Design and synthesis of a new Sialyl Lewis X Mimetic: How selective are the selectin receptors?

Article abstract of DOI:10.1016/S0960-894X(01)00130-5

The present paper reports the molecular modeling-based design and synthesis of an optically pure noncarbohydrate mimetic of sialyl Lewis X to inhibit E-selectin. Biological evaluation of the designed substance as well as that of its enantiomer gave, contrary to expectations, comparable IC₅₀ values. Results are discussed in terms of receptor binding specificity and the molecular modeling protocol used.

Full text of DOI:10.1016/S0960-894X(01)00130-5

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1109–1112
Design and Synthesis of a New Sialyl Lewis X Mimetic:
How Selective Are the Selectin Receptors?
Marc De Vleeschauwer,a Marc Vaillancourt,b,y Nathalie Goudreau,a,{ Yvan Guindona,b
and Denis Gravela,*
a
´
Departement de chimie, Universite de Montreal, C.P. 6128, succ. Centre-ville, Montreal, Quebec, Canada, H3C 3J7
Institut de recherches cliniques de Montreal (IRCM), 110 ave des Pins Ouest, Montreal, Quebec, Canada, H2W 1R7
´
´
´
´
´
´
b
´
Received 12 January 2001; accepted 12 February 2001
Abstract—The present paper reports the molecular modeling-based design and synthesis of an optically pure noncarbohydrate
mimetic of sialyl Lewis X to inhibit E-selectin. Biological evaluation of the designed substance as well as that of its enantiomer gave,
contrary to expectations, comparable IC50 values. Results are discussed in terms of receptor binding specificity and the molecular
modeling protocol used. # 2001 Elsevier Science Ltd. All rights reserved.
The selectins are key players in the fight against patho-
gens. They play an essential role in the immune system
as they are involved in the migration of leukocytes from
the blood vessels to the sites of inflammation.1 They are
responsible for the rolling of leukocytes on the endo-
thelial cells of post-capillary veinules leading to leuko-
cyte extravasation from the blood stream. E- and P-
selectins are expressed on activated endothelial cells
while L-selectins are constitutively expressed on leuko-
cytes. It is known that overrecruitment of leukocytes
can cause health problems such as rhumatoid arthritis,
septic shock, asthma, diabetes, psoriasis, and reperfu-
sion injury.2 Consequently, much effort has been
expended towards the study of these receptors and their
ligands.
calcium atom located within the receptor binding site,
and that the carboxyl groupof the sialic acid moiety is
involved in the formation of a salt bridge with a neigh-
boring arginine (Arg-97) side chain.4 In addition, it was
proposed that the 4-hydroxyl of the galactose unit binds
to the side chain of tyrosine (Tyr-94) through a hydro-
gen bond, but the energy involved seems less important
than that for the other two previously described inter-
actions.5 An examination of the literature revealed that
efforts in designing inhibitors were concentrated mainly
on the binding to the calcium atom and Arg-97, and
have been mostly directed towards modified carbo-
hydrates.6 This endeavor has given good results, but
carbohydrate derivatives, and particularly glycosides,
are not considered good drug candidates as they are
usually too sensitive to enzymatic hydrolysis to be orally
active and too hydrophilic to show good bioavailability.
The obvious solution then becomes the creation of small
One of the natural ligands of selectins is sialyl Lewis X
(sLex), a tetrasaccharide motif found on specialized
glycoproteins expressed on the surface of endothelial
cells (Fig. 1).3 Following a number of molecular biology
and medicinal chemistry studies on E-selectin such as
mutagenesis and ligand modification, it is now widely
recognized that the 2- and 3-hydroxyl groups on the
fucose unit of sLeX are essential for binding to the
*Corresponding author. Tel.: +1-514-343-6735; fax: +1-514-343-
7142; e-mail: graveld@magellan.unmontreal.ca
yPresent address: Departement de microbiologie et d’immunologie,
Universite de Montreal.
{Present address: Boehringer Ingelheim, 2100 Cunard, Laval, Quebec,
Canada, H7S 2G5
Figure 1. Structural drawing of the sLeX active conformation.9
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00130-5

