Synthesis of Clustered Glycoside - Antigen Conjugates by Two One-Pot, Orthogonal, Chemoselective Ligation Reactions: Scope and Limitations

Article abstract of DOI:10.1002/1521-3765(20010105)7:1<230::AID-CHEM230>3.0.CO;2-L

Major histocompatibility class II antigens have been bound to clustered glycosides for selective targeting of the dendritic cell mannose receptor. Di-, tetra-, and octavalent glycoside - antigen conjugates have been obtained after two, orthogonal, hydrazo

Full text of DOI:10.1002/1521-3765(20010105)7:1<230::AID-CHEM230>3.0.CO;2-L

FULL PAPER
Synthesis of Clustered Glycoside ± Antigen Conjugates by Two One-Pot,
Orthogonal, Chemoselective Ligation Reactions: Scope and Limitations
Cyrille Grandjean,* HeÂleÁne Gras-Masse, and Oleg Melnyk*[a]
Abstract: Major histocompatibility class II antigens have been bound to clustered
glycosides for selective targeting of the dendritic cell mannose receptor. Di-, tetra-,
and octavalent glycoside ± antigen conjugates have been obtained after two,
Keywords: antigens ´ carbohydrates
orthogonal, hydrazone/thioether ligations, performed by using thio derivatives of
´ chemoselectivity ´ dendrimers ´
d-mannose, d-galactose, or d(ꢀ)-quinic acid, glyoxylyl (or hydrazino)-N-chloroace-
hydrazones
tylated lysinyl trees, and N-terminal hydrazino (or glyoxylyl) peptide antigens.
Successful one-pot condensations have been developed to account for the nature of
the antigens and the valency of the trees.
the cluster effect.[7] Therefore, in the context of an ongoing
vaccine program, a work plan was devised to synthesize
Introduction
Selective targeting of dendritic cells promises much for
synthetic vaccine strategies. Unique among leukocyte pop-
ulations for their ability to initiate, stimulate, or inhibit the
immune response,[1] dendritic cells are also particularly
notable for their ability to activate helper-T cells, which in
turn affect both cytotoxic T lymphocyte and humoral re-
sponses, by a mechanism involving presentation of exogenous
antigen peptide fragments by the major histocompatibility
class II molecules.[2] Presented antigens are internalized by
the dendritic cells by fluid phase macropinocytosis or by an
uptake mediated by a mannose receptor, then entering into
the endocytic pathway for processing.[3] The latter mode of
internalization has been reported by Engering et al.[4] and Tan
et al.[5] to be between 200 and 10000 times more potent than
endocytosis in in vitro stimulation of T cell clone prolifer-
ation. Thus, specific antigen delivery to the dendritic cells, via
the mannose receptor, should overcome the intrinsic low
immunogenicity of peptides and improve vaccine efficacy.
However, it has been reported that the mannose receptor
can selectively bind and internalize molecules or microorgan-
isms coated with d-mannose, l-fucose, or N-acetyl-d-glucos-
amine residues[6] by means of simultaneous interaction with
several carbohydrate recognition domains, in accordance with
antigens linked to numerous glycosides and to optimize
antigen presentation by varying the nature, number, and
interresidue separations of the glycosides and the antigens. Of
the many natural or synthetic poly-mannosylated or -fucosy-
lated mannose receptor ligands thus far reported,[8] only a few
had been coupled to peptide antigens,[5] and these few, despite
a heavy emphasis upon the essential role played by the
mannose receptor, had been obtained by a strategy of a type
seemingly inappropriate for the repetitive approach desirable
for the preparation of a large set of ligands. Thus, the project
comprised: a) the preparation of dendrimeric cores bearing
two sets of chemocompatible functional groups, b) the syn-
thesis of glycosides and antigens derivatized with comple-
mentary functional groups, and c) the subsequent assembly,
after two orthogonal chemical ligations, of purified, fully
deprotected partners (Scheme 1).
Chemical ligation, which consists of the chemoselective
coupling of two unique and mutually reactive functional
groups in aqueous media, has been developed largely in
protein synthesis, scoring notable successes with both native
chemical ligation and expressed protein ligation.[9] These
methodologies have been extended to the preparation of
peptide ± oligonucleotide conjugates[10] and glycopeptides,
and even to the remodeling of cell surfaces.[11] Still, the
employment in biomolecular synthesis of multiple ligation is
exceedingly rare. Various protein or peptide conjugates have
been obtained using two sequential ligations of either oxime
or thioether.[12] Transposition of native chemical ligation to
solid-phase conditions enabled Canne et al.[13] and Camarero
et al.[14] to prepare proteins and peptides by performing up to
three ligations. Synthesis of template-assembled synthetic
[a] Dr. C. Grandjean, Dr. O. Melnyk, Prof. Dr. H. Gras-Masse
Á
Â
Laboratoire de Synthese, Structure et Fonction des Biomolecules
Â
UMR 8525 du CNRS, Universite de Lille II
Institut de Biologie et Institut Pasteur de Lille 1
rue du Professeur Calmette, BP245
Â
59021 Lille Cedex (France)
Fax : (‡ 33)320871233
E-mail: Cyrille.Grandjean@pasteur-lille.fr
230
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0230 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 1

230±239
group of the first lysine to be
introduced was deliberately not
incorporated into the dendri-
meric scaffold, so as to permit
its later transformation into a
hydrazino group by solid-phase
N-electrophilic amination with
N-Boc-3-(4-cyanophenyl)oxa-
ziridine (BCPO).[20] The reac-
tion between primary amines
and BCPO is known for low
yields, owing to the formation
of Schiff bases between re-
ꢀ
Scheme 1. Double ligation strategy for antigen-associated glycodendrimers. Present study: X ˆ ClCH2CO
;
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ ˆ
X' ˆ HS ; Yˆ NH2 NH (or HCOCO ); Y' ˆ HCOCO (or NH2 NH ); c ˆ SCH2CO ; y ˆ NH N (or
ˆ ꢀ
ꢀ
N NH ).
leased
4-cyanobenzaldehyde
and the unchanged amines.
Adaptation of the N-electrophilic amination to the solid
phase enabled the unwanted imines to be hydrolyzed
selectively back to the starting amines, which could then be
treated once more with BCPO. This cycle was repeated until
proteins (TASPs) provides further examples of stepwise
oxime or thioether ligations.[15] With the exception of solid-
phase chemical ligation, these syntheses required additional
steps to unmask the reactive functional groups. To date,
Canneꢁs preparation of the cMyc-Max factor (a heterodimeric
transcription factor) by means of thioester/oxime ligation,[16]
remains the only example in which a biomolecule has been
synthesized by a one-pot combination of two orthogonal
ligations. Such a strategy has also been suggested by Mutter
et al.[17] for TASP design.
We report here the synthesis of hydrazino (or glyoxylyl) N-
chloroacetyl-l-lysinyl trees and their coupling with thio
derivatives of d-mannose, d-galactose, or d-(ꢀ)-quinic acid,
and N-terminal glyoxylyl (or a-hydrazinoacetyl) peptide
antigens using novel, combined thioether/hydrazone ligation
reactions (Scheme 1).
Results and Discussion
Ligation with hydrazino-N-chloroacetyl-l-lysinyl trees:
As a first approach, an investigation was undertaken of the
reaction between N-chloroacetyl-l-lysinyl cores[18] bearing a
hydrazine function, N-terminal glyoxylyl antigen peptides,
and (2-thioethyl) a-d-mannopyranoside 10. Di- and tetrava-
lent chloroacetyl-lysinyl trees 2 and 5a were elaborated by
using the Fmoc/tert-butyl solid phase peptide synthesis
strategy[19] on a Rink amide resin (Scheme 2). The e-amino
   Á
Abstract in French: Des antig›nes de classe II ont ete lies a des
Scheme 2. Synthesis of hydrazino-(N-chloroacetyl)lysinyl cores. Com-
clusters dꢀhydrates de carbone afin de permettre le ciblage du
pound
1 was synthesized on a Rink amide Nleu AM PS resin;
a) ClCH2COOH (8 equiv/NH2), DIC (4 equiv/NH2), DMF, 1 h; b) TFA/
CH2Cl2 1:99, continuous flow; c) BCPO (1 equiv), CH2Cl2, room temper-
ature, 3 h, then N-benzylhydrazine dihydrochloride (1 equiv), DMF/
AcOH/H2O 85:15:5, 3 Â 10 min, then neutralization with iPrNEt2/CH2Cl2
5:95, 2 Â 2 min, (all operations repeated 4 times); d) TFA/anisole 10:1, 08C,
1 h; e) Fmoc-l-Lys(Fmoc)-OH (4 equiv/NH2), HBTU/HOBt/iPrNEt2
(4:4:12 equiv/NH2), NMP, 40 min; f) piperidine/NMP 20:80, 20 min; 20%
overall yield for 2 and 8‡4% overall yield for 5a and 5b, 3% for 6; Boc ˆ
tert-butyloxycarbonyl; Mtt ˆ 4-methyltrityl; DIC ˆ diisopropylcarbodi-
imide; DMF ˆ dimethylformamide; TFA ˆ trifluoroacetic acid; Fmoc ˆ 9-
fluorenylmethoxycarbonyl; HBTU ˆ N-[1H-benzotriazol-1-yl)-(dimethyl-
amino)methylene]-N-methylmethanaminium hexafluorophosphate N-ox-
ide; HOBt ˆ 1-hydroxybenzotriazole; NMP ˆ 1-methyl-2-pyrrolidone.
