Optimization of tether length in nonglycosidically linked bivalent ligands that target sites 2 and 1 of a Shiga-like toxin

Article abstract of DOI:10.1021/ja0258529

A series of bivalent ligands for a Shiga-like toxin have been synthesized, their experimentally determined inhibitory activities were compared with a simplified thermodynamic model, and computer simulations were used to predict the optimal tether length i

Full text of DOI:10.1021/ja0258529

Published on Web 02/20/2003
Optimization of Tether Length in Nonglycosidically Linked
Bivalent Ligands That Target Sites 2 and 1 of a Shiga-like
Toxin
Pavel I. Kitov,† Hiroki Shimizu,‡ Steven W. Homans,‡ and David R. Bundle*,†
Contribution from the Chemistry Department, UniVersity of Alberta,
Edmonton, Canada T6G 2G2 and School of Biochemistry and Molecular Biology,
UniVersity of Leeds, Leeds LS2 9JT, United Kingdom
Received February 7, 2002 ; E-mail: Dave.Bundle@ualberta.ca
Abstract: A series of bivalent ligands for a Shiga-like toxin have been synthesized, their experimentally
determined inhibitory activities were compared with a simplified thermodynamic model, and computer
simulations were used to predict the optimal tether length in bivalent ligands. The design of the inhibitors
exploits the proximity of the C-2′ hydroxyl groups of two Pk-trisaccharides when bound to two different,
neighboring carbohydrate recognizing binding sites located on the surface of Shiga-like toxin. NMR studies
of the complex between the toxin and bivalent ligands show that site 2 and site 1 of a single B subunit are
simultaneously occupied by a tethered Pk-trisaccharide dimer. A simplified thermodynamic treatment provides
the intrinsic affinities and binding energies for the intermolecular and intramolecular association events
and permits the deconvolution of the contributions to the relative binding energies for the set of bivalent
ligands. Conformational analysis based on MD simulations for bivalent galabioside dimers containing different
tethers demonstrated that the calculated local concentrations of the pendant ligand at the second binding
site correlate with the experimentally determined relative affinity values of the respective bivalent ligands,
thereby providing a predictive method to optimize tether length.
multiantennary glycans,4 for the most part short oligosaccharides
displayed randomly on various polymers achieve rather modest
Introduction
Intervention in the recognition of mammalian cell surface
oligosaccharides by secreted or surface bound lectins of
pathogenic organisms is a potentially attractive therapeutic
strategy to combat bacterial adherence1 or microbial toxin
binding and entry into specific cells.2 However, the realization
of this objective is frustrated by the low intrinsic affinity for
acceptance of a polar oligosaccharide in to what is nearly always
a hydrophilic, shallow depression in the protein surface of the
lectin. Tight binding is achieved through multivalent binding
facilitated by the occurrence of more than one carbohydrate
recognition site per lectin subunit, as well as the oligomeric
nature of many lectins, and the presence on mammalian cells
of very large numbers of glycolipids or glycoproteins that
display the relevant oligosaccharide structure. Numerous at-
tempts to emulate Nature’s multivalency by creating oligo- or
polymeric presentations of oligosaccharides have met with
varied success.3
gains in avidity, especially when compared to the number of
saccharide units attached to the polymer. The limited avidity
gains presumably arise because only a small fraction of the
available saccharides is able to bind with the lectin. Develop-
ment of a viable strategy that could reliably make more effective
use of every saccharide unit by adjusting the spatial display to
match the separation of the lectin sites is a challenging but
worthwhile approach, as illustrated by the recent reports of
spectacular avidity gains for penta- and decavalent inhibitors
that bind to the class of AB5 bacterial toxins.5,6 Translation of
such tailored multivalency to the recognition of cell surface
oligosaccharides by mammalian lectins would have potential
for intervention in inflammatory responses7 and a myriad of
other cell-cell interactions that are known to depend on
(4) (a) Lee, Y. C.; Townsend, R. R.; Hardy, M. R.; Lonngren, J.; Arnarp, J.;
Haraldsson, M.; Lonn, H. J. Biol. Chem. 1983, 258, 199-202. (b) Lee, R.
T.; Lin, P.; Lee, Y. C. Biochemistry 1984, 23, 4255-61. (c) Lee, R. T.;
Lee, Y. C. Glycoconjugate J. 2000, 17, 543-51.
(5) (a) Kitov, P. I.; Sadowska, J. M.; Mulvey, G.; Armstrong, G. D.; Ling, H.;
Pannu, N. S.; Read, R. J.; Bundle, D. R. Nature 2000, 403, 669-72. (b)
Bundle, D. R.; Kitov, P. I.; Read, R. J.; Ling, H.; Armstrong, G. D.
Treatment of Bacterial Dysentery. U.S. Patent 5,962,423, Oct 5, 1999. (c)
Bundle, D. R.; Kitov, P.; Read, R. J.; Ling, H.; Armstrong, G. Treatment
of Bacterial Infections. U.S. Patent 6,310,043, Oct 30, 2001.
Despite the landmark and spectacular example of hepatic
lectin binding to terminal galactose residues displayed on
† University of Alberta.
‡ University of Leeds.
(1) Aronson, M.; Medalia, O.; Schori, L.; Mirelman, D.; Sharon, N.; Ofek, I.
J. Infect. Dis. 1979, 139, 329-32.
(6) (a) Fan, E. K.; Zhang, Z. S.; Minke, W. E.; Hou, Z.; Verlinde, C. L. M. J.;
Hol, W. G. J. J. Am. Chem. Soc. 2000, 122, 2663-64. (b) Merritt, E. A.;
Zhang, Z. S.; Pickens, J. C.; Ahn, M.; Hol, W. G. J.; Fan, E. K. J. Am.
Chem. Soc. 2002, 124, 8818-8824. (c) Zhang, Z.; Merritt, E. A.; Ahn,
M.; Roach, C.; Hou, Z.; Verlinde, C. L. M. J.; Hol, W. G. J.; Fan, E. J.
Am. Chem. Soc. 2002, 124, 12991-8.
(2) (a) Karlsson, K.-A.; Milh, M. A.; A° ngstro¨m, J. In Molecular Recognition
in Host Parasite Interactions; Korhonen, T. K., Ed.; Plenum Press: New
York 1992; pp 115-32. (b) Karlsson, K.-A. Trends Pharmacol. Sci. 1991,
12, 265-72.
(3) Mammen, M.; Choi, S. K.; Whitesides, G. M. Angew. Chem., Int. Ed. 1998,
37, 2755-94.
(7) Springer, T. A. Nature 1991, 349, 196-7.
9
3284
J. AM. CHEM. SOC. 2003, 125, 3284-3294
10.1021/ja0258529 CCC: $25.00 © 2003 American Chemical Society

Tether Length Optimization Procedure
A R T I C L E S
Scheme 1
oligosaccharide-based recognition.8 Here, we report our attempts
to develop a predictive procedure for optimizing the spacing of
tethered oligosaccharides.
A large number of Pk-tri- and disaccharide analogues have
been synthesized by us18 and others,14 and their activities
evaluated. However, no viable univalent lead compound could
be identified. It appears that the plurality of the binding modes
compromises the chances to create a small molecular weight
inhibitor that would be equally efficient for blocking all three
possible binding pockets found on the surface of the toxin. On
the other hand, the multitude of binding modes provides an
opportunity to explore the effect of multivalent interaction with
both equivalent and nonequivalent binding sites as a target. The
arrangement of all three sites across the surface of the toxin
makes it reasonable to predict that tethering of several Pk-
trisaccharide units into one cluster will increase the inhibitory
activity due to what is commonly referred to as the cluster
glycoside effect.19
Design of Inhibitors. Examination of the crystal structure
suggested an opportunity for bridging two neighboring mol-
ecules of the trisaccharide bound to site 1 and site 2, respec-
tively. Despite its participation in water-mediated hydrogen
bonding, the hydroxyl group at the C-2 position of the â-Gal
moiety is not directly involved in any crucial interactions with
the protein in either site 1 or site 2.18 The distance between the
O-2′ atoms of two Pk-trisaccharides bound in site 1 and site 2
is only 11 Å, and the path between the two is not obstructed by
any protein residue. Previously, we have demonstrated that alkyl
substituents at the O-2′ position of Pk-trisaccharide do not
significantly affect binding to SLT-1.18 Therefore, bridging
between those two nonglycosidic hydroxyls presented an
attractive alternative to tethering via glycosidic positions, which
is more common in the design of multivalent glycoconjugates.
For the majority of model dimers, we have chosen a syntheti-
cally efficient, carbamate-containing bridge, the length of which
can be readily adjusted, and one tether is constructed using a
bifunctional squarate derivative. We present here the synthesis
of six such dimers 4-9 (Scheme 1) containing di- and
trisaccharide units.
Shiga-like toxins (SLTs) belong to a small but clinically
significant family of AB5 bacterial toxins.9 They consist of a
cytotoxic enzyme A subunit noncovalently linked to a ho-
mopentameric cell-adhesion carrier, B5. The entry of the
enzymic A subunit into the host cell relies on adhesion of the
B5 subunit to the cell membrane, an interaction that is highly
specific and mediated by Pk-trisaccharide, the carbohydrate
constituent of Gb3 glycolipid, also called CD77.10 Inhibition of
toxin binding to the host-cell surface is an attractive antiadhesive
therapy for bacterial infections.11
The solved crystal structure12 of the complex between Shiga-
like toxin type 1 (SLT-1) and a soluble analogue of Pk-
trisaccharide, R-D-Gal(1-4)-â-D-Gal(1-4)-â-D-GlcOR, reveals
15 Pk-trisaccharide binding sites of three distinct types. Although
all three types of binding sites recognize the same cell-surface
receptor, their relative affinity established by site-directed
mutagenesis13 decreases in the order: site 2 > site 1 > site 3
(the designation of binding sites is according to ref 14 and does
not necessarily follow the order of affinity). The intrinsic
affinities are very low, and even with the highest affinity, site
2 has an association constant of 500-1000 M-1 for Pk-
trisaccharide as determined by microcalorimetry.15 Site 1 is
estimated by NMR to have only 10-15% of the site 2 activity.16
We have previously demonstrated by mass spectrometry that
the equilibrium distribution of ligand bound and unbound SLT-1
B5 species indicates at least an order of magnitude difference
between the two highest binding constants.17
(8) Varki, A. Glycobiology 1993, 3, 97-130.