1110
M. De Vleeschauwer et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1109–1112
noncarbohydrate molecules that can effectively mimic
sLeX and bind efficiently and selectively to one of the
selectins.7 In the present study, we used a molecular
modeling approach to design a low molecular weight
noncarbohydrate inhibitor, which was then synthesized
and tested against sLeX.
the docking of the candidate molecule on the active site
of the E-selectin complex model by superimposing the
vicinal diol of the candidate structure with the calcium
binding diol of the fucose moiety of sLeX (Fig. 4). The
sialyl Lewis X ligand was then removed and the new E-
selectin–candidate structure complex minimized. It is
important to note here that the forcefield does not
properly take into account the interaction between the
ligand and the calcium atom, so a constraint is intro-
duced in the minimization process to keep the distance
and the angle of coordination of the diol of the candi-
date structure with the calcium identical to that in the
docked sLeX. The resulting calculated electrostatic and
van der Waals energies showed favorable interaction of
the mimic to the receptor (Fig. 5). It is interesting to
note that in Figure 5 the carboxyl of the mimic hydro-
gen bonds to both Arg-97 and Tyr-94 whereas in sLeX
(cf. Fig. 3), the neuraminic acid carboxyl binds only to
Arg-97 and from a different orientation. In summary,
the synthetic target validated by this process is the cis-
decalinic diol bearing a carboxylic acid side chain
represented in Figure 2.
The general approach used in the design of this new
inhibitor is similar to that of other research groups and
focused on exploiting some of the key interactions with
the selectin receptors such as the binding to the active
site calcium atom and binding to the side chain of Arg-
97 through the use of a vicinal diol and a carboxylic
acid moiety, respectively. In addition, our candidate
structure was also designed to mimic as much as possi-
ble the E-selectin bound conformation of sLeX as
determined by transfer NOE experiments.8 Therefore, in
order to access proper spatial orientation and distance
between the pharmacophoric groups, we used a con-
formationally semi-rigid cis-decalinic scaffold that
nicely mimics the stacked fucose and galactose rings in
the sLeX active conformation (Fig. 2).
Finally, in order to evaluate this mimic, we docked it on
to a model of the E-selectin receptor. Since there was no
published X-ray structure of sLex or any other molecule
complexed to E-selectin, we first followed the Kogan
approach to generate an E-selectin-sLeX model to serve
as template for the docking of our candidate inhibitor
(Fig. 3).5 This initial model was created using the pub-
lished X-ray structure of the lectin domain of apo E-
selectin and the published bound conformation of sLeX,
and is based on the homology between E-selectin and
the rat-mannose binding protein for which a co-crystal
structure with its natural ligand exists.10 Our modeled
ligand was then built and minimized while constraining
the essential binding groups in a conformation that
mimics the active conformation of sLeX.11 This allowed
A first attempt at elaborating an enantioselective synth-
esis of the target compound (Fig. 2 or structure 8;
Scheme 1) having proven inefficient, the target dihy-
droxyacid 8 was synthesized in nine steps starting with
the known 4-benzyloxycyclohexanone 1 via a racemic
Figure 4. Superimposition of the cis-decalinic mimic of Figure 2
(white) and sLeX (green) on E-selectin.
Figure 2. cis-Decalinic mimic of sLeX.
Figure 3. Docking of sLeX on E-selectin (blue). The pink sphere is the
calcium atom. Acetyl groupof GluNAc is deleted for clarity. Distance
of COOH to Arg-97: 2.22 A.
Figure 5. Docking of the cis-decalinic mimic of Figure 2 (white) on E-
selectin (blue). Distances of COOH to Arg-97 and Tyr-94: respectively
2.36 A and 2.23 A.

