Carbohydrate-based scaffolds for the generation of sortiments of bioactive compounds

Article abstract of DOI:10.1007/s007060200016

The polyfunctionality and conformational rigidity of carbohydrates make this class of compounds ideal scaffolds for the production of sortiments¹ of bioactive compounds. Examples of carbohydrate-derived peptidomimetics of biological interest, such as somatostatin agonists and integrin antagonists, are presented. In order to have access to solid phase supported sortiments of compounds, orthogonally protected or unprotected carbohydrates were linked to polymers and reacted in the solid phase employing different regioselective strategies. Original bicyclic and tricyclic glycidic scaffolds were easily obtained starting from natural sugars such as D-arabinose and D-fructose. Manipulation of these conformationally blocked compounds afforded different carbohydrate-based derivatives, among which azidoacids are useful precursors of β-turn peptidomimetics.

Full text of DOI:10.1007/s007060200016

Monatshefte fuÈr Chemie 133, 369±382 ꢀ2002)
Invited Review
Carbohydrate-Based Scaffolds
for the Generation of Sortiments
of Bioactive Compounds
Francesco Peri, Laura Cipolla, Eleonora Forni,
Ã
and Francesco Nicotra
Department of Biotechnology and Biosciences, University of Milano-Bicocca,
I-20126 Milano, Italy
Summary. The polyfunctionality and conformational rigidity of carbohydrates make this class of
1
compounds idealscaffodl s for the production of sortiments of bioactive compounds. Examples of
carbohydrate-derived peptidomimetics of biological interest, such as somatostatin agonists and in-
tegrin antagonists, are presented. In order to have access to solid phase supported sortiments of
compounds, orthogonally protected or unprotected carbohydrates were linked to polymers and reacted
in the solid phase employing different regioselective strategies. Original bicyclic and tricyclic glycidic
scaffolds were easily obtained starting from natural sugars such as D-arabinose and D-fructose. Ma-
nipulation of these conformationally blocked compounds afforded different carbohydrate-based de-
rivatives, among which azidoacids are usefulprecursors of ꢀ-turn peptidomimetics.
Keywords. Carbohydrates; Combinatorialchemistry; Scaffodl s; Peptidomimetics.
Introduction
The relevant role played by carbohydrates in biological recognition phenomena lies
in the unique feature of this class of natural compounds which combine high
combinatorialpotentialand conformationalrigidity. Monosaccharides, if compared
to amino acids, contain more than one stereocentre ꢀin general5 for hexoses), are
present in the pyranosidic or furanosidic form and, interestingly, the cyclic structure
induces a unique conformationalrigidity fundamentalfor biologicalinteractions of
oligosaccharides. It is now clear that the complex structure of oligosaccharides is
only partially involved in the weak binding with their receptors; the remaining part
appears to act as a scaffold that orients the binding determinants in the appropriate
conformation and provides a connection to the aglycons. An interesting example is
Ã
Corresponding author. E-mail: francesco.nicotra@unimib.it
1
Following the argumentation of Roald Hoffmann ꢀHoffmann R ꢀ2001) Angew Chem 113: 3439),
the terminus `sortiment' is to be preferred over `library' in this context

370
F. Peri et al.
Fig. 1. The tetrasaccharide sialyl Lewisx
represented by the tetrasaccharide sialyl Lewisx ꢀsLex, Fig. 1) that interacts with
selectins, the cell adhesion molecules expressed at the surface of vascular endo-
thelial cells after injury. This recognition triggers the anti-in¯ammatory response
and the metastasis formation, sLex being a determinant of neutrophils and metas-
tatic tumor cells. Extensive studies have shown that the structural requirements
for the biological activity of this tetrasaccharide are mainly the presence of the
carboxylic group of neuraminic acid and the hydroxyl groups of L-fucose ꢀas high-
lighted in Fig. 1). This observation has prompted the synthesis of a variety of
analogues in which the glycidic parts of the molecule not involved in the recog-
nition are replaced by spacers with lower molecular, and hence synthetic, com-
plexity and with improved metabolic stability. The studies have pointed out that the
conformationalrigidity induced by the gylcidic rings is fundamentalfor the
biological activity. The substitution of glycidic units with acyclic moieties causes a
drop in the biological activity due to the gain of conformational freedom; on the
contrary, the insertion of rigid scaffolds with limited conformational mobility
restores an adequate biological activity. For instance, the 3,4-disubstitututed N-
acetylglucosamine moiety of sLex can be substituted by 1,2-trans-cyclohexanediol
ꢀFig. 2) in the synthesis of a highly effective mimic [1]. Furthermore, the orien-
tation of the carboxylic group of neuraminic acid is quite important for an effective
adhesion to E-selectins. Compound 2 ꢀFig. 2) lacks the pyranosidic ring of neur-
aminic acid and has a lower biological activity than compound 3 in which the
methylcyclohexyl substituent decreases the mobility of the carboxylic group [2].
Fig. 2. Simpli®ed mimic of sialyl Lewisx

