A comparison of biological and calorimetric analyses of multivalent glycodendrimer ligands for concanavalin A

Article abstract of DOI:10.1016/S0957-4166(99)00589-3

The cluster glycoside effect - the observation that multivalent glycosides bind to their polyvalent protein receptors with apparent affinities greater than those that can be rationalized solely on the basis of valency - is by now a well established phenomenon. As part of a continuing effort to provide a molecular basis for the cluster glycoside effect, we report here the synthesis of two series of mannosylated dendritic ligands and their performance in a range of competitive and non-competitive binding assays, including hemeagglutination inhibition (HIA), enzyme-linked lectin assays (ELLA) and isothermal titration microcalorimetry (ITC). The first series of ligands contained a semi-rigid glycylglycine spacer and showed no significant performance enhancement in any binding studies. The second series of ligands contained a flexible tetraethylene glycol spacer; these ligands showed marked enhancements at tetravalent and hexavalent levels in both HIA (IC₅₀=3 and<0.8 μM, respectively) and ITC (K(A)=6.2x10⁴ and 1.5x10⁶ M^-1, respectively) studies. In all cases, the thermodynamic parameters of association are consistent with non-specific aggregation rather than enhanced lectin-ligand affinity. This conclusion is reinforced by the lack of enhancements in ligand activity observed in ELLA studies. Copyright (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0957-4166(99)00589-3

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 11 (2000) 95–111
A comparison of biological and calorimetric analyses of
multivalent glycodendrimer ligands for concanavalin A
Jayme B. Corbell,† Joseph J. Lundquist † and Eric J. Toone ∗
Departments of Biochemistry and Chemistry, Duke University, Durham, NC, 27708, USA
Received 24 November 1999; accepted 15 December 1999
Abstract
The cluster glycoside effect — the observation that multivalent glycosides bind to their polyvalent protein
receptors with apparent affinities greater than those that can be rationalized solely on the basis of valency —
is by now a well established phenomenon. As part of a continuing effort to provide a molecular basis for the
cluster glycoside effect, we report here the synthesis of two series of mannosylated dendritic ligands and their
performance in a range of competitive and non-competitive binding assays, including hemeagglutination inhibition
(HIA), enzyme-linked lectin assays (ELLA) and isothermal titration microcalorimetry (ITC). The first series
of ligands contained a semi-rigid glycylglycine spacer and showed no significant performance enhancement in
any binding studies. The second series of ligands contained a flexible tetraethylene glycol spacer; these ligands
showed marked enhancements at tetravalent and hexavalent levels in both HIA (IC50=3 and<0.8 µM, respectively)
and ITC (KA=6.2×104 and 1.5×106 M−1, respectively) studies. In all cases, the thermodynamic parameters of
association are consistent with non-specific aggregation rather than enhanced lectin–ligand affinity. This conclusion
is reinforced by the lack of enhancements in ligand activity observed in ELLA studies. © 2000 Elsevier Science
Ltd. All rights reserved.
1. Introduction
The role of protein–carbohydrate interactions in biological communication is by now widely appre-
ciated. Myriad processes in normal mammalian biology are mediated by such interactions including
neutrophil recruitment, fertilization, and immunological surveillance.1–7 Alternatively, carbohydrate-
based recognition facilitates the initiation of a variety of infectious diseases that commence with pathogen
recognition of and adhesion to a host through a protein–carbohydrate interaction.8 This common motif
has raised interest in the development of non-cytotoxic carbohydrate-based pharmaceuticals that would
prevent infection by inhibiting pathogen–host recognition.
∗
†
Corresponding author. Tel: (919) 681-3484; e-mail: toone@chem.duke.edu
These authors have contributed equally to this work.
0957-4166/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(99)00589-3
tetasy 3223

96
J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
The intellectual appeal of such a strategy notwithstanding, the development of carbohydrate-based
therapeutics has been thwarted by the exceptionally weak affinity of native saccharide ligands for
their protein receptors; individual interactions between lectins and their ligands typically proceed with
dissociation constants in the millimolar to micromolar range.9 Nature has overcome this weak affinity,
and cell to cell contacts mediated by protein–carbohydrate binding are frequently irreversible on a
biologically relevant timescale. Lectins are typically found as oligomeric structures; frequently many
copies of each oligomers are displayed on the surface of the cell. Together, these observations suggest
that nature has compensated for weak individual protein–carbohydrate affinities through multivalency.
Following this observation, several groups have sought to develop high affinity lectin ligands through
multivalency.3,8,10–23 Many such compounds exhibit superior performance in a range of competitive
binding assays relative to the corresponding monovalent ligand, with potencies greater than those
expected solely on the basis of their valency. This phenomenon has been termed the ‘cluster glycoside
effect’.24,25 Despite many observations of the phenomenon, the physical origin of the effect is still
unclear. Although most often interpreted in terms of an entropically assisted, or chelate-type, binding,
extrapolation of values of translational and rotational entropies for bimolecular complexation and
conformational entropy losses during the restriction of rigid rotors make this explanation unattractive.26
Here we continue work designed to provide an understanding of the molecular processes that underlie
the cluster glycoside effect through the construction and evaluation of the binding behavior of two series
of dendritic ligands.
2. Results and discussion
We have previously reported the synthesis of a series of mannosylated dendrimers and their perfor-
mance as ligands for the plant lectin concanavalin A.27 Our initial series of ligands consisted of mannosyl
units tethered to an aromatic core via propyl linkers. The binding of these ligands to concanavalin
A was evaluated by isothermal titration calorimetry (ITC), hemeagglutination inhibition assay (HIA),
and enzyme-linked lectin assay (ELLA). The study suggested that the apparent affinity of a ligand is
highly dependent on the type of assay used, and that observed enhancements are not due to alterations
in the thermodynamics of protein–ligand interaction. A body of evidence suggests that the behavior of
multivalent ligands in binding assays is highly dependent on the nature of the linker domain.28–31 Here
we extend our earlier studies through the preparation of ligands incorporating distinctly different linker
regions: a polypeptide spacer and a spacer based on tetraethylene glycol. The linkers were chosen to
extend the size of our previous ligands and evaluate the effect of linker flexibility on affinities.
2.1. Synthesis of ligands
Peptidic dendrimers were constructed on an azidoethyl glycylglycine linker. 2,3,4,6-Tetra-O-acetyl-
mannosyl trichloroacetimidate was coupled to 2-azidoethanol under TMSOTf promotion to form the
α-azidoethyl mannopyranoside 1. Reduction to the free amine preceded DCC/HOBT coupling to N-Cbz
glycylglycine which, following hydrogenolytic removal of the benzyl carbamate, provided the semi-
rigid mannosyl-spacer domain 2a. Alternatively, coupling of the trichloroacetimidate to 11-azido-3,6,9-
trioxaundecanol followed by hydrogenolysis yielded the flexible TEG-linked mannoside 2b (Scheme
1).
With glycosylated linkers in hand, dendrimer assembly proceeded according to our previously publish-
ed protocols. Coupling of the amino-terminated mannopyranosides to benzoyl chloride, 5-azidomethyl-

J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
97
Scheme 1. Synthesis of precursors for glycodendrimer assembly. Reagents: (i) Ac2O, pyridine; (ii) hydrazine acetate, 50°C; (iii)
CCl3CN, K2CO3; (iv) 2-azidoethanol, TMSOTf; (v) H2 (1 atm), Pd/C; (vi) Cbz-GlyGly, DCC, HOBT; (vii) H2 (1 atm), Pd/C;
(viii) TMSOTf, 11-azido-3,6,9-trioxaundecanol; (ix) H2 (1 atm), Pd/C
1,5-benzenedicarbonyl dichloride, and 1,3,5-benzenetricarbonyl trichloride yielded mono-, bi-, and
trivalent ligands 3a–5a and 3b–5b in the glycylglycine and TEG-linked series, respectively. Reduction of
bivalent dendrons 4a and 4b and coupling of the resulting amines to the N-hydroxysuccinimide deriva-
tives of 5-azidomethyl-1,5-benzenedicarboxylate 6 and 1,3,5-benzenetricarboxylate 7 yielded tetravalent
8a, 8b and hexavalent members 9a, 9b of both ligand series. Zemplen deprotection provided the fully
deprotected species 10a–14a and 10b–14b required for binding studies (Scheme 2).
2.2. Hemeagglutination inhibition assays (HIA)
The inhibition of hemeagglutination has been used extensively for the analysis of protein–carbohydrate
interactions, primarily due to the simplicity of the technique and the relatively low material
requirements.32 Previous assays have typically utilized chicken erythrocytes; because of their ready
availability, porcine erythrocytes were used here. To exclude effects based on variability of erythrocyte
surface composition, we first compared IC50 values for a series of previously determined carbohydrates
with chicken and porcine erythrocytes.33 Table 1 demonstrates that although agglutination of porcine
erythrocytes is inhibited at slightly lower levels of ligand than are chicken erythrocytes, the two cells
provide comparable IC50s, within experimental error.
Minimum inhibitory concentrations for both series of dendrimeric ligands were analyzed using porcine
erythrocytes. In accordance with previous observations, no significant benefit of multivalency was
observed at low ligand valency (Table 2). Rather, low-valency ligands bind in a similar fashion to
α-methyl mannose, a trend that continues into the second generation for the peptide-linked ligands.
Alternatively, the TEG-linked series differs markedly from the peptide-linked series. With this group of
ligands significant enhancements in inhibition are observed with second generation ligands: the tetra-
and hexavalent TEG dendrimers show valence-corrected inhibition enhancements of 50- and greater than
112-fold over the monovalent reference ligand.
2.3. Enzyme linked lectin assay
The enzyme linked lectin assay (ELLA) is a variation of the commonly used ELISA.34,35 ELLA
evaluates the ability of a soluble ligand to inhibit the adhesion of a multivalent lectin to a reference
ligand non-covalently immobilized on the surface of a microtitre plate. Readout is accomplished through

