
69.2 (C-20); 67.4 (C-6); 66.8 (C-40);61.1 (C-60); 53.4 (C-2); 40.8
(CH2Cl); 23.3, 20.7, 20.5 (CH3 acetates). HRESIMS calcd for
anes to pure EtOAc) yielded the known24 fucose hemiacetal 21
(2 mg, 15%), the pure disaccharide 22 (5 mg, 25%) and the impure
orthoester 20. The isomeric mixture of endo and exo orthoesters
(20) was obtained free of other contaminants (13 mg, 60%) upon
RP HPLC (65:35–90:10 CH3CN–H2O). 1H NMR showed that one iso-
mer assumed to be the less sterically hindered exo orthoester was
predominant (75%).
4.1.8.1. Analytical data for orthoester 20. 1H NMR (400 MHz,
CDCl3, 298 K, for the major isomer) d 7.42–7.25 (m, 5H, Ar);
6.29 (d, 1H, J = 3.6 Hz, H-10); 5.49 (d, 1H, J = 5.0 Hz, H-40); 5.36
(dd, 1H, J = 10.5, 3.1 Hz, H-30); 5.06 (d, 1H, J = 3.6 Hz, H-1);
4.97 (d, 1H, J = 11.6 Hz, CHHPh); 4.82–4.62 (m, 5H, CHHPh,
2 ꢄ CH2Ph); 4.29 (t, 1H, J = 6.2 Hz, H-5); 4.15 (dd, 1H, J = 10.6,
3.7 Hz, H-20); 4.09 (d, 2H, J = 5.5 Hz, H-6); 4.05 (dd, 1H,
J = 10.2, 4.8 Hz, H-2); 3.85 (m, 1H, H-50); 3.78 (dd, 1H, J = 10.1,
2.5 Hz, H-3); 3.65 (s, 1H, H-4); 2.19–2.02 (3 s, 3 ꢄ 3H, O-ace-
tates); 1.74 (s, 3H, orthoester CH3); 1.13 (d, 3H, J = 8.5 Hz,
H-60). 13C NMR (100 MHz, CDCl3, 298 K, for the major isomer):
170.4, 170.1, 169.4 (C@O); 138.7, 138.6, 138.5 (Ar quaternary);
128.5, 128.4, 128.3, 128.2, 128.0, 127.7, 127.5, 127.5, 127.3
(Ar); 99.7 (C-1); 91.4 (C-10); 78.7 (C-3); 78.6 (C-4); 75.8(C-2);
74.8,72.8, 72.7 (OCH2Ph); 71.5 (C-20); 69.4 (C-30); 68.4 (C-30);
67.7 (C-5); 67.4 (C-50); 61.3 (C-6); 21.0, 20.7, 20.5 (CH3 acetates
and orthoacetate); 16.5 (C-60). HRESIMS calcd for C47H53NO11
[M+NH4]+: 782.3388, found: 782.3370.
C
34H44ClNO16 [M+H]+: 758.2427, found: 758.2455.