(9) Merritt, E. A.; Hol, W. G. J. Curr. Opin. Struct. Biol. 1995, 5, 165-71.
(10) Lindberg, A. A.; Brown, J. E.; Stromberg, N.; Westerling-Ryd, M.; Schultz,
J. E.; Karlsson, K.-A. J. Biol. Chem. 1987, 262, 1779-85.
(11) Sharon, N.; Ofek, I. Glycoconjugate J. 2000, 17, 659-64.
(12) Ling, H.; Boodhoo, A.; Hazes, B.; Cummings, M. D.; Armstrong, G. D.;
Brunton, J. L.; Read, R. J. Biochemistry 1998, 37, 1777-88.
(13) (a) Bast, D. J.; Banerjee, L.; Clark, C.; Read, R. J.; Brunton, J. L. Mol.
Microbiol. 1999, 32, 953-60. (b) Soltyk, A. M.; MacKenzie, C. R.; Wolski,
V. M.; Hirama, T.; Kitov, P. I.; Bundle D. R.; Brunton, J. L. J. Biol. Chem.
2002, in press.
Synthesis of SLT-1 Inhibitors. The general assembly
strategy for synthesis of bivalent di- and trisaccharide ligands
(14) Nyholm, P. G.; Magnusson, G.; Zheng, Z.; Norel, R.; Binnington-Boyd,
B.; Lingwood, C. A. Chem. Biol. 1996, 3, 263-75.
(15) St. Hilaire, P. M.; Boyd, M. K.; Toone, E. J. Biochemistry 1994, 33, 14452-
63.
(16) Richardson, J. M.; Evans, P. D.; Homans, S. W.; Donohue-Rolfe, A. Nat.
Struct. Biol. 1997, 4, 190-3.
(17) Kitova, E. N.; Kitov, P. I.; Bundle, D. R.; Klassen, J. S. Glycobiology 2001,
11, 605-11.
(18) Kitov, P. I.; Bundle, D. R. J. Chem. Soc., Perkin Trans. 1 2001, 838-853.
(19) Lundquist, J. J.; Toone, E. J. Chem. ReV. 2002, 102, 555-8.
9
J. AM. CHEM. SOC. VOL. 125, NO. 11, 2003 3285

A R T I C L E S
Kitov et al.
Scheme 2
involves activation of a hydroxyl group of a partially protected
oligosaccharide derivative as its p-nitrophenyl carbonate, fol-
lowed by coupling with an amine to form a carbamate linkage.
Oxidation of the double bond of the allyl derivative 1418 with
4-methylmorpholine N-oxide in the presence of OsO4 followed
by the cleavage of the resulting diol with NaIO4 and subsequent
reduction of the resulting aldehyde with NaBH4 provided alcohol
15 (Scheme 2). Esterification of 10,18 12,18 and 15 with
p-nitrophenyl chloroformate in pyridine gave the corresponding
p-nitrophenyl carbonates 11, 13, and 16, which served as key
building blocks for installation of carbohydrate determinants
into the oligomeric inhibitors. Condensation of 11 with ethyl-
enediamine and propylenediamine afforded two homologous
bridged disaccharide dimers 18 and 19. Reaction of 13 with
p-phenylenediamine provided dimer 20 containing an aromatic
bridge. Because of the low nucleophilicity of the aromatic amine,
the reaction requires rather forcing conditions: extended time,
elevated temperature, and nucleophilic catalysis with DMAP.
The symmetric trisaccharide dimer 21 was obtained by the
reaction of 16 with propylenediamine, and the mixed di- and
trisaccharide dimer 22 was synthesized by the consecutive
treatment of propylenediamine with one equiv of 11 then with
16. Compound 17, which consists of a trisaccharide attached
to the tether fragment, was obtained by coupling of 16 with
one equiv of propylenediamine followed by the acetylation of
the product with acetic anhydride. Catalytic debenzylation of
17-22 afforded the target Pk-trisaccharide analogues 3-6, 8,
and 9.
The synthesis of squaric acid derivative 7 also starts from
the common precursor 10. Mesylation gave compound 23, which
was converted into azide 24 by nucleophilic displacement.
Catalytic hydrogenation of the azido functionality accompanied
by debenzylation afforded the deprotected amino derivative 25.
Condensation of 2 equiv of 25 with bifunctional squarate gave
target digalabioside 7.
Evaluation of Inhibitory Activity. A competitive solid-phase
assay was employed for evaluation of the inhibitory activities
of the Pk-trisaccharide tethered dimers 4-9 (Figure 1). We have
previously reported details of this assay.18 In brief, the recom-
binant B5 subunit of SLT-1 was immobilized on the wells of
microtiter plates, and then a solution of Pk-trisaccharide/BSA/
biotin glycoconjugate was coincubated in the microtiter well
in the absence or presence of various concentrations of inhibitor.
For each concentration of inhibitor, the assay was performed
in triplicate. After 18 h of equilibration, the plate was washed
with buffer and the amount of bound biotin was measured by
a standard biotin-streptavidine protocol. The results are presented
in the Table 1.
NMR Studies of the Complex between the SLT-1 B5
Subunit and Bivalent Ligand 9. To obtain direct structural
evidence for the simultaneous binding to two distinct sites, we
employed a “SAR-by-NMR” approach20 to examine the
9
3286 J. AM. CHEM. SOC. VOL. 125, NO. 11, 2003

Tether Length Optimization Procedure
A R T I C L E S
shift changes can be mapped onto the three-dimensional
structure of the protein.12 It is immediately obvious from these
data that the resonances derived from Thr 31, Asp 18, Asp 17,
Glu 65, and Lys 53 are shifted in a fashion that reflects those
observed in a previous study on the binding of Pk-trisaccharide.21
In that study, these shift changes were shown to arise from
binding at site 2. In the present study, these shift changes are
more substantial, because the affinity of 9 is greater than Pk-
trisaccharide, resulting in a larger fraction of complexed protein.
Additional shifts of comparable magnitude to those mentioned
previously can be observed in Figure 2. Shifts of resonances
assigned to Thr 19, Thr 21, Glu 28, and Lys 13 are of particular
note, since these were not detected previously in the SLT-1B-
Pk-trisaccharide complex.16,21 As the amino acids responsible
for these resonances are located in site 1 of the crystal structure,
the data taken together suggest that 9 simultaneously occupies
sites 1 and 2. In addition to the resonance shifts mentioned thus
far, a number of other significant shifts are observed in Figure
2 (e.g., Gly 61). The residues corresponding to these shifts are
common to both sites 1 and 2, and hence, we interpret these
shifts as arising from the combined effect of simultaneous
occupancy of both sites.
One formal alternative interpretation of these data is that 8
simultaneously occupies sites 1 and 2, in two different B subunit
molecules. However, in this hypothetical case, the resulting
complex would exhibit a large molecular correlation time with
significant broadening of resonances in 15N-1H HSQC spectra,
which is a characteristic of such large complexes (∼75 kDa).
Contrary to those expectations, the resonance line widths in the
complex of the B subunit with 9 are substantially smaller than
those for the B5 subunit alone, and hence, we can confidently
rule out this alternative interpretation. The basis of this line
narrowing is currently under investigation via 15N NMR
relaxation time measurements.
Molecular Dynamics Simulations. We constructed the
surrogate binding site {1,2} consisting of site 1 and site 2 by
modifying the crystal structure of the complex between SLT-1
and Pk-trisaccharide glycoside.12 Tether elements were added,
and the glucose moieties in bound Pk-trisaccharides were
replaced by methyl groups. Only side chains of selected amino
acids were retained in order to form a bumping surface that
limits the conformational space available for the pendant ligand
and tether. During simulations, the positions of the former
protein atoms were fixed and a harmonic tether of 100 kcal/
mol was applied to the disaccharide bound to site 2 (Figure 3).
Since relatively high barriers of rotation for single bonds would
preclude collection of representative conformational distributions
at room temperature in a reasonable period of time, all MD
experiments were conducted at a temperature of 1000 K.
Figure 1. (Panel A) Schematic representation of the competitive ELISA
used to measure inhibitor activity. (Panel B) Representative example of
binding data. Error bars represent the standard error for triplicates.
Table 1. Activities of Synthetic SLT-1 Inhibitors
number of
ligands
relative
activitiesa
compound
1/IC50 (M-1
)
118
218
3
4
5
6
7
8
9
1
1
1
2
2
2
2
2
2
33
1
4.7 × 102
7.1 × 102
7.7 × 102
1.03 × 103
2.5 × 102
1.56 × 103
3.9 × 103
1.8 × 104
23.3
31.2
7.57
47.3
a With respect to monovalent galabioside 1.
Discussion
interaction of 9 with the uniformly 15N-enriched Shiga-like toxin
B5 subunit. Shown in Figure 2 are overlaid 15N-1H HSQC
spectra of the B5 subunit alone (black contours) and in the
presence of a 4-fold molar excess (with respect to monomer B)
of 9 (red contours). Substantial chemical shift changes in the
complex provide direct evidence for binding of 9 to the protein.
By use of 15N and HN resonance assignments obtained in a
previous NMR study of the B subunit homopentamer,16 these
It has been previously shown that the IC50 values in this
competitive assay format with a labeled glycoconjugate as the
reporter molecule closely approximate the KD values between
the inhibitor of interest and the protein receptor.22 Indeed, the
data for methyl glycosides of galabiose 1 and Pk-trisaccharide
(21) Shimizu, H.; Field, R. A.; Homans, S. W.; Donohue-Rolfe, A. Biochemistry
1998, 37, 11078-82.
(22) (a) Vorberg, E.; Bundle, D. R. J. Immunol. Methods 1990, 132, 81-9. (b)
Sigurskjold, B. W.; Altman, E.; Bundle, D. R. Eur. J. Biochem. 1991, 197,
239-46.