Â
Á
mannose recepteur des cellules dendritiques. Des antigenes
Â
Á
 Â
Ã
Á
conjugues a 2, 4 ou 8 glycosides ont ete obtenus grace a deux
Â
Â
Â
Á
ligations orthogonales, hydrazone/thioether, realisees a partir
de derives soufres du d-mannose, du d-galactose ou de lꢀacide
Â
Â
Â
Â
Â
d-(ꢀ)-quinique, dꢀarbres de lysine N-chloroacetyles porteurs
Á
Â
d'un groupe glyoxylyle (ou hydrazino) et d'antigenes modifies
par un groupe hydrazino (ou glyoxylyle) en position N-
terminale. Des condensations en un seul pot efficaces ont pu
Ã
Â
Â
Á
etre developpees en fonction de la nature des antigenes et de la
valence des arbres.
Chem. Eur. J. 2001, 7, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0231 $ 17.50+.50/0
231

C. Grandjean, O. Melnyk, H. Gras-Masse
FULL PAPER
complete transformation was achieved. The procedure was
automated, and proved useful for the synthesis of various
peptides containing Lys(NH2), Orn(NH2), or a-hydrazino-
acetyl residues, which have already been employed success-
fully in hydrazone ligations.[21] Peptidyl resin 1 was first
deprotected by treatment with 1% TFA in CH2Cl2,[22] and
treated with BCPO (Scheme 2). After five cycles and cleavage
from the resin, core 2 was obtained in 20% overall yield. The
second generation core 5a was secured in a much lower yield
and proved to be contaminated by the unmodified, parent
lysinyl tree 5b(5a/5b: 2:1). Conversion was probably limited
as a consequence of steric hindrance. In this case, the overall
yield could not be improved by simply increasing the number
of BCPO cycles: the electrophilic amination process in fact
favored the formation of a more hydrophilic by-product (6),
the NMR and mass spectral data of which were suggestive of a
cyclic structure. Compound 6 arises from intramolecular
substitution of a chlorine atom by the Ne-end of the incoming
hydrazine group during the neutralization step of the N-ami-
nation process.[23] Nonetheless, to validate the double ligation
strategy, we synthesized the tetanus toxoid-derived, promis-
cuous immunogenic peptide TT830±846, augmented with an N-
terminal serine (7) (Scheme 3). Compound 7 was further
oxidized to the a-oxo-aldehyde 8 in 72% yield, upon treat-
ment with sodium metaperiodate.[24] (2-Thioethyl) a-d-man-
nopyranoside 10[23] was selected as the glycoside partner.
Stored as its disulfide derivative 9, this compound was
reduced by n-tributylphosphane prior to use (Scheme 4).
Scheme 4. Preparation of the thio derivatives. a) nBu3P (1 equiv), nPrOH/
H2O 1:1, room temperature, overnight; b) HS(CH2)2NH2, HCl (1 equiv),
K2CO3, H2O, 508C, overnight, then air, room temperature, 24 h; 71%.
Core 2 was first treated with a-oxo-aldehyde 8 in DMF/
citrate-phosphate buffer (pH 5.2) (Scheme 5).[25] After com-
pletion of the reaction (monitored by RP-HPLC), 10 was
introduced into the mixture and the pH was adjusted to 8 ± 8.5
(paper) by addition of solid K2CO3[26±28] to give construct 11 in
an isolated yield of 35% (Table 1, entry 1). A mixture of
hydrazino and amino cores, 5a and 5b, was treated similarly
to give the antigen-bearing cluster 12a, together with the
tetramannosylated l-lysinyl tree 12b, in 37 and 52% yield,
respectively (Figure 1, Table 1, entry 2). These encouraging
results led to the proposition of a novel route, which would
overcome the above-discussed difficulties associated with the
preparation of trees functionalized with a hydrazino group,
avoid the exposure of the antigen to sodium periodate, and be
applicable to a broader range of epitopes. For example, the
hydrazone ligation of peptides terminating in N-proline and
modified with an a-glyoxylyl function is highly problematic.
These derivatives in fact give rise, after nucleophilic intra-
molecular attack on the carbonyl group by the amide nitrogen
of the third amino acid in the chain, to stable cyclic hetero-
cycles.[29] This drawback might be circumvented by following a
reverse strategy: that is, the coupling of glyoxylyl N-(chloro-
acetyl)lysinyl cores with hydrazino peptide antigens.
Ligation with glyoxylyl-N-chloroacetyl-L-lysinyl trees:
Scheme 3. Preparation of the antigens. a) NaIO4 (1.2 equiv), AcOH/0.1m
aq KH2PO4, room temperature, 5 min, then HO(CH2)2OH in excess, 72%;
b) bromoacetic acid (8 equiv), DIC (4 equiv), DMF, 20 min, then added to
peptidyl resin, DMF, 1 h; c) BocNHNH2 (4 equiv), iPrNEt2 (6 equiv),
DMF, overnight; d) TFA/anisole/H2O 95:2.5:2.5, room temperature, 3 h,
40%.
The lysinyl cores 15, 16, and 17 were synthesized in 18 ± 31%
yields on a poly(ethyleneglycol) dimethylacrylamide copoly-
mer (PEGA) resin, using a novel linker[30] derived from
tartaric acid, which generates the glyoxylyl group upon
cleavage from the solid support.[31]
232
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0232 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 1

Clustered Glycoside ± Antigen Conjugates
230±239
Scheme 5. Orthogonal hydrazone/thioether ligation reactions between hydrazino-(chloroacetyl)lysinyl cores, the glyoxylyl peptide antigen 8 and the thio
sugar derivative 10.
Table 1. Assembling of conjugates by double orthogonal ligation.
Entry Thio
compound
Tree Antigen Procedure[a] Comments or
yield [%]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
10
2
8
8
20
20
20
20
20
20
A
A
A
A
A
B
B
B
B
B
35
37
43
40
18
51
41
35
10
10
10
10
10
26
29
10
10
10
10
10
10
26
29
5a
15
16
17
17
17
17
17
17
16
16
16
16
16
16
33
45
32a
32a
32a
32a
32a
32a
32a
aggregation
aggregation
aggregation
40% conversion[d]
59
42
35
B
[b]
[c]
C
C
C
[a] Procedure A: hydrazone ligation in DMF/citrate-phosphate buffer pH 5
3:2, then thioether ligation; procedure B: thioether ligation in DMF/H2O
90:10, K2CO3, then hydrazone ligation in DMF/citrate ± phosphate buffer;
procedure C: hydrazone ligation in 2-methyl-2-propanol/H2O 90:10, then
thioether ligation. [b] Thioether/hydrazone DMF/0.1m NaOAc pH 4.6.
[c] Thioether/hydrazone DMF/H2O/tBuOH 4:1:5. [d] Estimated yield
from RP-HPLC.
Figure 1. RP-HPLC chromatogram of the crude mixture from the
sequential hydrazone/thioether ligation of epitope 8, mixture of trees 5a/
5b and sugar derivative 10. Chromatographic conditions: TSK gel (Toso-
Haas) C18 (110 Š, 2 mm, 4.6  50 mm). Flow rate 1.5 mLminꢀ1, 508C.
Buffer A: 0.05% aqueous TFA. Buffer B: 0.05% TFA in CH3CN/H2O 80/
20. Gradient: 0% B for 5 min, 0 ± 100% B over 10 min, 100% B for 1 min,
0% B for 2 min; the upper inset and the lower inset correspond to the
positive ESI-MS of compound 12b and to ESI-MS true mass scale of
compound 12a, respectively.
