Constrained H-Type 2 Blood Group Trisaccharide Synthesized in a Bioactive Conformation via Intramolecular Glycosylation

Article abstract of DOI:10.1021/jo990979a

The methyl glycoside of the H-type 2 trisaccharide 1 was synthesized in a constrained, bioactive conformation via intramolecular aglycon delivery. Computer modeling of the crystal structure of the Ulex europaeus I lectin with a docked H-type 2 trisaccharide suggested that the disaccharide Galp(1→4)GlcpNAcl→OCH₃ could be tethered in a bioactive conformation if Gal O-6 and GlcNAc O-3 are linked via a three-carbon tether. The ethyl 1-thiogalactopyranoside 13 was used to alkylate the methyl 2-acetamido-2-deoxy glucopyranoside 7, and the resulting dimer was subjected to intramolecular glycosylation following protecting group manipulation. The tethered disaccharide 4 was glycosylated by the activated fucopyranosyl donor 3 to give the protected target molecule 17. Solid-phase binding assays showed that the tethered trisaccharide 2 was 3-fold less active than native H-type 2 trisaccharide 1 when assayed against the U. europaeus I lectin, whereas it was 250 times less active when assayed with the Psophocarpus tetragonolobus II lectin. The observed activities are consistent with published models for H-trisaccharide interactions with Ulex and Psophocarpus lectins and provide further evidence that suggests reduction of oligosaccharide flexibility by intramolecular tethering provides no significant gain in binding energy.

Full text of DOI:10.1021/jo990979a

9080
J . Org. Chem. 1999, 64, 9080-9089
Con str a in ed H-Typ e 2 Blood Gr ou p Tr isa cch a r id e Syn th esized in a
Bioa ctive Con for m a tion via In tr a m olecu la r Glycosyla tion
Shirley A. Wacowich-Sgarbi and David R. Bundle*
Department of Chemistry, The University of Alberta, Edmonton, AB T6G 2G2, Canada
Received J une 18, 1999
The methyl glycoside of the H-type 2 trisaccharide 1 was synthesized in a constrained, bioactive
conformation via intramolecular aglycon delivery. Computer modeling of the crystal structure of
the Ulex europaeus I lectin with a docked H-type 2 trisaccharide suggested that the disaccharide
Galp(1f4)GlcpNAc1fOCH3 could be tethered in a bioactive conformation if Gal O-6 and GlcNAc
O-3 are linked via a three-carbon tether. The ethyl 1-thiogalactopyranoside 13 was used to alkylate
the methyl 2-acetamido-2-deoxy glucopyranoside 7, and the resulting dimer was subjected to
intramolecular glycosylation following protecting group manipulation. The tethered disaccharide
4 was glycosylated by the activated fucopyranosyl donor 3 to give the protected target molecule
17. Solid-phase binding assays showed that the tethered trisaccharide 2 was 3-fold less active than
native H-type 2 trisaccharide 1 when assayed against the U. europaeus I lectin, whereas it was
250 times less active when assayed with the Psophocarpus tetragonolobus II lectin. The observed
activities are consistent with published models for H-trisaccharide interactions with Ulex and
Psophocarpus lectins and provide further evidence that suggests reduction of oligosaccharide
flexibility by intramolecular tethering provides no significant gain in binding energy.
In tr od u ction
tetragonolobus II (winged bean lectin),3 Galactia tenui-
flora,4 and Erythrina corallodendron.5 Each lectin em-
ploys a distinct mode of binding and recognizes different
topological features of the sugar epitope.6,7
In the search for carbohydrate-based agonists,1 the
moderate free energy of sugar-protein interactions
repeatedly confounds the discovery of high affinity car-
bohydrate ligands. In this study a constrained analogue
of the human blood group trisaccharide 1 was synthesized
by intramolecular glycosylation to yield the preordered
trisaccharide 2, which by reducing entropic losses might
The intrinsic affinity for oligosaccharide binding by
proteins, such as antibodies and lectins, is generally
characterized by association constants within the range
103-106 M-1. For many lectin-oligosaccharide interac-
tions, the modest affinity and free energy of interaction
are characterized by a strong enthalpic contribution offset
by unfavorable entropy.8-13 It has been suggested that
oligosaccharide conformational flexibility arising from
motion about each interresidue glycosidic bond may
account for most of the unfavorable entropy.14 Some
theoretical estimates place the entropic loss per im-
mobilized rotamer as high as 0.6-2.0 kcal mol-1 14,15
.
Attempts to verify this hypothesis by synthesizing oli-
(2) Allen, H. J .; J ohnson, A. A. Carbohydr. Res. 1977, 58, 253-265.
Frost, R. G.; Reitherman, R. W.; Miller, A. L.; O’Brien, J . S. Anal.
Biochem. 1975, 69, 170-179.
(3) Khan, M. I.; Sastry, M. V. K.; Surolia, A. J . Biol. Chem. 1986,
261, 3013-3018.
(4) Le Pendu, J .; Gerard, G.; Lambert, F.; Mollicone, R.; Oriol, R.
Glycoconjugate J . 1986, 3, 203-216.
(5) Gilboa-Garber, N.; Mizrahi, L. Can. J . Biochem. 1981, 59, 315-
320.
(6) Du, M.-H.; Spohr, U.; Lemieux, R. U. Glycoconjugate J . 1994,
11, 443-461.
(7) Adar, R.; Sharon, N. Eur. J . Biochem. 1996, 239, 668-674.
(8) Lemieux, R. U.; Delbaere, L. T. J .; Beierbeck, H.; Spohr, U. In
Host-Guest Molecular Interactions: From Chemistry to Biology, Ciba
Foundation Symposium; Wiley: Chichester, 1991; Vol. 158, pp 231-
248.
be expected to show higher affinity. The H-type 2 epitope
is a good model to probe the influence of oligosaccharide
flexibility on the thermodynamics of oligosaccharide-
lectin interactions, because the epitope is recognized by
four different lectins, Ulex europaeus I,2 Psophocarpus
(9) Lemieux, R. U. Acc. Chem. Res. 1996, 29, 373-380.
(10) Carver, J . P.; Michnick, S. W.; Imberty, A.; Cumming, D. A. In
Carbohydrate Recognition in Cellular Function, Ciba Foundation
Symposium; Wiley: Chichester, 1989; Vol. 145, pp 6-26.
(11) Toone, E. J . Curr. Opin. Struct. Biol. 1994, 4, 719-728.
(12) Lis, L.; Sharon, N. Chem. Rev. 1998, 98, 637-674.
(13) Spohr, U.; Hindsgaul, O.; Lemieux, R. U. Can. J . Chem. 1985,
63, 2644-2652.
* Phone: 403-492-8808. Fax: 403-492-7705. Email: dave.bundle@
ualberta.ca.
(1) Simanek, E. E.; McGarvey, G. J .; J ablonowski, J . A.; Wong, C.-
H. Chem. Rev. 1998, 98, 833-862. Wong, C.-H.; Moris-Varas, F.; Hung,
S.-C.; Marron, T. G.; Lin, C.-C.; Gong, K. W.; Weitz-Schmidt, G. J . Am.
Chem. Soc. 1997, 119, 8152-8158. Kolb, H. C.; Ernst, B. Chem. Eur.
J . 1997, 3, 1571-1578.
(14) Carver, J . P. Pure Appl. Chem. 1993, 65, 763.
(15) Searle, M. S.; Williams, D. H. J . Am. Chem. Soc. 1992, 114,
10690-10697.
10.1021/jo990979a CCC: $18.00 © 1999 American Chemical Society
Published on Web 11/20/1999

Constrained H-Type 2 Blood Group Trisaccharide
J . Org. Chem., Vol. 64, No. 25, 1999 9081
gosaccharides that are constrained in a conformation
preorganized for binding have failed to show substan-
tially higher free energy of binding.16-21 In two well
studied cases,17,18 the change in entropy was shown to
be alternatively favorable18 and unfavorable17 but not
decisive. Testing this hypothesis poses significant prob-
lems. The smaller of the two estimates for the energetic
penalty associated with freezing rotamers seems to be
the most reasonable figure and is based on freezing a
bond that is free to equilibrate between three rotamers
of equal energy. However, for the glycosidic bond, the exo-
anomeric effect skews the rotameric population between
two and not three rotamers. Consequently it seems
unlikely that the entropy associated with this process
would exceed 0.1-0.3 kcal mol-1 at 37 °C. Introduction
of a tether reduces the range of allowed glycosidic
torsional angles æ/ψ, but the tether itself is flexible as it
forms part of a macrocyclic ring, which even for short
tethers spans at least 10-13 bonds.17,18 Recent work by
Mammen et al.22 suggests that previous estimates of
torsional entropy are too high. Taken together these
considerations predict that gains in the free energy of
association from tethering oligosaccharides will be small.