(20) Shuker, S. B.; Hajduk, P. J.; Meadows, R. P.; Fesik, S. W. Science 1996,
274, 1531-4.
9
J. AM. CHEM. SOC. VOL. 125, NO. 11, 2003 3287

A R T I C L E S
Kitov et al.
Figure 2. Overlaid 15N-1H HSQC spectra of 15N-enriched Shiga-like toxin B subunit homopentamer in the absence (black contours) and presence (red
contours) of Pk-trisaccharide dimer 9 in 4-fold molar excess over protein monomer. Assignments of key resonances that exhibit shifts between the two
spectra are labeled in the usual notation. Assignments shown in red type correspond with those that were observed to shift in a previous study of the B
subunit-Pk-trisaccharide complex.16
When compared with univalent ligands (galabioside 1 and
Pk-trisaccharide 2), bivalent ligands 4-9 show substantially
higher activity. The activity of the Pk-trisaccharide 3, which
contains only a fragment of the propylenediamine tether, does
not differ significantly from that of unmodified trisaccharide 2.
This indicates that possible nonspecific interactions between the
protein surface and structural elements of the tether do not
contribute significantly to binding and the affinity of the dimeric
analogues results from protein-carbohydrate interactions. In the
series of dimeric ligands containing the propylenediamine tether
5, 8, and 9, the activities progressively increase with the number
and affinity of the ligands attached to both ends of the tether.
Thus, the weakest dimeric ligand is the dimer of Pk-disaccharides
5, and the strongest is the dimer of Pk-trisaccharides 9.
Multivalency effects observed for dimers 5, 8, and 9 that
contain the same propylendiamine-based tether can be inter-
preted in terms of a simplified thermodynamic model. It was
suggested by Jencks23 that the standard free energy of binding
for composite ligands A-B is the sum of ∆G° values for the
individual interactions A and B combined with a contribution
by the tether.
Figure 3. Structure of SLT-1-dimer 5 complex prepared for MD
simulation. Bound galabioside (left) in site 2 is constrained by a force of
100 kcal/mol to its crystal structure position. Free movements of tethered
pendant galabioside (shown on the right in its original position in site 1)
are limited by a bouncing surface formed by side chains of selected amino
acids.
∆G° ) ∆G°A + ∆G°B + ∆G°tether
(1)
2 obtained in our assay are in good agreement with KD values
measured by microcalorimetry.15 This is a rather fortuitous
coincidence due to two mutually compensating effects. Al-
though, generally, IC50 values obtained in a competitive assay
are systematically higher than the absolute values of KD, in the
case of the interaction of a monovalent ligand with a multivalent
receptor, the statistical factor leads to an increase of apparent
activity. However, the overriding importance of a solid-phase
competitive assay conducted under consistent conditions is the
reproducibility of the relatiVe values of KD. These are suf-
ficiently accurate for the comparative purposes discussed here.
In the case of symmetrical dimers, the statistical term
RT ln 2 must be introduced in order to account for the effectively
double concentration of ligand.
The relationship between the elementary interactions of
univalent and bivalent ligands with a composite binding site
{1,2} is delineated in Figure 4. This diagram allows a tentative
estimate of the individual contributions to the interaction of di-
and trisaccharides in binding sites 1 and 2. In conjunction with
NMR data for the complex of 9 with the toxin, this balance of
(23) Jencks, W. P. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 4046-50.
9
3288 J. AM. CHEM. SOC. VOL. 125, NO. 11, 2003

Tether Length Optimization Procedure
A R T I C L E S
(Clocal is the “local” or “effective” concentration), and the
corresponding intrinsic monoValent binding constant at the
second site (KII) as defined by the equation
K2 ) ClocalKII
(3)
Empirical rules that allow estimation of Clocal for the ends of
polymeric tethers6,25 are not applicable for short tethers of
irregular structure. However, for a system with a relatively small
number of atoms, a more straightforward approach can be taken
to evaluate the radial density bias introduced by the conforma-
tional flexibility of the tether. The model we employ assumes
that one ligand is bound in the protein site of highest affinity.
This location or rather an atom of the immobilized ligand serves
as a pivot or hinge point about which the second ligand and
tether move. In our case, the position of the hinge atom O-2 of
the â-Gal residue serves as the origin of internal coordinates.
Molecular dynamics simulations are used to monitor the
evolution of the system over a 5 ns time period.
Figure 4. Balance of the relative free energies of binding for a series of
monovalent and bivalent ligands. The binding energy for galabioside 1 is
taken as a reference. (/) The ∆G° values for symmetrical compounds 5
and 9 are corrected by the statistical term RT ln 2.
free energies presents compelling evidence that both binding
sites 1 and 2 are simultaneously engaged in interaction with
the bivalent ligands.
After establishing this fact, we turned our attention to the
conformational analysis of bivalent ligands in order to find
criteria that may facilitate a priori evaluation of different tethered
structures. The relative activities of bivalent galabiosides 4-7,
in which the tether structures differ only by the central structural
element, ethylene, propylene, phenylene, and squaric acid
diamide, are used to evaluate this model.
The local concentration of the unbound tethered ligand in a
volume circumscribed by two hemispheres with radii R and R
- ∆R is given as
3 × 1027 P(R,∆R)
Clocal
)
2πNA(R3 - (R - ∆R)3)
P(R,∆R)
(R3 - (R - ∆R)3)
) 792.73
(4)
The obvious advantage of Jencks’ concept is the additivity
of binding energies of individual interactions. However, with
the exception of A-B ligands containing particularly rigid
tethers, the interaction with a receptor will be a sequential rather
than concerted process. In agreement with this generalization,
it has been demonstrated that, for the case when two binding
sites are engaged in an interaction with a tethered dimeric ligand,
the increased binding constant results mainly from a decrease
in the off-rate of the dimer.24 Accordingly, the interaction of
our dimeric ligands with two different Pk-binding sites of SLT-1
may be better represented by two distinct steps: binding first
to one and then to another binding site (Scheme 3).
Obviously, the description of the system by Jencks’ formula
1 may only be possible under a condition of overwhelming
predominance of complete complex II among bound species.
In this case, the observed binding constant is approximately the
product of intermediate constants.
where R is the distance in Å between bound and unbound ligand;
P(R,∆R) is the probability that the pendant ligand is found
between R and R - ∆R from the bound ligand; and NA is
Avogadro’s number.
The radial probability function P(R,∆R) may be obtained by
analysis of the molecular dynamics simulation. For the dynamics
analysis, the internal coordinate is naturally chosen to be the
distance between the pivotal point, O-2 of the â-Gal of the
bound galabiose, and the ring oxygen of â-Gal of the pendant
galabiose, the latter being the closest to the center of gravity
for this monosaccharide. Figure 5 shows how the radial
concentrations of pendant ligands that are tethered by different
linkers change with the distance.
At separation distance R ) 10.9 Å, which corresponds to
the distance of interest in the crystal structure, both ethylene-
diamine- and propylenediamine-based linkers appear to be
substantially more favorable than the phenylenediamine-based
tether, and the highest affinity is predicted for the squaric-acid-
based tethered dimer. This tendency is in good agreement with
the observed affinities for compounds 4, 5, 6, and 7 (Table 1
and Figure 6). Evidently, the more subtle difference between
homologues 4 and 5 is beyond the limitations of this method.
The difference in binding energies between univalent and
bivalent Pk-trisaccharides 1 and 9 is 1.76 kcal/mol (see Figure
4), which corresponds to K2 ) 20 at ambient temperature.
Taking into account the value of Clocal ) 0.2 M from the
conformational analysis for the propylenediamine-based linker
and solving eq 3 with respect to KII, we can obtain the value
95 M-1 for the monovalent binding constant for the low affinity
Kapp ≈ [II]/[L][R] ) K1K2 ) K′1K′2
(2)
As indicated by previously published NMR16 and mass
spectrometry data,17 the difference between site 2 and site 1
affinities is at least an order of magnitude and no appreciable
binding with site 1 occurs until a high degree of site 2 occupancy
is achieved. Hence, for clarity, we may neglect the second route
via I′ as 10 times less probable.
We assume that the binding constant of intermolecular
interaction K1 equals 2KI, where KI is the binding constant of
the monovalent binding of Pk-di- or trisaccharide to site 2. The
intramolecular binding constant K2 is directly proportional to
the concentration of pendant ligand at the second binding site
(25) Gargano, J. M.; Ngo, T.; Kim, J. Y.; Acheson, D. W. K.; Lees, W. J. J.
Am. Chem. Soc. 2001, 123, 12909-10.
(24) Kramer, R. H.; Karpen, J. W. Nature 1998, 395, 710-3.
9
J. AM. CHEM. SOC. VOL. 125, NO. 11, 2003 3289

A R T I C L E S
Kitov et al.
Scheme 3
binding site 1. This is in good agreement with the magnitude
of the low affinity binding constant, which was estimated to be
the estimate of the population of conformations capable of
binding in the second site is based only on the radial distribution
function for a single interatomic distance.
I
10-15% that of the dominant binding site, Ka ) 500-1000
M-1. This independent evaluation of the method also suggests
that this approach can be used in the design of tethered
multimeric ligands as one of the a priori selection criteria. This
simple approach performs surprisingly well, considering that
An advantage of this approach in distinguishing relatively
small differences in activity of analogous bivalent ligands is
that it does not depend on the evaluation of binding energies.
It relies on estimation of the probability of supplementary
interaction rather than its strength and indirectly takes into
account all changes in conformational entropy (i.e., the sum of
torsional entropies of single bonds in a tether) upon binding
without actually calculating it. The magnitude of the radial
distribution function at the binding site separation distance
provides a convenient and computationally inexpensive means
for approximating the local concentration of a pendant ligand
at the secondary binding site for an arbitrary tether structure. It
also reflects the radial nature of the ligand movements about
the bound ligand. An alternative approach for evaluation of Clocal
involving calculation of the probability to find a pendant ligand
in the vicinity of the binding site would require a large training
set of bivalent ligands in order to define the capture distance,
whereas the radial density distribution function permits direct
comparison of any two ligands of interest.