Carbohydrate-Based Scaffold in Combinatorial Chemistry
371
From these observation it is clear that the structural diversity and conformational
rigidity of carbohydrates render these compounds interesting scaffolds capable
to link to different pharmacophoric groups and orient them in a precise spatial
arrangement.
Glycidic Scaffolds
An interesting example of the use of monosaccharides as scaffolds for the synthesis
of bioactive compounds is represented by the work of Hirschmann et al. [3]
describing the synthesis of bioactive somatostatin mimics from the naturalsugar D-
glucose. The peptide hormone somatostatin ꢀ4, Fig. 3) is a cyclic tetradecapeptide
which inhibits the release of several hormones including the growth hormone ꢀGH).
Since its biological activity is due to the interaction of the ꢀ-turn type I fragment
Phe-Trp-Lys-Thr ꢀhighlighted in Fig. 3) with the receptor, the aminoacidic side
chains responsible for the interaction have been attached to glycidic scaffolds
in a proper spatialarrangement. Figure 4 describes a prototype of a carbohydrate-
derived somatostatin mimic ꢀ5) prepared from commercially available 2,3,4,6-
tetra-O-acetyl-ꢁ-D-glucopyranosyl bromide. Employing different glycidic starting
materials, other analogues lacking the benzyloxy group alternatively in position 2,
3, or 4 of the sugar have also been synthesized [3]. More recently, this approach
has been extended to a model study in solution for the preparation of solid-phase
sortiments of compounds based on the ꢁ-D-glucose scaffold [4].
Another example of the use of carbohydrates to mimic non-carbohydrate
bioactive compounds has been reported by Nicolau et al. who synthesized mimics
such as compound 6 ꢀFig. 5) of the Arg-Gly-Asp ꢀRGD) sequence [5]. The RGD
sequence is responsible for the adhesion of extracellular matrix glycoproteins, such
as ®bronectin, ®brinogen, or vitronectin, to endothelial cells ꢀFig. 6), a phenomenon
that promotes, inter alia, angiogenesis in the case of ꢁVꢀ3-integrins and platelet
Fig. 3. The hormone somatostatin

372
F. Peri et al.
Fig. 4. A carbohydrate derived somatostatin mimic
Fig. 5. Carbohydrate derived mimic of the RGD sequence
Fig. 6. Schematic representation of RGD-Integrin interaction
aggregation in the case of ꢁIIbꢀ3-integrins [6]. Kessler and co-workers have
developed the synthesis of a series of amino acids derived from sugars ꢀSAA) with
different turn-inducing properties [7]. The SAA derived from glucosyluronic acid
methylamine ꢀGum) was incorporated in cyclic RGD peptides that proved active as
selective ꢁVꢀ3-integrin antagonists [8]. In all these examples, the potentials of the
glycidic scaffold have only partially been exploited, since only a couple of hydroxyl