98
J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
Scheme 2. Synthesis of multivalent glycodendrimers. (i) BzCl, Et3N; (ii) 5-azidomethyl benzene-1,3-dicarbonyl dichloride,
DMAP, pyridine; (iii) benzene-1,3,5-tricarbonyl trichloride, DMAP, pyridine; (iv) H2 (1 atm), Pd/C; (v) 5-azidomethylben-
zene-1,3-dicarboxylate di-N-hydroxysuccinimide ester, Et3N; (vi) 1,3,5-benzenetricarboxylate tri-N-hydroxysuccinimide ester,
Et3N; (vii) 0.1 M NaOMe, MeOH
Table 1
Comparison of hemeagglutination assay results for erythrocytes from chickens and pigs. Reported
IC50 values are from Gupta et al.33
the use of lectin directly labelled with peroxidase, eliminating washing steps that potentially lead to
the dissociation of weakly interacting systems. The most significant advantage of ELLA over the
hemeagglutination assay lies in the spectrometric analysis of binding: the assay thus provides a more
precise determination of IC50 values than does HIA.
Table 2 shows the results of ELLA experiments with the two series of ligands. Again, peptide-linked
ligands provide no significant enhancement in IC50 relative to the monovalent ligand. For this series,
increases in inhibitory power are solely a function of increased saccharide concentration. In contrast to

J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
99
Table 2
Biological assay results for peptide- and TEG-linked glycodendrimers. Parenthetical values are
corrected for carbohydrate concentration
the results of HIA, the TEG-linked series of ligands demonstrate no enhancement in performance with
increased valency; rather all five ligands behave essentially identically.
2.4. Isothermal titration microcalorimetry
A third technique that has found extensive utility in ligand binding studies during the last decade
is isothermal titration microcalorimetry.36–43 In this experiment, soluble ligand is titrated into a solution
containing the binding protein, and the heat evolved or absorbed during binding is measured as a function
of ligand concentration. Deconvolution of these data yields a binding constant, an enthalpy of binding,
and the stoichiometry of the interaction (Fig. 1). The binding constant is simply related to the free energy
of association, and the entropy of the association is thus available by subtraction. Because the assay is
soluble and homogeneous, ITC reports protein–carbohydrate interactions solely in terms of reversible
binding processes. Effects contributing to IC50s arising from, for example, surface display and phase
change are disregarded.
The results of ITC evaluation of the binding of both ligand series are presented in Table 3. The
reduction of ITC data requires an assumed model of ligand binding. The nature of this model is extremely
important; values of binding constants and enthalpies vary roughly linearly as ligand concentrations, in
turn dependent on the model of binding. Here we have used the simplest possible model of binding, that
of individual saccharides binding to individual, discrete binding sites in a non-cooperative fashion. From
this model the ITC data are fitted by Eq. (1), where Vo is the cell volume, 1/r=[P]totK, [L]r=[P]tot/[L]tot,
and requires the use of the total saccharide concentration for [L]tot and the total binding site concentration
for [P]tot.

100
J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
Fig. 1. Representative calorimetric data for the titration of peptide-linked trivalent 5a (6 mM) with concanavalin A (0.57 mM)
at 25°C. (A) Raw data for forty 2.2 µL injections and (B) the integrated curve showing experimental points (ꢀ) and the least
squares fit (−) to the integrated data. The buffer was 50 mM dimethyl glutaric acid, 250 mM NaCl, 1 mM CaCl2, and 1 mM
MnCl2, at pH 5.2. The fit gives the following values: n=1.02, K=7900 M−1, ∆H=−7.8 kcal mol−1
(1)
Table 3
Isothermal titration microcalorimetry results for peptide- and TEG-linked glycodendrimers. All values
listed above are corrected for the valency of the ligand in question
In accordance with the previous results, multivalent peptide-linked ligands behave similarly to the
monovalent parent compound. Our model of binding is validated by the excellent fit of the data by Eq.

J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
101
(1). Only slight changes in thermodynamic parameters occur through the second generation, notably
a diminution in the enthalpy of binding accompanied by a compensating increase in the entropic
term, a trend we have observed previously.27 The effect is more pronounced for higher valent ligands.
Unfortunately, the insolubility of the hexavalent ligand precluded calorimetric evaluation of binding.
The TEG-linked ligands follow similar trends in the first generation of ligands, with thermodynamic
parameters of binding essentially equivalent to those of the monovalent ligand. As was previously
observed for the propyl-linked dendritic ligands, the trivalent ligand shows a weaker interaction than
would be expected solely on the basis of saccharide concentration.27 Crystallographic studies in that
instance demonstrated the apparent loss in affinity was in fact due to the binding of only two of the
three saccharide ligands available, because of constraints applied by the lattice structure of the cross-
linked macroscopic structure. Thus, the apparent binding constant of the trivalent ligand is reduced
because of an overestimate of the available ligand, rather than from any alteration in the nature of
the protein–carbohydrate couple. Significant enhancements in affinity are observed during binding of
the second-generation TEG-linked ligands, although the enhancement is less than that observed in
hemeagglutination assays. Additionally, the decrease in enthalpy of association is greatly exaggerated; at
the hexavalent ligand fully 85% of the apparent free energy of association arises from a favorable entropy
of binding.
The thermodynamic pattern of association observed here is most consistent with an initial pro-
tein–carbohydrate interaction, which presumably proceeds with thermodynamic parameters equivalent to
those of a simple monovalent saccharide, followed by a reversible aggregation process. While exceptions
certainly exist, specific associations in aqueous solution are typified by an enthalpy of association greater
than the free energy and an unfavorable entropy of binding.9,39,44,45 In contrast, non-specific aggregation
processes typically proceed as entropically driven process, characterized by a small negative or positive
enthalpy. For example, the micellization of pure octyl-β-D-glucopyranoside proceeds spontaneously with
an enthalpy of +1.5 kcal mol−1 46,47
.
Several other studies of aqueous micelle formation reinforce these
trends. Thus, micelle formation of octyl-linked tetra- or pentaethylene glycol (∆Hmic +3.6 kcal mol−1
in both cases), sodium decyl sulfate (∆Hmic +5.5 kcal mol−1), and TX-100 (∆Hmic +3.1 kcal mol−1
)
are all endothermic processes.48–50 We suggest that the steadily diminishing enthalpy of binding is the
thermodynamic signature of an endothermic aggregation process superimposed on an exothermic ligand
binding event.
This model of association is consistent with the lack of enhancement in ligand performance by ELLA.
The origin of the discrepancy between ligand performance in ELLA and HIA can be readily explained
by the nature of the lectin involved. The lectin utilized in ELLA is directly labelled with horseradish
peroxidase, a 40 kDa protein. Our crystallographic study of the aggregates formed during concanavalin
A binding of dendritic ligands demonstrates the need for tight protein–protein packing.27 The attachment
of a large protein label presumably prevents the formation of large ordered aggregates during ligand
binding. Although very long linkers might overcome this effect, these observations suggest that ELLA
might more faithfully report protein–carbohydrate affinities in the traditional sense of the term than other
assays.
The thermodynamic contribution of aggregation to apparent ligand affinities is readily deducible by
assuming initial protein–carbohydrate interactions proceed with parameters equivalent to those of the
parent, monovalent compound. This assumption is justified on the basis of our previously reported
crystallographic studies of ligand cross-linking, work that demonstrates that the orientation of individual
saccharide recognition domains of a multivalent ligand within the concanavalin A binding site is identical
to that of methyl α-D-mannopyranoside. Overall thermodynamic parameters can then be separated as:
∆J=∆Ji+∆Ja