Figure 5. Theoretical distribution of Clocal for galabioside dimers 4-7
predicted on the basis of eq 4. The atom O-2′ belongs to the ligand held in
site 2, while the atom O-5′ is part of the pendant ligand.
Conclusions
A predictive method to optimize tether length has been
developed to facilitate the design of bivalent oligosaccharide
ligands. MD simulations provide the local concentration of
pendant ligands attached via a tether to one oligosaccharide that
is locked in one sugar binding site. The tether length, at which
the pendant ligand is most highly populated at the additional
binding site, expressed as a radial distribution function is shown
to correspond to optimal tether lengths determined experimen-
tally. In addition, a simplified thermodynamic treatment provides
the intrinsic affinities and binding energies for the intermolecular
and intramolecular association events and permits the decon-
volution of the contributions to the relative free binding energies
for the set of bivalent ligands. We will show in a subsequent
paper how this structure-based approach has been applied to
tailoring larger architectures, which, starting from the original
univalent ligand with millimolar activity, resulted in the design
of a potent sub-nanomolar inhibitor.5
Figure 6. Correlation between experimental intramolecular binding
constants K2 for symmetrical dimers 4-7 and calculated local concentration
of the pendant ligand at the low affinity binding site 1. (In the estimation
of K2, the statistical coefficient of 2 is taken into account.)
9
3290 J. AM. CHEM. SOC. VOL. 125, NO. 11, 2003

Tether Length Optimization Procedure
A R T I C L E S
PBST. An inhibitor at decreasing concentrations (starting from ∼50
mg/mL, 3.16 dilution factor, 50 µL/well) was mixed with biotinylated
Pk-BSA glycoconjugate (2 µg/mL, 50 µL/well), and the mixtures were
applied to the plates in triplicate. After incubation at room temperature
for 18 h, the plates were washed (6×) with PBST, incubated with
streptavidine-horseradish peroxidase in PBST (1 µg/mL, 100 µL/well)
for 1 h at room temperature, and then washed (6×) with PBST. 3,3′,5,5′-
Tetramethylbenzidine (TMB) peroxidase substrate was added (both the
streptavidine-horseradish peroxidase and its TMB substrate were
purchased from Gibco BRL Inc.). After 2-5 min, the reaction was
quenched by addition of H3PO4 (100 µL/well). The plates were read at
450 nm. Nonlinear data fitting into a logistic sigmoidal curve based
on the Levenberg-Marquardt algorithm was performed using Microcal
Origin 5.0 (Microcal Software, Inc., Northampton, MA 01060, USA).
Experimental Section
General Methods. Optical rotations were measured on a Perkin-
Elmer 241 polarimeter for samples in a 10 cm cell at ambient
temperature (22 ( 2 °C). Analytical TLC was performed on silica gel
60-F254 (Merck) with detection by quenching of fluorescence and/or
by charring with 10% H2SO4 in ethanol solution followed by heating
at 180 °C. Column chromatography was performed on silica gel 60
(Merck, 40-60 µm), and solvents were distilled prior to use. Sep-Pak
C18 cartridges (Waters) were conditioned prior to use by washing with
methanol (10 mL) and water (20 mL). 1H NMR spectra were recorded
at 300 and 600 MHz (Varian) in CDCl3 (referenced to residual CHCl3
at δH 7.24 ppm), CD3OD (referenced to residual CD2HOD at δH 3.3
ppm), or in D2O (referenced to external acetone at δH 2.225 ppm).
First order J values are given in Hz. All commercial reagents were
used as supplied; solvents were distilled from appropriate desiccants
prior to use. After extraction, solutions in DCM were filtered through
a cotton plug.
Methyl 3,6-Di-O-benzyl-2-O-[2-(4-nitrophenoxycarbonyloxy)-
ethyl]-4-O-(2,3,4,6-tetra-O-benzyl-r-D-galactopyranosyl)-â-D-galac-
topyranoside (11). To a solution of 1018 (490 mg, 0.521 mmol) in
pyridine (3 mL) 4-nitrophenylchloroformate (158 mg, 1.5 equiv) was
added. The mixture was stirred overnight at 40 °C, and then water (10
mL) was added. The mixture was diluted with DCM, washed with
water, and the organic layer was concentrated. Chromatography of the
residue on silica gel with pentane-ethyl acetate (85:15-75:25) gave
11 (542 mg, 94%), [R]D +40.7° (c 0.8; CHCl3).1H NMR (CDCl3): δ
NMR Sample Preparation. The B5 homopentamer of Shiga-like
toxin was prepared in 15N-enriched form as described.16 A sample was
prepared for NMR studies by dissolution of 2.5 mg of 15N-enriched
subunit in 250 µL of 100 mM phosphate buffer, pH 7.0, in a Shigemi
5 mm microcell. After acquisition of a 15N-1H HSQC spectrum of the
protein, 2.3 mg of 9 in solid form was dissolved in the sample and a
second 15N-1H HSQC spectrum was acquired under otherwise identical
conditions to those of the first.
NMR Experiments. All spectra were acquired on a Varian Unity
Inova 600 MHz spectrometer at a probe temperature of 303 K. Spectral
widths of 5500 and 2400 Hz were utilized with 2048 and 256 complex
points in the F2 and F1 dimensions, respectively. In total, 32 transients
were acquired per t1 increment. Solvent suppression was achieved using
the standard WATERGATE suppression scheme.26
′
′
7.37-7.17 (m, 34 H, arom.), 5.04 (d, 1 H, J1 ,2 2.0 Hz, H-1′), 4.89 (d,
2
2
2
1 H, J 11.2 Hz, Bn), 4.88 (d, 1 H, J 11.8 Hz, Bn), 4.80 (d, 1 H, J
12.0 Hz, Bn), 4.73 (s, 2 H, Bn), 4.67 (d, 1 H, 2J 11.8 Hz, Bn), 4.57 (d,
2
2
1 H, J 12.9 Hz, Bn), 4.53 (d, 1 H, J 11.3 Hz, Bn), 4.43-4.38 (m, 3
H, H-5′, CH2), 4.27 (d, 1 H, 2J 11.8 Hz, Bn), 4.22 (d, 1 H, 2J 11.8 Hz,
Bn), 4.19 (d, 1 H, J1,2 7.5 Hz, H-1), 4.14 (s, 2 H, Bn), 4.10-3.94 (m,
7 H, H-2′, H-3′, H-4′, H-4, H-6a, CH2), 3.58-3.45 (m, 7 H, H-2, H-5,
H-6b, H-6′a, Me), 3.35 (dd, 1 H, J3,2 9.9 Hz, J3,4 2.9 Hz, H-3), 3.26
′
′
′
′
(dd, 1 H, J6 b,5 4.9 Hz, J6 b,6 a 8.4 Hz, H-6′b). Anal. Calcd. for C64H67-
Molecular Dynamics Simulations. All MD simulations were
performed using Discover and the associated suite of software programs
from Molecular Simulations Inc. (San Diego). The consistent valency
force field (CVFF), without Morse and cross terms, was used with the
effective dielectric constant set to 80. Each run consisted of 1 fs time
step intervals for a total of 5 ns, with the temperature set to 1000 K.
The resulting trajectories for each molecule were sampled every 250
fs, and the structural properties of interest evaluated as an average over
20 000 configurations.
NO16: C, 69.49; H, 6.10; N, 1.27. Found: C, 69.52; H, 6.15; N, 1.29.
Methyl 3,6-Di-O-benzyl-2-O-(4-nitrophenoxycarbonyl)-4-O-(2,3,4,6-
tetra-O-benzyl-r-D-galactopyranosyl)-â-D-galactopyranoside (13).
To a solution of 1218 (393 mg, 0.438 mmol) in pyridine (3 mL)
4-nitrophenylchloroformate (106 mg, 1.2 equiv) was added. The mixture
was stirred for 3 h at 50 °C, then diluted with DCM, washed with
water, and concentrated. Chromatography of the residue on silica gel
with pentane-ethyl acetate (80:20) gave 13 (396 mg, 85%), [R]D
1
+38.1° (c 0.5; CHCl3). H NMR (CDCl3): δ 7.32-7.01 (m, 34 H,
The B5 subunit-bivalent ligand complex was prepared from the
coordinates of the crystal structure for SLT-1-Pk-trisaccharide (PDB
entry 1BOS). One of the B subunits with complete trisaccharide
occupancy of site 1 and site 2 was chosen. The positions for all atoms
for side chains of amino acids numbers 13, 15, 16, 17, 19, 21, 28, 30,
32, 58, 60, 61, and 62 were retained, and the rest of the protein was
deleted. The terminal glucose moiety was replaced by a methyl group,
and the tether was constructed from standard fragments. Potentials were
automatically assigned followed by manual correction, formal charges
were set to 0, and partial charges for the ligand were automatically
assigned, whereas for the protein atoms they were set to 0. The energy
of the tether was minimized for 10 000 iterations using a conjugate
gradient with fixed positions for the carbohydrates and protein prior to
MD. The MD calculations were performed with fixed positions of the
protein atoms and the positions of the bound methyl galabioside in
site 2 tethered by a harmonic oscillator constrain force of 100 kcal/
mol per atom.
arom.), 5.13 (dd, 1 H, J2,1 7.9 Hz, J2,3 10.3 Hz, H-2), 5.03 (broad s, 1
2
H, H-1′), 4.96-4.87 (m, 3 H, Bn), 4.74 (s, 2 H, Bn), 4.69 (d, 1 H, J
2
11.8 Hz, Bn), 4.52 (d, 1 H, J 11.1 Hz, Bn), 4.48-4.41 m, 3 H, H-5′,
Bn), 4.28 (d, 1 H, 2J 11.7 Hz, Bn), 4.21 (d, 1 H, 2J 11.7 Hz, Bn), 4.18
′
′
(d, 1 H, J4 ,3 2.7 Hz, H-4′), 4.14-3.98 (m, 6 H, H-1, H-4, H-6a, H-2′,
H-3′, Bn), 3.61-3.50 (m, 7 H, H-3, H-5, H-6b, H-6′b, OMe), 3.17
′
′
′
′
(dd, 1 H, J6 b,5 4.8 Hz, J6 b,6 a 8.6 Hz, H-6′b). Anal. Calcd. for C62H63-
NO15: C, 70.11; H, 5.98; N, 1.32. Found: C, 70.08; H, 5.99; N, 1.32.