Carbohydrate-Based Scaffold in Combinatorial Chemistry
373
Fig. 7. Different pharmacophores linked to a solid phase supported carbohydrate
groups out of ®ve were involved in the derivatization. Examples in which the whole
potential diversity of the sugar is exploited in the preparation of glycidic scaffolds
have been reported by Kunz [9], Hirschmann [4], and Hanessian [10] who were able
to differentiate each hydroxyl group of the sugar. The problem consists mainly in the
orthogonal protection of numerous hydroxyl groups; once this problem will be
completely solved, orthogonally protected carbohydrates could be used in combina-
torialchemistry, and compound sortiments in which the diversity consists in the
different arrangement of pharmacophores linked to the hydroxyl groups of
carbohydrates will be ef®ciently produced ꢀFig. 7). It is interesting to note that in
addition to hydroxylgroups, naturalsugars contain other functionalities such as the
amino or the carboxylic group which further enhance the potential diversity.
The development of orthogonalities among the numerous hydroxyl protecting
groups of sugars is a particularly tedious task. Furthermore, if the combinatorial
approach is developed in the solid phase, the linkage of the sugar to the solid
support must be retained throughout the synthesis and cleaved selectively at the
end of the process. As each derivatization step requires two reactions ꢀdeprotection
and derivatization), a hexose linked to the resin via a hydroxylgroup has stilfour
positions to be manipulated, which means eight reactions. In oligosaccharide syn-
thesis the protective groups are classi®ed as temporary and permanent protections,
the latter being removed at the end of the synthesis without affecting the glycosidic
linkages. A permanent protection should link the glycidic scaffold to the resin and
ensure complete orthogonality to the commonly used temporary protecting groups
and to the derivatization conditions.
A variety of groups including silanes [11], thioethers [12], benzylidene acetals
[13], succinamides [14], photolabile esters [15], p-acylaminobenzyl esters [16],
alkenes [17], tris-ꢀalkoxy)-benzyl amines ꢀBAL) [18], and phenoxyacetates [19]
have been employed to anchor monosaccharides to polymers in the ®rst step of the
solid-phase oligosaccharide synthesis. Most of these linkers interfere with some
common deprotection or derivatization conditions, thus limiting the versatility and
¯exibility in synthetic planning.
Examples of solid phase supported orthogonally protected glucoderivatives are
reported in Schemes 1±3. Starting from D-glucose ꢀ7) and exploiting different
protections as shown in the Schemes, phenylthio 3-O-allyl-4-O-p-methoxybenzyl-ꢀ-
D-glucopyranoside ꢀ19) and phenylthio 2-O-allyl-4-O-p-methoxybenzyl-ꢀ-D-
glucopyranoside ꢀ20) linked to a polystyrene-divinylbenzene-NH2 resin through a

374
F. Peri et al.
Scheme 1
Scheme 2

Carbohydrate-Based Scaffold in Combinatorial Chemistry
375
Scheme 3
succinamide linker have been synthesized [20]. In these compounds, a hydroxyl
group ꢀin position 2 or 3) is deprotected and a second one ꢀ3 or 2) is protected as allyl
ether which was selectively hydrolyzed on solid phase by transition metal catalyzed
isomerization of the double bond followed by acid treatment. The third OH ꢀat
position 4) is protected as p-methoxybenzyl ether that was cleaved selectively on the
solid phase either with oxidants such as DDQ or under acidic conditions ꢀ50% TFA in
CH2Cl2). Finally, the anomeric position was selectively functionalized on the solid
phase by reaction of the thiophenyl group activated as sulfoxide with an alcohol in
the presence of tri¯ic anhydride and a base. After regioselective decoration of the
four independent positions, the succinamide linker, which turned out to be stable
under the deprotection and functionalization conditions, was hydrolyzed with bases.
Schemes 1±3 clearly indicate the complexity underlying the preparation of
orthogonally protected substrates to be linked to a solid support. It is clear that the
possibility of a regioselective derivatization of the free hydroxyl groups of unpro-
tected sugars bound to solid supports would be much more interesting and effective.
In order to explore this opportunity, we investigated the behaviour of the mono-
saccharides methyl ꢁ-D-glucopyranoside ꢀ21), methyl ꢁ-D-mannopyranoside ꢀ22),
and methyl ꢀ-D-galactopyranoside ꢀ23) linked to a trityl resin by the primary
hydroxylgroup. When treated with Bu SnO and then with benzoylcholride, the
2
benzoylation of the stannylidene intermediates ꢀScheme 4) occurred in positions 2,
3, and again 3, respectively, with a regioselectivity above 98% [21]. These results
are promising in the perspective of decorating polymer-bound monosaccharides
regioselectively with a series of functional groups ꢀincluding other monosacchar-
ides in the stannylene-assisted regioselective glycosylation). In this context, new