4.1.7. Allyl 2-acetamido-4-O-(2,3,4,6-tetra-O-acetyl-b-
pyranosyl)-3-O-(2,3,4-tri-O-benzyl- -fucopyranosyl)-6-O-
benzyl-2-deoxy-b- -glucopyranoside (19)
D-galacto-
a-L
D
The disaccharide acceptor 16 (22 mg, 0.029 mmol) and the tri-
chloroacetimidate 17 (70 mg, 0.14 mmol, 4.9 equiv) were dissolved
in anhyd CH2Cl2 (2 mL) under N2. BF3ꢀOEt2 (7.2
2.0 equiv) was added and the mixture was stirred for 40 min at
rt. The reaction was quenched with Et3N (10 L, 0.072 mmol,
lL, 0.057 mmol,
l
2.5 equiv) and the mixture diluted with CH2Cl2 (30 mL) was
washed with satd aq NaHCO3 (30 mL). The aq phase was re-ex-
tracted with CH2Cl2 (20 mL) and the combined organic layers were
dried and concentrated. Flash chromatography (1:1–6:4 EtOAc–
hexanes) followed by RP HPLC (65:35–90:10 CH3CN–H2O) gave
the pure trisaccharide 19 (19.9 mg, 64%) as a white amorphous so-
lid. [
a
]
D = ꢂ27.2 (c 1.0, CHCl3). 1H NMR (600 MHz, CDCl3, 310 K) d
4.1.8.2. Analytical data for 22. [
a]
D = ꢂ6.6 (c 0.5, CHCl3). 1H NMR
7.38–7.21 (m, 20H, Ar); 5.91 (d, 1H, J = 7.9 Hz, NH); 5.84–5.76 (m,
(400 MHz, CDCl3, 294 K) d 7.39–7.21 (m, 5H, Ar); 5.92–5.81 (m, 1H,
–CH@CH2); 5.57 (d, 1H, J = 7.7 Hz, NH); 5.33 (s, 1H, J = 3.1 Hz, H-40);
5.29–5.11 (m, 3H, H-20, –CH@CH2); 4.90 (dd, 1H, J = 10.5, 3.4 Hz,
H-30); 4.78 (d, 1H, J = 8.2 Hz, H-1); 4.70 (d, 1H, J = 12.1 Hz, CHHPh);
4.50–4.43 (2 d, 2H, J = 12.0, 7.9 Hz, H-1, CHHPh); 4.31 (dd, 1H,
J = 12.9, 5.0 Hz, OCHH allyl); 4.13–3.98 (m, 4H, H-3, H-6a0, H-6b0,
OCHH allyl); 3.99 (s, 1H, OH); 3.86 (t, 1H, J = 6.2 Hz, H-50); 3.69–
3.59 (m, 3H, H-4, H-6a, H-6b); 3.52–3.37 (m, 2H, H-2, H-5);
2.15–1.94 (5 s, 5 ꢄ 3H, O-acetates and N-acetate). 13C NMR
(100 MHz, CDCl3, 295 K): 170.5, 170.1, 170.0, 169.2 (C@O); 138.0
(Ar quaternary); 133.8 (–CH@CH2); 128.5, 128.4, 128.3, 127.9,
127.8, 127.4 (Ar); 117.7 (–CH@CH2); 101.2 (C-1); 99.1 (C-10);
80.9 (C-4); 73.9 (C-5); 73.6 (OCH2Ph); 71.3 (C-3); 71.1 (C-50);
70.7 (C-30); 69.9 (OCH2 allyl); 68.7 (C-20); 68.0 (C-6); 66.8 (C-40);
61.3 (C-60); 57.2 (C-2); 23.7, 20.7, 20.6, 20.5 (CH3 acetates). HRE-
SIMS calcd for C32H43NO15 [M+H]+: 682.2711, found: 682.2709.
1H, –CH@CH2); 5.29–5.26 (d, 1H, J = 3.2 Hz, H-400); 5.25–5.20 (dd,
1H, J = 1.4, 17.2 Hz, –CH@CHH); 5.11–5.07 (m, 2H, H-10,
–
CH@CHH); 5.04 (dd, 1H, J = 8.2, 10.4 Hz, H-200); 4.93 (d, 1H,
J = 11.8 Hz, CHHPh); 4.87–4.70 (m, 6H, H-1, H-300, 2 ꢄ CH2Ph);
4.66 (d, 1H, J = 11.8 Hz, CHHPh); 4.58 (d, 1H, J = 12.0 Hz, CHHPh);
4.53 (s, 1H, J = 8.1 Hz, H-100); 4.41 (d, 1H, J = 12.0 Hz, CHHPh);
4.25 (m, 1H, H-50); 4.19 (dd, 1H, J = 5.1, 13.0 Hz, OCHH allyl);
4.14–4.07 (m, 3H, H-3, H-20, H-6a00); 4.02–3.96 (m, 2H, H-6b00,
OCHH allyl); 3.91 (t, 1H, J = 6.5 Hz, H-4); 3,87 (dd, 1H, J = 2.5,
10.1 Hz, H-30); 3.86–3.73 (m, 2H, H-6a, H-6b); 3.64–3.58 (m, 3H,
H-2, H-40, H-500); 3.56 (m, 1H, H-5); 2.0–1.77 (5 s, 5 ꢄ 3H, O-ace-
tates and N-acetate); 1.13 (d, 3H, J = 6.5 Hz, H-60). 13C NMR
(150 MHz, CDCl3, 310 K): 170.0, 169.9 (C@O); 138.9, 138.9, 138.7,