Methyl 4-O-[3,6-Di-O-benzyl-4-O-(2,3,4,6-tetra-O-benzyl-r-D-ga-
lactopyranosyl)-2-O-(2-hydroxyethyl)-â-D-galactopyranosyl]-2,3,6-
tri-O-benzyl-â-D-glucopyranoside (15). To a solution of 1418 (570 mg,
0.411 mmol) and 4-methylmorpholine N-oxide (96 mg, 0.82 mmol) in
an acetone-water (10 mL to 0.5 mL) mixture OsO4 (0.5 mL, 0.1 g/mL
solution in t-BuOH) was added. The mixture was stirred for 2 days
then taken up in DCM, washed with water, and concentrated. To a
solution of the residue in THF (15 mL) a solution of NaIO4 (174 mg)
in water (2 mL) was added. After stirring at 50 °C for 3 h, NaBH4
(110 mg) was added and the suspension was stirred for 0.5 h, neutralized
with AcOH, taken up in DCM, washed with water, and concentrated.
Chromatography of the residue on silica gel in hexanes-ethyl acetate
(70:30-65:35) gave 15 (490 mg, 87%), [R]D +36.1° (c 0.26; CHCl3).
1H NMR (CDCl3): δ 7.4-7.1 (m, 45 H, arom.), 5.02-4.99 (m, 2 H,
H-1′′, Bn), 4.83-4.62 (m, 8 H, Bn), 4.50-4.32 (m, 6 H, H-1′, Bn),
Inhibition ELISA Assay. PVC microtiter plates (Gibco BRL Inc.)
were incubated with Shiga-like toxin Type 1 (2.5 µg/mL, 100 µL/well)
at room temperature overnight and then washed (6×) with PBST (0.05%
Tween 20 in phosphate buffer saline, PBS). Blocking solution (2.5%
skimmed milk in PBS, 100 µL/well) was added, and the plates were
incubated for 1 h at room temperature and then washed (6×) with
′′ ′′
4.28-4.09 (m, 7 H, H-1, H-4′, H-5′′, Bn), 4.04 (dd, 1 H, J2 ,1 3.3 Hz,
(26) Sklenar, V.; Piotto, M.; Leppik, R.; Saudek, V. J. Magn. Reson., Ser. A
1993, 102, 241-5.
′′ ′′
J2 ,3 10.3 Hz, H-2′′), 4.02-3.70 (m, 8 H, H-4, H-6a, H-6b, H-5′, H-3′′,
9
J. AM. CHEM. SOC. VOL. 125, NO. 11, 2003 3291

A R T I C L E S
Kitov et al.
H-4′′, CH2), 3.52 (s, 3 H, Me), 3.62-3.35 (m, 7 H, H-3, H-5, H-2′,
H-6′a, H-6′′a, CH2), 3.33 (dd, 1 H, J2,1 7.7 Hz, J2,3 9.0 Hz, H-2), 3.27-
3.12 (m, 3H, H-3′, H-6′b, H-6′′b). Anal. Calcd. for C84H92O17: (1373.62)
C, 73.45; H, 6.75. Found: C, 73.48; H, 6.74.
3.00 (m, 4 H, CH2N), 1.48-1.39 (m, 2 H, CH2). Anal. Calcd. for
C
119H134N2O26: C, 71.17; H, 6.72; N, 1.39. Found: C, 70.92; H, 6.61;
N, 1.39.
N,N′-Bis{[methyl 3,6-di-O-benzyl-4-O-(2,3,4,6-tetra-O-benzyl-r-
Methyl 4-O-[3,6-Di-O-benzyl-4-O-(2,3,4,6-tetra-O-benzyl-r-D-ga-
lactopyranosyl)-2-O-[2-(4-nitrophenyloxycarbonylethyl)ethyl]-â-D-
galactopyranosyl]-2,3,6-tri-O-benzyl-â-D-glucopyranoside (16). A
mixture of 15 (434 mg, 0.316 mmol) and 4-nitrophenylchloroformate
(74 mg, 1.2 equiv) in dry pyridine (3 mL) was stirred for 4 h at 50 °C,
and then the reaction was quenched with water, concentrated, and
chromatographed on silica gel with pentane-ethyl acetate (80:20-75:
D-galactopyranosyl)-â-D-galactopyranoside-2-yloxy]ethyl}oxycarbonyl-
(p-phenylenediamine) (20). A mixture of 11 (162 mg, 152 µmol), 1,4-
phenylenediamine (8.2 mg, 76 µmol), and DMAP (20 mg) in DMF (2
mL) was stirred for 2 days at 50 °C, then diluted with ethyl acetate
(30 mL), washed with water, and concentrated. Chromatography of
the residue on silica gel with pentane-ethyl acetate (1:1) gave 20 (93
1
mg, 63%), [R]D +62.5° (c 0.3; CHCl3). H NMR (CDCl3): δ 7.36-
7.16 (m, 64 H, arom.), 6.38 (s, 2 H, NH), 5.19 (dd, 2 H, J2,1 8.0 Hz,
1
25) to give 16 (406 mg, 85%), [R]D +29.0° (c 0.5; CHCl3). H NMR
2
′
′
(CDCl3): δ 8.17 (m, 2 H, arom.), 7.40-7.10 (m, 47 H, arom.), 5.02-
5.00 (m, 2 H, H-1′′, Bn), 4.85-4.62 (m, 8 H, Bn), 4.50-4.40 (m, 6 H,
H-1′, Bn), 4.31-4.20 (m, 6 H, H-1, H-4′, CH2, Bn), 4.15-3.82 (m, 12
H, H-4, H-6a, H-6b, H-5′, H-2′′, H-3′′, H-4′′, H-5′′, CH2, Bn), 3.54 (t,
1 H, J3,2 ≈ J3,4 9.0 Hz, H-3), 3.50 (s, 3 H, Me), 3.50-3.40 (m, 4 H,
H-5, H-2′, H-6′a, H-6′′a), 3.34 (dd, 1 H, J2,1 7.8 Hz, H-2), 3.29-3.20
J2,3 10.2 Hz, H-2), 5.03 (d, 2 H, J1 ,2 3.2 Hz, H-1′), 4.94 (d, 2 H, J
2
2
12.0 Hz, Bn), 4.88 (d, 2 H, J 11.1 Hz, Bn), 4.79 (d, 2 H, J 11.8 Hz,
Bn), 4.77 (s, 4 H, Bn), 4.69 (d, 2 H, J 11.8 Hz, Bn), 4.56-4.51 (m,
2
4 H, H-5′, Bn), 4.38-4.01 (m, 22 H, H-1, H-4, H-6a, H-2′, H-3′, H-4′,
Bn), 3.58-3.53 (m, 6 H, H-5, H-6b, H-6′a), 3.50 (s, 6 H, OMe), 3.43
′
′
(dd, 2 H, J3,4 3.1 Hz, H-3), 3.21 (2 H, J6 b,6 a 8.9 Hz, H-6′b).). Anal.
Calcd. for C118H124N2O24: C, 72.52; H, 6.40; N, 1.43. Found: C, 72.23;
H, 6.44; N, 1.45.
′
′
′ ′
(m, 2 H, H-6′b, H-6′′b), 3.16 (dd, 1 H, J3 ,2 8.4 Hz, J3 ,4 4.7 Hz, H-3′).
Anal. Calcd. for C91H95NO21: C, 71.03; H, 6.22; N, 0.91. Found: C,
71.03; H, 6.29; N, 0.91.
N,N′-Bis{{methyl 4-O-[3,6-di-O-benzyl-4-O-(2,3,4,6-tetra-O-ben-
zyl-r-D-galactopyranosyl)-â-D-galactopyranosyl]-2,3,6-tri-O-benzyl-
â-D-glucopyranoside}-2′-yloxyethyl}oxycarbonyl-propylenedi-
amine (21). To a solution of 16 (108 mg, 70 µmol) in THF (1.5 mL)
1,3-diaminopropane (2.6 mg, 35 µmol) was added. The mixture was
stirred for 3 h at 50 °C, then concentrated, and chromatographed on
silica gel with hexane-ethyl acetate (70:30-60:40) to give 21 (93 mg,
Methyl 4-O-{2-O-[2-(3-Acetamidopropylaminocarbonyl)ethyl]-
3,6-di-O-benzyl-4-O-(2,3,4,6-tetra-O-benzyl-r-D-galactopyranosyl)-
â-D-galactopyranosyl}-2,3,6-tri-O-benzyl-â-D-glucopyranoside (17).
A solution of 16 (64.5 mg, 42 µmol) in THF (2 mL) was added to a
solution of 1,3-diaminopropane (15.5 mg, 210 µmol) in THF (0.5 mL).
After 15 min, Ac2O (0.3 mL) was added and the mixture was
concentrated. Chromatography of the residue on silica gel with hexane-
ethyl acetate (1:2) gave 17 (62.7 mg, 98%), [R]D +31.3° (c 0.2; CHCl3).
1H NMR (CDCl3): δ 7.4-7.1 (m, 45 H, arom.), 6.14 (broad t, 1 H,
1
92.5%), [R]D +25.8° (c 0.25; CHCl3). H NMR (CDCl3): δ 7.4-7.1
3
(m, 90 H, arom.), 5.53 (t, 2 H, J 6.3 Hz, NH), 5.04-5.01 (m, 4 H,
2
2
H-1′′, Bn), 4.81 (d, 2 H, J 11.2 Hz, Bn), 4.79 (d, 2 H, J 11.1 Hz,
Bn), 4.72-4.60 (m, 16 H, OCH2, Bn), 4.49-4.24 (m, 22 H, H-1, H-1′,
H-4′, Bn), 4.13-3.73 (m, 26 H, H-4, H-6a, H-6b, H-5′, H-6′a, H-2′′,
H-3′′, H-4′′, H-5′′, OCH2, Bn), 3.52 (t, 2 H, J2,3 ≈ J3,4 9.0 Hz, H-3),
3.51 (s, 6 H, Me), 3.49-3.21 (m, 12 H, H-2, H-5, H-2′, H-6′b, H-6′′a,
3
NHAc), 5.80 (t, 1 H, J 6.4 Hz, NH), 5.06-5.02 (m, 2 H, H-1′′, Bn),
3.52 (s, 3 H, OMe), 4.86-3.26 (m, 41 H, H-1, H-2, H-3, H-4, H-5,
H-6a, H-6b, H-1′, H-2′, H-4′, H-5′, H-6′a, H-6′b, H-2′′, H-3′′, H-4′′,
H-5′′, H-6′′a, H-6′′b, Bn, OCH2), 3.16-3.06 (m, 3 H, H-3′, CH2NAc),
2.84-2.72 (m, 2 H, CH2NCO2), 1.92 (s, 3 H, Ac), 1.34-1.24 (m, 2 H,
CH2CH2N). Anal. Calcd. for C90H102N2O19: C, 71.31; H, 6.78; N, 1.85.