The synthesis of an N-proline-terminated epitope HA307±319
was undertaken in parallel (Scheme 3). Modification of the
hydrazino moiety was achieved by acylation, with bromo-
acetic acid anhydride, of the peptidyl resin 18 at its N-termi-
nus to give 19, followed by nucleophilic substitution with
commercial tert-butylcarbazate.[32]
cleavage from the resin by treatment with acid. It proved
necessary to adopt the Fmoc/tert-butyl synthesis strategy
and not to use acetonitrile as an RP-HPLC purification
solvent.[33]
Modified antigen 20 was finally obtained, as the sole
product and in 40% overall yield, after deprotection and
Chem. Eur. J. 2001, 7, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0233 $ 17.50+.50/0
233

C. Grandjean, O. Melnyk, H. Gras-Masse
FULL PAPER
to incomplete substitution, while simultaneously favoring
the disulfide formation that consumes the starting material.
Thus, we considered performing the thioetherification
before the hydrazone ligation in almost dry DMF, since the
glyoxylyl group, being stable, would not be expected to
interfere during the substitution reaction. The thioether-
ification proceeded smoothly to furnish intermediate 23
(Figure 2c), which was further treated with 10. The presence
in the crude mixture of the desired tree 24, together with
reformed disulfide 9, was confirmed by RP-HPLC, in marked
contrast with the profile observed when the alternative
route was applied (compare Figure 2d and b). We were
actually able to obtain the fully assembled octameric con-
struct 24 in a considerably improved yield of 51% (Table 1,
entry 6).[35]
The same conditions were applied to the ligation of antigen
20, lysinyl core 17, and 2-thioethyl a-d-galactopyranoside
26[23] and the thio derivative of d-(ꢀ)-quinic acid 29. The
latter compound was obtained in 71% yield from quinide
27,[36] which was treated with cysteamine in degassed H2O.
Compound 29 was purified and stored as its disulfide
derivative 28, and reduced prior to use (Scheme 4). Quinic
acid was chosen since, apart from its use in evaluating the
scope of the methodology, it can act as a potential mannopyr-
anose mimic to the mannose receptor,[37] whereas galactose
should provide conjugates unrecognizable, a priori, by this
receptor. Octameric constructs 30 and 31 were obtained by
With these building blocks in hand, the next step was to
examine their ability to convert into the target molecules. A
one-pot double orthogonal reaction was used. Compounds 15,
16, and 17 were coupled with hydrazino antigen 20 to furnish
the corresponding ligated fragments (as detected by RP-
HPLC). Then, the second ligation was accomplished by
addition of 10 and adjustment of the pH to 8 ± 8.5, to give
constructs 21, 22, and 24 in 43, 40, and 18% yields,
respectively (Schemes 6 and 7, Table 1, entries 3 ± 5). How-
ever, these buffer conditions were detrimental to the thio-
etherification, which was very sluggish for the preparation of
octavalent conjugate 24, as reflected in the actual yield and
the complex RP-HPLC profile (Figure 2a, b). As explained
previously,[26, 34] the presence of water reduces the nucleophi-
licity of the thiol, thus lowering the reaction rate and leading
Scheme 6. Products of orthogonal hydrazone/thioether ligation reactions between glyoxylyl-(chloroacetyl)lysinyl cores 15 and 16, the hydrazino peptide
antigens, and the thio sugar derivatives.
234
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0234 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 1

Clustered Glycoside ± Antigen Conjugates
230±239
following the sequential thioether/hydrazone procedure,
in 41 and 35% isolated yield, respectively (Table 1, entries 7
and 8).
The next study was of variation in the nature of the antigen.
Thus, compounds 32a and 33 (epitopes TT830±846 and
HSP653±14, modified at their N-termini by an a-hydrazino-
acetyl group) were synthesized as described for compound 20
and obtained in 15 and 32% overall yields (Scheme 3).
Although the hydrazone ligation step was slower than
observed previously, 33 reacted under mild conditions with
lysinyl core 17 and mannose derivative 10 to furnish the
double-ligated construct 34 (Table 1, entry 9) in 45% yield
(60% based on recovered epitope) (Scheme 7). However, 32a
led to a complex mixture when treated under the same
conditions. Tetanus toxin derivative 32a proved to be stable in
a DMF/citrate-phosphate buffer for at least two hours (Fig-
ure 3, entry a), but aggregated in a few minutes upon coupling
either with tree 17 (Figure 3b, Table 1, entry 10) or with
glycosylated trees. The same reaction profile was observed by
analytical RP-HPLC when the reaction was conducted with
the smaller tree-like compound 16 (Table 1, entry 11), or when
switched from citrate ± phosphate to the occasional substitu-
te,[25b±d, 25g] 0.1m acetate buffer (pH 4.6), (Table 1, entry 12). To
enhance the hydrophilic character of the difficult epitopic
sequence for 32a, trifluoroacetate counterions were substi-
tuted with chloride moieties to give 32b. No improvement in
the reaction profile was observed, however. The influence of
the solvent was then examined. The use of DMSO, known to
disrupt aggregation of peptides and to increase the hydrazone
ligation rate considerably (compared to DMF)[25f] was pre-
cluded by the thioether ligation. As an alternative, we had
found that 2-methyl-propan-2-ol had a beneficial effect upon
hydrazone ligations that involved hydrophobic partners such
as lipopeptides.[38] And indeed, performance of the hydrazone
ligation in a tBuOH/DMF/citrate ± phosphate buffer reduced
aggregation and permitted partial coupling (Table 1, entry
13). The condensation of epitope 32a with core 16 in tBuOH/
H2O 90/10 (i.e., in the absence of DMF and salts) was
attempted next (Table 1, entry 14). Hydrazone ligation went
to completion overnight under these conditions, although the
material proved not entirely
Scheme 7. Products of orthogonal hydrazone/thioether ligation reactions
between glyoxylyl(chloroacetyl)lysinyl core 17, the thio sugar derivatives
and the hydrazino peptide antigens.
soluble (Figure 3c and d). The
crude mixture was treated fur-
ther with 2-thioethyl a-d-man-
nopyranoside 10 dissolved in
H2O, in the presence of K2CO3
at room temperature, to give a
clean polysubstitution (Fig-
ure 3e).
Tetramannosylated
epitope conjugate 35 was ob-
tained in 59% yield after RP-
HPLC purification (Scheme 6).
The double ligation was not as
efficient when performed using
a reverse procedure; that is,
thioether followed by hydra-
zone condensation. Nor was it
when the experiments were re-
peated with epitope 32b, with
Figure 2. Synthesis and characterization of octavalent compound 24. RP-HPLC chromatogram of the crude
mixtures from: a) the hydrazone ligation of tree 17 and hydrazino antigen 20; b) the thioether ligation of the
hydrazone intermediate with sugar derivative 10; c) the thioether ligation of tree 17 with 2-thioethyl a-d-
mannopyranoside 10 after 36 h; d) the second ligation with hydrazinoantigen 20 overnight; chromatographic
conditions: TSK gel (TosoHaas) C18 (110 Š, 2 mm, 4.6  50 mm). Flow rate 1.5 mLminꢀ1, 508C. Buffer A: 0.05%
aqueous TFA. Buffer B: 0.05% TFA in CH3CN/H2O 80/20. Gradient: 0% B for 5 min, 0 ± 100% B over 10 min,
100% B for 1 min, 0% B for 2 min; the inset corresponds to the ESI-MS true mass scale of compound 24.
Chem. Eur. J. 2001, 7, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0235 $ 17.50+.50/0
235

C. Grandjean, O. Melnyk, H. Gras-Masse
FULL PAPER
Experimental Section
General: Analytical reverse-phase
high performance liquid chromatogra-
phy (RP-HPLC); separations were
performed either upon
a Beckman
System Gold or on a Shimadzu LC-
9A system, using a TSK super ODS gel
(TosoHaas) (110 Š, 2 mm, 4.6 Â
50 mm) column at
a flow rate of
1.5 mLminꢀ1 (monitoring and analy-
sis), a Vydac C18 (100 Š, 2 mm, 4.6 Â
300 mm) or a Beckman Ultrapore C18
(100 Š, 5 mm, 4.6  250 mm) at a flow
rate of 1 mLminꢀ1 (analysis), with
detection at 215 or 230 nm at 508C.