138.0 (Ar quaternary); 134.1 (–CH@CH2); 128.5, 128.4, 128.3,
128.3, 128.2, 127.9, 127.8, 127.7, 127.6, 127.5, 127.3, 127.2 (Ar);
117.0 (–CH@CH2); 99.6 (C-100); 98.7 (C-1); 97.3 (C-10); 79.8 (C-30);
77.2 (C-500); 76.5 (C-20); 74.5 (OCH2Ph); 74.2 (C-4); 74.1 (C-5);
73.4, 73.4 (OCH2Ph); 73.3 (C-3); 72.7 (OCH2Ph); 70.6 (C-50, C-300);
69.5 (OCH2 allyl); 68.9 (C-200); 68.9 (C-6); 66.8 (C-400); 66.6 (C-50);
60.5 (C-600); 23.2, 20.7, 20.5, 20.5 (CH3 acetates); 16.7 (C-60). HRE-
SIMS calcd for C59H71NO19 [M+H]+: 1098.4699, found: 1098.4678.
4.1.9. Preparation of the glycoconjugate Lex-BSA (23)
A solution (10
lobutene-1,2-dione (diethyl squarate) in freshly distilled MeOH
was added to a solution of cysteamine Lex 10 (4 mg, 6.1
mol) in
freshly distilled MeOH (600 L). The reaction mixture was left at
lL/mL, 90 lL, 1 equiv) of 3,4-diethyoxy-3-cyc-
l
l
rt for 3 h and TLC (6:3:1 EtOAC–MeOH–H2O) showed that the
starting amine 10 was quantitatively converted to the desired
squarate compound. After evaporation of the methanol, a solution
4.1.8. Allyl 2-acetamido-4-O-(2,3,4,6-tetra-O-acetyl-b-
galactopyranosyl)-6-O-benzyl-2-deoxy-b- -glucopyranoside
-fucose (21) and 3,4,6-tri-O-acetyl-
-fucos-10-yl orthoacetyl)-
galactopyranose (20)
The disaccharide 16 (22 mg, 0.029 mmol) and the trichloroace-
timidate 17 donor (70 mg, 0.14 mmol, 4.9 equiv) were dissolved in
D-
D
(22); 2,3,4-tri-O-benzyl-
1,2-O-(2,3,4-tri-O-benzyl-
a
,b-
L
L
of BSA (8.1 mg, 0.120
100 L) was added to the squarate adduct obtained above
(4.3 mg, 5.9 mol, 50 equiv). The vial that contained the BSA solu-
tion was washed with more buffer (100 L) which was added to
lmol) in a pH 10 carbonate buffer (0.1 M,
a
-
a
-
D
-
l
l
l
the reaction mixture and the reaction was left to proceed for 7 days
at rt. The glycoconjugate was filtered against distilled H2O
(6 ꢄ 8 mL) using an Amicon ultrafiltration cell equipped with a
Diaflo membrane (25 mm, 30 kDa cut-off). The residue was then
taken up in water and lyophilized to give the pure Lex-BSA conju-
gate 23 (9.0 mg). The level of incorporation was measured by the
Dubois test26 that gave an average hapten loading (n) of 35 hapten
molecules per BSA carrier protein.
anhyd CH2Cl2 (2 mL) under N2. BF3ꢀOEt2 (7.2
2.0 equiv) was added and the mixture was stirred for 2 h 20 min
at rt. The reaction was quenched with Et3N (10 L, 0.072 mmol,
lL, 0.057 mmol,
l
2.5 equiv), the reaction mixture diluted with CH2Cl2 (30 mL) and
washed with satd aq NaHCO3 (30 mL). The aq phase was re-ex-
tracted with CH2Cl2 (20 mL) and the combined organic layers were
dried and concentrated. Flash chromatography (6:4 EtOAc–hex-

7184
J.-W. Wang et al. / Bioorg. Med. Chem. 18 (2010) 7174–7185
4.2. Competitive binding experiments
atoms only. In both searches the calculations terminated at the end
of the 150,000 iterations as the number of failures criteria to find
new conformations (10,000 iterations in a row) was never met.
