Synthetic analogs of the lewisb-tetrasaccharide and their binding to Griffonia simplicifolia lectin-part I

Article abstract of DOI:10.1055/s-2003-40329

The lewis b tetrasaccharide binds specifically to the Griffonia simplicifolia lectin GS-IV. In an effort to obtain tighter binding derivatives, we have synthesized a series of analogs of this tetrasaccharide and studied their binding to the lectin. This paper (part I in a series of two) describes the synthesis of and binding studies with analogs where the galactose unit (b) has been altered at the 6-position. Even though this position does not make direct contact with the protein in the crystal structure, the binding was very sensitive to these substitutions.

Full text of DOI:10.1055/s-2003-40329

LETTER
1327
Synthetic Analogs of the Lewisb-Tetrasaccharide and their Binding to
Griffonia Simplicifolia Lectin-Part I
Synthetic
Analogsiof the
v
ewisb-Tetraesaccharidekanand Kamath, Joanna Sadowska, Shahul Nilar, David R. Bundle, Ole Hindsgaul*
Department of Chemistry, University of Alberta, Edmonton, Alberta, T6G 2G2, Canada
E-mail: ole.hindsgaul@ualberta.ca
Received 12 March 2003
Dedicated to the memory of Professor Raymond U. Lemieux.
OH
Me
Abstract: The lewis b tetrasaccharide binds specifically to the
Griffonia simplicifolia lectin GS-IV. In an effort to obtain tighter
binding derivatives, we have synthesized a series of analogs of this
tetrasaccharide and studied their binding to the lectin. This paper
(part I in a series of two) describes the synthesis of and binding stud-
ies with analogs where the galactose unit (b) has been altered at the
6-position. Even though this position does not make direct contact
with the protein in the crystal structure, the binding was very sensi-
tive to these substitutions.
Key words: lewisb-tetrasaccharide, lectins, galactose unit
O
d
OH
OH
Asn135
MeO
O
NHAc
O
HO
H
O
a
N
H
O
b
O
O
O
H
HO
O
6b
O
HO
O
c
Me
H
HO
O
Asp89
Carbohydrates play a very important role in many biolog-
ical processes including inflammation, cellular differenti-
ation, hormone-cell recognition and bacterial-host cell
attachment, and have been implicated in several life
threatening diseases such as cancer, inflammatory disor-
ders, and microbial infection.1–8 One of the most impor-
O
H
H
O
Ser49
tant present drawbacks of carbohydrate-based therapeutic
agents is the low intrinsic affinity of carbohydrates to their
protein receptors (normally in the millimolar range).
Hence, it is important to investigate the factors which in-
fluence carbohydrate-protein binding strength with a view
to designing better inhibitors.
Figure 1 Schematic of the crystalline complex of Leb-OMe bound
to lectin GS-IV.10,13 Essential hydrogen bonding with the protein is
shown. OH-6 (unit b), which has been derivatized in the present work,
is highlighted.
actions, OH-3 (unit c) and OH-2 (unit d) also contribute
significantly though the corresponding deoxy-compounds
still bind.11–13
The binding of the Lewis-b (Leb) tetrasaccharide, a-L-
Fuc-(1→2)-b-D-Gal-(1→3)-[a-L-Fuc-(1→4)]-b-D-GlcN-
Ac-OMe (Leb-OMe), to the lectin IV of Griffonia
simplicifolia (GS-IV)9 is one of the most systematically
studied systems in carbohydrate-protein recognition
through the pioneering contributons of R. U. Lemieux and
his collaborators.10–20 GS-IV-Leb recognition therefore
represents an ideal system to evaluate approaches to the
design of better inhibitors of carbohydrate-protein bind-
ing. Numerous tetrasaccharide analogs have been chemi-
cally synthesized,12–18 and the crystal structure of a
tetrasaccharide-protein complex is known.10,13 Briefly, the
X-ray crystal structure of GS-IV bound to Leb-OMe in-
volves both polar and extensive nonpolar interactions
(Figure 1). Essential polar interactions occur through hy-
drogen bonding of OH-3 and OH-4 of galactosyl residue
(unit b) with Asp-89 and Asn-135. The 4-OH group of the
fucosyl residue c is also critically involved through a hy-
drogen bond to Ser-49. In addition to these essential inter-
Since a large portion of the tetrasaccharide resides in the
aqueous phase but is close to the protein, several opportu-
nities should exist for strengthening of the binding by add-
ing groups to the this part of the tetrasaccharide. The
crystal structure data10,13 shows that the 6b-hydroxy group
is accepted into a nonpolar region of the combining site
through without any direct hydrogen bonding to the pro-
tein. In fact, the 6b-O-methyl analog which was found to
bind 0.5 kcal/mol more strongly than the parent com-
pound, Leb-Ome.19 The crystal structure of the complex
also showed the internuclear distances between the oxy-
gen atom of the hydroxy group at position 6b and Tyr223,
Trp133 and Arg88 (residues not shown in Figure 1) to be
in the range of 4–7 Å units.20 We therefore decided to ex-
amine if substitution of OH-6b by groups other than me-
thyl might further enhance the binding. To accomplish
this, we replaced the 6b hydroxy group by an amino
group, and attached various substituents by N-acylation.