Semipreparative RP-HPLC separa-
tions were performed on a Shimad-
zu LC-4A system, on a Hypersil hy-
perprep C-18 (300 Š, 8 mm, 10 Â
260 mm) column at a flow rate of 3
or 5 mLminꢀ1, with detection at either
215 or 230 nm at 508C. Solvent system
A: 0.05% TFA in water; solvent
system B: 0.05% TFA in 80% aceto-
nitrile/20% water; solvent system C:
0.05% TFA in 60% acetonitrile/40%
water; solvent system D: 0.03% HCl
in water; solvent system E: 0.03%
HCl in 80% acetonitrile/20% water;
solvent system F: 0.05% TFA in 40%
iPrOH/60% water; solvent system G:
0.03% HCl in 40% iPrOH/60% wa-
ter. ESI-MS spectra were recorded on
a Micromass Quatro II Electrospray
Figure 3. Stability of epitope 32a and synthesis and characterization of tetravalent construct 35. RP-HPLC
chromatograms from: a) and b) epitope 32a in DMF/citrate ± phosphate buffer (pH 5) (3:2, room temperature),
2 h alone and in the presence of tree 17, 10 min, respectively; c) and d) hydrazone ligation of tree 16, epitope 32a
in tBuOH/H2O 90/10, after 1 h and overnight, respectively; chromatographic conditions: TSK gel (Toso-
Haas) C18 (110 Š, 2 mm, 4.6  50 mm). Flow rate 1.5 mLminꢀ1, 508C. Buffer A: 0.05% aqueous TFA. Buffer B:
0.05% TFA in CH3CN/H2O 80/20. Gradient: 0% B for 5 min, 0 ± 100% B over 10 min, 100% B for 1 min, 0% B
for 2 min; the inset corresponds to the ESI-MS true mass scale of compound 35.
Mass Spectrometer. H and 13C NMR spectra were recorded on a Bruker
DRX 300 spectrometer in H2O/D2O 90:10. Chemical shifts are given in
ppm and referenced to internal 3-trimethylsilyl-[D4-2,2,3,3]propionic acid
sodium salt (TMSP).
1
chloride counter-ions. Finally, conjugates 36 and 37 (Table 1,
entries 15 and 16) were assembled, following the hydrazone/
thioether sequence in tBuOH, from core 16, epitope 32a, and
either compound 26 or 29, in 42 and 35% yield, respectively
(Scheme 6).
Compounds were checked for homogeneity by analytical RP-HPLC, using
two different solvent systems (solvent systems A/B and A/F), and by
analytical capillary zone electrophoresis in a 75 mm  500 mm fused silica
capillary, with a 28 mA current and a 30 kV field in an Applied Biosystems
Model 270A-HT system (Foster City, USA). Separations were performed
either at 508C using a 50 mm sodium borate migration buffer at pH 9.2, or a
20 mm citrate buffer at pH 2.47 and at 408C. Lysinyl trees and epitopes
were also characterized by amino acid analysis after total acidic hydrolysis
of the peptidyl resin with 1 mL of propionic acid/12n HCl (1:1) at 1108C
for 24 h.
Conclusion
This study provides new insights into chemoselective ligation.
In particular, the reliability of combined thioether/hydrazone
ligation for the preparation of conjugates has been demon-
strated. N-chloroacetylated lysinyl trees, modified with either
a hydrazino or a glyoxylyl group, proved to be valuable
intermediates. Moreover, the latter are very promising since
they are readily accessible and bear the two sets of electro-
philic groups, thus preventing side reactions. In the context of
our studies, the sequential procedure, using buffered DMF or
2-methyl-propan-2-ol, or employing the reversed order of
ligation reactions, facilitated the preparation of antigens
linked to up to four glycosides. To date, no limitation has
been found regarding the thio-functionalized segment. This
procedure can also circumvent the difficulties related to the
coupling of epitopes terminating in N-proline. Further studies
will be required to extend this methodology to the prepara-
tion of octameric trees linked to difficult epitopic sequences.
However, vectorization of antigens by tetrameric glycoside
clusters seems to be sufficient for targeting the mannose
receptor.[5a]
Synthesis of 2-{Na,Ne-di-[(1Sn,3R,4Sn,5R)-1,3,4,5-tetrahydroxycyclohex-
ane-1-carboxamido-1-yl]}ethyl disulfide 28: A mixture of quinic acid 1,5-
lactone 27 (52 mg, 0.3 mmol) and cysteamine hydrochloride (34 mg,
1 equiv) in degassed H2O containing solid K2CO3 (pH 8 ± 9, paper) was
stirred overnight at 508C. The mixture was further stirred at room
temperature in the presence of air for another 24 h before being purified by
RP-HPLC [gradient: 100 ± 0 to 80 ± 20 (A/C), 15 min, then isocratic] to
furnish 28 (70 mg, 71%) as a white powder after lyophilization: 1H NMR:
d ˆ 8.24 (t, 3J(H,H) ˆ 5.8 Hz, 2H; 2 NH), 4.05 (ddd, 3J(H,H) ˆ 3, 3, and
6.2 Hz, 2H; 2 H-3), 3.90 (ddd, 3J(H,H) ˆ 4.7, 9, and 11.6 Hz, 2H; 2 H-5),
3.40 (t, 3J(H,H) ˆ 6.4 Hz, 4H; 2 CH2), 3.38 (dd, 3J(H,H) ˆ 3 and 9 Hz, 2H;
2 H-4), 2.73 (t, 3J(H,H) ˆ 6.4 Hz, 4H; 2 CH2), 1.94 ± 1.83 (m, 6H; 4 H-2 and
2 H-6), 1.77 (dd, 3J(H,H) ˆ 11.6 and 13.4 Hz, 2H; 2 H-6'); 13C NMR: d ˆ
177.5 (2 CON), 77.8 (2 C-1), 75.4 (2 C-4), 70.8 (2 C-3), 66.7 (2 C-5), 40.7
(2 C-6), 38.5, 37.5, and 37.1 (4 CH2 and 2 C-2); TOF-PDMS: m/z: 469
‡
[M‡H] .
Solid-phase peptide synthesis: All peptidic compounds were synthesized
using the Fmoc/tert-butyl strategy. The coupling steps were performed
using a fourfold excess of amino acid per amine subjected to HBTU/HOBt/
DIEA activation in NMP.
236
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0236 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 1

Clustered Glycoside ± Antigen Conjugates
230±239
Synthesis of the hydrazino-N-chloroacetylated lysinyl trees: These com-
pounds were synthesized manually on a Rink amide-Nleu-AM-PS resin
(substitution, 0.38 mmolgꢀ1) (Senn Chemicals), by using the Fmoc/tert-
butyl strategy. Lysinyl core of valency 2 as synthesized on a 0.25 mmol scale.
Lysinyl core of valency 4 was synthesized on a 0.1 mmol scale. Typically,
with coupling monitored by the TNBS test,[39] HBTU dissolved in NMP
(4 equiv) was added to a mixture of the amino acid (4 equiv), HOBt
(4 equiv), and DIEA (8 equiv) in NMP. After stirring for 1 min, the mixture
was added to the peptidyl resin (1 equiv) solvated in NMP containing
DIEA (4 equiv), and mechanically shaken for 40 min. After filtration, the
peptidyl resin was washed with NMP (3 Â 2 min), and CH2Cl2 (3 Â 2 min).
Each coupling was followed by a capping with Ac2O/DIEA/CH2Cl2 10/5/85
for 10 min, followed by washing with CH2Cl2 (3 Â 1 min). Cleavage of the
Fmoc protecting groups was achieved by treatment with 20% piperidine in
NMP (1 Â 2 min, 1 Â 20 min), followed by washing first with NMP (2 Â
2 min), and then with CH2Cl2 (2 Â 1 min). Fmoc-b-Ala-OH was anchored
to the resin first (for the preparation of the second generation lysinyl cores,
0.26 equiv of Fmoc-b-Ala-OH was coupled, to diminish the load to
0.1 mmol). Chain protection of the first lysinyl residue introduced was
effected by means of a 4-methyltrityl moiety. The other lysinyl residues
were added as their Fmoc-l-Lys(Fmoc)-OH derivatives. After the third
and fourth coupling steps, the peptidyl resin was deprotected and acylated
using a fourfold excess of preformed chloroacetic anhydride, prepared by a
DCC activation, to provide the first and second level carrier cores. The
methyltrityl group of the e-NH2 of the first lysine residues was removed by
a continuous flow of 1% TFA in CH2Cl2. The three cores were then
subjected to the N-electrophilic amination procedure. The peptidyl resins
were placed in an Applied Biosystems 431A peptide synthesizer (Foster
City, CA), and subjected five times to the following cycles: 1) N-Boc-3-(4-
cyanophenyl)oxaziridine (Acros Organics, Noisy le Grand, France)
(1 equiv) in CH2Cl2 (6 mL) for 30 min (Boc-3-(4-cyanophenyl)oxaziridine
was introduced as a solid in the cartridges normally used for peptide
synthesis. The reagent was dissolved in the cartridge with CH2Cl2 and the
solution transferred immediately to the reaction vessel); 2) CH2Cl2 and
DMF washes; 3) N-benzylhydrazine dihydrochloride (50 mg) in DMF/
Ac2O/H2O 85/15/10 (5 mL) for 30 min, 3 times; 4) DMF and CH2Cl2
washes; 5) 5% DIEA in CH2Cl2 for 2 min, twice; 6) CH2Cl2 washes. A
Kaiser test was performed after these cycles to monitor conversion. The
cores were then cleaved from the resin and deprotected by TFA/anisole 10/
1 (11 mL per gram of peptidyl resin), for 1 h at 08C, precipitated in cold
diethyl ether/heptane 1/1, centrifuged, dissolved in water and lyophilized.