A NUNC 96-well ELISA microtiter plate was coated with a dilu-
tion of the Lex-BSA glycoconjugate 23 (100
l
L per well, 3
lg/mL) in
Structures that did not have the 4C1 conformation for
D
-glucosa-
-galactose or those that did not maintain a 4C1 conforma-
-fucose were eliminated. The results of the two searches
a 0.05 M carbonate-bicarbonate buffer at pH 9.6. The plate was
covered with Saran wrap and incubated at 4 °C for 16 h. The anti-
gen solution was discarded and the plate was washed (3 ꢄ 5 min
incubation period with buffer) with a PBS buffer at pH 7.4 contain-
ing 0.05% Tween 20. The plate was then rinsed with PBS and
mine or
tion for
D
L
were combined for analysis and conformations with energies over
10 kcal molꢂ1 above the global minima were rejected.
blocked with skim milk (DIFCO, 5% in PBS-0.05% Tween, 400
per well). After a 1 h incubation at 37 °C, the plate was rinsed
(3ꢄ) with PBS. Each well received a solution containing 50 L of
competitor (concentration range from 0.2 g/mL to 20 mg/mL in
2% skim milk in PBS) and 50 L of a solution of SH1 antibodies
(1:125 dilution in 2% skim milk in PBS). The plate was incubated
for 4 h at 23 °C. The plate was subsequently washed (3 ꢄ 5 min
incubation period with buffer) with PBS-0.05% Tween and rinsed
with PBS. A dilution of commercially available alkaline phospha-
tase-labeled goat anti-mouse antibody (1:1500 in 2% skim milk
l
L
4.3.2. NMR experiments
NMR experiments were recorded in 5 mm NMR tubes at
400 MHz (HSQC) and 600 MHz (1H, 13C, COSY, HMBC and NOESY).
All spectra were recorded at 300 K in D2O. One-dimensional ROESY
spectra were acquired with selective excitations of protons: H-10,
H-100 and H-50 using a 5 s relaxation delay between scans. The opti-
mum irradiation range for each signal was used to determine the
best length of the soft pulse applied. This value was automatically
calculated by Bruker software (BUTSEL-NMR). 1D ROESY spectra
were recorded at nine or ten mixing times from 20 ms to 300 or
400 ms. Before Fourier transformation, the FIDs were zero-filled
once and multiplied with a 1 Hz line broadening factor. Spectra
were phase and baseline corrected and integrated. The integrals
measured for the irradiated signals were plotted against mixing
time and the obtained curves were fitted to a double exponential
decaying function:
l
l
l
in PBS, 100 lL per well) was added to each well. After 1 h of incu-
bation at 37 °C, the plate was washed (3 ꢄ 5 min incubation period
with buffer) with PBS-0.05% Tween, and rinsed with PBS. A solu-
tion of the chromogenic substrate p-nitrophenyl phosphate
(100 lL per well) at 1 mg/mL in 0.05 M sodium carbonate buffer
pH 9.8 containing 10 mM MgCl2 was added. After a 14 h incubation
period at 23 °C, the absorbance values were read at 405 nm
employing a PowerWave XS plate reader. All samples were pre-
pared in triplicate. The absorbance values were plotted as the per-
centage inhibition against an increasing concentration of
competitor, calculated using wells containing no competitor as
the reference point. The values were fitted to a 4 parameter logis-
tics sigmoidal equation:29 y = y0 + a/[1 + (x/x0)b] using Sigma Plot
10.0.
fðsmÞ ¼ ꢂA½expðBsmÞ þ expðCsmÞꢅ
where sm is the mixing time and A, B and C are adjustable parame-
ters. The values of these integrals were extrapolated to 0 ms mixing
time, and the integrals from cross-relaxation peaks were normal-
ized through division by these extrapolated values. The normalized
cross-relaxation integrals were plotted against the mixing times
and the buildup curves were fitted to a double exponential equation
of the form
4.3. Conformational analysis
fðsmÞ ¼ A½expðBsmÞ ꢂ expðCsmÞꢅ
and the initial slopes at 0 ms mixing times were determined from
the calculated first derivatives,
4.3.1. Stochastic searches
The trisaccharide was subjected to all atom and heavy atom sto-
chastic conformational searches using the AMBER94 force field39 in
the Molecular Operating Environment (MOE2005)32 program suite.