102
J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
Where ∆Ji refers to the intrinsic contribution, or contribution of protein–carbohydrate interaction,
and ∆Ja refers to the contribution arising from aggregation effects. Table 4 shows that the enthalpy of
aggregation is significantly endothermic in the second-generation dendrimers for both series. In the case
of the TEG-linked series, the increase in the entropic component is significantly larger than the enthalpic
penalty, leading to an overall increase in the binding constant.
Table 4
Thermodynamic effects of aggregation. All values are corrected for ligand valency
3. Conclusions
The performance of multivalent saccharide ligands is strongly a function of the assay used to evaluate
them. It is inappropriate to draw conclusions regarding ligand binding based on either agglutination
or microtitre plate inhibition assays. Observed enhancements in apparent affinity appear to be the
result of entropically driven protein aggregation phenomena that are highly dependent on the precise
molecular details of the ligand, including valency and linker composition, and protein concentration. In
some circumstances the contribution of aggregation events to overall thermodynamic parameters rivals
that of protein–carbohydrate interaction. We are continuing our studies of the role of multivalency in
protein–carbohydrate interaction and will report our results in due course.
4. Experimental
4.1. General
Concanavalin A was purchased from Sigma. D-Mannose and glycylglycine were purchased from
Aldrich and used without further purification. Dichloromethane was distilled from calcium hydride prior
to use. BioGel P2 and SX-1 resins were obtained from BioRad. Trichloroacetonitrile was fractionally
distilled prior to use. Porcine erythrocytes were collected fresh and stored in Alsever’s solution. All
other chemicals were reagent grade and used without further purification. Column chromatography was
carried out using flash grade silica gel. Thin layer chromatography slides were silica gel 60 F254 (Merck)
1
and detected by using Hanessian’s dip and UV detection. All H NMR spectra were obtained on GE
QE300, Varian Mercury, or Varian Unity spectrometers operating at 300.015, 300.07 and 399.956 MHz,
respectively. All 13C NMR spectra were taken on a GE QE300 spectrometer, operating at 75.48 MHz or
a Varian Mercury NMR spectrometer operating at 75.452 MHz. All spectra were obtained in CDCl3 or
D2O and referenced to TMS, an internal solvent peak, or sodium acetate. FAB-MS was carried out on a

J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
103
JEOL SX-102 mass spectrometer. Protein concentrations were determined by the method of Edelhoch,51
and carbohydrate concentrations were determined by phenol–sulfuric acid chars according to DuBois.52
4.2. Syntheses
4.2.1. 1-O-(2-Azidoethyl)-2,3,4,6-tetra-O-acetyl-α-D-mannopyranoside 1
2,3,4,6-Tetra-O-acetyl-α-D-mannosyl trichloroacetimidate (1.4 g, 2.4 mmol) was dissolved in di-
chloromethane (10 mL) with 2-azidoethanol (1.1 mL, 12 mmol) and 4 Å molecular sieves (1 g) under
a nitrogen atmosphere. The mixture was cooled to −30°C, TMSOTf (150 µL, 0.96 mmol) was added,
and the reaction was stirred for 5 h at 25°C. The mixture was filtered over Celite, concentrated in vacuo,
and purified by silica gel chromatography (petroleum ether:ethyl acetate 2:1) to yield 1 (570 mg, 53%
yield) as a colorless oil. 1H NMR (300 MHz, CDCl3): δ 2.00 (s, 3H), 2.06 (s, 3H), 2.12 (s, 3H), 2.17 (s,
3H), 3.45–3.51 (m, 2H), 3.65–3.71 (m, 1H), 3.84–3.91 (m, 1H), 4.04–4.15 (m, 2H), 4.3 (dd, 1H, J=5.28,
12.30 Hz), 4.88 (d, J=1.63 Hz), 5.27–5.39 (m, 3H) ppm. 13C NMR (75 MHz, CDCl3): δ 20.59, 20.64,
20.67, 20.80, 50.24, 62.33, 65.84, 66.96, 68.73, 69.26, 97.60, 169.73, 169.93, 170.54 ppm. IR (neat): ν
2954, 2955, 2893, 2106, 1744, 1368, 1228, 1136, 1048, 979 cm−1
.
4.2.2. 1-O-[2-N-(N-Benzyloxycarbonyl glycylglycyl)ethyl]-2,3,4,6-tetra-O-acetyl-α-D-mannopyrano-
side 2a
Azidosugar 1 (150 mg, 0.31 mmol) was dissolved in ethanol (10 mL). Degussa Pd/C (30 mg) was
added, and H2 was bubbled through the solution for 3 h. The mixture was filtered over Celite and
concentrated in vacuo. The residue was dissolved in anhydrous DMF (10 mL) and Cbz-glycylglycine
(90 mg, 0.3 mmol), DCC (80 mg, 0.3 mmol), and HOBT (45 mg, 0.33 mmol) were added. The solution
was stirred under a nitrogen atmosphere at 25°C for 12 h, filtered over Celite, concentrated in vacuo, and
purified by silica gel chromatography (petroleum ether:ethyl acetate:ethanol 4:3:1) to yield 2a (120 mg,
63% yield) as a colorless oil. 1H NMR (300 MHz, CDCl3): δ 1.97 (s, 3H), 2.04 (s, 3H), 2.09 (s, 3H), 2.12
(s, 3H), 3.42–3.47 (m, 2H), 3.55–3.63 (m, 1H), 3.89–3.97 (m, 4H), 3.98–4.02 (m, 1H), 4.10–4.15 (m,
1H), 4.22–4.28 (dd, 1H, J=5.50, 12.23 Hz), 4.81–4.82 (d, 1H, J=1.13 Hz), 5.11 (s, 2H), 5.22–5.32 (m,
3H), 6.0 (broad t, 1H), 7.1 (broad t, 1H), 7.2 (broad t, 1H), 7.07–7.34 (m, 5H) ppm. 13C NMR (75 MHz,
CDCl3): δ 20.65, 20.69, 20.82, 20.80, 38.85, 43.08, 62.42, 66.00, 66.78, 67.10, 68.53, 69.08, 69.31,
97.50, 127.98, 128.17, 128.46, 136.04, 156.23, 169.08, 169.68, 170.04, 170.21, 170.34, 170.79 ppm. HR
FAB-MS (pos) calcd for C28H37N3O14: 639.2276, obsd MH+: 640.2342.
4.2.3. 1-O-(11-Azido-3,6,9-trioxaundecyl)-2,3,4,6-tetra-O-acetyl-α-D-mannopyranoside 2b
2,3,4,6-Tetra-O-acetyl-α-D-mannosyl trichloroacetimidate (500 mg, 1.0 mmol) was dissolved in
dichloromethane (5 mL) with 11-azido-3,6,9-trioxaundecanol (1.0 mL, 5.01 mmol) and 4 Å molecular
sieves (0.2 g) under a nitrogen atmosphere. TMSOTf (80 µL, 0.4 mmol) was added, and the reaction
was stirred at 25°C for 6 h. Solid sodium bicarbonate (200 mg) was added. The mixture was filtered over
Celite, concentrated in vacuo, and purified by silica gel chromatography (petroleum ether:ethyl acetate
2:1) to yield 2b (570 mg, 53% yield) as a colorless oil. 1H NMR (300 MHz, CDCl3): δ 2.00 (s, 3H), 2.05
(s, 3H), 2.11 (s, 3H), 2.16 (s, 3H), 3.40 (t, 2H, J=5.0 Hz), 3.62–3.84 (m, 14H), 4.07–4.12 (m, 2H), 4.30
(dd, 1H, J=5.0, 12.3 Hz), 4.88 (d, 1H, J=1.6 Hz), 5.26–5.35 (m, 3H) ppm. 13C NMR (75 MHz, CDCl3):
δ 20.67, 20.74, 20.88, 50.61, 62.33, 66.06, 67.33, 68.31, 69.03, 69.50, 69.92, 70.60, 70.64, 70.68, 97.65,
169.69, 169.86, 170.01, 170.67 ppm. IR (neat) ν 2922, 2105, 1746, 1369, 1224, 1135, 1086, 1047, 978
cm−1. HR FAB-MS (pos) calcd for C22H35N3O13: 549.2171, obsd (M−N2+3H)+: 524.2361.