Found: C, 71.13; H, 6.82; N, 1.81.
′
′
′ ′
H-6′′b), 3.12 (dd, 2 H, J3 ,2 8.2 Hz, J3 ,4 4.7 Hz, H-3′), 2.75 (m, 4 H,
CH2N), 1.17 (t, 2 H, 3J 6.2 Hz, CH2CH2N). Anal. Calcd. for
C
173H190N2O36: C, 72.31; H, 6.61; N, 0.97. Found: C, 72.32; H, 6.69;
N, 0.96.
N-{[Methyl 3,6-di-O-benzyl-4-O-(2,3,4,6-tetra-O-benzyl-r-D-lac-
N,N′-Bis{[methyl 3,6-di-O-benzyl-4-O-(2,3,4,6-tetra-O-benzyl-r-
D-galactopyranosyl)-â-D-galactopyranoside-2-yloxy]ethyl}oxycarbonyl-
ethylenediamine (18). A mixture of 13 (121 mg, 110 µmol) and
ethylenediamine (3.3 mg, 55 µmol) in DMF (1 mL) was stirred for 1
h at 50 °C, then diluted with ethyl acetate (30 mL), washed with water,
and concentrated. Chromatography of the residue on silica gel with
pentane-ethyl acetate (1:1) gave 18 (87.6 mg, 80%), [R]D +37.4° (c
0.23; CHCl3). 1H NMR (CDCl3): δ 7.36-7.16 (m, 60 H, arom.), 5.03
(s, 2 H, H-1′), 4.91-4.85 (m, 6 H, Bn, NH), 4.76 (d, 2 H, 2J 12.5 Hz,
topyranosyl)-â-D-galactopyranoside-2-yloxy]ethyl}oxycarbonyl-N′-
{{methyl 4-O-[3,6-di-O-benzyl-4-O-(2,3,4,6-tetra-O-benzyl-r-D-
galactopyranosyl)-â-D-galactopyranosyl]-2,3,6-tri-O-benzyl-â-D-
glucopyranoside}-2′-yloxyethyl}oxycarbonyl-propylenediamine (22).
A solution of 16 (64.5 mg, 42 µmol) in THF (2 mL) was added to a
solution of 1,3-diaminopropane (15.5 mg, 210 µmol) in THF (0.5 mL).
After 15 min, the mixture was concentrated, taken up into ethyl acetate,
washed with water, and concentrated. To a solution of the residue in
THF (2 mL) a solution of 13 (50 mg, 45 µmol) was added. The mixture
was incubated overnight at 40 °C and then concentrated. Chromatog-
raphy of the residue on silica gel with hexane-ethyl acetate (1:1) gave
22 (93.7 mg, 90.8%), [R]D +34.3° (c 0.2; CHCl3). 1H NMR (CDCl3):
2
2
Bn), 4.74 (s, 4 H, Bn), 4.66 (d, 2 H, J 11.7 Hz, Bn), 4.66 (d, 2 H, J
11.7 Hz, Bn), 4.58-4.51 (m, 4 H, Bn), 4.41-4.37 (m, 2 H, H-5′),
4.28-3.92 (m, 26 H, H-1, H-4, H-6a, H-2′, H-3′, H-4′, Bn, CH2), 3.88-
3.82 (m, 2 H, CH2), 3.55-3.43 (m, 14 H, H-2, H-5, H-6b, H-6′a, OMe),
′
′
3.31 (dd, 2 H, J3,2 9.9 Hz, J3,4 2.8 Hz, H-3), 3.22 (dd, 2 H, J6 b,5 4.9
3
Hz, J6 b,6 a 8.3 Hz, H-6′b), 3.10 (broad s, 4 H, CH2N). Anal. Calcd. for
118H132N2O26: C, 71.07; H, 6.67; N, 1.40. Found: C, 70.89; H, 6.72;
N, 1.42.
N,N′-Bis{[methyl 3,6-di-O-benzyl-4-O-(2,3,4,6-tetra-O-benzyl-r-
δ 7.4-7.1 (m, 75 H, arom.), 5.64 (t, 1 H, J 6.3 Hz, (NH)3), 5.05-
′
′
C
5.00 (m, 3 H, H-(1′′)3, H-(1′)2, Bn3), 4.89-3.70 (m, 53 H, H-13, H-43,
H-(1′)3, H-(3′)3, H-(4′)3, H-(5′)3, H-(2′′)3, H-(3′′)3, H-(4′′)3, H-(5′′)3,
H-12, H-42, H-(2′)2, H-(3′)2, H-(4′)2, H-(5′)2, Bn, CH2O), 3.52 (s, 3 H,
OMe), 3.49 (s, 3 H, OMe), 3.57-3.19 (m, 21 H, H-23, H-33, H-53,
H-(2′)3, H-(6′a)3, H-(6′b)3, H-(6′′a)3, H-(6′′b)3, H-22, H-32, H-(6a)2,
D-galactopyranosyl)-â-D-galactopyranoside-2-yloxy]ethyl}oxycarbonyl-
propylenediamine (19). Prepared as described for 18 from 13 (113
mg, 102 µmol) and 1,3-diaminopropane (3.78 mg, 51 µmol). Yield
H-(6b)2, H-52, H-(6′a)2, H-(6′b)2), 3.12 (dd, 2 H, J3 ,2 8.2 Hz, J3 ,4 4.7
Hz, H-3′), 3.13 (dd, 1 H, J(3 )3,(2 )3 8.25 Hz, J(3 )3,
′
′
′ ′
4.6 Hz, H-(3′)3),
1
′
′
′
′
(4 )3
100 mg (97%), [R]D +43.0° (c 0.33; CHCl3). H NMR (CDCl3): δ
2.99-2.92 (m, 2 H, (CH2N)2), 2.82-2.73 (m, 2 H, (CH2N)3), 1.3-1.2
(m, 2 H, CH2CH2N). Anal. Calcd. for C146H162N2O31: C, 71.84; H,
6.69; N, 1.15. Found: C, 71.78; H, 6.61; N, 1.15.
7.36-7.16 (m, 60 H, arom.), 5.00 (s, 2 H, H-1′), 4.92 (broad s, 2 H,
NH), 4.86 (d, 2 H, 2J 11.2 Hz, Bn), 4.83 (d, 2 H, 2J 11.7 Hz, Bn), 4.73
2
2
(d, 2 H, J 12.0 Hz, Bn), 4.72 (s, 4 H, Bn), 4.63 (d, 2 H, J 11.9 Hz,
Bn), 4.56-4.48 (m, 4 H, Bn), 4.38-4.34 (m, 2 H, H-5′), 4.26-3.78
(m, 28 H, H-1, H-4, H-6a, H-2′, H-3′, H-4′, Bn, CH2O), 3.54-3.40
(m, 14 H, H-2, H-5, H-6b, H-6′a, OMe), 3.28 (dd, 2 H, J3,2 9.7 Hz, J3,4
Methyl 4-O-{2-O-[2-(3-Acetamidopropylaminocarbonyl)ethyl]-
4-O-(r-D-galactopyranosyl)-â-D-galactopyranosyl}-â-D-glucopyra-
noside (3). The title compound was obtained in 89% yield by
hydrogenation of 17 as described for the preparation of 4, [R]D +46.4°
′
′
′
′
2.4 Hz, H-3), 3.20 (dd, 2 H, J6 b,5 5.1 Hz, J6 b,6 a 8.7 Hz, H-6′b), 3.04-
9
3292 J. AM. CHEM. SOC. VOL. 125, NO. 11, 2003

Tether Length Optimization Procedure
A R T I C L E S
1
′′ ′′
(c 0.3; H2O). H NMR (D2O): δ 4.97 (d, 1 H, J1 ,2 3.7 Hz, H-1′′),
ranosyl)-â-D-galactopyranosyl]- â-D-glucopyranoside}-2′-yloxyethyl}-
oxycarbonyl-propylenediamine (8). The per-benzyl ether 21 (91 mg,
37.3 µmol) was hydrogenated in AcOH in the presence of Pd/C (10%,
20 mg) for 24 h. The catalyst was filtered out. Supernatant was
concentrated. The residue was passed through a Sep-Pak (C-18)
cartridge in 10% MeOH to yield 8 (34.4 mg, 85%), [R]D +63.8° (c
0.4; H2O). 1H NMR (D2O): δ 4.99 (m, 2 H, H-(1′′)3, H-(1′)2), 4.54 (d,
′
′
4.53 (d, 1 H, J1 ,2 7.8 Hz, H-1′), 4.38 (d, 1 H, J1,2 8.0 Hz, H-1), 4.35
′′ ′′
′′ ′′
(t, 1 H, J5 ,6 a ≈ J5 ,6 b 6.4 Hz, H-5′′), 4.23-4.21 (m, 2 H, CH2O), 4.10-
3.57 (m, 16 H, H-3, H-4, H-5, H-6a, H-6b, H-3′, H-4′, H-5′, H-6′a,
H-6′b, H-3′′, H-4′′, H-6′′a, H-6′′b, CH2O), 3.58 (s, 3 H, Me), 3.42 (dd,
′
′
1 H, J2 ,3 9.9 Hz, H-2′), 3.30 (m, 1 H, H-2), 3.23-3.14 (m, 4 H, CH2-
CH2CH2N), 1.99 (s, 3 H, Ac), 1.76-1.66 (m, 2 H, CH2CH3N). 13C
NMR (D2O): δ 174.88 (CdO), 159.24 (CdO), 103.92 (C-1â), 103.25
(C-1â), 101.27 (C-1R), 80.65, 78.67, 76.09, 75.88, 75.17, 73.76, 72.92,
72.15 (CH2), 71.88, 69.98, 69.82, 69.37, 65.48 (CH2), 61.46 (CH2),
61.26 (CH2), 60.79 (CH2), 58.00 (CH3), 38.68 (CH2), 37.56 (CH2), 29.17
(CH2), 22.68 (CH3). Electrospray ionization MS for C27H48N2O19Na:
calcd 727.2749, found 727.2746.