In this implementation of AMBER94, any parameters missing for a
class of compound are approximated by MOE based on parameter
values for similar molecular fragments. The starting structure was
built using the carbohydrate builder in MOE2005 and submitted to
the stochastic search. This search is similar to the random incre-
mental pulse search (RIPS) method48 in which the coordinates of
each atom are randomly perturbed, after which the entire mole-
cule is minimized to attempt to generate a new conformer. In these
stochastic searches, new conformations are generated by the rota-
tion of all bonds (including ring bonds) to random dihedral angles.
In our searches a bias of 30° was used. This means that dihedral an-
gles were selected with a sum-of-Gaussians distribution with
peaks at multiples of 30°. In addition to bond rotations, atom posi-
tions were randomly perturbed by 0.4 Å. These random perturba-
tions were repeated 150,000 times per conformational search
and each generated conformer was subjected to 500 steps of min-
imization with full degrees of freedom with implicit solvation by
the Generalized Born/Surface Area (GB/SA) continuum solvation
model34 using a dielectric constant of 78. Two separate searches
were performed and the databases were combined for analysis. Ini-
tially each new conformation was checked against the previously
found conformations using a root-mean-square (RMS) tolerance
of 0.1 Å on all atoms, including the hydrogen atoms to account
for possible hydrogen bonding. The second search was carried
out using a more stringent RMS value (0.01 Å) applied to the heavy
F0ð0Þ ¼ AðB ꢂ CÞ:Ref:43
Interproton distances were calculated based on the isolated spin
pair approximation (ISPA):
1=6
rij ¼ rref ðSref =SijÞ
where S is the initial slope at sm = 0, and r is the proton–proton
distance.
The intra-residue cross-peaks used as reference for distance
determinations were: H-50/H-30 (2.50 Å) and H-100/H-300 (2.64 Å) for
compounds 5 and 6; H-100/H-300 (2.64 Å) for compound 7; H-50/H-30
(2.50 Å) and H-100/H-500 (2.39 Å) for compound 8; H-10/H-20
(2.51 Å), H-50/H-30 (2.59 Å) and H-100/H-300 (2.64 Å) for compound 9.
Acknowledgments
The authors thank the National Sciences and Engineering Re-
search Council of Canada for financial support. We thank Professor
S.-I. Hakomori from the Pacific Northwest Research Institute
(Seattle, USA) for the generous gift of mAb SH1. A.A. acknowledges
the World bank for a fellowship under the DUE program.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2010.08.040.

J.-W. Wang et al. / Bioorg. Med. Chem. 18 (2010) 7174–7185
7185
20. Hendel, J. L.; Wang, J-W.; Jackson, T. A.; Hardmeier, K.; De Los Santos, R.;
Auzanneau, F.-I. J. Org. Chem. 2009, 74, 8321.
References and notes
21. Wang, A.; Auzanneau, F.-I. Carbohydr. Res. 2010, 345, 1216.
22. Aguilera, B.; Jiménez-Barbero, J.; Fernández-Mayoralas, A. Carbohydr. Res. 1998,
308, 19.
23. Amvam-Zollo, P.-H.; Sinaÿ, P. Carbohydr. Res. 1986, 150, 199; Toepfer, A.;
Schmidt, R. R. J. Carbohydr. Chem. 1993, 12, 809; Schmidt, R. R.; Stumpp, M.
Liebigs Ann. Chem. 1983, 1249.
24. Dejter-Juszynski, M.; Flowers, H. M. Carbohydr. Res. 1971, 18, 219.
25. Perlin, A. S. Can. J. Chem. 1963, 41, 399; Auzanneau, F.-I.; Bundle, D. R. Can. J.
Chem. 1993, 71, 534.