4.24 ± 4.01 (m, 4H; 4 Â lysyl a-H), 3.99, 3.97, and 3.94 (3 Â s, 3 Â 2H; 3 Â
COCH2Cl), 3.52 and 3.47 (2 AB systems, 3J(H,H) ˆ 14.2, 2  1H;
ꢀ
ꢀ
ꢀ
COCH2NH ), 3.37 ± 3.17 (m, 2H; b-alanyl b-H), 3.11 ± 2.97 (m, 6H, 6 Â
lysyl e-H), 2.70 ± 2.90 (m, 2H; 2  lysyl e-H), 2.33 (t, 3J(H,H) ˆ 6.7 Hz, 2H;
b-alanyl a-H), 1.68 ± 1.52 (m, 8H, 8 Â lysyl b-H), 1.42 ± 1.27 (m, 8H, 8 Â
lysyl d-H), 1.25 ± 1.13 (m, 8H, 8 Â lysyl g-H); positive ESI-MS: m/z: 886
[M‡H] , 462 [M‡H‡K]2‡
.
‡
Synthesis of the glyoxylyl-N-chloroacetylated lysinyl trees: Synthesis and
full characterization of these constructs will be reported elsewhere.[31]
Synthesis of the antigen peptides: All epitopes were synthesized on a
0.25 mmol scale on
a Rink amide-Nleu-AM-PS resin (substitution,
0.38 mmolgꢀ1) (Senn Chemicals), on an Applied Biosystems A431 peptide
synthesizer. The side chain protecting group for Fmoc-l-Lys-OH was tert-
butoxycarbonyl; the side chain protecting group for Fmoc-l-Ser-OH,
Fmoc-l-Glu-OH, Fmoc-l-Thr-OH, Fmoc-l-Tyr-OH, and Fmoc-l-Asp-OH
was a tert-butyl group; the side chain protecting group for Fmoc-l-Asn-
OH, Fmoc-l-Gln-OH, and Fmoc-l-His-OH was a trityl group; the side
chain protecting group for Fmoc-l-Arg-OH was a 2,2,4,6,7-pentamethyl-
dihydrobenzofuran-5-sulfonyl group.
HCO-CO-TT830±846 (8): The peptide was cleaved from the resin and
deprotected using TFA/H2O/iPr3SiH 95/2.5/2.5 (10 mL per gram of peptidyl
resin), for 3 h at room temperature, precipitated in cold diethyl ether,
centrifuged, dissolved in water and lyophilized. Peptide Ser-TT830±846 (7)
(38.5 mg, 6%) was obtained as a white powder after semi-preparative RP-
HPLC [gradient: 100:0 to 75:25 (A/B), 50 min] purification and lyophili-
‡
zation; TOF-PDMS: m/z: 2068 [M‡H] .
Peptide Ser-TT830±846 (7) (9 mg, 3.56 mmol) was dissolved in aqueous 0.1m
KH2PO4, and the pH adjusted to 7 by adding 1n NaOH. NaIO4 (3.3 mg,
1.2 equiv) dissolved in H2O (300 mL) was then added, and the mixture was
stirred at room temperature for 10 min and the reaction quenched by
adding ethylene glycol (15 mL). The mixture was diluted with 50% aqueous
AcOH, filtered, and purified by RP-HPLC [gradient: 100:0 to 80:20 (D/E),
7 min, then 80:20 to 75:25 (D/E), 18 min], to furnish 8 (5.5 mg, 72%) as a
‡
white powder after lyophilization; TOF-PDMS: m/z: 2059.1 [M‡Na] .
H2N-Gly-HA307±319 (20): Following automated synthesis, one half of the
peptidyl resin (0.125 mmol) was deprotected and coupled twice manually
using a fourfold excess of preformed bromoacetic anhydride (prepared by
DIC activation (20 min) in DMF), followed by washing with DMF (3 Â
1 min) and CH2Cl2 (3 Â 2 min). Then, tert-butylcarbazate (66 mg, 4 equiv)
was added to the peptidyl resin, solvated in DMF containing DIEA
(174 mL, 8 equiv). The mixture was stirred mechanically overnight and then
washed with DMF (3 Â 1 min), CH2Cl2 (3 Â 2 min), and Et2O (2 Â 1 min).
Finally, peptide was cleaved from the resin and deprotected with TFA/
anisole/H2O 95/2.5/2.5 (10 mL) for 3 h at room temperature, precipitated in
cold diethyl ether, centrifuged, dissolved in water, and lyophilized. H2N-
Gly-HA307±319 (20) (101 mg, 40%) was obtained as a white powder after
semi-preparative RP-HPLC purification [gradient: 100:0 to 50:50 (A/F),
Compound 2 (26 mg, 20%) was obtained as a white powder after RP-
HPLC purification [gradient: 100:0 to 90:10 (A/B), 20 min], followed by
lyophilization; 1H NMR: d ˆ 8.37 (d, 3J(H,H) ˆ 6.3 Hz, 1H; lysyl aNH),
8.28 (d, 3J(H,H) ˆ 7.1 Hz, 1H; lysyl aNH), 8.14 (brt, 1H; lysyl eNH), 8.00 (t,
3J(H,H) ˆ 5.7 Hz, 1H; alanyl NH), 7.41 and 6.71 (2 brs, 2  1H; NH2),
4.15 ± 4.06 (m, 2H; 2 Â lysyl a-H), 4.00 and 3.95 (2 s, 2 Â 3H; 2 Â CH2),
3
3.38 ± 3.27 (m, 2H; b-alanyl b-H), 3.10 (dt, J(H,H) ˆ 6.6 and 6.6 Hz, 2H;
‡
lysyl e-H), 2.98 (t, 3J(H,H) ˆ 7.6 Hz, 2H; lysyl e-H), 2.33 (t, 3J(H,H) ˆ
6.4 Hz, b-alanyl a-H), 1.67 ± 1.46 (m, 6H, 4 Â lysyl b-H and 2 Â lysyl d-H),
90 min], and lyophilization; positive ESI-MS: m/z: 1575 [M‡H] .
H2N-Gly-TT830±846 (32a, 32b): After automated synthesis, peptidyl resin
was deprotected and coupled twice manually using a fourfold excess of
preformed bromoacetic anhydride (prepared by DIC activation (20 min) in
DMF), followed by washing with DMF (3 Â 1 min) and CH2Cl2 (3 Â 2 min).
Then, tert-butylcarbazate (66 mg, 4 equiv) was added to the peptidyl resin,
solvated in DMF containing DIEA (174 mL, 8 equiv). The mixture was
stirred mechanically overnight, and then washed with DMF (3 Â 1 min),
CH2Cl2 (3 Â 2 min), and Et2O (2 Â 1 min). Finally, the peptide was cleaved
from the resin and deprotected with TFA/iPr3SiH/H2O 95/2.5/2.5 (10 mL)
for 1.5 h at room temperature, precipitated in cold diethyl ether,
centrifuged, dissolved in water, and lyophilized. One half of the crude
H2N-Gly-TT830±846 was purified by semi-preparative RP-HPLC [gradient:
100:0 to 75:25 (A/F), 25 min, then isocratic (A/F)] to furnish 32a (51 mg,
15%) as a white powder after lyophilization; positive ESI-MS: found:
2052.0; calcd 2052.5; m/z: 1026.7 [M‡2H]2‡, 684.8 [M‡3H]3‡, 514.0
[M‡4H]4‡. The other half of the crude peptide was purified by semi-
preparative RP-HPLC [gradient: 100:0 to 80:20 (D/G), 50 min, then
isocratic (D/G)] to furnish 35b (30 mg, 11%) as a white powder after
lyophilization.
1.43 ± 1.30 (m, 2H; 2 Â lysyl d-H),1.29 ± 1.17 (m, 4H; 4 Â lysyl g-H);
13
ˆ
C NMR: d ˆ 177.1, 174.2, 173.9, 170.2, and 170.0 (5 NC O), 54.6 and
54.0 (2 Â lysyl a-C), 51.5 (lysyl e-C), 39.8 (lysyl e-C), 36.2 and 34.9 (b-alanyl
a-C and b-C), 42.7 and 42.4 (2 Â CH2), 30.8 and 30.7 (2 Â lysyl b-C) 28.0
(lysyl d-C), 24.0 (lysyl d-C), 22.5 and 22.4 (2 Â lysyl g-C); TOF-PDMS: m/z:
‡
‡
576 [M‡(CH3)2C‡Na] , 554 [M‡(CH3)2C] .