The suggestion that glycosidic torsions are the major
contributor to loss of conformational entropy on complex-
ation remains an unproven hypothesis. An alternative
interpretation of the thermodynamic data emphasizes
that the exo-anomeric effect imposes conformational
preferences on glycosidic linkages and stresses the
importance of reordering solvent water about polyam-
phiphilic surfaces.8,9
U. europaeus I,2 the primary lectin target of this
synthetic study, binds to the H-type 2 human blood group
determinant R-L-Fuc(1f2)-â-D-Gal(1f4)-â-D-GlcNAcOR.6
The lectin is used clinically in the identification of blood
group “O” individuals because it agglutinates human red
blood cells that display the H-type 2 determinant as part
of cell membrane glycolipids or glycoproteins. The crystal
structure of the H-type 2 oligosaccharide-specific lectin
from U. europaeus without bound saccharide has been
solved.23 Thermodynamic data from van’t Hoff plots for
the binding of H-type 2 by Ulex lectin show large
favorable enthalpy opposed by large unfavorable entropy
(∆G° ) -8 kcal/mol, ∆H° ) -29 kcal/mol, T∆S° ) -21
kcal/mol).26 The size of the entropic penalty suggests that
a conformational change may accompany the binding of
the sugar.14,30-34
To study the magnitude and origin of the entropic
penalty, we have synthesized a H-type 2 trisaccharide
with a tether that bridges the Gal and GlcNAc monosac-
charides. Although this tethered saccharide 2 has been
designed primarily for binding with U. europaeus I lectin,
the bioactivity of 1 and 2 are also determined with P.
tetragonolobus II lectin (winged bean lectin),3,35 another
H-type 2-specific lectin. The activities will also eventually
be measured for G. tenuiflora4 and E. corallodendron,5
lectins. In the search for affinity gains from preorgani-
zation of oligosaccharide epitopes, the H-type 2 epitope
thus maximizes the probability for observing an improve-
ment in ∆G° due to a decrease in the entropy penalty.
The work has general and practical relevance to the
search for high affinity carbohydrate-based agonists.
Preorganization of an oligosaccharide epitope in its
bound conformation poses several difficulties, especially
in relation to the ability of the tether to constrain the
oligosaccharide in a bioactive conformation. By selecting
an oligosaccharide such as the human blood group H-type
2 epitope for such studies, it is possible to access a range
of distinct lectin binding sites and increase the potential
for identifying a recognition system in which entropic
gains may be observed. Furthermore, the crystal struc-
tures of two lectins that bind H-antigen have been
solved,23,24 and with the exception of E. corallodendron
lectin,25 all of the lectin sites have been extensively
mapped by epitope congeners,6,26-29 thereby facilitating
the design of tethered oligosaccharides.
Resu lts
The design of a tethered H-type 2 trisaccharide is based
on a crystal structure of the Ulex lectin23 and published
studies of congener binding to three lectins, U. europaeus
I, G. tenuiflora, and P. tetragonobolus II.6,26-29 Although
there is currently no reported structure for the complex
with oligosaccharide, the Ulex lectin cocrystallizes with
a molecule of 2-methyl-2,4-pentanediol, the precipitant
used for protein crystallization, and this complex provides
a reference point for saccharide docking. Inhibition data
for a series of monodeoxy, monomethyl, and monodeoxy-
fluoro analogues of the H-type 2 trisaccharide have
identified the functional groups essential for binding.6,26-29
In the case of U. europaeus I, the O-2, O-3, and O-4
hydroxyl groups of fucose were essential. It was further
concluded that these hydroxyls must be buried within
the binding site of Ulex, while the hydroxyl group at the
O-3 position of galactose is hydrogen bonded to the lectin
at the periphery of the combining site. A pentose conge-
ner of H-type 2 in which the methyl group of fucose was
replaced by hydrogen was also found to have a signifi-
(16) Wilstermann, M.; Balogh, J .; Magnusson, G. J . Org. Chem.
1997, 62, 3659-3665.
(17) Alibes, R.; Bundle, D. R. J . Org. Chem. 1998, 63, 6288-6301.
Bundle, D. R.; Alibes, R.; Nilar, S.; Otter, A.; Warwas, M.; Zhang, P.
J . Am. Chem. Soc. 1998, 120, 5317-5318.
(18) Navarre, N.; van Oijen, A. H.; Boons, G. J . Tetrahedron Lett.
1997, 38, 2023. Navarre, N.; Amiot, N.; van Oijen, A. H.; Imberty, A.;
Poveda, A.; J imenez-Barbero, J .; Cooper, A.; Nutley, M. A.; Boons, G.
J . Chem. Eur. J . 1999, 5, 2281-2294.
(19) Kabat, E. A.; Liao, J .; Burzynska, M. H.; Wong, T. C.; Thøger-
son, H.; Lemieux, R. U. Immunol. 1981, 18, 873-881.
(20) Lindh, I.; Hindsgaul, O. J . Am. Chem. Soc. 1991, 113, 216-
223.
(21) Kolb, H. C. Bioorg. Med. Chem. 1997, 7, 2629-2634.
(22) Mammen, M.; Shakhnovich, E. I.; Whitesides, G. M. J . Org.
Chem. 1998, 63, 3168-3175.
(29) Lemieux, R. U.; Du, M.-H.; Spohr, U.; Acharya, S.; Surolia, A.
Can. J . Chem. 1994, 72, 158-163.
(30) Bush, A. C. Curr. Opin. Struct. Biol. 1992, 2, 655.
(31) Homans, S. W. In Molecular Glycobiology; Fukuda, M., Hinds-
gaul, O., Eds.; Oxford University Press: Oxford, 1994; p 230.
(32) Homans, S. W.; Rutheford, T. Biochem. Soc. Trans. 1993, 21,
449.
(33) Imberty, A.; Bourne, Y.; Cambillau, C.; Rouge, P.; Perez, S. Adv.
Biophys. Chem. 1993, 3, 61.
(34) Rice, K. G.; Wu, P.; Brand, L.; Lee, Y. C. Curr. Opin. Struct.
Biol. 1993, 3, 669.
(23) Vandonselaar, M.; Delbaere, L. T. J . J . Mol. Biol. 1994, 243,
345-346.
(24) Shaanan, B.; Lis, H.; Sharon, N. Science 1991, 254, 862-866.
(25) Surolia, A.; Sharon, N.; Schwarz, F. P. J . Biol. Chem. 1996,
271, 17697-17703.
(26) Hindsgaul, O.; Khane, D. P.; Bach, M.; Lemieux, R. U. Can. J .
Chem. 1985, 63, 2653-2658.
(27) Spohr, U.; Paszkiewicz-Hnatiw, E.; Morishima, N.; Lemieux,
R. U. Can. J . Chem. 1992, 70, 254-271.
(35) Patanjali, S. R.; Sajjan, S. U.; Surolia, A. Biochem. J . 1988,
252, 625-631: Schwarz, F. P.; Puri, K.; Surolia, A. J . Biol. Chem. 1991,
266, 24344-24350.
(28) Cromer, R.; Spohr, U.; Khare, D. P.; LePendu, J .; Lemieux, R.
U. Can. J . Chem. 1992, 70, 1511-1530.

9082 J . Org. Chem., Vol. 64, No. 25, 1999
Wacowich-Sgarbi and Bundle
Com p u ter Mod elin g. The model of the tethered
H-type 2 Ulex complex was developed in several stages.
The coordinates for the Ulex structure with bound
2-methyl-2,4-pentanediol provided the starting point for
positioning the ligand in the lectin site. Following an
unpublished procedure of Lemieux (personal communica-
tion), coordinates for R-L-fucopyranose were used to dock
fucose in a manner such that the O-3 and O-4 atoms of
R-L-fucopyranose were overlaid on the two hydroxyl
groups of the pentanediol. Because epitope mapping of
the H-type 2 ligand established that O-3 and O-4 of fucose
are essential to binding6 and possess a synclinal relation-
ship, this mode of docking seemed most consistent with
the binding motifs of several published saccharide-lectin
complexes.12 Having docked the monosaccharide residue,
the coordinates of this sugar provided a key feature onto
which the more complex trisaccharide 1 and 2 could be
overlaid.
The conformation of trisaccharide 1 was modeled using
the GEGOP force field,36 and the global minimum energy
conformation was selected. Because most oligosaccha-
ride-protein complexes reveal bound oligosaccharides in
conformations close to their energy minima, the fucose
residue of trisaccharide 1 in its global energy conformer
was overlaid on the fucose residue that was docked in
the pentanediol binding site. This gave a model of
trisaccharide 1 docked with lectin (Figure 2). The ap-
proach was subsequently justified as transferred NOE
studies (Milton and Bundle, unpublished results) are
consistent with a bound H trisaccharide epitope in this
low energy conformation.
Trisaccharide 1 in its low energy conformation was
investigated for tethering. Molecular modeling revealed
that a three-carbon linker between Gal O-6 and GlcNAc
O-3 would create a tethered H-type structure 2, which
is effectively constrained to a narrow range of conformers
that closely approximate the lowest energy forms of 1.
However, the tethered structure retains sufficient flex-
ibility to adopt several low energy conformations that are
similar to those of bound trisaccharide 1. To confirm that
the bound form of the tethered H-trisaccharide 2 could
adopt realistic conformations while avoiding tether-
protein contacts, energy minimizations for 2 were per-
formed with the CVFF force field of the Discover software
and then the minimized structure was docked to give a
lectin-trisaccharide 2 complex. The docked complex was
created by overlaying 2 on the bound form of the H-type
2 trisaccharide 1.