376
F. Peri et al.
Scheme 4
opportunities can arise from lipase-catalyzed acylation of sugars on solid phase,
once the problems related to the permeability of the resin to the enzymes and to
solid-phase kinetics are solved.
Conformationally Constrained Glycidic Scaffolds
In the development of glycidic scaffolds for the production of sortiments of com-
pounds of pharmaceuticalinterest, two main aspects must be taken into account. First
of all, the originality of the scaffold is important in order to allow patenting, and
naturalsugars are of course not idealsubstrates from this point of view. Secondly, the
conformational rigidity should be ideally enhanced; this strategy affords much more
active compounds once the correct conformation has been achieved. Having both of
these aspects in mind, we studied the possibility to synthesize original polycyclic
scaffolds starting from perbenzylated sugars by an easy debenzylation-iodocycliza-
tion procedure [22]. Starting from polybenzylated sugars with an anomeric append-
age with a CˆC double bond, a cyclic iodoether was obtained by treatment with
iodine. The mechanism of this iodocyclization consists in the attack of the inter-
mediate iodonium ion by the benzyloxy group in ꢂ-position with respect to the
double bond with loss of the benzyl group ꢀScheme 5). Two examples of application
of this reaction are reported in Schemes 6 [23] and 7 [24]. The allylic substituent has
been introduced at the anomeric centre of polybenzylated acetyl arabinofuranoside
30 and methylfructofuranoside 36 affording the corresponding allyl C-glycosides 31
and 37. Treatment of 31 and 37 with iodine results in the iodocyclization. In the case
of the arabinoderivative 31, in which a mixture of ꢁ- and ꢀ-C-glycosides is obtained,
it is possible to effect a kinetic resolution due to the easier formation of the cis-fused
bicyclic structure 32. Besides the formation of a second cycle, the iodocyclization

Carbohydrate-Based Scaffold in Combinatorial Chemistry
377
Scheme 5
Scheme 6
reaction gives rise to a new usefulfunctionalgroup, the alkyliodide, which can be
exploited for further functionalizations. In the examples reported in Schemes 6 and 7,
treatment of the iodides 32 and 38 with tetrabutylammonium azide afforded the
azides 33 or 39, precursors of the corresponding amines. A carboxylic function
was subsequently introduced in these glycidic scaffolds in order to obtain sugar-
aminoacids. The primary hydroxylgroup of 33 and 39 was selectively deprotected by
acetolysis of the benzyl ether followed by basic hydrolysis of the acetate. The
primary hydroxylgroup of the obtained compounds was oxidized using the Jones
reagent affording the azidoacids 35 and 40.
The conformationalanaylsis of azidoacidic scaffodl ꢀ S)-35 and ꢀR)-35 was
performed by NMR spectroscopy and molecular modelling. Both compounds

378
F. Peri et al.
Scheme 7
Scheme 8
turned out to be good isosters of the dipeptide in position i ‡ 1 and i ‡ 2 of a
ꢀ-turn. Compound ꢀS)-35 was incorporated as amino acid in a cyclic RGD peptide
according to the synthetic procedure shown in Scheme 8 [23]. Compound 44 was

Carbohydrate-Based Scaffold in Combinatorial Chemistry
379
tested as inhibitor of the adhesion of fetal bovine aortic endothelial GM7373 cells
to their ligands: ®broblast growth factor 2 ꢀFGF-2), vitronectin ꢀVN), and ®bro-
nectin ꢀFN). A selective inhibition of the adhesion of FGF-2 and VN was observed,
indicating a role of 44 as ꢁVꢀ3 integrin antagonist [25].
Alternatively, the cyclic iodoethers obtained by iodocyclization-debenzylation
undergo reductive elimination if treated with Zn and acids, restoring the original
double bond ꢀScheme 9). The ®nal result in this case is the selective debenzylation
on the hydroxylgroup in ꢂ-position to the double bond. The free hydroxyl group
can therefore be manipulated in order to perform further transformations. In
the case reported in Scheme 9, the free hydroxylgroup of compound 45 has
Scheme 9