Compound 5a (25 mg, 12%) was obtained as a white powder (contami-
nated (ca. 33%) with the unreacted lysinyl core) after RP-HPLC
purification [gradient: 100:0 to 80:20 (A/B), 20 min, then 80:20 to 70:30
(A/B), 25 min], followed by lyophilization together with a more hydro-
philic cyclic by-product, 6 (6 mg, 3%), formed by intramolecular nucleo-
philic substitution of one chloride by the NHe of the hydrazino group (the
ability of this by-product to react with a-oxo-aldehyde to furnish hydrazone
was used to confirm the integrity of the NHe); for 5a and 5b: positive ESI-
‡
‡
MS: m/z: 922 [M‡H] , 480 [M‡H‡K]2‡ and 907 [M‡H] , 473
[M‡H‡K]2‡, respectively. for 6 (3%), 1H NMR: d ˆ 8.34 (d, 3J(H,H) ˆ
7.2 Hz, 1H; lysyl aNH), 8.29 (d, 3J(H,H) ˆ 6.4 Hz, 1H; lysyl aNH), 8.28 (d,
3J(H,H) ˆ 8.2 Hz, 1H; lysyl aNH), 8.19 (d, 3J(H,H) ˆ 8.4 Hz, 1H; lysyl
aNH), 8.10 (t, J(H,H) ˆ 5.9 Hz, 1H; lysyl NH), 8.03 (t, J(H,H) ˆ 6.0 Hz,
1H; lysyl eNH), 8.00 (t, 3J(H,H) ˆ 5.9 Hz, 1H; lysyl eNH), 7.91 (t,
3J(H,H) ˆ 6.0 Hz, 1H; b-alanyl NH), 6.71 and 7.42 (2  brs, 2  H; NH2),
H2N-Gly-HSP653±14 (33): Peptide 33 was cleaved from the resin and
deprotected with TFA/anisole/iPr3SiH/H2O 90/2.5/2.5/2.5 (10 mL) for 2 h at
room temperature, precipitated in cold diethyl ether, centrifuged, dissolved
3
e
3
Chem. Eur. J. 2001, 7, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0237 $ 17.50+.50/0
237

C. Grandjean, O. Melnyk, H. Gras-Masse
FULL PAPER
in H2O/AcOH, and lyophilized. One quarter of the crude H2N-Gly-
HSP653±14 was purified by semipreparative RP-HPLC [gradient: 100:0 to
85:15 (A/F), 15 min, then 85:15 to 80:20 (A/F), 10 min, then isocratic (A/
F)] to furnish 33 (39 mg, 32%) as a white powder after lyophilization;
positive ESI-MS: found: 1479.0; calcd 1479.6; m/z: 740.5 [M‡2H]2‡, 494.1
then 85:15 to 75:25 (A/B), 25 min, then isocratic]; positive ESI-MS: found:
3391; calcd 3392.0; m/z: 1696.5 [M‡2H]2‡
,
1131.5 [M‡3H]3‡
, 848.9
[M‡4H]4‡
.
Octavalent conjugate 24: This compound (1.99 mg, 18%) or (5.14 mg,
51%) was obtained following procedure A or B, respectively; RP-HPLC
[gradient 100:0 to 85:15 (A/B), 15 min, then 85:15 to 75:25 (A/B), 20 min,
then isocratic]; positive ESI-MS: found: 5025; calcd 5025.9; m/z: 1676.0
[M‡3H]3‡
.
General procedures for the double ligation: All solvents were degassed by
bubbling N2 through them before use.
[M‡3H]3‡, 1257.3 [M‡4H]4‡
.
Octavalent conjugate 30: This compound (0.90 mg, 41%) was obtained
following procedure B; RP-HPLC [gradient 100:0 to 85:15 (A/B), 15 min,
then 85:15 to 72:28 (A/B), 25 min, then isocratic]; positive ESI-MS: found:
Procedure A: Hydrazino- or a-oxo-aldehyde lysinyl tree (0.5 ± 2 mmol) was
treated overnight at room temperature under nitrogen with modified
antigen peptide (1 equiv) in DMF/citrate-phosphate buffer pH 5 3:2 (500 ±
1000 mL), with monitoring by RP-HPLC. In parallel, the disulfide
derivative (1.5 equiv per chloroacetyl moiety to be substituted), dissolved
in nPrOH/H2O 1/1 (500 mL), was treated at room temperature overnight
under nitrogen with tri-n-butylphosphane (1 equiv per disulfide). Com-
pletion of the reaction was monitored by RP-HLPC. The mixture was
concentrated under reduced pressure for 20 min, and then dissolved in H2O
(100 mL) and added to the former reaction mixture, diluted with DMF
(900 ± 1400 mL, so as to obtain 80% aqueous DMF), and the apparent pH
adjusted to 8 ± 8.5 (paper) by addition of solid potassium carbonate. The
mixture was stirred at room temperature for 24 ± 48 h (monitoring by RP-
HPLC), diluted with H2O, frozen, and lyophilized. Compounds were
obtained as white powders after RP-HPLC and lyophilization.
5024.0; calcd 5025.9; m/z: 1257.5 [M‡4H]4‡, 1006.1 [M‡5H]5‡
.
Octavalent conjugate 31: This compound (1.52 mg, 35%) was obtained
following procedure B; RP-HPLC [gradient 100:0 to 90:10 (A/B), 10 min,
then 90:10 to 75:25 (A/B), 50 min]; positive ESI-MS: found: 5112; calcd
5114.1; m/z: 1704.9 [M‡3H]3‡, 1278.9 [M‡4H]4‡, 1030.9 [M‡5H‡K]5‡
,
859.3 [M‡5H‡K]6‡
.
Octavalent conjugate 34: This compound (0.40 mg, 45%) was obtained
following procedure B; RP-HPLC [gradient 100:0 to 75:25 (A/B), 15 min,
then 75:25 to 70:30 (A/B), 30 min]; positive ESI-MS: found: 4927.5; calcd
4928.6; m/z: 1643.7 [M‡3H]3‡, 1233.2 [M‡4H]4‡, 999.3 [M‡3H‡Na‡K]5‡
.
Tetravalent conjugate 35: This compound (0.45 mg, 59%) was obtained
following procedure C; RP-HPLC [gradient 100:0 to 60:40 (A/B), 70 min];
positive ESI-MS: found: 3867; calcd 3869.5; m/z: 1290.1 [M‡3H]3‡, 968.0
Procedure B: The disulfide derivative (1.5 equiv per chloracetyl moiety to
be substituted), dissolved in nPrOH/H2O 1/1 (500 mL), was treated
overnight with tri-n-butylphosphane (1 equiv per disulfide) at room
temperature under nitrogen. Completion of the reaction was monitored
by RP-HLPC. The mixture was concentrated under reduced pressure for
20 min, and then dissolved in H2O (50 mL) and added to 1 ± 2 mmol of the
glyoxylyl-lysinyl tree (1 equiv) dissolved in DMF (600 mL). The apparent
pH was adjusted to 8 ± 8.5 (paper) by addition of solid potassium carbonate.
The mixture was stirred under nitrogen at room temperature for 24 ± 48 h
and monitored by RP-HPLC. Citrate/phosphate buffer (350 mL) and
modified antigen peptide were then introduced. The pH was adjusted to 5.2
by addition of 1n aqueous HCl. The mixture was stirred for a further 12 h at
room temperature, diluted with H2O, frozen, and lyophilized. Compounds
were obtained as white powders after RP-HPLC purification and
lyophilization.
[M‡4H]4‡, 774.6 [M‡5H]5‡
.
Tetravalent conjugate 36: This compound (0.55 mg, 35%) was obtained
following procedure C; RP-HPLC [gradient 100:0 to 75:25 (A/B), 20 min,
then 75:25 to 65:35 (A/B), 20 min, then isocratic]; positive ESI-MS: found:
3868.5; calcd 3869.5; m/z: 1290.6 [M‡3H]3‡, 968.0 [M‡4H]4‡, 783.6
[M‡5H]5‡
.
Tetravalent conjugate 37: This compound (0.71 mg, 42%) was obtained
following procedure C; RP-HPLC [gradient 100:0 to 75:25 (A/B), 20 min,
then 75:25 to 60:40 (A/B), 30 min]; positive ESI-MS: found: 3912; calcd
3913.6; m/z: 1305.0 [M‡3H]3‡, 979.1 [M‡4H]4‡, 783.6 [M‡5H]5‡
.