The rationale for assuming that the pentanediol oc-
cupied the saccharide binding site of the Ulex lectin is
based on several factors. Several crystal structures of
legume lectins and their substrates show large sequence
homologies including certain invariant amino acids that
participate in binding site hydrogen bonding and hydro-
phobic interactions and others that are involved in the
coordination of the calcium and manganese ions.8,9,12,37
From the crystal structure of the Ulex lectin, it is found
that the amino acids usually involved in coordinating the
metal ions, asparagine and aspartic acid, are present and
are in close vicinity to the ions.23 The pentanediol binding
site contains the amino acids asparagine, aspartic acid,
F igu r e 1. Conclusions of epitope mapping studies for H-type
2 trisaccharide with 3 lectins: (a) Ulex europaeus I, (b)
Psophocarpus tetragonolobus II (winged bean lectin), and (c)
Galactia tenuiflora. Three types of interactions are noted: (i)
buried hydroxyl groups involved in hydrogen bonds deep
within the binding site are identified within a square, (ii)
essential hydrophobic interactions are identified by a triangle,
and (iii) hydroxyl groups implicated in hydrogen bonds at the
periphery of the binding site are identified by a circle.
cantly decreased activity, indicating that the methyl
group fulfills an important hydrophobic interaction with
the lectin.26 The results of binding site mapping with
trisaccharide congeners for the three H-type 2-specific
lectins, U. europaeus I, G. tenuiflora, and P. tetragonolo-
bus II are summarized (Figure 1).
Design of Con str a in ed H-typ e 2 Der iva tives. To
design tethered derivatives of H-type 2 trisaccharide
preorganized in its bound conformation, the following
criteria had to be fulfilled:
• The tether should not be attached to key polar groups
that are involved in protein-carbohydrate interactions.
• The conformation of the bioactive epitope should be
well approximated by tethering.
• When bound, the tether should not interfere sterically
with the surface of the protein.
To ensure that the tether met the first requirement,
reference was made to the epitope mapping studies6,26-29
summarized in Figure 1.
Because the principle lectin of interest is the Ulex
lectin, the binding topography of the H-type 2 epitope
indicated that the O-3 position of N-acetylglucosamine
and the O-6 position of galactose would be suitable sites
for tethering; neither site was involved in obligatory
sugar-protein interactions and both occupied solvent
exposed position in the complex (compound 2). Tethers
at these positions could interfere with the protein surface
of the winged bean lectin and very probably would cause
problems for Galactia, which requires O-3 of GlcNAc for
a buried hydrogen bond.6
(36) Stuike-Prill, R.; Meyer, B. Eur. J . Biochem. 1990, 194, 903-
919.
(37) Delbaere, L. T. J .; Vandonselaar, M.; Prasad, L.; Quail, J . W.;
Nikrad, P. V.; Pearlstone, J . R.; Carpenter, M. R.; Smillie, L. B.; Spohr,
U.; Lemieux, R. U. Can. J . Chem. 1990, 68, 1116-1121.

Constrained H-Type 2 Blood Group Trisaccharide
J . Org. Chem., Vol. 64, No. 25, 1999 9083
F igu r e 2. Computer modeling: stereo plot of the binding site amino acids (mid gray lines) of the Ulex lectin cocrystallized with
a molecule of (R)-2,4-dihydroxy-2-methylpentane (light gray lines) (Delbaere et al., unpublished data). The 3- and 4-hydroxyl
groups of the fucose residue of H-type 2 trisaccharide 1 (black lines) are superimposed over those of the dihydroxymethylpentane.
and tyrosine, those often involved in buried carbohydrate-
lectin hydrogen bonding and hydrophobic interactions.
In this context, U. europaeus I fits the general binding
site motif of legume lectins.12,37 In comparing the X-ray
structure of Ulex with that of other legume lectins, the
location of the 5-, 6-, and 7-stranded â-sheets and the
metal ions are superimposeable. In addition, the binding
site of 2-methyl-2,4-pentanediol is found to overlap with
the binding site of the carbohydrate substrates.
Syn th esis. Str a tegy. Thioglycosides are “versatile”
glycosyl donors because they may be activated in various
ways or readily converted into other glycosyl donors.38
They also allow a number of protective group manipula-
tions. Therefore, a thioglycoside was the donor of choice
for the glycosylation reactions. Retrosynthetic analysis
(Figure 3) suggested tethering the ethyl thiogalactoside
5 and the 2-acetamido-2-deoxy-glycopyranoside 7 via a
1,3-propanediol. The order of assembly of these three
components may vary. Glycosylation of the two sugars
followed by tethering of the disaccharide may be envi-
sioned. A second route to the tethered disaccharide 4
would involve initial tethering of the two sugars with the
1,3-propanediol linker, followed by an intramolecular
glycosylation. The second method of assembly was the
method of choice because it has been demonstrated that
intramolecular glycosylations generally proceed in high
yield,18,39-42 whereas the tethering of preassembled oli-
gosaccharides and formation of large rings are low yield
processes.17 Computer modeling showed that even with-
F igu r e 3. Retrosynthesis of the tethered H-type 2 trisaccha-
ride 2.
out neighboring group participation a tethered structure
would favor the delivery of the galactopyranosyl group
leading to â-glycoside formation in 4. This preference may
be attributed to a combination of ring strain and a steric
mismatch between 6-O-benzyl and 2′-O-p-methoxybenzyl
groups. Other groups have reported that “prearranged”
molecules favor the formation of one glycosidic form,
depending on the relative configuration of the tethered
donor and acceptor.18,39
(38) Garegg, P. J . Adv. Carbohydr. Chem. Biochem. 1997, 52, 179-
205.
A number of strategies were considered for tethering
two monosaccharides prior to intramolecular glycosyla-
tion. Monosaccharide 3-hydroxypropyl ethers could be
employed to displace sulfonate esters from the second
monosaccharide residue. However, as depicted in routes
1 and 2 (Figure 4) such reactions would most likely lead
to elimination products, because even primary sulfonates
(39) Ziegler, T.; Ritter, A.; Hurttlen, J . Tetrahedron Lett. 1997, 38,
3715-3718. Ziegler, T.; Lau, R. Tetrahedron Lett. 1995, 36, 1417-
1420.
(40) Hindsgaul, O.; Bols, M. Tetrahedron Lett. 1996, 37, 4211-4212.
Barresi, F.; Hindsgaul, O. Synlett 1992, 759-761. Hansen, H. C.;
Barresi, F.; Hindsgaul, O. J . Am. Chem. Soc. 1991, 113, 9376-9377.
(41) Stork, G.; La Clair, J . J . J . Am. Chem. Soc. 1996, 118, 247-
248. Stork, G.; Kim, G. J . Am. Chem. Soc. 1992, 114, 1087-1088.
(42) Ito, Y.; Ogawa, T. J . Am. Chem. Soc. 1997, 119, 5562-5566.

9084 J . Org. Chem., Vol. 64, No. 25, 1999
Wacowich-Sgarbi and Bundle
F igu r e 4. Potential problems in tethering the galactose and N-acetylglucosamine monosaccharides via displacement reactions.
of galactose are notoriously resistant to displacement
reactions. At a secondary ring carbon, displacement is
even more disfavored and complicated by inversion of
configuration. Consequently, the tethered molecule 14
was approached by sequential substitution reactions
using a monosaccharide alkoxide as the nucleophile, with
the sulfonate leaving group attached to the tether moiety
(Schemes 2 and 3).
Subsequent steps involved the selective deprotection
of the O-2 position of galactose, followed by glycosylation
with fucopyranosyl donor 3 and deprotection, would then
lead to the desired tethered trisaccharide 2.
sodium cyanoborohydride and hydrochloric acid in ether
and THF gave compound 15. A poor yield was obtained
from the sodium cyanoborohydride reaction as a result
of the acid sensitivity of the p-methoxybenzyl group and
the ability of the acetonide to undergo reductive ring
opening. Intermediate 15, possessing both a donor site
and an acceptor site, is set up to undergo an intramo-
lecular glycosylation. Its structure, as well as that of the
acetylated derivative, was verified by one-dimensional 1H
NMR, two-dimensional GCOSY, and HMQC experi-
ments.
Two methods of activation were tried for the intramo-
lecular glycosylation of 15. The first method used NIS
and silver triflate and gave a yield of 35%. However, this
was accompanied by partial iodination of the p-methoxy-
benzyl group, a fatal side reaction because the iodinated
p-methoxybenzyl group was found to be stable to DDQ
oxidation. The second method of glycosylation using
methyltriflate and 2,6-di-tert-butyl-4-methylpyridine (DT-
BMP) resulted in a higher yield of 64%, and a â/R ratio
of 8:1 was observed.