26. Dubois, M.; Gilles, K. H.; Hamilton, J. K.; Rebers, P. A.; Smith, F. Anal. Chem.
1956, 28, 350.
1. Hakomori, S.-I.; Jeanloz, R. W. J. Biol. Chem. 1964, 239, 3606; Hakomori, S.-I.;
Koscielak, J.; Bloch, K. J.; Jeanloz, R. W. J. Immunol. 1967, 98, 31.
2. Fukushi, Y.; Hakomori, S.-I.; Nudelman, E.; Cochran, N. J. Biol. Chem. 1984, 259,
4681.
3. Fukushi, Y.; Hakomori, S.-I.; Shepard, T. J. Exp. Med. 1984, 159, 506; Fukushi, Y.;
Kannagi, R.; Hakomori, S.-I.; Shepard, T.; Kulander, B. G.; Singer, J. W. Cancer
Res. 1985, 45, 3711; Itzkowitz, S. H.; Yuan, M.; Fukushi, Y.; Palekar, A.; Phelps, P.
C.; Shamsuddin, A. M.; Trump, B. T.; Hakomori, S.-I.; Kim, Y. S. Cancer Res. 1986,
46, 2627; Nakasaki, H.; Mitomi, T.; Noto, T.; Ogoshi, K.; Hanaue, H.; Tanaka, Y.;
Makuuchi, H.; Clausen, H.; Hakomori, S.-I. Cancer Res. 1989, 49, 3662.
4. Singhal, A. K.; Ørntoft, T. F.; Nudelman, E.; Nance, S.; Schibig, L.; Stroud, M. R.;
Clausen, H.; Hakomori, S.-I. Cancer Res. 1990, 50, 1375.
27. Hirayama, K.; Akashi, S.; Furuya, K.-I. Biochem. Biophys. Res. Commun. 1990, 173,
639.
5. Kobata, A.; Ginsburg, V. J. Biol. Chem. 1969, 244, 5496; Yang, H.-J.; Hakomori, S.-
I. J. Biol. Chem. 1971, 246, 1192.
28. Bundle, D. R. Pure Appl. Chem. 1989, 61, 1171; Miekle, P. J.; Young, N. M.;
Bundle, D. R. J. Immunol. Methods 1990, 132, 255; Reimer, K. B.; Gidney, M. A. J.;
Bundle, D. R.; Pinto, B. M. Carbohydr. Res. 1992, 232, 131.
29. Harrington, C. R.; Edwards, P. C.; Wischik, C. M. J. Immunol. Methods 1990, 134,
261; Plikaytis, B. D.; Turner, S. H.; Gheesling, L. L.; Carlone, G. M. J. Clin.
Microbiol. 1991, 29, 1439.
30. Bundle, D. R. Recognition of Carbohydrate Antigens by Antibody Binding Sites.
In Bioorganic Chemistry: Carbohydrates; Hecht, S. M., Ed.; Oxford University
Press: New York, 1999; pp 370–440.
31. Auzanneau, F.-I.; Sourial, E.; Schmidt, J. M.; Feher, M. Can. J. Chem. 2001, 80,
1088.
32. Chemical Computing Group Inc., Montreal, QC, Canada.
33. Halgren, T. A. J. Comp. Chem. 1996, 17, 490.
6. Zhang, S.; Zhang, H. S.; Cordon-Cardo, C.; Reuter, V. E.; Singhal, A. K.; Lloyd, K.
O.; Livingston, P. O. Int. J. Cancer 1997, 73, 50; Satoh, J.; Kim, S. U. J. Neurosci. Res.
1994, 37, 466; Croce, M. V.; Isla-Larrain, M.; Rabassa, M. E.; Demichelis, S.;
Golussi, A. G.; Crespo, M.; Lacunza, E.; Segal-Eiras, A. Pathol. Oncol. Res. 2007, 13,
130.