Acknowledgements
Procedure C: The disulfide derivative (1.5 equiv per chloroacetyl moiety to
be substituted), dissolved in nPrOH/H2O 1/1 (500 mL), was treated
overnight with tri-n-butylphosphane (1 equiv per disulfide) at room
temperature under nitrogen. Completion of the reaction was monitored
by RP-HLPC. The mixture was concentrated under reduced pressure for
20 min. In parallel, 0.4 mmol of the glyoxylyl-lysinyl tree 16 was treated with
the modified antigen (1 equiv) in tBuOH/H2O (180 ± 20 mL) at room
temperature under nitrogen, and monitored by RP-HPLC. On completion,
the reaction mixture was transferred to the crude thio derivative solution
and the apparent pH adjusted to 8 ± 8.5 (paper) by addition of solid
potassium carbonate. After completion of the reaction, the crude mixture
was diluted with H2O, lyophilized and frozen, and lyophilized once more.
Compounds were obtained as white powders after RP-HPLC and
lyophilization.
This work was supported financially by the CNRS, the Pasteur Institute, the
University of Lille, the ANRS, and the Fondation pour la Recherche
Â
Medicale and the Ensemble contre le Sida (Sidaction grant to CG). We are
grateful to B. Coddeville and G. Montagne for recording ES-MS and NMR
spectra, respectively, and also to S. Brooks for proof-reading the manu-
script and to C. Rommens for her dedication to the project.
[1] J. Blanchereau, R. M. Steinman, Nature 1998, 392, 245 ± 252.
[2] R. N. Germain, Cell 1994, 76, 287 ± 299.
[3] a) R. M. Steinman, J. Swansson, J. Exp. Med. 1995, 182, 283 ± 287; b) F.
Sallusto, M. Cella, C. Danieli, A. Lanzavecchia, J. Exp. Med. 1995,
182, 389 ± 400.
[4] A. J. Engering, M. Cella, D. Fluitsma, M. Brockhaus, E. C. M.
Hoefsmit, A. Lanzavecchia, J. Pieters, Eur. J. Immunol. 1997, 27,
2417 ± 2425.
[5] a) M. C. A. A. Tan, A. M. Mommaas, J. W. Drijfhout, R. Jordens,
J. J. M. Onderwater, D. Verwoerd, A. A. Mulder, A. N. van der Hei-
den, D. Scheidegger, L. C. J. M. Oomen, T. H. M. Ottenhoff, A. Tulp,
J. J. Neefjes, F. Koning, Eur. J. Immunol. 1997, 27, 2426 ± 2435;
b) M. C. A. A. Tan, R. Jordens, A. Geluk, B. O. Roep, T. H. M.
Ottenhoff, J. W. Drijfhout, F. Koning, Int. Immunol. 1998, 10, 1299 ±
1304.
Divalent conjugate 11: This compound (0.55 mg, 35%) was obtained
following procedure A; RP-HPLC [gradient 100:0 to 80:20 (A/B), 10 min,
‡
then 80:20 to 70:30 (A/B), 40 min]; positive ESI-MS: m/z: 2976.2 [M‡K] ,
1469.8 [M‡2H]2‡, 980.2 [M‡3H]3‡, 735.4 [M‡4H]4‡
.
Tetravalent conjugate 12a: The double ligation was carried out on 5a,
which contained about 33% of untransformed lysinyl core 5b, following
procedure A. Compound 12a (1.15 mg, 37%) was obtained, together with a
more hydrophilic compound (12b, 0.91 mg, 52%) after RP-HPLC [gra-
dient 100:0 to 80:20 (A/B), 10 min, then 80:20 to 70:30 (A/B), 40 min]; 12a,
positive ESI-MS: found: 3754; calcd 3755.43; m/z: 1878.7 [M‡2H]2‡
,
1252.6 [M‡3H]3‡, 939.6 [M‡4H]4‡; 12b, positive ESI-MS: m/z: 1722.4
Â
[6] a) A. Avrameas, D. McIlroy, A. Hosmalin, B. Autran, P. Debre, M.
‡
[M‡H] , 861.5 [M‡2H]2‡
.
Monsigny, A.-C. Roche, P. Midoux, Eur. J. Immunol. 1996, 26, 394 ±
400; b) P. D. Stahl, Curr. Opin. Immunol. 1992, 4, 49 ± 52.
[7] a) Y. C. Lee, R. T. Lee, Acc. Chem. Res. 1995, 28, 321 ± 327; b) M.
Mammen, S.-K. Choi, G. M. Whitesides, Angew. Chem. 1998, 110,
2908 ± 2953; Angew. Chem. Int. Ed. 1998, 37, 2754 ± 2794.
Divalent conjugate 21: This compound (1.18 mg, 43%) was obtained
following procedure A; RP-HPLC [gradient 100:0 to 85:15 (A/B), 15 min,
then 85:15 to 75:25 (A/B), 20 min, then isocratic]; positive ESI-MS: found:
2574; calcd 2575; m/z: 1288.2 [M‡2H]2‡, 859.3 [M‡3H]3‡, 644.7 [M‡4H]4‡
.
Tetravalent conjugate 22: This compound (1.49 mg, 40%) was obtained
[8] For natural ligands, see: a) V. Kery, J. J. F. Krepinsky, C. D. Warren, P.
Capek, P. D. Stahl, Arch. Biochem. Biophys. 1992, 298, 49 ± 55; b) R. T.
following procedure A; RP-HPLC [gradient 100:0 to 85:15 (A/B), 10 min,
238
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0238 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 1

Clustered Glycoside ± Antigen Conjugates
230±239
Lee, Y. Ichikawa, M. Fay, K. Drickamer, M.-C. Shao, Y. C. Lee, J. Biol.
Chem. 1991, 266, 4810 ± 4815; for glycosylated neoglycoprotein, see:
c) K. Sarkar, H. Sekhar Sarkar, L. Kole, P. K. Das, Mol. Cell. Biochem.
1996, 156, 109 ± 116; d) C. A. Hoppe, Y. C. Lee, J. Biol. Chem. 1983,
258, 14193Ð14199; for mannosylated poly-l-lysine, see: e) D. Der-
[23] Similar intramolecular cyclization has been observed during the
labeling of (chloroacetyl)lysinyl trees with fluorescein isothiocyanate,
see: C. Grandjean, C. Rommens, H. Gras-Masse, O. Melnyk,
Tetrahedron Lett. 1999, 40, 7235 ± 7238. Attack by the Ne end rather
than the Nz was confirmed by the ability of the cyclic compound to
form a hydrazone with antigen 8.
Á
rien, P. Midoux, C. Petit, E. Negre, R. Mayer, M. Monsigny, A.-C.
Roche, Glycoconjugate J. 1989, 6, 241 ± 255; f) C. A Hoppe, Y. C. Lee,
Biochemistry 1984, 23, 1723 ± 1730; for mannosylated clusters, see:
g) J. C. Robbins, M. Hanh Lam, C. S. Tripp, R. L. Bugianesi, M. M.
Ponpipom, T. Y. Shen, Proc. Natl. Acad. Sci. USA 1981, 78, 7294 ±
7298; h) M. M. Ponpipom, R. L. Bugianesi, J. C. Robbins, T. W.
Doebber, T. Y. Shen, J. Med. Chem. 1981, 24, 1388 ± 1395; i) Y.
Ohsumi, C. A. Hoppe, T. Ogawa, Y. C. Lee, Archiv. Biochem.
Biophys. 1988, 260, 241 ± 249; j) E. A. L. Biessen, F. Noorman, M. E.
van Teijlingen, J. Kuiper, M. Barrett-Bergshoeff, D. C. Rijken, T. J. C.
van Berkel, J. Biol. Chem. 1996, 271, 28024 ± 28030.
[24] K. F. Geoghegan, J. G. Stroh, Bioconjugate Chem. 1992, 3, 138 ± 146.
[25] For examples of hydrazone chemical ligations, see: a) T. P. King, S. W.
Zhao, T. Lam, Biochemistry 1986, 25, 5774 ± 5779; b) I. Fisch, G.
Künzi, K. Rose, R. E. Offord, Bioconjugate Chem. 1992, 3, 147 ± 153;
c) H. F. Gaertner, K. Rose, R. Cotton, D. Timms, R. Camble, R. E.
Offord, Bioconjugate Chem. 1992, 3, 262 ± 268; d) H. F. Gaertner,
R. E. Offord, R. Cotton, D. Timms, R. Camble, K. Rose, J. Biol. Chem.
1994, 269, 7224 ± 7230; e) J. C. Spetzler, J. P. Tam, Int. J. Peptide Res.