P r ep a r a tion of th e Lin k er . The linker synthon 8
was prepared in a one-pot synthesis. Treatment of 1,3-
propanediol with 1 equiv of sodium hydride in THF
(Scheme 1) followed by addition of tert-butyldiphenylsilyl
Sch em e 1a
Syn t h esis a n d Dep r ot ect ion of t h e Tet h er ed
Tr isa cch a r id e. Tethered disaccharide 16 was subjected
to selective deprotection with DDQ to give 4 in 68% yield
(Scheme 4), accompanied by partial loss of the isopropy-
lidene acetal. The disaccharide acceptor 4 was glycosy-
lated by the fucose thioglycoside 3, which was activated
by dimethyl(methylthio)sulfonium triflate (DMTST). This
gave the tethered trisaccharide 17 in 55% yield. The
protected tethered trisaccharide 17 was then deprotected
by hydrolysis of the acetal, followed by hydrogenolysis
of the benzyl groups to give the tethered H-type 2
trisaccharide 2 in 90% yield over two steps.
Biologica l Activity. The biological activity of the
tethered trisaccharide 2 relative to that of the native
trisaccharide 1 was determined by solid-phase immu-
noassay for the U. europaeus I and P. tetragonolobus II
lectins. Affinity purified lectin adsorbed to microtiter
plates was allowed to compete for biotin-labeled H-type
2 glycoconjugate in the presence and absence of the
inhibitor 2. In this assay format the affinity constant of
an intermediate affinity ligand is well approximated by
the inverse of the IC50 value.44 The tethered trisaccharide
2 was a strong inhibitor of glycoconjugate binding to the
Ulex lectin; however, the untethered H-type 2 ligand 1
exhibited binding that was 3 times higher (Figure 5a).
In sharp contrast, trisaccharide 1 was 250 times more
active than 2 when each was assayed for binding to P.
tetragonolobus lectin (Figure 5b).
a
(a) NaH, THF, TBDMSCl; (b) Et3N, MsCl.
chloride43 gave the silyl ether, which was not isolated but
reacted with methanesulfonyl chloride and triethylamine.
This gave the desired linker 8 in 81% yield. Preparation
of the linker in two distinct steps lead to a lower overall
yield.
P r ep a r a tion of th e Ga la ctose Don or . To synthesize
a galactose donor, with two different groups at the O-2
and O-6 position, mixed acetal 10 was used (Scheme 2).
This mixed acetal allowed the introduction of a p-
methoxybenzyl group at the O-2 position of galactose;
selective removal of the mixed acetal under mild condi-
tions of hydrolysis then gave monosaccharide 5. Reaction
of the monosaccharide alkoxide generated from 5 with
the linker 8 gave 11, from which the TBDMS protecting
group was removed with TBAF in THF. The resulting
alcohol 12 was treated with triethylamine and methane-
sulfonyl chloride to give the desired galactose donor 13.
Assem bly of th e Teth er ed Disa cch a r id e. Tether
coupling of the sugars was executed in dry THF. Once
the alkoxide of 7 was formed and after the addition of
methanesulfonate 13, a few drops of DMSO were added,
and the desired linked product 14 was obtained in 83%
yield (Scheme 3). DMSO facilitates a faster reaction
presumably by dissociation of the sodium alkoxide ion
pair of 7. Reductive opening of the benzylidene ring with
(43) McDougal, P. G.; Rico, J . G.; Oh, Y.-I.; Condon, B. D. J . Org.
Chem. 1986, 51, 3388-3390.
(44) Meikle, P. J .; Young, N. M.; Bundle, D. R. J . Immunol. Methods
1990, 132, 255-261.

Constrained H-Type 2 Blood Group Trisaccharide
J . Org. Chem., Vol. 64, No. 25, 1999 9085
Sch em e 2a
a
(a) (CH3)2C(OCH3)2, TsOH; (b) PMBCl, NaH, DMF; (c) H+; (d) NaH, TBDMSO(CH2)3OMs, THF; (e) TBAF, THF; (f) MsCl, Et3N.
Sch em e 3a
Ulex protein. Binding of the tethered target molecule 2
to Ulex lectin showed decreased affinity, ∆(∆G°) ) + 0.6
kcal mol-1 relative to H-type 2 trisaccharide 1. The
preservation of the binding energy suggests tethering has
provided a bioactive conformation for 2 and is consistent
with the molecular modeling and the topology of the
recognition surface proposed by Lemieux.6 In these
studies the possible involvement of the O-6′ hydroxyl
group in hydrogen bonding to solvent water at the
periphery of the Ulex binding site was proposed, and the
small loss of binding energy for the tethered trisaccharide
2 most likely arises from the functional group alterations
at this center. The strong inhibitory power of 2 also
provides support for the computer generated model for
Ulex lectin complexed with the H-type 2 trisaccharide
(Lemieux personal communication, Figure 2).
The dramatic reduction of inhibitory power for 2 with
the P. tetragonolobus lectin (∆(∆G°) ) +3.3 kcal mol-1
)
points to substantial destabilization of the ligand-protein
complex. Lemieux’s model for the topology of the binding
surface of 1 with the lectin proposed that O-6 of galactose
is involved in hydrogen bonding at the periphery of the
binding site.6 Nevertheless, it should be noted that
epitope mapping established that substitution of a 6-O-
methyl galactose residue for galactose resulted in a ∆-
(∆G°) ) +0.1 kcal mol-1. This would suggest that there
are no immediate steric or electronic problems in alkyl-
ation of this hydroxyl group. Introduction of the tether
may result in several changes that could account for the
loss of binding energy. These would include changes in
the rotamer distribution of the hydroxymethyl group,
adverse affects on an extended network of hydrogen
bonds that involve structured water molecules, or a
simple steric clash between the tether and protein.
a
(a) NaH, THF, DMSO; (b) NaCNBH3, HCl in ether, 3 Å MS,
THF; (c) AgOTf, NIS, 4 Å MS, CH2Cl2; (d) MeOTf, 4 Å MS,
DTBMP, CH2Cl2.
Con clu sion
These data highlight the difficulty in designing a tether
that can effectively constrain an oligosaccharide in a
bioactive conformation while avoiding unfavorable pro-
tein contacts. Apparently, even modification of hydroxyl
groups that are thought to lie at the periphery of the wing
bean lectin binding site and which should be involved in
relatively weak hydrogen bonds may cause sufficient
unfavorable interactions to destabilize the complex by
The synthesis of a tethered H-type 2 trisaccharide 2
has been accomplished via an intramolecular glycosyla-
tion reaction that utilized the tether to control the
stereoselectivity of galactose-glucosamine glycosidic bond
formation.
Ulex lectin is classified with the group of lectins that
bind L-fucose; however, it binds the H-type 2 trisaccha-
ride 900-fold tighter than L-fucose,12 which indicates a
protein-oligosaccharide contact surface that extends
beyond the fucose residue. Epitope mapping by Lemieux
et al. concluded that fucose and portions of the galactose
residue make the most important polar contacts with
ca.+3 kcal mol-1
.
Interactions in the Ulex lectin site, which is the main
focus of this study, are more promising. The solid-phase
assays reported here show that the free energy of binding
changes by only 0.6 kcal mol-1. Published data from van’t

9086 J . Org. Chem., Vol. 64, No. 25, 1999
Wacowich-Sgarbi and Bundle
Sch em e 4a
a
(a) DDQ, CH2Cl2, H2O; (b) DMTST, 4 Å MS, DTBMP, CH2Cl2; (c) 90% AcOH 65 °C; (d) H2, Pd(OH)2/C, EtOH.
determine the effect of tethering on the entropy of
binding. These measurements combined with conforma-
tional studies of the solution conformation of the bound
trisaccharide are the subject of further studies.
Exp er im en ta l Section
All commercial reagents were used as supplied. Solvents
used in reactions were dried according to standard methods.45
Molecular sieves used in the experiments were flame-dried and
then cooled under high vacuum immediately prior to use. TLC
was performed on silica gel 60-F254 plates (E. Merck, Darms-
tadt) with detection by charring with 5% sulfuric acid in
ethanol. Column chromatography was performed on silica gel
60 (E. Merck 40-60 µM, Darmstadt) with redistilled solvents.
Optical rotations were measured with a Perkin-Elmer 241
1
polarimeter at 22 °C. H NMR spectra were recorded at 300,
500, or 600 MHz and are referenced to internal CHCl3 (δ 7.24
ppm, CDCl3) or to 0.1% external acetone (δ 2.225 ppm, D2O).
13C NMR spectra (HMQC or APT) were recorded at 75, 120,
or 150 MHz and are referenced to internal CHCl3 (δ 77.0 ppm,
CDCl3) or to 0.1% external acetone (δ 31.07 ppm, D2O).
Coupling constants are expressed in Hz. Assignments of
resonances for the target compound 2 were made by two-
dimensional homonuclear and heteronuclear shift correlation
experiments (GCOSY and HMQC). Verification of the position
of glycosidic linkages were made by homonuclear 2D offset
compensated rotating-frame nuclear Overhauser effect (ROE-
SY) experiments. Microanalyses were carried out by the
analytical service at this department, and all samples submit-
ted for elemental analyses were dried overnight under vacuum
with phosphorus pentoxide or Drierite at 56 °C (refluxing
acetone). Electrospray high-resolution mass spectra (ES HRMS)
of all samples were recorded on a Micromass ZabSpec Hybrid
Sector-TOF mass spectrometer.