104
J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
4.2.4. 1-O-[2-(N-Benzylamidoglycylglycyl)ethyl]-2,3,4,6-tetra-O-acetyl-α-D-mannopyranoside 3a
Protected glycopeptide 2a (135 mg, 0.21 mmol) was dissolved in absolute ethanol (25 mL). Degussa
Pd/C (50 mg) was added, and H2 was bubbled through the mixture for 2.5 h. The solution was filtered
over Celite and concentrated in vacuo. The residue was dissolved in dichloromethane (20 mL) and
benzoyl chloride (70 µL, 0.6 mmol) and triethylamine (20 µL) were added. The reaction was stirred
under a nitrogen atmosphere at 25°C for 24 h then washed with water (1×20 mL), 1 M HCl (1×20 mL),
saturated sodium bicarbonate (1×20 mL), and brine (1×20 mL). The organic layer was dried (MgSO4),
filtered, and concentrated in vacuo. Silica gel chromatography (ethyl acetate:ethanol 2:1) yielded 3a (86
1
mg, 70% yield) as a colorless oil. H NMR (300 MHz, CDCl3): δ 1.96 (s, 3H), 2.02 (s, 3H), 2.09 (s,
3H), 2.11 (s, 3H), 3.4 (m, 2H), 3.6 (m, 1H), 3.95–3.97 (m, 2H), 4.08–4.14 (m, 3H), 4.11–4.16 (m, 1H),
4.22–4.28 (dd, 1H, J=5.3, 12.0 Hz), 4.82–4.83 (d, 1H, J=1.24 Hz), 5.25–5.29 (m, 3H), 7.2 (broad t, 1H),
7.28–7.50 (m, 4H), 7.7 (br t, 1H), 7.76–7.85 (m, 2H) ppm. 13C NMR (75 MHz, CDCl3): δ 20.61, 20.66,
20.79, 25.05, 25.52, 38.92, 53.68, 62.34, 65.89, 66.68, 68.44, 69.2, 69.30, 97.54, 130.37, 133.69, 136.73,
166.60, 169.60, 169.73, 170.17, 170.27, 170.60, 170.73 ppm. HR FAB-MS (pos) calcd for C27H35N3O13:
609.2171, obsd MH+: 610.2227.
4.2.5. 1-O-(11-Benzylamido-3,6,9-trioxaundecyl)-2,3,4,6-tetra-O-acetyl-α-D-mannopyranoside 3b
Glycoconjugate 2b (65 mg, 0.12 mmol) was dissolved in absolute ethanol (10 mL). Degussa Pd/C
(20 mg) was added, and H2 was bubbled through the mixture for 1.5 h. The solution was filtered over
Celite and concentrated in vacuo. The residue was dissolved in dichloromethane (10 mL) and benzoyl
chloride (28 µL, 0.25 mmol), triethylamine (165 µL, 10 equiv.), and DMAP (20 mg) were added. The
reaction was stirred under a nitrogen atmosphere at 25°C for 18 h then washed with 1 M HCl (1×10 mL),
saturated sodium bicarbonate (1×10 mL), and brine (1×10 mL). The organic layer was dried (MgSO4),
filtered, and concentrated in vacuo. Silica gel chromatography (ethyl acetate:petroleum ether:methanol
1
4:3:1) yielded 3b (53 mg, 71%) as a colorless oil. H NMR (300 MHz, CDCl3): δ 1.98 (s, 3H), 2.03
(s, 3H), 2.10 (s, 3H), 2.15 (s, 3H), 3.63–3.79 (m, 16 H), 4.03–4.11 (m, 2H), 4.30 (dd, 1 H, J=5.0, 12.3
Hz), 4.86 (d, 1H, J=1.6 Hz) 5.26–5.35 (m, 3H), 6.86 (br s, 1 H), 7.40–7.53 (m, 3H), 7.81 (d, 2H, J=7.1
Hz) ppm. 13C NMR (75 MHz, CDCl3): δ 20.67, 20.74, 39.71, 62.34, 67.33, 69.02, 69.56, 69.77, 69.85,
69.92, 70.21, 70.49, 70.68, 70.81, 97.66, 127.02, 128.44, 131.34, 134.56, 167.43, 169.92, 170.71 ppm.
IR (neat) ν 3408, 2090, 1647, 1371, 1227, 1048 cm−1. FAB-MS (pos) calcd for C29H41NO14: 627.25,
obsd MNa+: 650.25.
4.2.6. 5-Azidomethyl-N,N0-bis[2-O-(2,3,4,6-tetra-O-acetyl-α-D-mannopyranosyl)-N-ethylglycylglycyl]-
benzene-1,3-dicarboxamide 4a
Protected glycopeptide 2a (810 mg, 1.21 mmol) was dissolved in absolute ethanol (25 mL). Degussa
Pd/C (50 mg) was added, and H2 was bubbled through the mixture for 5 h. The solution was filtered
over Celite and concentrated in vacuo. The residue was dissolved in dichloromethane (35 mL) and 5-
azidomethylbenzene-1,3-dicarbonyl dichloride (110 mg, 0.4 mmol), HOBT (190 mg, 1.4 mmol), and
DCC (285 mg, 1.4 mmol) were added. The mixture was stirred under a nitrogen atmosphere at 25°C for
48 h then filtered over Celite and washed with water (1×40 mL), 1 M HCl (1×40 mL), saturated sodium
bicarbonate (1×40 mL), and brine (1×40 mL). The organic layer was dried (MgSO4), filtered, and
concentrated in vacuo. Silica gel chromatography (ethyl acetate:ethanol 2:1) yielded 4a (314 mg, 66%
yield) as a colorless oil. 1H NMR (400 MHz, CDCl3): δ 1.90 (s, 6H), 1.96 (s, 6H), 2.02 (s, 6H), 2.06 (s,
6H), 3.3 (m, 4H), 3.5 (m, 2H), 3.7 (m, 2H), 3.8 (m, 4H), 3.95–3.97 (m, 4H), 4.01–4.07 (m, 2H), 4.17–4.21
(dd, 1H, J=5.1, 12.0 Hz), 4.30 (s, 2H), 4.77 (br s, 2H), 5.17–5.24 (m, 3H), 7.4 (broad t, 2H), 8.0 (broad
t, 2H), 8.4 (broad t, 2H), 7.83 (s, 2H), 8.20 (s, 1H) ppm. 13C NMR (75 MHz, CDCl3): δ 19.69, 19.73,