1995, 45, 78 ± 85; f) J Shao, J. P. Tam, J. Am. Chem. Soc. 1995, 117,
3893 ± 3899; g) K. Rose, W. Zeng, P.-O. Regamey, I. V. Chernushevich,
K. G. Standing, H. F. Gaertner, Bioconjugate Chem. 1996, 7, 552 ± 556;
h) see ref. [21].
[9] For reviews, see: a) T. W. Muir, P. E. Dawson, S. B. H. Kent, Methods
Enzymol. 1997, 289, 266 ± 298; b) J. P. Tam, J. C. Spetzler, Biomed.
Pept. Proteins Nucleic Acids 1995, 1, 123 ± 132; c) J. P. Tam, Q. Yu, Z.
Miao, Biopolymers (Peptide Science) 2000, 51, 311 ± 332; d) T. C.
Evans Jr., M.-Q. Xu, Biopolymers (Peptide Science) 2000, 51, 333 ±
342; e) B. Ayers, U. K. Blaschke, J. A. Camarero, G. J. Cotton, M.
Holford, T. W. Muir, Biopolymers (Peptide Science) 2000, 51, 343 ±
354.
[26] For the precise thioetherification procedure, see: C. Grandjean, C.
Rommens, H. Gras-Masse, O. Melnyk, J. Chem. Soc. Perkin Trans. 1
1999, 2967 ± 2976.
[27] For reviews about glycodendrimer synthesis, see: a) N. Jayaraman,
S. A. Nepogodiev, J. F. Stoddart, Chem. Eur. J. 1997, 3, 1193 ± 1199;
b) D. Zanini, R. Roy in Carbohydrate Mimics. Concepts and Methods
(Ed.: Y. Chapleur), Wiley-VCH, Weinheim, 1998, pp. 385 ± 415.
[28] For examples of nucleophilic displacement of chloroacetylated trees
with protected thiosugar derivatives, see: a) ref. [18b]; b) D. Zanini,
W. K. C. Park, R. Roy, Tetrahedron Lett. 1995, 36, 7383 ± 7386; c) R.
Roy, W. K. C. Park, Q. Wu, S.-N. Wang, Tetrahedron Lett. 1995, 36,
[10] H. McPherson, M. C. Wright, P. A. Lohse, Synlett 1999, 978 ± 980, and
references therein.
[11] G. A. Lemieux, C. R. Bertozzi, Trends Biotechnol. 1998, 16, 506 ± 513.
[12] a) L. A. Vilaseca, K. Rose, R. Werlen, A. Meunier, R. E. Offord, C. L.
Nichols, W. L. Scott, Bioconjugate Chem. 1993, 4, 515 ± 520; b) S. E.
Cervigny, P. Dumy, M. Mutter, Angew. Chem. 1996, 108, 1325 ± 1328;
Angew. Chem. Int. Ed. Engl. 1996, 35, 1230 ± 1232; c) D. Lelievre, H.
Â
4377 ± 4380; d) D. Page, D. Zanini, R. Roy, Bioorg. Med. Chem. 1996,
4, 1949 ± 1961; e) D. Zanini, R. Roy, J. Org. Chem. 1996, 61, 7348 ±
7354; f) D. Zanini, R. Roy, Bioconjugate Chem. 1997, 8, 187 ± 192;
g) D. Zanini, R. Roy, J. Am. Chem. Soc. 1997, 119, 2088 ± 2095;
h) M. M. Palcic, H. Li, D. Zanini, R. S. Bhella, R. Roy, Carbohydr. Res.
1998, 305, 433 ± 442.
Á
Chabane, D. Delmas, Abstracts of Papers, 11eme reunion peptides et
proteines, Aussois, France 1999, p. 30; d) Z. Wei, C.-H. Tung, T. Zhu,
S. Stein, Bioconjugate Chem. 1994, 5, 468 ± 474.
[13] L. E. Canne, P. Botti, R. J. Simon, Y. Chen, E. A. Dennis, S. B. H.
Kent, J. Am. Chem. Soc. 1999, 121, 8720 ± 8727.
[14] J. A. Camarero, G. J. Cotton, A. Adeva, T. W. Muir, J. Peptide Res.
1998, 51, 303 ± 316.
[15] a) O. Nyanguile, M. Mutter, G. Tuchscherer, Lett. Peptide Science
1994, 1, 9 ± 16; b) G. Tuchscherer, D. Grall, M. Mathieu, M. Mutter, J.
Peptide Res. 1999, 54, 185 ± 194; c) H. K. Rau, W. Haehnel, J. Am.
Chem. Soc. 1998, 120, 468 ± 476; d) S. Futaki, M. Aoki, T. Ishikawa, F.
Kondo, T. Asahara, M. Niwa, Y. Nakaya, T. Yagami, K. Kitagawa,
Bioorg. Med. Chem. 1999, 7, 187 ± 192.
[29] K. Rose, J. Chen, M. Dragovic, W. Zeng, D. Jeannerat, P. Kamalaprija,
U. Burger, Bioconjugate Chem. 1999, 10, 1038 ± 1043.
[30] J.-S. Fruchart, H. Gras-Masse, O. Melnyk, Tetrahedron Lett. 1999, 40,
6225 ± 6228.
[31] J.-S. Fruchart, C. Grandjean, C. Rommens, H. Gras-Masse, O. Melnyk,
unpublished results.
[32] L. J. Wilson, M. Li, D. E. Portlock, Tetrahedron Lett. 1998, 39, 5135 ±
5138.
[33] The hydrazine function might have not survived the HF cleavage
conditions required by the Boc/benzyl strategy: D. Bonnet, F. Samson,
C. Rommens, H. Gras-Masse, O. Melnyk, Int. J. Peptide Res. 1999, 54,
270 ± 278; some a-hydrazino peptides decomposed when purified in
CH3CN/TFA mixtures: see also reference [38].
Â
Â
[16] L. E. Canne, A. R. Ferre-DꢁAmare, S. K. Burley, S. B. H. Kent, J. Am.
Chem. Soc. 1995, 117, 2998 ± 3007.
[17] M. Mutter, P. Dumy, P. Garrouste, C. Lehmann, M. Mathieu, C.
Peggion, S. Peluso, A. Razaname, G. Tuchscherer, Angew. Chem.
1996, 108, 1587 ± 1591; Angew. Chem. Int. Ed. Engl. 1996, 35, 1482 ±
1485.
[34] T. W. Muir, M. J. Williams, M. H. Ginsberg, S. B. H. Kent, Biochem-
istry 1994, 33, 7701 ± 7708.
[18] a) R. G. Denkewalter, J. F. Kolc, W. J. Lukasavage, US Patent
4410688, 1983 [Chem. Abstr. 1984, 100, 103907p]; b) R. Roy, D.
Zanini, J. Meunier, A. Romanowska, J. Chem. Soc. Chem. Commun.
1993, 1869 ± 1873.
[35] Preliminary communication: C. Grandjean, C. Rommens, H. Gras-
Masse, O. Melnyk, Angew. Chem. 2000, 112, 1110 ± 1114; Angew.
Chem. Int. Ed. Engl. 2000, 39, 1068 ± 1072.
[36] P. Q. Huang, K. Sabbe, M. Pottie, M. Vandewalle, Tetrahedron 1995,
36, 8299 ± 8302.
[19] G. B. Fields, R. L. Noble, Int. J. Pept. Protein Res. 1990, 35, 161 ± 214.
Â
[20] a) J. Vidal, L. Guy, S. Sterin, A. Collet, J. Org. Chem. 1993, 58, 4791 ±
[37] C. Grandjean, O. Melnyk, H. Gras-Masse, G. Angyalosi, C. Rommens,
C. Auriault, FR 9909928.
[38] D. Bonnet, C. Rommens, H. Gras-Masse, O. Melnyk, Tetrahedron
Lett. 2000, 41, 45 ± 48.
4793; b) C. Klinguer, O. Melnyk, E. Loing, H. Gras-Masse, Tetrahe-
dron Lett. 1996, 37, 7259 ± 7262.
[21] a) O. Melnyk, M. Bossus, D. David, C. Rommens, H. Gras-Masse, J.
Peptide Res. 1998, 52, 180 ± 184; b) F. Samson, D. Bonnet, C. Rom-
mens, H. Gras-Masse, O. Melnyk, J. Pept. Science 1999, 5, 352 ± 359.
[22] L. Bourel, O. Carion, H. Gras-Masse, O. Melnyk, J. Pept. Science 2000,
6, 264 ± 270.
[39] W. S. Hancock, J. E. Battersby, Anal. Biochem. 1976, 71, 260 ± 264.
Received: May 8, 2000 [F2471]
Chem. Eur. J. 2001, 7, No. 1
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0701-0239 $ 17.50+.50/0
239