Molecular modeling studies using the structure-based
drug design technique LUDI21–24 were carried out focused
on modifications at the 6b position of galactose. A data-
Synlett 2003, No. 9, Print: 11 07 2003.
Art Id.1437-2096,E;2003,0,09,1327,1330,ftx,en;S02403ST.pdf.
© Georg Thieme Verlag Stuttgart · New York

1328
V. Kamath et al.
LETTER
AcO
AcO
OAc
AcO OAc
AcO
Ph
Ph
O
BnO
HO
O
O
O
O
O
O
O
O
O
+
a
b
OMe
OMe
NHAc
BnO
OMe
O
HO
AcO
NHAc
NHAc
AcO
Br
2
3
5
4
OBn
OBn
O
HO
HO
O
OH
R
OBn
OMe
O
O
O
N3
O
f
c
d
O
O
O
O
Br
O
O
OH
OH
6
7. R = OH
O
NHAc
O
OBn
9
e
OBn
BnO
8. R = N3
O
OBn
10
OBn
BnO
HO
BnO
OBn
OH
OBn
OH
O
O
OBn
OH
HO
HO
HO
HO
NH2
O
N3
g
h
O
O
O
O
O
O
OMe
O
OMe
O
NHAc
NHAc
O
O
O
OBn
OH
11
1
OH
OBn
BnO
HO
Scheme 1 Reagents: a) Hg(CN)2–nitromethane–toluene (89%); b) NaCNBH3–HCl (75%); c) NaOMe–MeOH (95%); d) Acetone/TsOH
(78%); e) Ph3P–DEAD–DPPA (72%); f) 9–Bu4NBr (68%); g) HOAc/H2O; h) Na–NH3 (64%, two steps).
base of chemically relevant fragments was searched for ity of the tetrasaccharide were within this distance. The re-
suitable candidates for incorporation into the Lewis b tet- sult from the search was pruned down to a subset of seven
rasaccharide structure so as to enhance the binding. candidates (Scheme 2) which were then synthesized and
Changes resulting from substitutions at the 6b position of evaluated as inhibitors.
the galactosyl residue was the focus of the design, and the
lectin active site was defined as including all amino acids
of the surrounding protein within a radius of 15 Å from
this atom. This value for the radius was considered to be
The synthesis (Scheme 1) began with the condensation of
the peracetylated galactosyl bromide 2 with the blocked
GlcNAc derivative 3 using Helferich conditions to furnish
the blocked disaccharide 4 (89%).25 The benzylidene ring
sufficient as the essential residues important for the activ-
NHCO(CH2)2COOH
HO
HO
a
b
c
O
2. IC50 = 0.44mM
O
O
NHCOCH2OCH2COOH
HO
HO
O
3. IC50 = 0.64mM
4. IC50 = 0.44mM
O
HO
O
OH
OH
O
HO
HO
NHCOCH2CH3
OH
O
HO
HO
NH2
O
O
O
O
O
O
OMe
NHAc
HO NHCOCH2NHAc
O
d
O
5. IC50 = 0.18mM
HO
O
O
OH
1. IC50 = 0.2 mM
O
OH
HO
HO
HO
NHCO(CH2)2NHAc
O
e
f
6. IC50 = 0.45mM
7. IC50 = 0.074mM
O
O
HOOC
NH
O
HO NH
COOH
O
O
HO
O
O
Scheme 2 Reagents: a) Succinic anhydride, DMF/MeOH; b) Diglycolic acid anhydride; c) Propionic acid anhydride; d) 1. PfpO-Gly-NH-
Fmoc, 2. morpholine, 3. 10% Ac2O/MeOH; e) 1. PfpO-b-Ala-NH-Fmoc, 2. morpholine, 3. 10% Ac2O/MeOH; f) 1. PfpO-Gly-NH-Fmoc, 2.
morpholine, 3. 1,2,4-benzenetricarboxylic acid anhydride.