J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
105
19.88, 24.13, 24.58, 33.02, 37.99, 42.04, 52.78, 61.44, 65.00, 65.76, 67.56, 68.28, 68.37, 96.62, 129.44,
132.80, 135.84, 165.92, 168.71, 168.84, 169.28, 169.39, 169.70, 169.84 ppm. IR (neat): diagnostic ν
2102 cm−1 azide. HR FAB-MS (pos) calcd for C45H65O26N9: 1195.4042, obsd MH+: 1196.4136.
4.2.7. 5-Azidomethyl-N,N0-bis[11-O-(2,3,4,6-tetra-O-acetyl-α-D-mannopyranosyl)-3,6,9-trioxa-
undecyl]benzene-1,3-dicarboxamide 4b
Glycoconjugate 2b (80 mg, 0.14 mmol) was dissolved in absolute ethanol (10 mL). Degussa Pd/C
(20 mg) was added, and H2 was bubbled through the mixture for 1.5 h. The solution was filtered over
Celite and concentrated in vacuo. The residue was dissolved in tetrahydrofuran (10 mL). Pyridine (100
µL), DMAP (20 mg), and 5-azidomethylbenzene-1,3-dicarbonyl dichloride (12 mg, 0.054 mmol) were
added, and the solution was stirred under a nitrogen atmosphere at 25°C for 16 h. The solvent was
removed in vacuo, and the crude reaction was dissolved in ethyl acetate (15 mL) and washed with 1 M
HCl (1×10 mL), saturated sodium bicarbonate (1×10 mL), and brine (1×10 mL). The organic layer was
dried (MgSO4), filtered, and concentrated in vacuo. Silica gel chromatography (ethyl acetate:petroleum
1
ether 2:1) yielded 4b (43 mg, 63%) as a colorless oil. H NMR (300 MHz, CDCl3): δ 1.98 (s, 6H),
2.04 (s, 6H), 2.10 (s, 6H), 2.16 (s, 6H), 3.57–3.78 (m, 32H), 4.05–4.14 (m, 4H), 4.23–4.27 (dd, 2H,
J=4.6, 12.1 Hz), 4.44 (s, 2H), 4.84 (d, 2H, J=1.4 Hz), 5.22–5.37 (m, 6H), 7.65 (s, 2H), 8.01 (s, 1H) ppm.
13C NMR (75 MHz, CDCl3): δ 19.77, 19.84, 21.44, 21.51, 67.26, 69.94, 68.41, 68.59, 69.20, 69.57,
69.85, 69.86, 70.22, 70.42, 70.45, 70.49, 72.26, 96.43, 128.59, 135.34, 166.12, 169.90, 169.92 ppm. HR
FAB-MS (pos) calcd for C53H77N5O28: 1231.4755, obsd MH+: 1232.4891.
4.2.8. N,N0,N00-Tris[2-O-(2,3,4,6-tetra-O-acetyl-α-D-mannopyranosyl)-N-ethylglycyglycyl]benzene-1,
3,5-tricarboxamide 5a
Protected glycopeptide 2a (135 mg, 0.21 mmol) was dissolved in absolute ethanol (25 mL). Degussa
Pd/C (50 mg) was added, and H2 was bubbled through the mixture for 2.5 h. The solution was filtered
over Celite and concentrated in vacuo to yield deprotected glycopeptide. The deprotected glycopeptide
(100 mg, 0.2 mmol) was dissolved in dichloromethane (10 mL). 1,3,5-Benzenetricarbonyl trichloride
(12 mg, 0.05 mmol) and triethylamine (25 µL) were added, and the reaction was stirred under a nitrogen
atmosphere at 25°C for 24 h. The solution was washed with water (1×20 mL), 1 M HCl (1×20 mL),
saturated sodium bicarbonate (1×20 mL), and brine (1×20 mL). The organic layer was dried (MgSO4),
filtered, and concentrated in vacuo. Silica gel chromatography (ethyl acetate:ethanol 2:1) yielded 5a (58
mg, 69%) as a colorless oil. 1H NMR (400 MHz, CDCl3): δ 1.99 (s, 9H), 2.06 (s, 9H), 2.11 (s, 9H), 2.15
(s, 9H), 3.37 (br m, 3H), 3.48–3.57 (br m, 6H), 3.70–3.77 (m, 6H), 4.01–4.15 (m, 18H), 4.26–4.30 (dd,
3H, J=5.3, 12.4 Hz), 4.82 (br s, 3H), 5.25–5.35 (m, 9H), 7.1 (br s, 3H), 8.4 (s, 3H), 8.6 (br s, 3H), 9.22
(br s, 3H) ppm. 13C NMR (75 MHz, CDCl3): δ 20.66, 20.75, 20.84, 20.92, 29.62, 39.12, 43.05, 44.45,
44.47, 62.36, 65.93, 66.99, 68.62, 69.04, 69.25, 97.77, 129.09, 133.41, 165.88, 169.67, 170.05, 170.11,
170.40, 170.46, 170.51, 170.86 ppm. HR FAB-MS (pos) calcd for C69H93N9O39: 1671.5572, obsd MH+:
1672.5642.
4.2.9. N,N0,N00-Tris[11-O-(2,3,4,6-tetra-O-acetyl-α-D-mannopyranosyl)-3,6,9-trioxaundecyl]benzene-
1,3,5-tricarboxamide 5b
Glycoconjugate 2b (161 mg, 0.29 mmol) was dissolved in absolute ethanol (10 mL). Degussa Pd/C (30
mg) was added, and H2 was bubbled through the mixture for 1.5 h. The solution was filtered over Celite
and concentrated in vacuo. The residue was dissolved in tetrahydrofuran (10 mL), and pyridine (100
µL), DMAP (20 mg), and 1,3,5-benzenetricarbonyl trichloride (25 mg, 0.097 mmol) were added. The
reaction mixture was stirred under a nitrogen atmosphere at 25°C for 16 h. The solution was concentrated

106
J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
in vacuo, and the residue was dissolved in ethyl acetate (15 mL). The solution was washed with 1 M HCl
(1×10 mL), saturated sodium bicarbonate (1×10 mL), and brine (1×10 mL). The organic layer was
dried (MgSO4), filtered, and concentrated in vacuo. Purification by silica gel chromatography (ethyl
acetate:petroleum ether 5:1) yielded 5b (43 mg, 63%) as a colorless oil. 1H NMR (300 MHz, CDCl3): δ
1.96 (s, 9H), 2.02 (s, 9H), 2.07 (s, 9H), 2.13 (s, 9H), 3.54–3.78 (m, 48H), 4.00–4.13 (m, 6H) 4.24–4.29
(dd, 3H, J=5.1, 12.3 Hz), 4.82 (d, 3H, J=1.4 Hz), 5.22–5.33 (m, 9H), 7.19 (br s, 3H), 8.39 (s, 3H) ppm.
FAB-MS (pos) calcd for C75H111N3O42: 1725.7, obsd MH+: 1726.7.
4.2.10. 5-Azidomethylbenzene-1,3-dicarboxylate di-N-hydroxysuccinimide ester 6
5-Azidomethylbenzene-1,3-dicarboxylic acid (200 mg, 0.9 mmol), N-hydroxysuccinimide (230 mg,
2.22 equiv.), and DCC (413 mg, 2 mmol) were dissolved in tetrahydrofuran (25 mL). The reaction was
stirred under a nitrogen atmosphere at 25°C for 18 h. The mixture was filtered, and dichloromethane
(25 mL) was added. The solution was concentrated in vacuo, and the residue was dissolved in 1:1
ethanol:dichloromethane. Upon cooling, a white precipitate formed which was separated by gravity
filtration. The precipitate 6 (330 mg, 88% yield) was dried in vacuo and used without further purification.
1H NMR (300 MHz, CDCl3): δ 2.94 (s, 8H), 4.58 (s, 2H), 8.38 (s, 2H), 8.84 (s, 1H) ppm. 13C NMR (75
MHz, CDCl3): δ 24.65, 130.91, 134.27, 159.34, 167.67 ppm. HR FAB-MS (pos) calcd for C17H13N5O8:
415.0764, obsd MH+: 416.0824.
4.2.11. 1,3,5-Benzenetricarboxylate tri-N-hydroxysuccinimide ester 7
Benzene 1,3,5-tricarboxylate (300 mg, 1.4 mmol), N-hydroxysuccinimide (540 mg, 4.7 mmol), and
DCC (1.2 g, 6 mmol) were dissolved in tetrahydrofuran (25 mL). The reaction was stirred under a nitro-
gen atmosphere at 25°C for 18 h. The mixture was filtered, and dichloromethane (25 mL) was added. The
solution was concentrated in vacuo, and the residue was dissolved in 1:1 ethyl acetate:dichloromethane.
Upon cooling, a white precipitate formed which was separated by gravity filtration. The precipitate 7
(420 mg, 59.8% yield) was concentrated in vacuo and used without further purification. 1H NMR (300
MHz, CDCl3): δ 2.94 (s, 12H), 9.14 (s, 3H) ppm. 13C NMR (75 MHz, CDCl3): δ 25.76, 137.32, 168.26
ppm. FAB-MS (pos) calcd for C21H15N3O12: 501.0656, obsd MH+: 502.0720.
4.2.12. 5-Azidomethyl-N,N0-bis[N,N0-bis[2-O-(2,3,4,6-tetra-O-acetyl-α-D-mannopyranosyl)-N-ethyl-
glycylglycyl]-1-phenylmethyl-3,5-dicarboxamide]benzene-1,3-dicarboxamide 8a
Protected bivalent 4a (100 mg, 0.08 mmol) was dissolved in methanol (15 mL). Degussa Pd/C (30
mg) was added, and H2 was bubbled through the mixture for 4.5 h. The solution was filtered over Celite,
washed with hot methanol, and concentrated in vacuo. The residue was dissolved in tetrahydrofuran (10
mL) and triethylamine (50 µL). Compound 6 (8 mg, 0.023 mmol) was added, and the reaction was stirred
under a nitrogen atmosphere at 25°C for 1 week. The reaction was concentrated in vacuo and purified by
size-exclusion chromatography (BioGel SX-1 resin in tetrahydrofuran) to yield 8a (27 mg, 47% yield)
as a colorless oil. 1H NMR (300 MHz, CDCl3): uninterpretable. 13C NMR: unable to obtain. FAB-MS
(pos) calcd for C107H137N17O54: 2524, obsd MH+: 2525.
4.2.13. 5-Azidomethyl-N,N0-bis[N,N0-bis[11-O-(2,3,4,6-tetra-O-acetyl-α-D-mannopyranosyl)-3,6,9-
trioxaundecyl]-1-phenylmethyl-3,5-dicarboxamide]benzene-1,3-dicarboxamide 8b
Protected bivalent 4b (43 mg, 0.035 mmol) was dissolved in absolute ethanol (10 mL). Degussa Pd/C
(60 mg) was added, and H2 was bubbled through the mixture for 2 h. The solution was filtered over
Celite, washed with hot ethanol (40 mL), and concentrated in vacuo. The residue was dissolved in
tetrahydrofuran (10 mL) and triethylamine (50 µL). Compound 6 (5.7 mg, 0.014 mmol) was added,