Computational studies were performed using software from
Biosym/MSI (San Diego). Minimization calculations for the
tethered disaccharide 2 were done with the Discover program,
using the CVFF force field, and graphical displays were printed
out from the Insight II molecular modeling program.
Rigid body potential energy calculations were performed for
the H-type 2 trisaccharide 1 using the CVFF force field, as
well as the program GEGOP36 that is based upon the HSEA
force field.46 Monosaccharide coordinates were generated from
X-ray or neutron diffraction data. Potential energy maps (φ,ψ)
were calculated for each glycosidic linkage at 5° intervals. For
F igu r e 5. Biological activity of the tethered H-type 2 trisac-
charide 2 relative to the native H-type 2 trisaccharide 1
assayed with the lectins (a) Ulex europaeus I and (b) Psopho-
carpus tetragonolobus II winged bean lectin.
Hoff plots26 show a strong enthalpic contribution to
binding (∆H° ) 29 kcal mol-1). Consequently titration
calorimetry, which depends on a large enthalpic term for
its sensitivity, should be able to provide useful data to
(45) Perrin, D. D.; Armarego, W. L. F.; Perrin, D. R. In Purification
of Laboratory Chemicals, 2nd ed.; Pergamon Press: New York, 1980.
(46) Thøgersen, H.; Lemieux, R. U.; Bock, K.; Meyer, B. Can. J .
Chem. 1982, 60, 44-57.

Constrained H-Type 2 Blood Group Trisaccharide
J . Org. Chem., Vol. 64, No. 25, 1999 9087
a given linkage, the map was obtained by choosing the
minimum energy for each point (rigid body relaxed map).
Solid-phase immunoassay (EIA) were performed in duplicate
as described below using a variant of previously described
methods.44,47 U. europaeus I lectin was purified by affinity
chromatography on a fucose-agarose column (Sigma). Pure
lectin was eluted by 1% fucose in PBS buffer. Psophocarpus
tetragonolobus II lectin was purified on a lactose affinity
column and eluted with 1% lactose in PBS buffer. The dialyzed
lectin solutions were used to coat EIA plates. Biotin-labeled
H-type 2 BSA glycoconjugate48 was allowed to bind to solid-
phase lectin in the presence and absence of inhibitors.
Affinity purified lectin dissolved in phosphate buffered
saline (PBS) (1 µg/mL) was coated on 96-well ELISA plates
overnight at 4°. The plate was washed five times with PBS
containing Tween 20 (0.05% v/v), blocked with bovine serum
albumin (BSA) (Sigma) 2.0% in PBS for 1 h, and then washed
three times with PBS containing Tween 20 (PBST). A glyco-
conjugate consisting of H-type 2 trisaccharide conjugated to
BSA and biotinylated (0.02-0.3 µg/mL) was mixed with
inhibitor at concentrations in the range 0.1 nM to 10 mM. The
duplicate mixtures were added to the coated microtiter plate
and incubated at room temperature for 18 h. Then the plate
was washed five times with PBST, and streptavidin horserad-
ish peroxidase was added and incubated for 1 h at room
temperature. This step was followed by washing five times
with PBST. TMB horseradish peroxidase substrate was added,
and after 2 min the color reaction was stopped with 1 M
phosphoric acid. Absorbance was read at 450 nm, and percent
inhibition was calculated using wells containing no inhibitor
as the reference. The data were plotted with Origin software
(Microcal Software, Northampton, MA).
1-O-ter t-Bu tyld im eth ylsilyloxy-3-O-m eth a n esu lfon yl-
oxy-p r op a n e (8). Diol 6 (3.06 g, 40.3 mmol) was added
dropwise to a solution of NaH (0.967 g, 40.3 mmol) in dry THF
(60 mL) purged with argon. After the mixture stirred for 1 h,
TBDMSCl (6.06 g, 40.3 mmol) was added portionwise, and the
reaction was stirred overnight. Triethylamine (8.3 mL, 61.4
mmol) was added followed by slow addition of MsCl (5 mL,
56.0 mmol) in dry THF (5 mL). After 90 min, the reaction was
diluted with ether; washed with water, 1 M NaOH, and brine;
dried over anhydrous Na2SO4; and concentrated. The residue
was chromatographed (10:1 hexanes-EtOAc) to give a clear
yellow liquid (8.83 g, 81%). 1H NMR (300 MHz, CDCl3): δ 4.29
(t, 2H, MsOCH2), 3.67 (t, 2H, TBDMSOCH2), 2.94 (s, 3H, CH3-
SO3), 1.88 (quintet, 2H, CH2CH2CH2), 0.88 (s, 9H, tBu), 0.05
(s, 6H, (CH3) 2Si). 13C NMR (75 MHz, CDCl3): δ 67.2, 58.4,
37.2, 32.2, 25.9, -5.4. Anal. Calcd for C10H24O4SSi (268.46):
C, 44.74; H, 9.01; S, 11.94. Found: C, 44.61; H, 9.22.
EtOAc) had been completely hydrolyzed to 5 (Rf 0.35). The
organic layer was then washed with water, NaHCO3, and brine
and finally dried over anhydrous Na2SO4. After removal of the
solvent, the crude syrup was chromatographed on silica (5:2
hexanes-EtOAc containing 5% triethylamine) to give 5 that
crystallized from ether (3.65 g, 48%) mp 90-91 °C. [R]D -2.9°
1
(c 2.8, CHCl ). H NMR (300 MHz, CDCl3): δ 7.32 (d, 2H, J
3
8.8, CH3OC6H4), 6.85 (d, 2H, CH3OC6H4), 4.75 (d, 1H, J gem 11.0,
C6H4CH2O), 4.67 (d, 1H, C6H4CH2O), 4.40 (d, 1H, J 1,2 9.5, H-1),
4.22 (t, 1H, J 2,3 6.1, J 3,4 5.9, H-3), 4.17 (dd, 1H, J 4,5 2.0, H-4),
3.88-3.96 (m, 1H, H-6a or H-6b), 3.74-3.80 (m, 5H, H-5, H-6a
or H-6b, CH3OC6H4), 3.42 (dd, 1H, H-2), 2.70 (m, 2H, SCH2-
CH3), 1.43, 1.33 (2s, 3H each, CCH3), 1.28 (t, 3H, J 7.4,
SCH2CH3). 13C NMR (75 MHz, CDCl3): δ 159.3, 130.3, 130.0,
129.9, 113.8, 113.7, 109.9, 83.8, 79.6, 78.8, 75.7, 74.0, 73.3, 69.4,
55.3, 27.9, 26.4, 24.8, 15.0. Anal. Calcd for C19H28O6S
(384.50): C, 59.35; H, 7.34; S, 8.34. Found: C, 59.30; H, 7.41.
Eth yl 3,4-O-Isop r op ylid en e-2-O-p-m eth oxyben zyl-6-O-
(3′-O-ter t-bu tyld im eth ylsilyloxyp r op yl)-1-th io-â-D-ga la c-
top yr a n osid e (11). NaH (25 mg, 1.0 mmol) was added to a
solution of alcohol 5 (378 mg, 1.0 mmol) in dry THF (4 mL).
After the mixture stirred for 30 min at 60 °C, a solution of
1-O-tert-butyldimethylsilyloxy-3-O-methanesulfonyloxy-pro-
pane (360 mg, 1.0 mmol) in dry THF (1 mL) was added
dropwise. The reaction mixture was stirred at 60 °C for 3 h,
and then 1 equiv of both NaH and linker were added. The
reaction mixture was allowed to concentrate by evaporation
of solvent to approximately one-third of its original volume.
TLC indicated complete conversion of the alcohol. The reaction
was quenched with EtOAc containing residual water; diluted
with CH2Cl2; washed with water, NaHCO3, and brine; and
then dried over anhydrous Na2SO4. After concentration,
column chromatography of the crude material (8:1 hexanes-
EtOAc) gave 11 (501 mg, 91%) as an oil. [R]D -8.8° (c 14.7,
CHCl ). 1H NMR (300 MHz, CDCl3): δ 7.32 (d, 2H, J 8.4, CH3-
OC6H43), 6.84 (d, 2H, CH3OC6H4), 4.75 (d, 1H, J gem 11.0,
C6H4CH2O), 4.66 (d, 1H, C6H4CH2O), 4.38 (d, 1H, J 1,2 9.7, H-1),
4.16 (m, 2H, H-3, H-4), 3.81 (H-5), 3.77 (s, CH3OC6H4), 3.66
(Hlink-1), 3.60-3.70 (H-6a, H-6b), 3.55 (m, 2H, Hlink-3), 3.40 (dd,
1H, J 2,3 6.1, H-2), 2.69 (m, 2H, SCH2CH3), 1.75 (quintet, 2H,
H
link-2), 1.42, 1.32 (2s, 3H each, CCH3), 1.28 (t, 3H, J 7.5,
SCH2CH3), 0.86 (s, 9H, (CH3) 3CSi), 0.02 (s, 6H, (CH3) 2Si). 13
C
NMR (300 MHz, CDCl3): δ 130.0, 113.8, 83.6 (1J C,H 152.2),
79.2, 78.4, 75.1, 73.6, 72.9, 69.7, 67.9, 59.7, 55.1, 32.5, 27.5,
26.0, 25.6, 24.4, 14.8, -5.6. Anal. Calcd for C28H48O7SSi
(556.85): C, 60.39; H, 8.69; S, 5.76. Found: C, 60.37; H, 8.81.