Synlett 2003, No. 9, 1327–1330 ISSN 1234-567-89 © Thieme Stuttgart · New York

LETTER
Synthetic Analogs of the Lewisb-Tetrasaccharide
1329
Table 1 Selected Low-Resolution MS and 360 MHz NMR Data
(a295 K, D2O, dacetone = 2.225 ppm) for the Leb-OMe Analogs
in 4 was reductively opened with sodium cyanoborohy-
dride-hydrogen chloride26 to provide compound 5 (75%).
Deacetylation of 5 followed by isopropylidination using
standard conditions furnished 7 (78%). Treatment of 7
with triphenyl phosphine (Ph3P), DEAD, and diphe-
nylphosphoryl azide (DPPA)27 gave the azido compound
8 (72%). Finally, introduction of the two fucosyl residues
by reacting 8 with 9 and tetrabutylammonium bromide
gave 10 (68%). The isopropylidene group was selectively
removed using 80% acetic acid followed by debenzyla-
tion and reduction of the azide using Birch conditions28 to
furnish the desired amino analog 1 in 64% yield over the
2 steps (Scheme 1).
Analogs 1H NMR data (360 MHz)a
[M + H]+
1
2
3
4
5
6
7
1.25 and 1.26 (d × 2, 6 H, H-6c, H-6d); 2.07 (s, 689.2
3 H, NHAc); 3.47 (s, 3 H, OMe); 4.33 (q, 1 H,
H-5d); 4.32 (d, 1 H, J = 8 Hz, H-1a); 4.66 (d, 1
H, J = 8 Hz, H-1b); 4.80 (q, 1 H, J = 6.5 Hz, H-
5c); 5.01 (d, 1 H, J = 3.5 Hz, H-1c); 5.15 (d, 1
H, J = 4 Hz, H-1d)
1.25 and 1.26 (d × 2, 6 H, H-6c, H-6d); 2.44–
2.56 [m, 4 H, -(CH2)2-]; 3.47 (s, 3 H, OMe);
4.35 (q, 1 H, H-5d); 4.42 (d, 1 H, J = 8 Hz, H-
1a); 4.61 (d, 1 H, J = 8 Hz, H-1b); 4.84 (q, 1 H,
J = 6.5 Hz, H-5c); 5.02 (d, 1 H, J = 3.5 Hz, H-
1c); 5.14 (d, 1 H, J = 4 Hz, H-1d)
789.1
Scheme 2 describes the synthesis of six N-acyl derivatives
of 1. They were generated from 1 and readily available N-
acylating reagents under standard conditions. All analogs
(compounds 1–7) were purified by column chromatogra-
1.24 and 1.27 (d × 2, 6 H, H-6c, H-6d); 3.48 (s, 805.2
3 H, OMe); 4.04 (s, 4 H, -CH2-O-CH2-); 4.33
(q, 1 H, H-5d); 4.34 (d, 1 H, J = 8 Hz, H-1a);
4.64 (d, 1 H, J = 8 Hz, H-1b); 4.84 (q, 1 H, J =
6.5 Hz, H-5c); 5.00 (d, 1 H, J = 3.5 Hz, H-1c);
5.14 (d, 1 H, J = 4 Hz, H-1d)
1
phy, lyophilized and structurally characterized by H
NMR and mass spectroscopy (see Table 1). The reaction
yields were in the 80–90% range.
The binding of GS IV lectin to compounds 1–7 and to the
parent compound Leb-OMe was then studied by an ELISA
inhibition technique. Briefly, GS IV lectin, purified using
a Synsorb-Lewisb affinity column,9,14,29,30 was coated on
96-well ELISA plates. Lewis-b-BSA glycoconjugate pre-
mixed with inhibitor solution was then added to the wells.
The effective inhibitor concentrations were in the range 1
nM–1 mM. After incubation at r.t. for 18 h and washing,
the Lewis-b-BSA-glycoconjugate that remained bound to
the wells was detected by adding first mouse monoclonal
anti Lewis-b antibody, then goat anti-mouse IgG-horse-
radish peroxidase conjugate, and finally substrate
(3,3¢5,5¢-tetramethylbenzidine). Absorbance was read at
450 nm and percent inhibition was calculated using wells
containing no inhibitor as the reference points.