3
1 H, J1 ,2 7.7 Hz, H-(1′) ), 4.43 (d, 1H, J1,2 7.7 Hz, H-12), 4.39 (d, 1 H,
′
′
J1,2 8.0 Hz, H-13), 4.35 (t, 2 H, J5 ,6 ≈ J5 ,6 ≈ J5 ,6 a ≈ J5 ,6 b 6.6 Hz,
H-(5′′)3, H-(5′)2), 4.25-4.20 (m, 4 H, CH2O), 4.10-3.56 (m, 28 H,
H-33, H-43, H-53, H-(6a)3, H-(6b)3, H-(3′)3, H-(4′)3, H-(5′)3, H-(6′a)3,
H-(6′b)3, H-(3′′)3, H-(4′′)3, H-(6′′a)3, H-(6′′b)3, H-32, H-42, H-52, H-(6a)2,
H-(6b)2, H-(2′)2, H-(3′)2, H-(4′)2, H-(6′a)2, H-(6′b)2, CH2O), 3.58 (s, 6
′′ ′′
′′ ′′
′′ ′′
′′ ′′
H, Me), 3.43 (dd, 1 H, J2 ,3 9.9 Hz, H-(2′)3), 3.38 (dd, 1 H, J2,3 10.1
′
′
N,N′-Bis{[methyl 4-O-(r-D-galactopyranosyl)-â-D-galactopyra-
noside-2-yloxy]ethyl}oxycarbonyl-ethylenediamine (4). A mixture of
18 (78 mg, 39 µmol) and 10% Pd/C (30 mg) was suspended in AcOH
(4 mL) and stirred overnight under hydrogen flow, then filtered,
concentrated, dissolved in water, and applied to Sep-Pak (C-18). The
cartridge was washed with water and then with 10% MeOH. Sugar
containing fractions were concentrated to furnish 4 (30 mg, 84%), [R]D
3
Hz, H-22), 3.31 (m, 1 H, H-23), 3.19 (broad t, 4 H, J 6.7 Hz, CH2N),
1.72 (m, 2 H, CH2CH3N).13C NMR (D2O): δ 159.12 (CdO), 104.64
(C-1â), 103.90 (C-1â), 103.22 (C-1â), 101.25 (C-1R), 101.11 (C-1R),
80.67, 80.58, 78.70, 78.30, 76.11, 75.91, 75.75, 75.19, 73.78, 72.97,
72.78, 72.20 (CH2), 71.91, 71.86, 71.51 (CH2), 70.02, 69.91, 69.55,
69.40, 65.51 (CH2), 65.40 (CH2), 61.52 (CH2), 61.30 (CH2), 61.06
(CH2), 60.84 (CH2), 58.14 (CH3), 58.06 (CH3), 38.72 (CH2), 29.89
(CH2). Electrospray ionization MS for C41H72N2O31Na: calcd 1111.4016,
found 1111.4002.
1
′
′
+84.5° (c 0.26; H2O). H NMR (D2O): δ 4.98 (d, 2 H, J1 ,2 3.8 Hz,
′
′
′ ′
H-1′), 4.42 (d, 2 H, J1,2 7.7 Hz, H-1), 4.35 (t, 2 H, J5 ,6 a ≈ J5 ,6 b ) 6.4
Hz, H-5′), 4.25-4.18 (m, 4 H, CH2O), 4.05 (d, 2 H, J4,3 3.1 Hz, H-4),
′
′
4.04 (d, 2 H, J4 ,3 3.3 Hz, H-4′), 4.05-4.02 (m, 2 H, CH2O), 3.94-
3.89 (m, 6 H, H-6a, H-3′, CH2O), 3.86-3.83 (m, 4 H, H-6a, H-2′),
3.77 (dd, 2 H, J3,2 10.1 Hz, H-3), 3.75-3.69 (m, 6 H, H-5, H-6′a, H-6′b),
3.59 (s, 6 H, OMe), 3.48 (dd, 2 H, H-2), 3.26 (broad s, 4 H, CH2N).
13C NMR (D2O): δ 159.23 (CdO), 104.62 (C-1â), 101.09 (C-1R),
80.61, 78.28, 75.74, 72.76, 71.86, 71.44 (CH2), 70.02, 69.90, 69.54,
65.47 (CH2), 61.51 (CH2), 61.05 (CH2), 58.12 (CH3), 41.07 (CH2).
Electrospray ionization MS for C34H60N2O26Na: calcd 935.3332, found
935.3338.
N,N′-Bis{{methyl 4-O-[4-O-(r-D-galactopyranosyl)-â-D-galacto-
pyranosyl]-â-D-glucopyranoside}-2′-yloxyethyl}oxycarbonyl-propyl-
enediamine (9). The per-benzyl derivative 22 (83 mg, 28.9 µmol) was
hydrogenated in AcOH in the presence of Pd/C (10%, 20 mg) for 24
h. The catalyst was filtered out, and the supernatant was concentrated.
The residue was passed through a Sep-Pak (C-18) cartridge in 10%
MeOH to yield 9 (29.3 mg, 81%), [R]D +51.7° (c 0.2; H2O). 1H NMR
′′ ′′
′ ′
(D2O): δ 4.97 (d, 2 H, J1 ,2 3.7 Hz, H-1′′), 4.53 (d, 2 H, J1 ,2 7.8 Hz,
′′ ′′
′′ ′′
H-1′), 4.39 (d, 2 H, J1,2 8.0 Hz, H-1), 4.35 (t, 2 H, J5 ,6 a ≈ J5 ,6 b 6.6
Hz, H-5′′), 4.23-4.21 (m, 4 H, CH2O), 4.10-3.57 (m, 34 H, H-3, H-4,
H-5, H-6a, H-6b, H-3′, H-4′, H-5′, H-6′a, H-6′b, H-2′′, H-3′′, H-4′′,
N,N′-Bis{[methyl 4-O-(r-D-galactopyranosyl)-â-D-galactopyra-
noside-2-yloxy]ethyl}oxycarbonyl-propylenediamine (5). The title
compound (31.2 mg, 78%) was prepared from 19 (87 mg, 43.3 µmol)
as described for 4, [R]D +78.6° (c 0.3; H2O). 1H NMR (D2O): δ 4.97
′
′
H-6′′a, H-6′′b, CH2O), 3.58 (s, 6 H, Me), 3.43 (dd, 2 H, J2 ,3 10.1 Hz,
3
H-2′), 3.30 (m, 2 H, H-2), 3.18 (t, 4 H, J 6.7 Hz, CH2N), 1.72 (m, 2
H, CH2CH2N). 13C NMR (D2O): δ 159.11 (CdO), 103.89 (C-1â),
103.22 (C-1â), 101.24 (C-1R), 80.67, 78.69, 76.19, 75.90, 75.27, 73.78,
72.96, 71.91 (CH2), 70.00, 69.85, 69.39, 65.56 (CH2), 61.52 (CH2),
61.30 (CH2), 60.84 (CH2), 58.06 (CH3), 38.78 (CH2), 29.92 (CH2).
Electrospray ionization MS for C47H82N2O36Na: calcd 1273.4545, found
1273.4535.
′
′
(d, 2 H, J1 ,2 3.8 Hz, H-1′), 4.42 (d, 2 H, J1,2 7.9 Hz, H-1), 4.35 (t, 2
′
′
′ ′
H, J5 ,6 a ≈ J5 ,6 b 6.4 Hz, H-5′), 4.24-4.17 (m, 4 H, CH2O), 4.04 (d, 2
′
′
′
H, J4,3 3.1 Hz, H-4), 4.03 (d, 2 H, J4 ,3 3.3 Hz, H-4), 4.04-4.01 (m, 2
H, CH2O), 3.94-3.89 (m, 6 H, H-6a, H-3′, CH2O), 3.85-3.83 (m, 4
H, H-6a, H-2′), 3.76 (dd, 2 H, J3,2 10.3 Hz, H-3), 3.75-3.69 (m, 6 H,
H-5, H-6′a, H-6′b), 3.57 (s, 6 H, OMe), 3.38 (dd, 2 H, H-2), 3.18 (t, 4
H, J 6.6 Hz, CH2N), 175-170 (m, 2 H, NCH2CH2CH2N). 13C NMR
(D2O): δ 159.27 (CdO), 104.66 (C-1â), 101.13 (C-1R), 80.55, 78.27,
75.74, 72.75, 71.83, 71.49 (CH2), 69.99, 69.88, 69.53, 65.36 (CH2),
61.46 (CH2), 61.01 (CH2), 58.08 (CH3), 38.64 (CH2). Electrospray
ionization MS for C35H62N2O26Na: calcd 949.3488, found 949.3490.