Eth yl 3,4-O-Isop r op ylid en e-2-O-p-m eth oxyben zyl-6-O-
(3′-h ydr oxypr opyl)-1-th io-â-D-galactopyr an oside (12). Com-
pound 11 (218 mg, 0.4 mmol) was stirred for 1 h with a solution
of TBAF in THF (1 M, 5 mL). The reaction mixture was
evaporated, diluted with CH2Cl2, washed with brine, dried over
anhydrous Na2SO4, and concentrated. The crude syrup was
chromatographed (1:1 hexanes-EtOAc) to give 12 (172 mg,
100%) as an oil. [R]D -9.7° (c 3.2, CHCl3). 1H NMR (300 MHz,
CDCl3): δ 7.31 (d, 2H, J 8.7, CH3OC6H4), 6.83 (d, 2H, CH3-
OC6H4), 4.73 (d, 1H, J gem 11.1, C6H4CH2O), 4.65 (d, 1H,
C6H4CH2O), 4.38 (d, 1H, J 1,2 9.6, H-1), 4.17 (H-3), 4.13 (H-4),
E t h yl 3,4-O-Isop r op ylid en e-2-O-p -m et h oxyb en zyl-1-
th io-â-D-ga la ctop yr a n osid e (5). To a mixture of thioglycoside
9 (7.86 g, 20.0 mmol) and 2,2-dimethoxypropane (80 mL) was
added toluenesulfonic acid (20 mg). The reaction mixture was
stirred overnight, neutralized with triethylamine, concen-
trated, and evaporated twice with toluene. The residue was
diluted with CH2Cl2, washed with brine, dried with anhydrous
Na2SO4, and concentrated. The dried syrup was dissolved in
DMF (70 mL) and purged with argon. To the ice-cooled solution
was added NaH (0.58 g, 24.2 mmol) portionwise. After stirring
for 45 min, p-methoxybenzyl chloride (4.10 mL, 30.3 mmol)
was added dropwise. After 2 h, another 0.5 equiv of NaH and
p-methoxybenzyl chloride were added to the reaction mixture.
After an additional 2 h, the reaction was complete, and
diethylamine (1.5 mL) was added to the mixture to destroy
excess p-methoxybenzyl chloride. The reaction was then
quenched with EtOAc containing residual water, diluted with
CH2Cl2, washed with NaHCO3 and water, and then stirred
with 1 M HCl until the mixed acetal (Rf 0.79; 1:1 hexanes-
3.83 (m, 1H, H-5), 3.76 (s, CH3OC6H4), 3.74, 3.68 (Hlink-1, Hlink
-
3), 3.66-3.74 (H-6a, H-6b), 3.40 (dd, 1H, J 2,3 6.0, H-2), 2.58-
2.80 (m, 2H, SCH2CH3), 1.79 (m, 2H, Hlink-2), 1.40, 1.32 (2s,
3H each, CCH3), 1.27 (t, 3H, J 7.4, SCH2CH3). 13C NMR (75
MHz, CDCl3): δ 159.3, 130.0, 113.7, 129.9, 110.0, 83.7 (1J C,H
152.2), 79.6, 78.6, 75.4, 73.9, 73.2, 70.3, 70.9, 61.9, 55.3, 31.8,
27.9, 26.3, 24.6, 14.9. Anal. Calcd for C22H34O7S (442.58): C,
59.70; H, 7.74; S, 7.25. Found: C, 59.51; H, 7.90; S, 7.18.
Eth yl 3,4-O-Isop r op ylid en e-2-O-p-m eth oxyben zyl-6-O-
(3′-O-m eth a n esu lfon yloxyp r op yl)-1-th io-â-D-ga la ctop yr a -
n osid e (13). Triethylamine (0.11 mL, 0.8 mmol) was added
to a solution of alcohol 12 (237 mg, 0.5 mmol) in dry CH2Cl2
(4 mL). Methanesulfonyl chloride (0.07 mL, 0.9 mmol) in CH2-
Cl2 (1 mL) was added dropwise. After stirring for 2 h at room
temperature, the solution was diluted with CH2Cl2; washed
with water, 1 M NaOH, water, and brine; and dried over
anhydrous Na2SO4. Evaporation and chromatography (3:2
hexanes-EtOAc) gave 13 (263 mg, 93%) as a white solid. [R]D
(47) Bundle, D. R.; Eichler, E.; Gidney, M. A. J .; Meldal, M.;
Ragauskas, A.; Sigurskjold, B. W.; Sinnot, B.; Watson, D. C.; Yaguchi,
M.; Young, N. M. Biochemistry 1994, 33, 5172-5182.
(48) Wilchek, M.; Bayer, E. A. In Methods in Enzymology: Avidin-
Biotin Technology; Academic Press: New York, 1990; Vol. 184, pp 138-
149.

9088 J . Org. Chem., Vol. 64, No. 25, 1999
Wacowich-Sgarbi and Bundle
1
-6.6° (c 28.8, CHCl3). H NMR (300 MHz, CDCl3): δ 7.32 (d,
61.07; H, 7.60; N, 1.95; S, 4.23. The acetylated form of 15
provided a characteristic NMR spectrum showing a downfield
shift of the 4-OH signal (4.88 ppm).
2H, J 8.7, CH3OC6H4), 6.85 (d, 2H, CH3OC6H4), 4.74 (d, 1H,
J gem 11.0, C6H4CH2O), 4.66 (d, 1H, C6H4CH2O), 4.38 (d, 1H,
J 1,2 9.6, H-1), 4.31 (t, 2H, J 6.2, Hlink-3), 4.17 (J 2,3 6.0, J 3,4 5.8,
H-3), 4.15 (J 4,5 2.0, H-4), 3.83 (ddd, 1H, J 5,6ab 5.1 & 6.9, H-5),
3.77 (s, 3H, CH3OC6H4), 3.64-3.70 (m, 2H, H-6a, H-6b), 3.52-
3.64 (m, 2H, Hlink-1), 3.40 (dd, 1H, J 2,3 6.0, H-2), 2.98 (s, 3H,
MsO), 2.58-2.78 (m, 2H, SCH2CH3), 1.98 (quintet, 2H, J 6.13,
Met h yl 2-Acet a m id o-6-O-b en zyl-2-d eoxy-4-O-(3,4-O-
isop r op ylid en e-2-O- p-m eth oxyben zyl-â-D-ga la ctop yr a -
n osyl)-3,6′-di-O-(pr opan -1,3-diyl)-â-D-glu copyr an oside (16).
Meth od 1. To a dry mixture of starting material 15 (287 mg,
0.4 mmol) and powdered 4 Å molecular sieves purged with
argon was added dry CH2Cl2 (17 mL). Once the reaction flask
was covered with aluminum foil, NIS (129 mg, 0.6 mmol) and
silver triflate (57 mg, 0.2 mmol) were added. After 30 min,
the reaction mixture was diluted with CH2Cl2, and the
molecular sieves were filtered off. The filtrate was washed with
Na2S2O3, NaHCO3, and brine and dried. The concentrated
material was chromatographed (2:1 toluene-acetone) to give
impure 16 as a white solid that contained the iodonated and
noniodonated p-methoxybenzyl group (90.2 mg, 35%).
Meth od 2. To a dry mixture of starting material 15 (97 mg,
0.1 mmol), DTBMP (93.1 mg, 0.5 mmol), and powdered 4 Å
molecular sieves purged with argon was added dry CH2Cl2 (10
mL). Once the reaction had been stirred for 15 min, methyl
triflate (44 µL, 0.4 mmol) was added. After 45 min, the reaction
mixture was diluted with CH2Cl2 and filtered through Celite.
The filtrate was washed with NaHCO3, 0.5 M HCl, NaHCO3,
and brine; dried; and concentrated. The crude material was
chromatographed (2:1 toluene-acetone) to give 16 (44 mg,
50%) as a white solid. The R-anomer of 16 was also isolated
(5.6 mg, 7%). 1H NMR (500 MHz, CDCl3): δ 6.82-7.32 (9H,
ArH), 5.25 (d, 1H, J 9.2, NHAc), 4.69, 4.58 (2 d, 1H each, J gem
11.1, CH3OC6H4CH2O), 4.43, 4.36 (2 d, 1H each, J gem 12.1,
PhCH2O), 4.28 (d, 1H, J 1,2 7.9, H-1), 4.20 (quartet, 1H, J 2,3
∼8.5, H-2), 4.14 (d, 1H, J 1′,2′ 8.4, H′-1), 4.03 (t, 1H, J 3,4 ∼6.3,
H′-3), 3.96 (t, 1H, J 3,4 9.5, J 4,5 9.5, H-4), 3.90 (dd, 1H, J 3′,4′ 5.5,
J 4′,5′ 2.0, H′-4), 3.84 (dd, 1H, J 5,6a, 4.3, J 6a,6b 10.5, H-6a), 3.77
H
link-2), 1.42, 1.33 (2s, 3H each, CCH3), 1.28 (t, 3H, J 7.4,
SCH2CH3). 13C NMR (300 MHz, CDCl3): δ 129.6, 113.5, 83.5,
79.3, 78.4, 75.2, 73.6, 72.9, 69.9, 66.8, 66.3, 55.0, 37.0, 29.1,
27.5, 25.9, 24.4, 14.6. Anal. Calcd for C23H36O9S2 (520.67): C,
53.06; H, 6.97; S, 12.32. Found: C, 52.92; H, 7.07; S, 12.59.