1.13 (t, 3 H, CH3); 1.24 and 1.26 (d × 2, 6 H, H- 745.1
6c, H-6d); 2.30 (q, 2 H, -CH2-); 3.48 (s, 3 H,
OMe); 4.33 (q, 1 H, H-5d); 4.34 (d, 1 H, J = 8
Hz, H-1a); 4.61 (d, 1 H, J = 8 Hz, H-1b); 4.82
(q, 1 H, J = 6.5 Hz, H-5c); 5.01 (d, 1 H, J = 3.5
Hz, H-1c); 5.15 (d, 1 H, J = 4 Hz, H-1d)
1.25 and 1.27 (d × 2, 6 H, H-6c, H-6d); 2.05 and 788.2
2.07 (2 × s, 6 H, NHAc); 3.48 (s, 3 H, OMe);
4.33 (q, 1 H, H-5d); 4.33 (d, 1 H, J = 8 Hz, H-
1a); 4.61 (d, 1 H, J = 8 Hz, H-1b); 4.81 (q, 1 H,
J = 6.5 Hz, H-5c); 5.01 (d, 1 H, J = 3.5 Hz, H-
1c); 5.14 (d, 1 H, J = 4 Hz, H-1d).
1.25 and 1.27 (d × 2, 6 H, H-6c, H-6d); 2.04 and 802.2
2.06 (2 × s, 6 H, NHAc); 3.49 (s, 3 H, OMe);
4.31 (d, 1 H, J = 8 Hz, H-1a); 4.34 (q, 1 H, H-
5d); 4.61 (d, 1 H, J = 8 Hz, H-1b); 4.82 (q, 1 H,
J = 6.5 Hz, H-5c); 5.01 (d, 1 H, J = 3.5 Hz, H-
1c); 5.14 (d, 1 H, J = 4 Hz, H-1d)
The IC50 values for 1–7 were as shown in Scheme 2. The
best new inhibitor was compound 7 (0.074 mM). Howev-
er, this value is not as good as that of the parent com-
pound, Leb-OMe (0.036 mM). Thus, none of the analogs
synthesized were any better inhibitors than the parent
compound. Of the substituted amino compounds 2–7,
only 7 (0.074 mM) and 5 (0.18 mM) were better inhibitors
than the amino compound itself 1 (0.2 mM). These results
further confirm the importance of the 6b position for bind-
ing, but clearly groups other than those investigated here
1.25 and 1.27 (d × 2, 6 H, H-6c, H-6d); 2.04 and 938.2
2.06 (2 × s, 6 H, NHAc); 4.20 (d, 1 H, J = 8 Hz,
H-1a); 4.35 (q, 1 H, H-5d); 4.61 (d, 1 H, J = 8
Hz, H-1b); 4.82 (q, 1 H, J = 6.5 Hz, H-5c); 4.95
(d, 1 H, J = 3.5 Hz, H-1c); 5.15 (d, 1 H, J = 4
Hz, H-1d); 7.50–8.40 (m, aromatic)
will have to be used to obtain a tighter binding to the References
GS-IV lectin.
(1) Varki, A. Glycobiology 1993, 3, 97.
(2) Drickamer, K. Nature (London) 1992, 360, 183.
(3) Giannis, A. Angew. Chem. Int. Ed. Engl. 1994, 33, 178.
(4) Weis, W. I.; Drickamer, K.; Hendrickson, W. A. Nature
(London) 1992, 360, 127.
(5) Dennis, J. W. In Cell Surface Carbohydrates and Cell
Development; Fukuda, M., Ed.; CRC Press: Boca Raton,
1992, 161–194.
(6) Karlson, K. A.; Milh, M. A.; Ångström, J. In Molecular
Recognition in Host Parasite Interactions; Korhonen, T. K.,
Ed.; Plenum Press: New York, 1992, 115–132.
(7) Karlsson, K. A. Annu. Rev. Biochem. 1991, 12, 265.
(8) Carver, J. P. Pure and Appl. Chem. 1993, 65, 763.
Acknowledgement
We thank Dr. A. Otter for carrying out the high-field NMR spectral
analysis, and Dr. A. Morales for mass spectrometric characterizati-
on. This work was supported by grants from the Natural Sciences
and Engineering Research Council of Canada (NSERC). We are in-
debted to Professor Thomas Norberg of the Swedish Agricultural
University for his critical review of this manuscript.
Synlett 2003, No. 9, 1327–1330 ISSN 1234-567-89 © Thieme Stuttgart · New York

1330
V. Kamath et al.
LETTER
(9) Shibata, S.; Goldstein, I. J.; Baker, D. A. J. Biol. Chem.
1982, 25, 9324.
(29) 100 g of peeled and finely ground seeds were stirred with
200 mL of MeOH (5 × 15min) and CH2Cl2 (3 × 15 min). The
filtrates were discarded and the powder was dried overnight
at r.t. All subsequent procedures were carried out at 4 °C.