J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
107
and the reaction was stirred under a nitrogen atmosphere at 25°C for 48 h. The reaction was concentrated
in vacuo and purified by size-exclusion chromatography (BioGel SX-1 resin in tetrahydrofuran) to yield
8b (23 mg, 63%) as a colorless oil. 1H NMR (300 MHz, CDCl3): δ 1.98 (s, 12H), 2.03 (s, 12H), 2.09 (s,
12H), 2.14 (s, 12H), 2.27 (s, 2H), 3.59–3.74 (m, 56H), 4.02–4.10 (m, 14H), 4.13–4.30 (dd, 4H, J=4.9,
12.3 Hz), 4.41 (s, 2H), 4.60 (s, 4H), 4.84 (d, 4H, J=1.5 Hz), 4.99 (s, 2H), 5.24–5.30 (m, 18H), 6.98 (s,
2H), 7.56 (s, 2H), 7.80 (s, 1H), 7.87 (s, 2H), 7.91 (s, 1H), 8.40 (s, 1H) ppm. HR FAB-MS (pos) calcd for
C115H161N9O58: 2595.9926, obsd MNa+: 2618.9912.
4.2.14. N,N0,N00-Tris-[N,N0-bis[2-O-(2,3,4,6-tetra-O-acetyl-α-D-mannopyranosyl)-N-ethylglycyl-
glycyl]-1-phenylmethyl-3,5-dicarboxamide]benzene-1,3,5-tricarboxamide 9a
Protected bivalent 4a (80 mg, 0.06 mmol) was dissolved in methanol (15 mL). Degussa Pd/C (30
mg) was added, and H2 was bubbled through the mixture for 4 h. The solution was filtered over Celite,
washed with hot methanol, and concentrated in vacuo. The residue was dissolved in tetrahydrofuran (10
mL) and triethylamine (50 µL). Compound 7 (4 mg, 0.008 mmol) was added, and the reaction was stirred
under a nitrogen atmosphere at 25°C for 1 week. The reaction was concentrated in vacuo and purified
by size-exclusion chromatography (BioGel SX-1 resin in tetrahydrofuran) to yield 9a (18 mg, 61%) as a
colorless oil. Both 1H and 13C NMR spectra showed only broad featureless resonances. FAB-MS (pos)
calcd for C156H201O81N21: 3664, obsd MNa+: 3687.
4.2.15. N,N0,N00-Tris-[N,N0-bis[11-O-(2,3,4,6-tetra-O-acetyl-α-D-mannopyranosyl)-3,6,9-trioxa-
undecyl]-1-phenylmethyl-3,5-dicarboxamide]benzene-1,3,5-tricarboxamide 9b
Protected bivalent 4b (95 mg, 0.08 mmol) was dissolved in absolute ethanol (10 mL). Degussa Pd/C
(50 mg) was added, and H2 was bubbled through the mixture for 2.5 h. The solution was filtered over
Celite, washed with hot ethanol (30 mL), and concentrated in vacuo. The reduced bivalent compound
was dissolved in tetrahydrofuran (10 mL) and triethylamine (50 µL). Compound 7 (10 mg, 0.02
mmol) was added, and the reaction was stirred under a nitrogen atmosphere at 25°C for 72 h. The
reaction was concentrated in vacuo and purified by size-exclusion chromatography (BioGel SX-1 resin
in tetrahydrofuran) to yield 9b (36 mg, 48%) as a colorless oil. 1H NMR (300 MHz, CDCl3): δ 1.98 (s,
18H), 2.03 (s, 18H), 2.09 (s, 18H), 2.14 (s, 18H), 3.60–3.65 (m, 84H), 3.75 (s, 6H), 4.05–4.16 (m, 21H),
4.25–4.31 (dd, 6H, J=4.9, 12.3 Hz), 4.84 (d, 6H, J=1.5 Hz), 5.24–5.34 (m, 21H), 7.90 (br m, 12H) ppm.
FAB-MS (pos) calcd for C163H227N9O87: 3774, obsd MNa+: 3797.
4.2.16. 1-O-[2-(N-Benzylamidoglycylglycyl)ethyl]-α-D-mannopyranoside 10a
Protected monovalent 3a (50 mg, 0.08 mmol) was dissolved in 0.1 M sodium methoxide in methanol
(15 mL), and the reaction stirred at 25°C for 18 h. The solution was neutralized with Dowex H+ resin,
filtered, and concentrated in vacuo. The residue was dissolved in deionized water and purified by size-
1
exclusion chromatography (BioRad P2 gel in water) to yield 10a (quant.) as a white powder. H NMR
(400 MHz, D2O): δ 3.41–3.55 (m, 2H), 3.58–3.65 (m, 3H), 3.70–3.82 (m, 3H), 3.84–3.87 (m, 1H),
3.91–3.93 (dd, 1H, J=1.7, 3.4 Hz), 3.96 (s, 2H), 4.16 (s, 2H), 4.84 (d, 1H, J=1.71 Hz), 7.46–7.68 (m,
3H), 7.88–7.84 (m, 2H) ppm. 13C NMR (75 MHz, D2O): δ 38.88, 42.42, 43.12, 60.80, 65.58, 66.66,
69.90, 70.38, 72.68, 99.54, 127.20, 128.20, 128.67, 128.75, 131.13, 132.50, 171.31, 171.41, 172.29 ppm.
FAB-MS (pos) calcd for C19H27O9N3: 441.11, obsd MNa+: 464.1.
4.2.17. 1-O-(11-Benzylamido-3,6,9-trioxaundecyl)-α-D-mannopyranoside 10b
Protected monovalent 3b (30 mg, 0.033 mmol) was dissolved 0.1 M sodium methoxide in methanol (5
mL), and the reaction stirred at 25°C for 18 h. The solution was neutralized with Dowex H+ resin, filtered,