380
F. Peri et al.
Scheme 10
been oxidized with Swern reagent, and the aldehyde 46 was treated with vinyl
magnesium bromide in order to introduce a further CˆC double bond. Compound
47 is obtained in 97.5% de, the attack of the Grignard reagent occurring from the
re-face of the carbonylgroup. The two C ˆC double bonds of compound 47 are in
the correct position to be involved in two different iodocyclizations; in fact, the
double bond of the allylic appendage at the anomeric centre can generate a cyclic
iodoether with the free hydroxylgroup, whereas the double bond of the vinylgroup
can originate a cycle with the oxygen of the benzyl group in position 3 of the parent
fructoside. Interestingly, the stereochemistry of the vinylation is correct for two
cyclization in the same molecule. We have been able to obtain alternatively each
bicyclic scaffold 48 or 49 depending on the experimentalconditions. For short
reaction times in THF with 1 equivalent of iodine, compound 48 is formed chemo-
and stereoselectively, whereas in dichloromethane compound 49 was obtained in
70% yield and 71% de. By treatment with an excess of iodine in dichloromethane,
as expected both 48 and 49 afforded the tricyclic scaffold 50. Alternatively, the
polyfunctional glycidic scaffold 52, bearing ®ve different functionalities ꢀa car-
boxylic acid, a ketone, an azide, a primary and a secondary protected hydroxyl
group), has been obtained ꢀScheme 10).
Conclusions
The present review focuses on the use of carbohydrates or carbohydrate-derived
compounds as scaffolds. Carbohydrates present an attractive option for non-peptide
scaffolding as they contain readily convertible functional groups on a pyran or
furan ring with a well-de®ned spatial orientation. Playing with carbohydrates, it
is possible to produce an unexpected variety of new conformationally rigid

Carbohydrate-Based Scaffold in Combinatorial Chemistry
381
polyfunctional molecules. Searching for new pharmacologically active compounds,
sortiments of molecules have been obtained in solid phase from regioselective
decoration of carbohydrates with pharmacophores. The bioactive compounds
containing a carbohydrate or carbohydrate-derived core have generally a reduced
conformational mobility. Consequently, they are also excellent candidates for
structure-activity investigations.
References
[1] Kolb HC, Ernst B ꢀ1997) Chem Eur J 3: 1571
[2] a) Simanek EE, McGarvey GJ, Jablonowski JA, Wong C-H ꢀ1998) Chem Rev 98: 833;
b) Sibata K, Hiruma K, Kante O, Wong C-H ꢀ2000) J Org Chem 65: 2393; c) Jahnke W,
Kolb HC, Blommers MJJ, Magnani JL, Ernst B ꢀ1997) Angew Chem Int Ed 36: 2603;
d) Thoma G, Kinzy W, Bruns C, Patton JT, Magnani JL, Bantli R ꢀ1999) 42: 4909
[3] a) Hirschmann R, Nicolau KC, Pietranico S, Salvino J, Laehy EM, Splengeler PA, Furst G,
Smith III AB, Strader CD, Cascieri MA, Candelore MR, Donaldson C, Vale W, Maechler L
ꢀ1992) J Am Chem Soc 114: 9217; b) Hirschmann R, Nicolau KC, Pietronico S, Leay EM,
Salvino J, Arison B, Cichy MA, Spoors PG, Shakespeare WC, Sprengeler PA, Hamley P,
Smith III AB, Reisine T, Raynor K, Maechler L, Donaldson C, Vale W, Freidinger RM,
Cascieri MR, Strader CD ꢀ1993) J Am Chem Soc 115: 12550; c) Hirschmann R,
Wenquing Y, Cascieri M, Strader CD, Maechler L, Cichy-Knight MA, Hynes J Jr,
van Rijn RD, Sprengeler P, Smith III A ꢀ1996) J Med Chem 39: 2441; d) Hirschmann R,
Hynes J Jr, Cichy-Knight MA, van Rijn RD, Sprengeler PA, Spoors PG, Shakespeare WC,
Pietronico-Cole S, Barbosa J, Liu J, Yao W, Rohrer S, Smith III AB ꢀ1998) J Med Chem 41: 1382
[4] Hirschmann R, Ducry L, Smith AB III ꢀ2000) J Org Chem 65: 8307
[5] Nicolau KC, Trujillo JI, Chibale K ꢀ1997) Tetrahedron 53: 8751

[6] a) Aumalley M, Gurrah M, Muller G, Calvete J, Timpl R, Kessler H ꢀ1991) FEBS Lett 291: 50;
b) Muller G, Gurrath M, Kessler H, Timpl R ꢀ1992) Angew Chem 31: 326; c) Haubner R,