The dry meal (54 g) was extracted with 400 mL P1 buffer for
2 h with gentle stirring. The extract was centrifuged for 1 h
at 12k rpm and the sediment was re-extracted with P1 buffer
(400 mL) overnight. The aqueous extracts were combined
and solid ammonium sulfate was added to 30% saturation.
The mixture was gently stirred for 2 h. The precipitated
proteins were spun (12k rpm, 20 min) and then discarded.
More ammonium sulfate was added to 75% saturation and
left stirring overnight. The precipitate was collected by
centrifugation, suspended in 25 mL P1 buffer and dialyzed
against P1 buffer. The extract was loaded on the Synsorb-Leb
affinity column (30 g) and the column was washed with P1
buffer until the OD280 reading was less than 0.02. The lectin
was eluted with 0.05 M carbonate/bicarbonate buffer pH 10
with 0.015M NaCl (yield 26mg). The lectin was dialyzed
against P2 buffer, and then concentrated using a
(10) Delbaere, L. T. J.; Vadonselaar, M. P. L.; Quail, J. W.;
Wilson, K. S.; Dauter, Z. J. Mol. Chem. 1993, 230, 950.
(11) Lemieux, R. U. ACS Symp. Ser. 1993, 519, 5.
(12) Nikrad, P. V.; Beierbeck, H.; Lemieux, R. U. Can. J. Chem.
1992, 70, 241.
(13) Delbaere, L. T. J.; Vandonselaar, M. P. 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.
(14) Spohr, U.; Hindsgaul, O.; Lemieux, R. U. Can. J. Chem.
1985, 63, 2644.
(15) Lemieux, R. U.; Ming-Hui, D.; Spohr, U. J. Am. Chem. Soc.
1994, 116, 9803.
(16) Lemieux, R. U.; Szweda, R.; Paszkiewicz, E.; Spohr, U.
Carbohydr. Res. 1990, 205, c12.
(17) Spohr, U.; Morishima, N.; Hindsgaul, O.; Lemieux, R. U.
Can. J. Chem. 1985, 63, 2659.
(18) Lemieux, R. U.; Cromer, R.; Spohr, U. Can. J. Chem. 1988,
66, 3083.
(19) Lemieux, R. U. Chem. Soc. Rev. 1989, 18, 347.
(20) Lemieux, R. U.; Delbaere, L. T. J.; Beierbeck, H.; Spohr, U.
In Ciba Foundation Symposium No. 158, Host-Guest
Molecular Interactions: From Chemistry to Biology; Wiley:
London, 1991, 231–248.
(21) Böhm, H. J. J. Comp. Aided Molec. Design 1992a, 6, 69.
(22) Böhm, H. J. J. Comp. Aided Molec. Design 1992b, 6, 593.
(23) Böhm, H. J. J. Comp. Aided Molec. Design 1994a, 8, 243.
(24) Böhm, H. J. J. Comp. Aided Molec. Design 1994b, 8, 623.
(25) Spohr, U.; Lemieux, R. U. Carbohydr. Res. 1988, 174, 211.
(26) Garegg, P. J.; Hultberg, H. Carbohydr. Res. 1981, 93, c10.
(27) Varasi, M.; Walker, K. A.; Maddox, M. L. J. Org. Chem.
1987, 52, 4235.
Centriprep10 (Amicon). The concentrated lectin was loaded
on Synsorb A affinity column (30 g) and recovered flow
through was then run through Synsorb B affinity column (30
g). In both cases P2 buffer was used as eluent. The purity of
lectin obtained was confirmed on SDS-PAGE with 2% 2ME
where two protein bands were observed of apparent
molecular weights of approximately 28 and 26 KD. (P1
buffer: 0.08 M Na2HPO4, 0.02 M KH2PO4, 0.15 M NaCl,
0.015 M NaN3, 0.5 mM Na2S2O4, 0.14 mM CaCl2. P2 buffer:
0.072 M Na2HPO4, 0.028 M NaH2PO4, 0.15 M NaCl, 3 mM
NaN3, 0.1 mM CaCl2, and 0.1 mM MnCl2). pH of both
buffers was adjusted to 7.2.
(30) Delmotte, F. M.; Goldstein, I. J. Eur. J. Biochem. 1980, 112,
219.
(28) Schön, I. Chem. Rev. 1984, 84, 287.
Synlett 2003, No. 9, 1327–1330 ISSN 1234-567-89 © Thieme Stuttgart · New York