108
J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
and concentrated in vacuo. The residue was dissolved in deionized water and purified by size-exclusion
chromatography (BioRad P2 gel in water) to yield 10b (quant.) as a colorless oil. 1H NMR (400 MHz,
D2O): δ 3.39–3.73 (m, 21H), 3.77 (dd, 1H, J=1.7, 3.3 Hz), 4.68 (d, 1H, J=1.5, H-1, Hz), 7.38 (t, 2H,
J=7.5 Hz), 7.47 (t, 1H, J=7.5 Hz), 7.62 (d, 2H, J=7.0 Hz) ppm. FAB-MS (pos) calcd for C21H33NO10:
459.21, obsd MNa+: 482.20.
4.2.18. 5-Azidomethyl-N,N0-bis[2-O-(α-D-mannopyranosyl)-N-ethylglycylglycyl]benzene-1,3-di-
carboxamide 11a
Protected bivalent 4a (75 mg, 0.06 mmol) was dissolved in 0.1 M sodium methoxide in methanol (10
mL), and the reaction stirred at 25°C for 18 h. The solution was neutralized with Dowex H+ resin, filtered,
and concentrated in vacuo. The residue was dissolved in deionized water and purified by size-exclusion
1
chromatography (BioRad P2 gel in water) to yield 11a (quant.) as a white solid. H NMR (400 MHz,
D2O): δ 3.41–3.55 (m, 4H), 3.58–3.65 (m, 6H), 3.70–3.81 (m, 6H), 3.83–3.87 (m, 2H), 3.91–3.92 (dd,
2H, J=1.7, 3.5 Hz), 3.97 (s, 4H), 4.20 (s, 4H), 4.60 (s, 2H), 4.85 (d, 2H, J=1.71 Hz), 8.04 (d, 2H, J=1.68
Hz), 8.23 (t, 1H, J=1.68 Hz) ppm. 13C NMR (75 MHz, D2O): δ 24.42, 24.47, 38.86, 42.42, 43.10, 51.57,
60.78, 65.56, 66.64, 67.43, 69.87, 70.35, 72.65, 72.74, 99.51, 125.94, 130.53, 133.79, 137.43, 169.66,
171.35, 172.02 ppm. FAB-MS (pos) calcd for C33H49N9O18: 859.32, obsd MNa+: 882.3.
4.2.19. 5-Azidomethyl-N,N0-bis[11-O-(α-D-mannopyranosyl)-3,6,9-trioxaundecyl]benzene-1,3-di-
carboxamide 11b
Protected bivalent 4b (56 mg, 0.05 mmol) was dissolved in 0.1 M sodium methoxide in methanol
(5 mL), and the reaction stirred at 25°C for 18 h. The solution was neutralized with Dowex H+ resin,
filtered, and concentrated in vacuo. The residue was dissolved in deionized water and purified by size-
exclusion chromatography (BioRad P2 gel in water) to yield 11b (quant.) as a white solid. 1H NMR (400
MHz, D2O): δ 3.60–3.97 (m, 28H), 4.59 (s, 2H), 4.82 (d, 2H, J=1.6 Hz), 7.96 (s, 2H), 8.13 (s, 1H) ppm.
FAB-MS (pos) for C37H61N5O20: 895.39, obsd MNa+: 918.38.
4.2.20. N,N0,N00-Tris[2-O-(α-D-mannopyranosyl)-N-ethylglycylglycyl]benzene-1,3,5-tricarboxamide
12a
Protected trivalent 5a (20 mg, 0.02 mmol) was dissolved in 0.1 M sodium methoxide in methanol (10
mL), and the reaction stirred at 25°C for 18 h. The solution was neutralized with Dowex H+ resin, filtered,
and concentrated in vacuo. The residue was dissolved in deionized water and purified by size-exclusion
1
chromatography (BioRad P2 gel in water) to yield 12a (quant.) as a white solid. H NMR (400 MHz,
D2O): δ 3.41–3.53 (m, 6H), 3.58–3.65 (m, 9H), 3.70–3.81 (m, 9H), 3.84–3.87 (m, 3H), 3.91–3.93 (dd,
3H, J=1.7, 3.5 Hz), 3.98 (s, 6H), 4.23 (s, 6H), 4.85 (d, 3H, J=1.71 Hz), 8.47 (s, 1H) ppm. 13C NMR
(75 MHz, D2O): δ 27.29, 32.42, 39.32, 42.91, 43.60, 61.19, 65.97, 67.04, 70.28, 73.03, 99.85, 129.92,
134.23, 169.16, 171.54, 172.16 ppm. FAB-MS (pos) calcd for C45H69N9O27: 1167, obsd MNa+: 1190.
4.2.21. N,N0,N00-Tris[11-O-(α-D-mannopyranosyl)-3,6,9-trioxaundecyl]benzene-1,3,5-tricarboxamide
12b
Protected trivalent 5b (58 mg, 0.05 mmol) was dissolved in 0.1 M sodium methoxide in methanol
(5 mL), and the reaction stirred at 25°C for 18 h. The solution was neutralized with Dowex H+ resin,
filtered, and concentrated in vacuo. The residue was dissolved in deionized water and purified by size-
exclusion chromatography (BioRad P2 gel in water) to yield 12b (quant.) as a white solid. 1H NMR (400
MHz, D2O): δ 3.60–3.97 (m, 54H), 4.82 (d, 3H, J=1.6 Hz), 8.37 (s, 3H) ppm. FAB-MS (pos) calcd for
C51H87N3O30: 1222, obsd MNa+: 1245.

J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
109
4.2.22. 5-Azidomethyl-N,N0-bis[N,N0-bis[2-O-(α-D-mannopyranosyl)-N-ethylglycylglycyl]-1-phenyl-
methyl-3,5-dicarboxamide]benzene-1,3-dicarboxamide 13a
Protected tetravalent 8a (20 mg, 0.008 mmol) was dissolved in 0.1 M sodium methoxide in methanol
(10 mL), and the reaction stirred at 25°C for 18 h. The solution was neutralized with Dowex H+ resin,
filtered, and concentrated in vacuo. The residue was dissolved in deionized water and purified by size-
exclusion chromatography (BioRad P2 gel in water) to yield 13a (quant.) as a white solid. 1H NMR (400
MHz, D2O, 60°C): δ 3.40–3.48 (m), 3.55–3.65 (m), 3.69–3.81 (m), 3.82–3.84 (m), 3.87–3.89 (m), 3.96
(s), 4.17 (s, 2H), 4.40 (s), 4.55 (br s), 8.02 (m), 8.17 (m) ppm.
4.2.23. 5-Azidomethyl-N,N0-bis[N,N0-bis[11-O-(α-D-mannopyranosyl)-3,6,9-trioxaundecyl]-1-phenyl-
methyl-3,5-dicarboxamide]benzene-1,3-dicarboxamide 13b
Protected tetravalent 8b (23 mg, 0.008 mmol) was dissolved in 0.1 M sodium methoxide in methanol
(10 mL), and the reaction stirred at 25°C for 24 h. The solution was neutralized with Dowex H+ resin,
filtered, and concentrated in vacuo. The residue was dissolved in deionized water and purified by size-
exclusion chromatography (BioRad P2 gel in water) to yield 13b (quant.) as a white solid. 1H NMR (400
MHz, D2O): δ 3.60–3.97 (m, 56H), 4.60 (s, 2H), 4.82 (d, 4H, J=1.6 Hz), 7.97 (br m, 6H), 8.13 (br m,
3H) ppm. FAB-MS (pos) calcd for C83H129N9O42: 1924, obsd MNa+: 1947.
4.2.24. N,N0,N00-Tris[N,N0-bis[2-O-(α-D-mannopyranosyl)-N-ethylglycylglycyl]-1-phenylmethyl-3,5-
dicarboxamide]benzene-1,3,5-tricarboxamide 14a
Protected hexavalent 9a (18 mg, 0.005 mmol) was dissolved in 0.1 M sodium methoxide in methanol
(10 mL), and the reaction stirred at 25°C for 18 h. The solution was neutralized with Dowex H+ resin,
filtered, and concentrated in vacuo. The residue was dissolved in deionized water and purified by size-
1
exclusion chromatography (BioRad P2 gel in water) to yield 14a (quant.) as a colorless oil. H NMR
(400 MHz, D2O): δ 3.41–3.51 (m), 3.59–3.68 (m), 3.72–3.82 (m), 3.84–3.87 (m), 3.92–3.94 (m), 3.98
(s), 4.20 (s, 2H), 7.43–7.44 (m), 7.70–7.72 (m), 7.82–7.84 (m) ppm.
4.2.25. N,N0,N00-Tris[N,N0-bis[11-O-(α-D-mannopyranosyl)-3,6,9-trioxaundecyl]-1-phenylmethyl-3,
5-dicarboxamide]benzene-1,3,5-tricarboxamide 14b
Protected hexavalent 9b (18 mg, 0.005 mmol) was dissolved in 0.1 M sodium methoxide in methanol
(10 mL), and the reaction stirred at 25°C for 18 h. The solution was neutralized with Dowex H+ resin,
filtered, and concentrated in vacuo. The residue was dissolved in deionized water and purified by size-
exclusion chromatography (BioRad P2 gel in water) to yield 14b (quant.) as a colorless oil. Both 1H and
13C NMR spectra showed only broad featureless resonances. FAB-MS (pos) calcd for C115H179N9O63:
2694, obsd MNa+: 2717.
4.2.26. Hemeagglutination inhibition assay
Prior to each assay, the minimum lectin concentration for agglutination of a 2% porcine erythrocyte
solution was determined, and a sixfold excess of lectin in phosphate-buffered saline pH 7.3 (PBS) was
utilized during inhibition assays. Serial twofold dilutions of ligand were incubated with concanavalin A
for 2 h at 37°C. A 2% solution of porcine erythrocytes in phosphate was added to the lectin solution before
subsequent incubation at 37°C. The minimum concentration of ligand required to inhibit agglutination
was determined and reported relative to α-D-methyl mannopyranoside.