M. De Vleeschauwer et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1109–1112
1111
approach (Scheme 1).12 Conversion to the correspond-
ing enone 2 was effected by the Trost procedure using
diphenyldisulfide and the magnesium salt of mono-
peroxyphthalic acid (MMPP).13,14 The next stepwas the
most critical one in this sequence and involved carrying
out a Lewis acid catalyzed Diels–Alder reaction with
1,3-butadiene that established the cis ring juncture of
the decalin and the exo position of the benzyloxy group.
This reaction was carefully monitored to maximize the
selectivity towards the endo product, and to minimize
epimerization of one of the hydrogens at the ring junc-
tion in the Diels–Alder adduct. In order to prevent this
epimerization, we avoided using an aqueous work-up;
instead, we quenched the Diels–Alder keto-adduct by
cannulating the reaction mixture into a lithium alumi-
num hydride suspension in THF to reduce the ketone,
thus securing the ring junction by generating the oxy-
adduct 3. The optimal conditions for this two-step one-
pot reaction were obtained by using SnCl4 as catalyst,
which resulted in a 51% yield of a 5:2 ratio of the cis and
trans junction of the desired endo hydroxy-adduct 3.15
The next stepin the synthesis involved the introduction
of the exo-cis vicinal diol functionality. This reaction
proceeded using the Sharpless AD-mix a reagent where
the bulky ligand ensured exo dihydroxylation.16 The
resulting exo-diol was converted to acetonide 4 and then
deoxygenated by the Barton procedure.17,18 Debenzyla-
tion of 5 required two sequential treatments with palla-
dium on charcoal to obtain the free alcohol 6, probably
because the starting material was contaminated with
sulfur impurities from the deoxygenation reaction. It
should be mentioned here that deacetonization also
occurred during this reaction, so the diol had to be
reprotected before pursuing the sequence.19
At this point in the sequence, resolution of racemic 6
was carried out by chromatographic separation of the
O-acetylmandelate diastereoisomeric mixture prepared
by coupling alcohol 6 with O-acetylmandelic acid using
DCC and DMAP. The optically pure alcohols 6 were
then obtained by hydrolyzing the diastereoisomeric pair
of esters with aqueous potassium carbonate.20 Finally,
O-alkylation of each enantiomer 6 under phase transfer
catalysis21 yielded the two enantiomers of the product 7
which were totally deprotected, in one step, using tri-
fluoroacetic acid in dichloromethane.22 Final purifica-
tion of the resulting enantiomeric dihydroxyacids 8 by
chromatography on Sephadex1 G-10 yielded the pure
enantiomeric pair of 8 for testing.23
The compounds were tested in competition assays.
Expression vectors coding for E-selectin-Ig and P-selec-
tin-Ig24 were used to produce the chimeric proteins in
COS cells. The proteins were then coated to wells and
used in a cell assay based on previously described
methodology.25 Briefly, HL-60 cells containing radio-
active tritium and expressing the P- and E-selectin
ligands with the natural sLeX on their surface are added.
The binding, which is relevant to the natural interaction
between the selectins and the cells leads to the adhesion
of HL-60 cells to the coated wells. The ability of the
mimic to displace the interaction of selectins and HL-60
cells is monitored and IC50’s are evaluated. If the mimic
has a better or similar affinity to the selectins compared
to sLeX, cells are not able to adhere to the wells in a
dose-dependent concentration manner. The activity of
the mimics is measured by counting the change in
radioactivity as a function of concentration. A parallel
control experiment is also performed by measuring the
activity of sLeX in this assay.
Much to our surprise and as indicated in Table 1, both
enantiomers gave, more or less, the same activity as
sLeX for both E- and P- selectins. This result was at the
same time rewarding and puzzling! It was rewarding
because it justified the modeling approach to generate
simple, noncarbohydrate and low-molecular weight
mimics of sLeX. On the other hand, it was puzzling
because, as indicated above, our modeling protocol fol-
lowed that of Kogan which involved a superimposition
of the vicinal diol of the candidate structure with the
calcium binding diol of the fucose moiety of sLeX. In
our case, it happens that such a superimposition of the
vicinal diol moiety is possible for one enantiomer only
Scheme 1. Synthesis of sLeX mimic 8.
Table 1. Biological activity of the enantiomeric pair
Compound
IC50 (mM)
E-selectin
P-selectin
sLeX
Enantiomer A
Enantiomer B
5.0
7.0
5.0
5.5
5.0
4.5