Eth yl 3,4-O-Isop r op ylid en e-2-O-p-m eth oxyben zyl-6-O-
(3′-O-(m eth yl 2-a ceta m id o-4,6-O-ben zylid en e-2-d eoxy-â-
D-glu cop yr a n os-3-oxy)p r op yl)-1-th io-â-D-ga la ctop yr a n o-
sid e (14). Sodium hydride (9.7 mg, 0.4 mmol) was added to a
solution of glucopyranoside 7 (135 mg, 0.4 mmol) in dry THF
(3 mL). The mixture was stirred for 30 min at 60 °C. The
methanesulfonate 13 (146 mg, 0.3 mmol) in dry THF (2 mL)
was added, and the reaction was stirred at 60 °C for 20 min.
Dry Me2SO (0.5 mL) was added, and after an additional 2-4
h of heating at 60 °C, the reaction was complete. The cooled
reaction mixture was diluted with CH2Cl2, washed with
NaHCO3 and brine, and dried over Na2SO4. The concentrated
residue was chromatographed (2:1 hexane-acetone) and gave
14 (174 mg, 83%) as a white solid. [R]D 17.5° (c 5.1, CHCl ).
3
1H NMR (500 MHz, CDCl3): δ 6.84-7.50 (9H, ArH), 5.93 (d,
1H, J 8.0, NHAc), 5.51 (s, 1H, PhCHO), 4.70-4.82 (m, 2H,
J
gem 11.4, J 1,2 8.3, C6H4CH2O and H-1), 4.65 (d,1H, C6H4CH2O),
4.36 (J 1′,2′ 9.6, H′-1), 4.16 (t, 1H, J 2′,3′ ∼6.0, J 3′,4′ ∼6.0, H′-3),4.09
(dd, 1H, J 4′,5′ 1.9, H′-4), 3.98 (J 3,4 ∼9.6, H-3), 3.94 (m, 2H, Hlink
-
1), 3.76 (H-5), 3.72-3.82 (m, 4H, CH3OC6H4 and H′-5), 3.60-
3.70 (H-6a, H-6b), 3.59-3.79 (H′-6a, H′-6b), 3.56 (H-4), 3.50-
3.70 (Hlink-3), 3.50 (s, OCH3), 3.50 (H-2), 3.40 (H′-2), 2.56-2.80
(m, 2H, SCH2CH3), 2.01 (s, 3H, NHCOCH3), 1.66-1.82 (m, 2H,
(s, CH3OC6H4), 3.76 (H′-5), 3.73 (H-6b), 3.73, 3.42 (Hlink-1, Hlink
-
3), 3.68 (H-3), 3.65-3.80 (H′-6a, H′-6b), 3.46 (s, 3H, OCH3),
3.40-3.45 (m, 1H, H-5), 3.37 (dd, 1H, J 2′,3′ 7.2, H′-2), 1.99 (s,
3H, NHCOCH3), 1.60-1.76 (m, 2H, Hlink-2), 1.35, 1.27 (2 s,
3H each, CCH3). 13C NMR (500 Hz, CDCl3): δ 129.7, 113.4,
128.2, 127.6, 104.3 (1J C′,H′ 158.7), 102.4 (1J C,H 158.7), 80.2, 79.4,
73.9, 75.2, 73.5, 72.8, 72.8, 78.73, 50.96, 74.9, 68.9, 68.1, 56.9,
56.86, 55.0, 56.3, 30.5, 27.5, 26.1, 23.4. ES HRMS: m/z
H
link-2), 1.39, 1.30 (2s, 3H each, CCH3), 1.26 (t, 3H, J 7.3,
SCH2CH3). 13C NMR (300 MHz, CDCl3): δ 129.7, 113.6, 128.8,
128.0, 125.8, 101.7, 101.0, 83.8, 82.3, 79.4, 78.3, 77.5, 75.5, 73.6,
72.8, 69.9, 68.9, 68.5, 68.9, 67.7, 65.7, 55.1, 56.9, 30.1, 27.6,
26.1, 23.5, 24.8, 14.5. Anal. Calcd for C38H53NO12S (747.92):
C, 61.03; H, 7.14; N, 1.87; S, 4.29. Found: C, 61.36; H, 7.28;
N, 1.85.
710.3152 [M + Na]+ ( 0.1 mDa and 688.3 [M + H]+ (C36H49
NO12 requires m/z 687.80).
-
Eth yl 3,4-O-Isop r op ylid en e-2-O-p-m eth oxyben zyl-6-O-
(3′-O-(m et h yl 2-a cet a m id o-6-O-b en zyl-2-d eoxy-â-D-glu -
copyr an os-3-oxy)pr opyl)-1-th io-â-D-galactopyr an oside (15).
Dry THF (5 mL) was added to a mixture of compound 14 (34
mg, 0.05 mmol), sodium cyanoborohydride (30 mg, 0.5 mmol),
dry powdered 3 Å molecular sieves, and methyl orange (1 mg).
The suspension was purged with argon. A freshly made
saturated solution of HCl in dry ether was added dropwise to
the reaction mixture until the evolution of gases ceased and
the indicator turned bright pink. A further 2-fold volume of
saturated HCl in ether was added, and the reaction was left
to stir for 30 min. The reaction mixture was diluted with CH2-
Cl2, filtered through Celite, washed with NaHCO3 and brine,
and dried over Na2SO4. After solvent evaporation, column
chromatography of the crude material (2:1 toluene-acetone)
gave a white solid 15 (11.9 mg, 35%). [R]D -12.1° (c 1.9, CHCl
Met h yl 2-Acet a m id o-6-O-b en zyl-2-d eoxy-4-O-(3,4-O-
isop r op ylid en e-â-D-ga la ctop yr a n osyl)-3,6′-d i-O-(p r op a n -
1,3-d iyl)-â-D-glu cop yr a n osid e (4). To compound 16 (36 mg,
0.05 mmol) dissolved in CH2Cl2 (1 mL) was added water (0.06
mL) and DDQ (17 mg, 0.08 mmol). The mixture was stirred
at room temperature for 1 h, and another 1.5 equiv of DDQ
(17 mg, 0.08 mmol) was added. After 1 h the reaction was
quenched with a saturated aqueous NaHCO3 solution. The
mixture was diluted with CH2Cl2, washed with NaHCO3 and
brine, dried over anhydrous Na2SO4, and concentrated. The
crude material was chromatographed (1:1 toluene-acetone)
and gave 4 (20 mg, 67%) as a white solid. [R]D 6.8° (c 3.7, CHCl
1
3
). H NMR (500 MHz, CDCl3): δ 7.26-7.36 (5H, Ph), 5.28 (d,
1H, J 9.0, NH), 4.65 (d, 1H, J gem 12.1, PhCH2O), 4.55 (d, 1H,
PhCH2O), 4.33 (d, 1H, J 1,2 7.9, H-1), 4.12 (d, 1H, J 1′,2′ 8.4, H′-
1), 4.06 (quartet, 1H, J 2,3 ∼9.3, H-2), 3.94 (t, J 3,4 ∼5.5, H′-3),
3.88 (H′-4), 3.87 (H-6a), 3.86 (H-4), 3.79 (H-6b), 3.73, 3.68, 3.63,
3.58 (Hlink-1, Hlink-3, H′-6a, H′-6b), 3.67 (H-3), 3.53 (ddd, 1H,
J H,OH 2.5, J 2′,3′ 6.5, H′-2), 3.45 (H-5, CH3O), 1.99 (s, 3H,
NHCOCH3), 1.60-1.78 (m, 2H, Hlink-2), 1.56, 1.47 (2 s, 3H each,
CCH3). 13C NMR (500 MHz, CDCl3): δ 128.0, 127.9, 105.0
(1J C′,H′ 153.1), 102.3 (1J C,H 163.2), 79.9, 79.2, 77.6, 74.1, 73.7,
73.6, 73.0, 69.3, 68.3, 66.2, 59.3, 56.4, 52.3, 30.7, 27.9, 26.1,
23.5. Anal. Calcd for C28H41NO11 (567.64): C, 59.25; H, 7.28;
N, 2.47. Found: C, 59.56; H, 7.57; N, 2.42.