Gratias R, Diefenbach B, Goodman SL, Jonczyk A, Kessler H ꢀ1996) JAm Chem Soc 118: 7461;
d) Giannis A, Rubsam F ꢀ1997) Angew Chem 109: 606; e) Bach ACII, Espina J, Jackson SA,
Stouten PFW, Duke JL, Mousa SA, De Grado WF ꢀ1996) J Am Chem Soc 118: 293
[7] a) Von Roedern EG, Kessler H ꢀ1994) Angew Chem Int Ed 33: 687; b) Von Roedern EG, Lohof G,
Hessler G, Hoffmann M, Kessler H ꢀ1996) J Am Chem Soc 118: 10165
[8] a) Lohof E, Planker E, Mang C, Burkhart F, Dechantsreiter MA, Haubner R, Wester HJ,
Schwaiger M, Holzemann G, Goodman SL, Kessler H ꢀ2000) Angew Chem Int Ed 39: 2761;
b) Haubner R, Wester HJ, Burkhart F, Senekowitsch-Schmidtke R, Weber W, Goodman SL,
Kessler H, Schwaiger M ꢀ2001) J Nuc Med 42: 326; c) Gibson C, Sulyok GAG, Hahn D,
Goodman SL, Holzemann G, Kessler H ꢀ2001) Angew Chem Int Ed 40: 165
[9] Kallus C, Opatz T, Wunberg T, Schmidt W, Henke S, Kunz H ꢀ1999) Tetrahedron Lett 40: 7783
[10] Hanessian S, Saavedra OM, Xie F, Amboldi N, Battistini C ꢀ2000) Bioorg Med Chem Lett 10:
439
[11] a) Danishefsky SJ, Mc Clure KF, Randolph JT, Ruggieri RB ꢀ1993) Science 260: 1307; b) Doi T,
Sugiki M, Yamada H, Takahashi T, Porco JA Jr ꢀ1999) Tetrahedron Lett 40: 2141
[12] Zhu T, Boons G-J ꢀ1998) Angew Chem Int Ed 37: 1898
[13] a) Chiu SHL, Anderson L ꢀ1976) Carbohydr Res 50: 227; b) Rademann J, Schmidt RR ꢀ1996)
Tetrahedron Lett 37: 3989
[14] Lampe TFJ, Weitz-Schmidt G, Wong CH ꢀ1998) Angew Chem Int Ed 37: 1707
[15] a) Rodebaugh R, Joshi S, Fraser-Reid B, Geysen HM ꢀ1997) J Org Chem 62: 5660; b) Zehavi T,
Patchornik A ꢀ1973) J Am Chem Soc 95: 5673
[16] Fukase K, Nakai Y, Egusa K, Porco JA ꢀ1999) Synlett 7: 1074

382
F. Peri et al.: Carbohydrate-Based Scaffold in Combinatorial Chemistry
[17] a) Knerr L, Schmidt RR ꢀ1999) Synlett 11: 1802; b) Melean LG, Haase W-C, Seeberger PH
ꢀ2000) Tetrahedron Lett 41: 4329; c) Andrade RB, Plante J, Melean LG ꢀ1999) Org Lett 11: 1811
[18] Tolborg JF, Jensen KJ ꢀ2000) Chem Commun 2: 147
[19] Wu X, GrathwohlM, Schmidt RR ꢀ2001) Org Lett 3: 747
[20] Peri F, Nicotra F ꢀunpublished results)
[21] Peri F, Cipolla L, Nicotra F ꢀ2000) Tetrahedron Lett 41: 8587
[22] a) Nicotra F, Panza L, Russo G ꢀ1987) J Org Chem 52: 5627; b) Cipolla L, Lay L, Nicotra F
ꢀ1997) J Org Chem 62: 6678
[23] Peri F, Cipolla L, La Ferla B, Nicotra F ꢀ2000) J Chem Soc Chem Commun 2303
[24] Forni E, Cipolla L, Caneva E, La Ferla B, Peri F, Nicotra F ꢀ2002) Tetrahedron Lett ꢀin press)
[25] Peri F, Bassetti R, Caneva E, De Gioia L, La Ferla B, Presta M, Tanghetti E, Nicotra F ꢀ2002)
J Chem Soc Perkin Trans 1 ꢀin press)
Received September 10, 2001. Accepted November 2, 2001