N,N′-Bis{[methyl 4-O-(r-D-galactopyranosyl)-â-D-galactopyra-
noside-2-yloxy]ethyl}oxycarbonyl-(p-phenylenediamine) (6). The title
compound 6 (28.6 mg, 78%) was prepared from 20 (82.1 mg, 42 µmol)
as described for 4. Fractions with 20% MeOH were collected from
Sep-Pak chromatography, [R]D +84.4° (c 0.34; H2O). 1H NMR
Methyl 3,6-Di-O-benzyl-4-O-(2,3,4,6-tetra-O-benzyl-r-D-galac-
topyranosyl)-2-O-(2-methylsulfonyloxyethyl)-â-D-galactopyrano-
side (23). To a solution of 1018 (920 mg, 0.98 mmol) in Py (5 mL)
MsCl (1.5 equiv) was added. After the reaction was stirred for 2 h at
0-15 °C, it was quenched with water, and the mixture was taken up
into DCM, washed with water, and concentrated. Chromatography of
the residue on silica gel in hexane-ethyl acetate (8:2-6:4) gave 23
1
(976 mg, 98%), [R]D +39.1° (c 1.0; CHCl3). H NMR (CDCl3): δ
′
′
7.4-7.1 (m, 30 H, arom.), 5.03 (d, 1 H, J1 ,2 2.1 Hz, H-1′), 4.91 (d, 1
H, 2J 11.1 Hz, Bn), 4.88 (d, 1 H, 2J 11.4 Hz, Bn), 4.78 (d, 1 H, 2J 11.5
Hz, Bn), 4.67 (d, 1 H, 2J 11.8 Hz, Bn), 4.76 (s, 2 H, Bn), 4.54 (d, 1 H,
′
′
(D2O): δ 7.48 (s, 4H, arom.), 4.99 (d, 2 H, J1 , 4.0 Hz, H-1′), 4.81
2
2J 11.2 Hz, Bn), 4.52 (d, 1 H, 2J 12.7 Hz, Bn), 4.38 (dd, 1 H, J5 ,6 a 5.0
′
′
(dd, 2 H, J2,1 7.9 Hz, J2, 9.9 Hz, H-2), 4.61 (d, 2 H, H-1), 4.39 (t, 2
3
′
′
′
′
′
′
Hz, J5 ,6 b 8.6 Hz, H-5′), 4.29-4.20 (m, 4 H, Bn, CH2OMs), 4.14 (d, 1
H, J1,2 7.5 Hz, H-1), 4.14 (s, 2 H, Bn), 4.09 (m, 3 H, H-4′, H-2′, H-3′),
4.04-3.91 (m, 4 H, H-4, H-6a, CH2), 3.56-3.44 (m, 4 H, H-2, H-5,
H-6b, H-6′a), 3.52 (s, 3 H, Me), 3.31 (dd, 1 H, J3,4 2.9 Hz, J3,2 9.9 Hz,
H, J5 , 6 a ≈ J5 , 6 b 6.4 Hz, H-5′), 4.11 (d, 2 H, J4, 3 2.9 Hz, H-4), 4.06 (d,
′
′
2 H, J4 , 3 3.5 Hz, H-4′), 3.99-3.97 (m, 4 H, H-3, H-3′), 3.95 (dd, 2 H,
J6a,5 7.3 Hz, J6a,6b 11.5 Hz, H-6a), 3.89-3.85 (m, 4 H, H-6b, H-2′),
′
′
′
′
3.82 (broad t, H-5), 3.72 (dd, 2 H, J6 a,5 6.0 Hz, J6 a,6 b 11.5 Hz, H-6′a),
3.68 (dd, 2 H, J6 b,5 6.9 Hz, H-6′b), 3.56 (s, 6 H, OMe).13C NMR
(D2O): δ 156.39 (CdO), 134.61 (arom.), 122.01 (arom.), 103.02 (C-
1â), 101.47 (C-1R), 78.73, 76.10, 74.27, 71.82, 71.76, 71.74, 71.72,
71.67, 70.07, 70.01, 70.00, 69.92, 69.65, 61.47 (CH2), 61.08 (CH2),
58.18 (CH3). Electrospray ionization MS for C34H52N2O24Na: calcd
895.2807, found 895.2813.
′
′
′
′
H-3), 3.26 (dd, 1 H, J6 b,6 a 8.4 Hz, H-6′b), 2.88 (s, 3 H, Ms). Anal.
Calcd. for C58H66O14S: C, 68.35; H, 6.53; S, 3.15. Found: C, 68.30;
H, 6.59; S, 3.30.
Methyl 3,6-Di-O-benzyl-4-O-(2,3,4,6-tetra-O-benzyl-r-D-galac-
topyranosyl)-2-O-(2-azidoethyl)-â-D-galactopyranoside (24). A sus-
pension of 23 (510 mg, 0.5 mmol), NaN3 (163 mg, 2.5 mmol), and
cat. Bu4NI (10 mg) in dry DMF (5 mL) was stirred for 3 h at 50 °C.
The mixture was diluted with ethyl acetate, washed with brine, and
N-{[Methyl 4-O-(r-D-galactopyranosyl)-â-D-galactopyranoside-
2-yloxy]ethyl}oxycarbonyl-N′-{{methyl 4-O-[4-O-(R-D-galactopy-
9
J. AM. CHEM. SOC. VOL. 125, NO. 11, 2003 3293

A R T I C L E S
Kitov et al.
concentrated. Chromatography on silica gel with hexanes-ethyl acetate
3,4-Di-{[methyl 4-O-(r-D-galactopyranosyl)-â-D-galactopyrano-
side-2-yloxy]ethylamino}-3-cyclobutene-1,2-dion (7). To a solution
of 25 (38 mg, 82 µmol) and 3,4-diethoxy-3-cyclobutene-1,2-dion (6.8
mg, 40 µmol) in MeOH Et3N (∼60 µL) was added. The mixture was
stirred for 24 h, then concentrated, and chromatographed on silica gel
with DCM-MeOH-water (8:2:0-500:500:5) to give title compound
1
(8:1) gave azid 24 (393 mg, 81%), [R]D +32.4° (c 0.4; CHCl3). H
′
′
NMR (CDCl3): δ 7.4-7.1 (m, 30 H, arom.), 5.02 (d, 1 H, J1 ,2 1 Hz,
2
2
H-1′), 4.90 (d, 1 H, J 11.2 Hz, Bn), 4.86 (d, 1 H, J 11.4 Hz, Bn),
2
2
4.78 (d, 1 H, J 11.5 Hz, Bn), 4.76 (s, 2 H, Bn), 4.67 (d, 1 H, J 11.8
Hz, Bn), 4.55 (d, 1 H, J 11.2 Hz, Bn), 4.54 (d, 1 H, J 12.7 Hz, Bn),
4.39 (dd, 1 H, J5 ,6 a 5.0 Hz, J5 ,6 b 8.6 Hz, H-5′), 4.26 (d, 1 H, J 11.9
Hz, Bn), 4.22 (d, 1 H, 2J 11.9 Hz, Bn), 4.16 (d, 1 H, J1,27.5 Hz, H-1),
4.13 (s, 2 H, Bn), 4.09 (m, 3 H, H-4′, H-2′, H-3′), 4.00-3.81 (m, 6 H,
H-4, H-6a, CH2), 3.56-3.44 (m, 3 H, H-2, H-5, H-6′a), 3.54 (s, 3 H,
2
2
2
1
′
′
′ ′
(35 mg, 85%), [R]D +82.1° (c 0.6; H2O). H NMR (D2O): δ 4.98 (d,
′
′
2 H, J1 ,2 3.8 Hz, H-1′), 4.39 (d, 2 H, J1,2 7.7 Hz, H-1), 4.33 (t, 2 H,
′
′
′ ′
J5 ,6 a ≈ J5 ,6 b ) 6.6 Hz, H-5′), 4.04-4.10 (m, 2 H, H-4, H-4′), 3.99-
3.70 (m, H, H-3, H-5, H-6a, H-6b, H-2′, H-3′, H-6′a, H-6′b, CH2),
3.54 (s, 6 H, OMe), 3.36 (dd, 2 H, J2,3 10.0 Hz, H-2). ), 13C NMR
(D2O): δ 181.93 (CdC), 168.75 (CdO), 104.00 (C-1â), 100.48 (C-
1R), 79.87, 77.80, 75.11, 72.34, 71.79 (CH2), 71.31, 69.39, 69.31, 68.91,
60.95 (CH2), 60.43 (CH2), 57.35 (CH3), 44.50 (CH2). Electrospray
ionization MS for C34H57N2O24: calcd 877.3301, found 877.3303.
′
′
Me), 3.34-3.28 (m, 2 H, H-3, H-6b), 3.26 (dd, 1 H, J6 b,6 a 8.4 Hz,
5 ,6 b 4.8 Hz, H-6′b). Anal. Calcd. for C57H63N3O11: C, 70.86; H, 6.57;
′
′
J
N, 4.35. Found: C, 71.04; H, 6.71; N, 4.35.
Methyl 4-O-(r-D-Galactopyranosyl)-2-O-(2-aminoethyl)-â-D-ga-
lactopyranoside Acetic Acid Salt (25). Hydrogen gas was passed
through a suspension of 24 (363 mg, 0.378 mmol) and Pd(OH)2 (100
mg) in acetic acid (5 mL) for 24 h. The catalyst was filtered out, and
the solution was concentrated to give yellow oil 25 (159 mg, 91%).1H
Acknowledgment. Authors are grateful to Ms. Joanna Sad-
owska for performing solid-phase assays. The studies were
supported by a grant to D.R.B. from the Canadian Bacterial
Diseases Network (CBDN) and by grants to S.W.H. from
BBSRC (Grant Number 24/B12993) and The Wellcome Trust
(Grant Number GA056981FR).
′
′
NMR (D2O): δ 4.98 (d, 1 H, J1 ,2 3.7 Hz, H-1′), 4.48 (d, 1 H, J1,2 7.7
′
′
′ ′
Hz, H-1), 4.34 (t, 1 H, J5 ,6 a ) J5 ,6 b 6.1 Hz, H-5′), 4.10-4.01 (m, 9 H,
H-4, H-4′, CH2O), 3.97-3.70 (m, H, H-3, H-5, H-6a, H-6b, H-2′, H-3′,
H-6′a, H-6′b, CH2), 3.54 (s, 3 H, Me), 3.37 (dd, 1 H, J2,3 7.8 Hz, H-2),
3.22 (t, 2 H, 3J 4.8 Hz, CH2N). Electrospray ionization MS for C15H30-
NO11: calcd 400.1819, found 400.1820
JA0258529
9
3294 J. AM. CHEM. SOC. VOL. 125, NO. 11, 2003