7. Kamath, V. P.; Diedrich, P.; Hindsgaul, O. Glycoconjugate J. 1996, 13, 15.
8. Auzanneau, F.-I.; Pinto, B. M. Bioorg. Med. Chem. 1996, 4, 2003.
ˇ
9. Zhang, J.; Yergey, A.; Kowalak, J.; Kovác, P. Carbohydr. Res. 1998, 313, 15.
ˇ
10. Hou, S.-J.; Saskena, R.; Kovác, P. Carbohydr. Res. 2008, 343, 196.
11. Asnani, A. Ph.D. Thesis, University of Guelph, August 2006.
12. Jacquinet, J.-C.; Sinaÿ, P. J. Chem. Soc. 1979, 314.
13. Hindsgaul, O.; Norberg, T.; Pendu, J. L.; Lemieux, R. U. Carbohydr. Res. 1982, 109,
109.
34. Still, W. C.; Tempczyk, A.; Hawley, R. C.; Hendrickson, T. J. Am. Chem. Soc. 1990,
112, 6127; Qiu, D.; Shenkin, P. S.; Hollinger, F. P.; Still, W. C. J. Phys. Chem. A.
1997, 101, 3005; Schaefer, M.; Karplus, M. A. J. Phys. Chem. 1996, 100, 1578.
35. Thøgersen, H.; Lemieux, R. U.; Bock, K.; Meyer, B. Can. J. Chem. 1982, 60, 44;
14. Sato, S.; Ito, Y.; Nukada, T.; Nakahara, Y.; Ogawa, T. Carbohydr. Res. 1987, 167,
197; Jain, R. K.; Matta, K. L. Carbohydr. Res. 1992, 226, 91; Toepfer, A.; Schmidt,
R. R. Tetrahedron Lett. 1992, 33, 5161; Windmüller, R.; Schmidt, R. R.
Tetrahedron Lett. 1994, 35, 7927; Numomura, S.; Iida, M.; Numata, M.;
Sugimoto, M.; Ogawa, T. Carbohydr. Res. 1994, 263, C1; Jain, R. K.; Vig, R.;
Rampal, R.; Chandrasekaran, E. V.; Matta, K. L. J. Am. Chem. Soc. 1994, 116,
12123; Yan, L.; Kahne, D. J. Am. Chem. Soc. 1996, 9239; Hummel, G.; Schmidt, R.
R. Tetrahedron Lett. 1997, 38, 1173; Lay, L.; Manzoni, L.; Schmidt, R. R.
Carbohydr. Res. 1998, 310, 157; Figueroa-Pérez, S.; Verez-Bencomo, V.
Tetrahedron Lett. 1998, 39, 9143; Kretzschmar, G.; Stahl, W. Tetrahedron
1998, 54, 6341; Manzoni, L.; Lay, L.; Schmidt, R. R. J. Carbohydr. Chem. 1998, 17,
769; Sato, K.-I.; Seki, H.; Yoshimoto, A.; Nanaumi, H.; Takai, Y.; Ishido, Y. J.
Carbohydr. Chem. 1998, 17, 703; Cao, S.; Gan, Z.; Roy, R. Carbohydr. Res. 1999,
318, 75; Zhang, Y.-M.; Esnault, J.; Mallet, J.-M.; Sinaÿ, P. J. Carbohydr. Chem.
1999, 18, 419; Zhu, T.; Boons, G.-J. J. Am. Chem. Soc. 2000, 122, 10222; Zhu, T.;
Boons, G.-J. Chem. Eur. J. 2001, 7, 2382; La Ferla, B.; Prosperi, D.; Lay, L.;
Giovanni, R.; Panza, L. Carbohydr. Res. 2002, 337, 1333; de la Fuente, J. M.;
Penadés, S. Tetrahedron: Asymmetry 2002, 13, 1879; Xia, J.; Alderfer, J. L.; Locke,
R. D.; Piskorz, C. F.; Matta, K. L. J. Org. Chem. 2003, 68, 2752; Mukherjee, D.;
Sarkar, S. K.; Chattopadhyay, P.; Chowdhury, U. S. J. Carbohydr. Chem. 2005, 24,
251; de Paz, J.-L.; Ojeda, R.; Barrientos, A. G.; Penadés, S.; Martín-Lomas, M.