110
J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
4.2.27. ELLA analysis
ELLA assays were carried out as previously described. All incubations were carried out at 37°C.34,35
Briefly, Falcon microtiter plates were coated overnight with yeast mannan (100 µL of a 10 µg/mL
solution in PBS). Excess mannan was removed by three washes with PBS containing 0.05% v/v Tween-
20 (PBST). Coated wells were blocked with 1% w/v bovine serum albumin in PBS (150 µL) for 1
h. During this period, serial twofold dilutions of ligand were also incubated with peroxidase labeled
concanavalin A (a dilution of a 0.1 mg/mL solution which had been determined to give A410 of 1.0–0.8
AU). The lectin solutions were then transferred to the blocked plate for an additional 1 h incubation.
These plates were subsequently washed twice with PBST and once in PBS prior to development with
ABTS (50 µL of a 250 µg/mL solution in citrate–phosphate buffer pH 4.0 with 0.02% v/v hydrogen
peroxide). The absorbance at 410 nm was monitored after a 20 min period. Fractional inhibitions are
based on A410 without inhibitor present and referenced to the inhibitory potential of an α-D-methyl
mannose standard determined for each plate. All assays were carried out in duplicate.
4.2.28. Titration microcalorimetry
Calorimetric experiments were carried out on a MicroCal Omega isothermal titration microcalorime-
ter; details on the instrument and mathematical analyses are described elsewhere.37 The reaction cell
(volume=1.3678 mL) contained a solution of concanavalin A (0.60 mM monomer) in buffer (50 mM
dimethyl glutaric acid, 250 mM sodium chloride, 1 mM calcium chloride, 1 mM manganese chloride,
at pH 5.2). A series of forty injections of dendrimer solution (4–28 mM mannoside) were made. The
volume of each injection was 2.2 µL, the duration of injection was 4.4 s, and the time between injections
was 3 min. The data from the resulting injections were integrated to generate a titration curve. A non-
linear least squares fit of the data was used to determine the binding constant, K, the enthalpy of binding,
∆H, and the stoichiometry of binding, n.
References
1. Varki, A. J. Clin. Invest. 1997, 99, 158–162.
2. Lowe, J. B.; Ward, P. A. J. Clin. Invest. 1997, 99, 822–826.
3. Kiessling, L. L.; Pohl, N. L. Chem. Biol. 1996, 3, 71–77.
4. Gimbrone, M. A.; Nagel, T.; Topper, J. N. J. Clin. Invest. 1997, 99, 1809–1813.
5. Shur, B. D. Curr. Opin. Cell Biol. 1993, 5, 854–63.
6. Firpo, P. P.; Axberg, I.; Scheibel, M.; Clark, E. A. AIDS Res. Hum. Retroviruses 1992, 8, 357–366.
7. McCluskey, J.; Boyd, L. F.; Highet, P. F.; Inman, J.; Margulies, D. H. J. Immunol. 1988, 141, 1451–1455.
8. Mammen, M.; Choi, S. K.; Whitesides, G. M. Angew. Chem., Int. Ed. 1998, 37, 2755–2794.
9. Toone, E. J. Curr. Opin. Struct. Biol. 1994, 4, 719–728.
10. Ashton, P. R.; Boyd, S. E.; Brown, C. L.; Jayaraman, N.; Nepogodiev, S. A.; Stoddart, J. F. Chem. Eur. J. 1996, 2,
1115–1128.
11. Biessen, E. A. L.; Noorman, F.; van Teijlingen, M. E.; Kuipert, J.; Barret-Bergshoeff, M.; Bijsterbosch, M. K.; Rijken, D.
C.; van Berkel, T. J. C. J. Biol. Chem. 1996, 271, 28024–28030.
12. Choi, S.-K.; Mammen, M.; Whitesides, G. M. Chem. Biol. 1996, 3, 97–194.
13. Lees, W. J.; Spaltenstein, A.; Kingery-Wood, J. E.; Whitesides, G. M. J. Med. Chem. 1993, 37, 3419–3433.
14. Page, D.; Roy, R. Bioconjug. Chem. 1997, 8, 714–723.
15. Page, D.; Roy, R. Bioorg. Med. Chem. Lett. 1996, 6, 1765–1770.
16. Quesenberry, M. S.; Lee, R. T.; Lee, Y. C. Biochemistry 1997, 36, 2724–2732.
17. Roy, R. Polymer News 1996, 6, 692.
18. Roy, R. Curr. Opin. Struct. Biol. 1996, 6, 692–702.
19. Sigal, G. B.; Mammen, M.; Dahmann, G.; Whitesides, G. M. J. Am. Chem. Soc. 1996, 118, 3789–3800.
20. Strong, L. E.; Kiessling, L. L. J. Am. Chem. Soc. 1999, 121, 6193–6196.

J. B. Corbell et al. / Tetrahedron: Asymmetry 11 (2000) 95–111
111
21. Sparks, M. A.; Williams, K. W.; Whitesides, G. M. J. Med. Chem. 1993, 36, 778–783.
22. Roy, R. Topics Curr. Chem. 1997, 187, 241–274.
23. Zanini, D.; Roy, R. Bioconjug. Chem. 1997, 8, 187–192.
24. Ozaki, K.; Lee, R. T.; Lee, Y. C.; Kawasaki, T. Glycoconjugate J. 1995, 12, 268–274.
25. Lee, Y. C.; Lee, R. T. Acc. Chem. Res. 1995, 28, 321–327.
26. Mammen, M.; Shakhnovich, E. I.; Whitesides, G. M. J. Org. Chem. 1998, 63, 3168–3175.
27. Dimick, S. M.; Powell, S. C.; McMahon, S. A.; Moothoo, D. N.; Naismith, J. H.; Toone, E. J. J. Am. Chem. Soc. 1999, 121,
10286–10296.
28. Schuster, M. C.; Mortell, K. H.; Hegeman, A. D.; Kiessling, E. L. J. Mol. Catal. A: Chem. 1997, 116, 209–216.
29. Inoue, M.; Nakada, H.; Oka, Y.; Tanaka, N.; Yamashina, I. Glycoconjugate J. 1997, 14, 147–153.
30. Roy, R.; Page, D.; Perez, S. F.; Bencomo, V. V. Glycoconjugate J. 1998, 15, 251–263.
31. Peerlings, H. W. I.; Nepogodiev, S. A.; Stoddart, J. F.; Meijer, E. W. Eur. J. Org. Chem. 1998, 1879–1886.
32. Gupta, R. K.; Sharma, S. B.; Ahuja, S.; Saxena, S. N. Acta Microbiol. Hung. 1991, 38, 81–90.
33. Gupta, D.; Oscarson, S.; Raju, T. S.; Stanley, P.; Toone, E. J.; Brewer, C. F. Eur. J. Biochem. 1996, 242, 320–326.
34. Page, D.; Zanini, D.; Roy, R. Bioorg. Med. Chem. 1996, 4, 1949–1961.
35. Zanini, D.; Roy, R. J. Am. Chem. Soc. 1997, 119, 2088–2095.
36. Freire, E.; Mayorga, O. L.; Straume, M. Anal. Chem. 1990, 62, A950–A959.
37. Wiseman, T.; Williston, S.; Brandts, J. F.; Lin, L. N. Anal. Biochem. 1989, 179, 131–137.
38. Chervenak, M. C.; Toone, E. J. Biochemistry 1995, 34, 5685–5695.
39. Hinz, H.-J. Annu. Rev. Biophys. Bioeng. 1983, 12, 285–317.
40. Schwarz, F. P.; Ahmed, H.; Bianchet, M. A.; Amzel, M.; Vasta, G. R. Biochemistry 1997, 37, 5867–5877.
41. Aki, H.; Goto, M.; Kai, M.; Yamamoto, M. J. Therm. Anal. 1999, 57, 361–370.
42. Graziano, G.; Catanzano, F.; Nappa, M. Int. J. Biol. Macromol. 1999, 26, 45–53.
43. Pluschke, G.; Mutz, M. J. Therm. Anal. 1999, 57, 377–388.
44. Williams, B. A.; Chervenak, M. C.; Toone, E. J. J. Biol. Chem. 1992, 267, 22907–22911.
45. Stites, W. E. Chem. Rev. 1997, 97, 1233–1250.
46. Antonelli, M. L.; Bonicelli, M. G.; Ceccaroni, G.; Lamesa, C.; Sesta, B. Colloid Polym. Sci. 1994, 272, 704–711.
47. Paula, S.; Sus, W.; Tuchtenhagen, J.; Blume, A. J. Phys. Chem. 1995, 99, 11742–11751.
48. Nusselder, J. J. H.; Engberts, J. B. F. N. J. Colloid Interface Sci. 1992, 148, 353.
49. Weckstrom, K.; Hann, K.; Rosenholm, J. B. J. Chem. Soc., Faraday Trans. 1994, 90, 733.
50. Biridi, K. S. Colloid Polym. Sci. 1983, 261, 45.
51. Edelhoch, H. Biochemistry 1967, 6, 1948–1954.
52. Dubois, M.; Gilles, K. A.; Hamilton, J. K.; Rebers, P. A.; Smith, F. Anal. Chem. 1956, 28, 350.