1112
M. De Vleeschauwer et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1109–1112
raising the possibility that chelation of the calcium by
the vicinal diol may adopt different orientations of
similar energies depending on the enantiomer tested. On
the other hand, there remains the more remote possi-
bility where the ‘wrong’ enantiomer could bind in a
totally different manner but fortuitously with the same
energy.
7. Wong, C.-H. Acc. Chem. Res. 1999, 32, 376.
8. (a) Poppe, L.; Brown, G. S.; Philo, J. S.; Nikrad, P. V.;
Shah, B. H. J. Am. Chem. Soc. 1997, 119, 1727. (b) Scheffler,
K.; Ernst, B.; Katopodis, A.; Magnani, J. L.; Wang, W. T.;
Weisemann, R.; Peters, T. Angew. Chem., Int. Ed. Engl. 1995,
34, 1841.
9. DeFrees, S. A.; Gaeta, F. C. A.; Lin, Y.-C.; Ichikawa, Y.;
Wong, C.-H. J. Am. Chem. Soc. 1993, 115, 7549.
10. (a) Graves, B. J.; Crowther, R. L.; Chandran, C.; Rum-
berger, J. M.; Li, S.; Huang, K.-S.; Presky, D. H.; Familletti,
P. C.; Wolitzky, B. A.; Burns, D. K. Nature 1994, 367, 532.
(b) Weis, W. I.; Drickamer, K.; Hendrickson, W. A. Nature
1992, 360, 127.
11. All the molecular graphics were performed using InsightII
software (Molecular Simulations Inc., San Diego, CA). The
minimization and docking protocols were done using Discover
and the cff91 forcefield (MSI, San Diego, CA). No water
molecules were included in this computation.
12. Bloss, A. S.; Brook, P. R.; Ellam, R. M. J. Chem. Soc,
Perkin Trans. 2. 1973, 2165.
Whatever the exact reason, the present experimental
results do not justify a choice at the moment but
they clearly demonstrate that low molecular weight
noncarbohydrate mimics of sLeX can be generated using
the simple cis-decalinic scaffold. Furthermore, because
of its simplicity and the diversity of access routes to a
large number of its analogues, it is expected that addi-
tional pharmacophoric groups can be introduced to
access additional binding and better selectivity.
13. Trost, B. M.; Salzmann, T. N.; Hiroi, K. J. Am. Chem.
Soc. 1976, 98, 4887.
Acknowledgements
14. Brougham, P.; Cooper, M. S.; Cummerson, D. A.; Hea-
ney, H.; Thompson, N. Synthesis 1987, 1015.
15. Oppolzer, W.; Petrzilka, M. J. Am. Chem. Soc. 1976, 98,
6722.
16. Sharpless, K. B.; Amberg, W.; Bennani, Y.; Crispino,
G. A.; Hartung, J.; Jeong, K.-S.; Kwong, H.-L.; Morikawa,
K.; Wang, Z.-M.; Xu, D.; Zhang, X.-L. J. Org. Chem. 1992,
57, 2768.
17. Anastasia, M.; Allevi, P.; Ciuffreda, P.; Fiecchi, A.; Scala,
A. J. Chem. Soc., Perkin Trans. 1. 1986, 2117.
18. (a) Hagiwara, H.; Uda, H. Chem. Commun. 1987, 1351.
(b) Barton, D. H. R.; McCombie, S. W. J. Chem. Soc., Perkin
Trans. 1. 1975, 1574.
The authors thank the Natural Sciences and Engineer-
ing Research Council of Canada (NSERC) and Bio-
Mega Boehringer Ingelheim Research Inc. for financial
support. The authors are also indebted to Dr. B. Seed
from the Massachusetts General Hospital (Boston, MA
02114, USA) for generously providing the expression
vectors for E-selectin-Ig and P-selectin-Ig.
References and Notes
1. Vestweber, D.; Blanks, J. E. Physiol. Rev. 1999, 79, 181.
2. Dıaz-Gonzalez, F.; Sanchez-Madrid, F. Immunol. Today.
1998, 169.
19. For an example of deacetonization by a palladium cata-
lyst: Schmeck, C.; Hegedus, L. S. J. Am. Chem. Soc. 1994, 116,
9927.
3. (a) Phillips, M. L.; Nudelman, E.; Gaeta, F. C. A.; Perez,
M.; Sinhgal, A. K.; Hakomori, S.-I.; Paulson, J. C. Science
1990, 250, 1130. (b) Walz, G.; Aruffo, A.; Kolanus, W.; Bev-
ilacqua, M.; Seed, B. Science 1990, 250, 1132.
20. Whitesell, J. K.; Reynolds, D. J. Org. Chem. 1983, 48,
3548.
21. Pietraszkiewicz, M.; Jurczak, J. Tetrahedron 1984, 40,
2967.
4. Ramphal, J. Y.; Zheng, Z.-L.; Perez, C.; Walker, L. E.;
DeFrees, S. A.; Gaeta, F. C. A. J. Med. Chem. 1994, 37, 3459.
5. Kogan, T. P.; Dupre, B.; Keller, K. M.; Scott, I. L.; Bui,
H.; Market, R. V.; Beck, P. J.; Voytuis, J. A.; Revelle, R. M.;
Scott, D. J. Med. Chem. 1995, 38, 4976.
6. (a) Simanek, E. E.; McGravey, G. J.; Jablonowski, J. A.;
Wong, C.-H. Chem. Rev. 1998, 98, 833. (b) Roy, R. In Car-
bohydrates in Drug Design; Witezak, Z. J., Nieforth, K. A.,
Eds.; Marcel Dekker: New York, 1997; pp 83–135.
22. Leblanc, Y.; Fitzsimmons, B. J.; Adams, J.; Perez, F.;
Rokach, J. J. Org. Chem. 1986, 51, 789.
1
23. All products gave satisfactory H NMR, 13C NMR, IR,
MS, and elemental analysis.
24. Aruffo, A.; Kolanus, W.; Walz, G.; Fredman, P.; Seed, B.
Cell 1991, 67, 35.
25. Kretzschmar, G.; Sprengard, U.; Kunz, H.; Bartnik, E.;
Schmidt, W.; Toepfer, A.; Horsch, B.; Krause, M.; Seiffge, D.
¨
Tetrahedron 1995, 51, 13015.