1
3). H NMR (500 MHz, CDCl3): δ 6.84-7.32 (9H, ArH), 5.77
(d, 1H, J 8.0, NHAc), 4.73 (d, 1H, J gem 11.0, CH3OC6H4CH2O),
4.69 (d, 1H, J 1,2 8.2, H-1), 4.65 (d, 1H, CH3OC6H4CH2O), 4.59
(d, 1H, J gem 12.1, PhCH2O), 4.57 (d, 1H, PhCH2O), 4.36 (d, 1H,
J 1′,2′ 9.5, H′-1), 4.15 (t, 1H, J 3′,4′ ∼6.0, H′-3), 4.11 (dd, 1H, J 4′,5′
1.8, H′-4), 3.81 (ddd, 1H, J 5′,6′a/b 6.0, H′-5), 3.80 (H-6a, H-6b),
3.77 (CH3OC6H4), 3.75, 3.61 (Hlink-1, Hlink-3), 3.73 (H-3), 3.66
(H′-6a, H′-6b), 3.50 (H-4), 3.48 (H-5), 3.46 (s, CH3O), 3.39 (dd,
1H, J 2′,3′ 6.4, H′-2), 3.29 (quartet, 1H, J 2,3 ∼8.2, H-2), 2.60-
2.78 (m, 2H, SCH2CH3), 1.97 (s, 3H, NHCOCH3), 1.70-1.83
(m, 2H, Hlink-2), 1.40, 1.31 (2 s, 3H each, CCH3), 1.28 (t, 3H, J
7.4, SCH2CH3). 13C NMR (500 Hz, CDCl3): δ 129.8, 113.5,
127.2, 100.9 (1J C,H 162.8), 83.7 (1J C′,H′ 154.8), 81.2, 79.7, 78.4,
75.2, 74.2, 74.0, 73.7, 73.2, 71.8, 70.2, 68.7, 68.5, 56.6, 56.5,
Met h yl 2-Acet a m id o-6-O-b en zyl-2-d eoxy-4-O-(3,4-O-
isopr opyliden e-2-O-(2,3,4-tr i-O-ben zyl-r-L-fu copyr an osyl)-
â-D-ga la ctop yr a n osyl)-3,6′-d i-O-(p r op a n -1,3-d iyl)-â-D-glu -
cop yr a n osid e (17). A mixture of disaccharide 4 (50 mg, 0.09
mmol) and thioglycoside 3 (44 mg, 0.09 mmol) in dry CH2Cl2
(2 mL) was prepared and purged with argon. The solution of
55.2, 29.56, 27.7, 26.2, 24.6, 23.3, 15.1. Anal. Calcd for C38H55
NO12S (749.93): C, 60.86; H, 7.39; N, 1.87; S, 4.28. Found: C,
-

Constrained H-Type 2 Blood Group Trisaccharide
J . Org. Chem., Vol. 64, No. 25, 1999 9089
sugars 3 and 4 was transferred via cannula to a mixture of
DMTST (60 mg, 0.2 mmol), DTBMP (57 mg, 0.3 mmol), and 4
Å molecular sieves in dry CH2Cl2 (2 mL) at -68 °C purged
with argon. The reaction was allowed to warm to room
temperature. After stirring for 2 h the reaction was diluted
with CH2Cl2, washed with NaHCO3 and brine, dried over Na2-
SO4, and concentrated. The crude material was chromato-
graphed (3:2 hexane-acetone) and gave 17 (47.6 mg, 55%) as
a white solid. 1H NMR (500 MHz, CDCl3): δ 7.26-7.43 (m,
20H, Ph), 5.59 (d, 1H, J 1′′,2′′ 3.8, H′′-1), 5.25 (d, 1H, J 9.2, NH),
4.92, 4.61 (2 d, 2H, J gem 11.6, PhCH2O), 4.77, 4.61 (2 d, 2H,
J gem 11.9, PhCH2O) 4.78, 4.74 (2 d, 2H, J gem 11.6, PhCH2O)
4.52, 4.48 (2 d, 2H, J gem 12.0, PhCH2O), 4.26 (d, 1H, J 1,2 7.9,
H-1), 4.18 (quartet, 1H, J 2.3 9.6, H-2), 4.11 (d, J 1′,2′ 8.6, H′-1),
4.09 (J 3′,4′ 6.6, H′-3), 4.05 (dd, J 2,3 10.1, H′′-2), 3.95 (quartet,
1H, H′′-5), 3.89 (t, 1H, J 3,4 9.2, J 4,5 9.3, H-4), 3.85 (dd, 1H, J 4,5
5.5, H′-4), 3.82 (dd, 1H, J 3′′,4′′ 2.6, H′′-3), 3.78 (H-6a, H-6b), 3.75
(H′-2), 3.70 (H′-5), 3.69-3.76 (Hlink-1, Hlink-3), 3.67 (H-3), 3.62
(d, 1H, J 4′′,5′′ 2.3, H′′-4), 3.47 (s, 3H, CH3O), 3.34-3.40 (m, 1H,
H′-6a), 3.25 (m, 1H, H-5), 1.99 (s, 3H, NHCOCH3), 1.64-1.74
(m, 2H, Hlink-2), 1.44, 1.28 (2 s, 3H each, CCH3), 1.09 (d, 3H,
J 5′′,6′′ 6.4, H′′-6). 13C NMR (500 MHz, CDCl3): δ 126.0-130.0,
102.5 (1J C′,H′ 156.6), 102.3 (1J C,H 159.0), 95.6 (1J C′′,H′′ 172.6), 80.5,
78.9, 78.3, 77.3, 76.1, 76.0, 75.1, 74.52 74.1, 73.6, 73.2, 72.5,
68.3, 68.0, 66.1, 56.3, 56.2, 50.4, 30.0, 27.6, 26.2, 23.1, 16.4.
ES HRMS: m/z 1006.4577 [M + Na]+ ( 1.2 mDa and 984.5
[M + H]+ (C55H69NO15 requires m/z 984.17).
then dissolved in ethanol (10 mL), to which 20% palladium
hydroxide on charcoal was added (10 mg). The mixture was
hydrogenated for 18 h under an atmosphere of hydrogen to
give 2 as a white solid. The suspension was filtered and
concentrated, and the crude material was purified by HPLC
(water-methanol gradient 0-10%) to give 2 (3.6 mg, 90%).
[R]D -88.7° (c 2.3, H2O). Rf 0.41 (14:6:1 CH2Cl2-methanol-
water). 1H NMR (600 MHz, D2O): δ 5.39 (d, 1H, J 1′′,2′′ 2.0, H′′-
1), 4.46 (d, 1H, J 1′,2′ 8.1, H′-1), 4.45 (d, 1H, J 1,2 8.2, H-1), 4.18
(quartet, 1H, H′′-5), 4.13 (dd, 1H, J 2,3 11.0, H-2), 4.09 (t, J 3,4
9.5, J 4,5 9.5, H-4), 4.07 (dd, J 6a,5 1.8, J gem 11.7, H-6a), 3.87 (H′-
6a), 3.87 and 3.44 (Hlink-1), 3.86 (H′-3), 3.85 (H-6b), 3.84 (H′′-
4), 3.82 (H′′-2), 3.81 (H′′-3), 3.79 (H′-5), 3.78 (H′-4), 3.78 and
3.65 (Hlink-3), 3.77 (H-3), 3.71 (dd, J 2′,3′ 9.5, H′-2), 3.70 (H′-6b),
3.52 (s, 3H, OCH3), 3.45 (H-5), 2.04 (s, 3H, NHCOCH3), 1.72-
1.79 (m, 2H, Hlink-2), 1.25 (d, 3H, J 5′′,6′′ 6.8, H′′-6). 13C NMR
1
(125 MHz, D2O): δ 175.1 (CdO, NHAc), 104.0 (C-1, J C,H
160.7), 103.0 (C′-1, 1J C′,H′ 161.7), 100.1 (C′′-1, 1J C′′,H′′ 175.2), 79.0
(C-3), 76.4 (C-5), 76.1 (C′-2), 75.6 (C′-5), 75.0 (C′-3), 74.4 (C-
4), 72.6, 70.4, 69.9 (C′-4), 69.4 (C′-6), 69.4 (CH2, Clink-3), 69.0
(C′′-2, C′′-3, C′′-4), 67.7 (C′′-5), 61.2 (C-6), 58.1 (CH3O), 57.7
(CH2, Clink-1), 51.2 (C-2), 29.5 (CH2, Clink-2), 22.9 (NHCOCH3),
16.1 (C′′-6). ES HRMS: m/z 606.2373 [M + Na]+ ( 0.1 mDa
and 584.3 [M + H]+ (C24H41NO15 requires m/z 583.60).
Ack n ow led gm en t. The authors thank Dr. R. U.
Lemieux for unpublished data on the docked H-type 2
complex with Ulex lectin and Professor L. Delbaere for
crystal structure coordinates of the Ulex lectin. The
research was made possible by Natural Science and
Engineering Research Council of Canada (NSERC)
research grants to D. R. Bundle.
Meth yl 2-Aceta m id o-2-d eoxy-4-O-(2-O-(r-L-fu cop yr a -
n osyl)-â-D-ga la ctop yr a n osyl)-3,6′-d i-O-(p r op a n -1,3-d iyl)-
â-D-glu cop yr a n osid e (2). The protected, tethered trisaccha-
ride 17 (7 mg, 0.007 mmol) was dissolved in 90% acetic acid
and heated at 65 °C for 6 h. The solvents were evaporated,
and the residue coevaporated with toluene. The partially
protected trisaccharide 18 (Rf 0.13 (2:3 hexane-acetone) was
J O990979A