Tetrahedron: Asymmetry 2005, 16, 149; Nagai, Y.; Ito, N.; Sultana, I.; Sugai, T.
Tetrahedron 2008, 64, 9599; Shimizu, H.; Yoshimura, Y.; Hinou, H.; Nishimura,
S.-I. Tetrahedron 2008, 64, 10091.
ˇ
Imberty, A.; Mikros, E.; Koca, J.; Mollicone, R.; Oriol, R.; Pérez, S. Glycoconjugate
J. 1995, 12, 331; Reynolds, M.; Fuchs, A.; Lindhorst, T. K.; Perez, S. Mol. Simul.
2008, 34, 447.
36. Miller, K. E.; Mukhopadhyay, C.; Cagas, P.; Bush, C. A. Biochemistry 1992, 31,
6703.
37. Haselhorst, T.; Weimar, T.; Peters, T. J. Am. Chem. Soc. 2001, 123, 10705.
38. Pérez, S.; Mouhous-Riou, N.; Nifant’ev, N. E.; Tsvetkov, Y. E.; Bachet, B.;
Imberty, A. Glycobiology 1996, 6, 537.
39. Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K. M.; Ferguson, D. M.;
Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.; Kollman, P. A. J. Am. Chem. Soc. 1995,
117, 5179.
40. Pérez, S.; Imberty, A.; Engelsen, S. B.; Gruza, J.; Mazeau, K.; Jimenez-Barbero, J.;
Poveda, A.; Espinosa, J.-F.; van Eyck, B. P.; Johnson, G.; French, A. D.; Kouwijzer,
M. E. C. E.; Grootenuis, P. D. J.; Bernardi, A.; Raimondi, L.; Senderowitz, H.;
Durier, V.; Vergoten, G.; Rasmussen, K. Carbohydr. Res. 1998, 314, 141.
41. Imberty, A.; Pérez, S. Chem. Rev. 2000, 100, 4567.
42. IUPAC-I.U.B. Commission of Biochemical Nomenclature Eur. J. Biochem. 1997,
243, 9.
43. Maaheimo, H.; Kosma, P.; Brade, L.; Brade, H.; Peters, T. Biochemistry 2000, 39,
12778.
44. Lemieux, R. U. Acc. Chem. Res. 1996, 29, 373.
45. Woods, R. J. Glycoconjugate J. 1998, 15, 209; Clarke, C.; Woods, R. J.; Gluska, J.;
Cooper, A.; Nutelt, M. A.; Boons, G.-J. J. Am. Chem. Soc. 2001, 123, 12238;
Laederach, A.; Reilly, P. J. Proteins: Struct. Funct. Bioinf. 2005, 60, 591.
46. Van Room, A.-M. M.; Van de Vijver, K. K.; Jacobs, W.; Van Marck, E. A.; Van
Dam, G. J.; Hokke, C. H.; Deedler, A. M. Microbes Infect. 2004, 6, 1125.
47. Gordon, A. J.; Ford, R. A. A Chemist’s Companion; John Wiley & Sons: New York,
1972. pp 429–436.
15. Asnani, A.; Auzanneau, F.-I. Carbohydr. Res. 2008, 343, 1653.
16. Gan, Z.; Cao, S.; Wu, Q.; Roy, R. J. Carbohydr. Chem. 1999, 18, 755.
17. Wang, A.; Hendel, J. L.; Auzanneau, F.-I. Beilstein J. Org. Chem. 2010, 6.
doi:10.3762/bjoc.6.17.
18. Ball, G. E.; O’Neill, R. A.; Schultz, J. E.; Lowe, J. B.; Weston, B. W.; Nagy, J. O.;
Brown, E. G.; Hobbs, C. J.; Bednarski, M. D. J. Am. Chem. Soc. 1992, 114, 5449;
Nagahori, N.; Nishimura, S.-I. Chem. Eur. J. 2006, 12, 6478.
48. Chang, G.; Guida, W. C.; Still, W. C. J. Am. Chem. Soc. 1989, 111, 4379.
19. Hendel, J. L.; Cheng, A.; Auzanneau, F.-I. Carbohydr. Res. 2008, 343, 2914.