Convergent synthesis of D-(-)-quinic and shikimic acid-containing dendrimers as potential C-lectin ligands by sulfide ligation of unprotected fragments

Article abstract of DOI:10.1039/a904679h

The preparation of D-(-)-quinic and (-)-shikimic acid-derived dendrimers with valencies of 4, 8 and 16, respectively, as potential C-lectin ligands is reported. D-(-)-Quinic and shikimic acids were branched to an (S-tert-butylthio-L-cysteine)-containing t

Full text of DOI:10.1039/a904679h

View Article Online / Journal Homepage / Table of Contents for this issue
Convergent synthesis of D-(؊)-quinic and shikimic acid-containing
dendrimers as potential C-lectin ligands by sulfide ligation of
unprotected fragments
Cyrille Grandjean,* Corinne Rommens, Hélène Gras-Masse and Oleg Melnyk*
UMR 8525 CNRS, Université de Lille II, Institut de Biologie et Institut Pasteur de Lille,
1 rue du Professeur Calmette, BP447, F-59021 Lille Cedex, France
Received (in Cambridge, UK) 11th June 1999, Accepted 7th September 1999
The preparation of -(Ϫ)-quinic and (Ϫ)-shikimic acid-derived dendrimers with valencies of 4, 8 and 16, respectively,
as potential C-lectin ligands is reported. -(Ϫ)-Quinic and shikimic acids were branched to an (S-tert-butylthio--
cysteine)-containing tripeptide on solid phase to furnish compounds 1 and 3. These intermediates were reduced upon
treatment with tri-n-butylphosphine and linked to N-chloroacetylated -lysinyl cores via a nucleophilic substitution
performed in aqueous DMF.
hexane or cyclohexene. Finally, the presence of a carboxylic
Introduction
acid group renders functionalisation more easy than in the
Lectins (carbohydrate-binding proteins) play essential roles in
mannose case, particularly when using a solid-phase strategy.
numerous biological events including clearance from the circu-
Shikimic acid 8 and the tetra-O-acetylated quinic acid deriv-
ative 9 have been condensed to peptidyl resins 7a or 7b to give
compounds 1 and 3, after acidic cleavage. These intermedi-
ates were deprotected and assembled with the preformed
latory system and intracellular routing of glycoproteins, cell–
cell adhesion and immune defence.1 However, the affinity of
lectins for carbohydrates, although highly specific, is generally
weak. High-affinity complexes can be obtained when a multi-
valent sugar ligand interacts with clustered receptors on the
dendrimeric cores via a sulfide linkage at the ultimate stage of
the synthesis. This strategy affords greater flexibility, minimises
lectins, a property known as the cluster effect.2 Among the
and simplifies the purification procedure and avoids side-
synthetic structures which can mimic natural oligosaccharide
reactions which might have occurred during deprotection steps
counterparts, glycodendrimers are especially well adapted since
(Scheme 1).
they satisfy the multivalency criteria and can be fully character-
ised in terms of shape, size and carbohydrate content.3 Such
constructions might be used as antibody, drug or molecule
carriers. Within the scope of a synthetic vaccine programme, we
have planned to deliver antigens to antigen presenting cells
(APCs), such as dendritic cells or macrophages by targeting
dendrimers at their mannose receptors.4 A similar approach
relying upon an antigen linked to lysine-based mannoside
clusters had been envisaged by Koning and co-workers5 and
met with substantial success. In this paper, the synthesis of
potential dendrimeric ligands of APCs mannose receptors is
presented (Scheme 1). These dendrimers are formed by modi-
fied N-chloroacetylated -lysinyl cores, easily amenable to a
ligation with peptide antigens. In this study, we postulated that
-(Ϫ)-quinic and shikimic acids could replace -mannose as
the glycosyl vector:6 In fact, the mannose receptor of APCs
belongs to one of the C-type superfamily lectins.7 For these
lectins, preponderant interactions involve a calcium ion
coordination by glutamic and asparagine residues in the
carbohydrate-recognition-domain and by two vicinal hydroxy
groups of the ligand (3- and 4-OH for -mannose), provided
they are in a trans di-equatorial relationship,8 as illustrated in
the X-ray structure of the rat mannose-binding protein com-
plexed with an oligosaccharide.9 Thus, mannose receptors are
able to bind not only -mannose but also N-acetyl--glucos-
amine or -fucose excluding -galactose and related sugar-
containing molecules. Adopting a glycomimetic approach, we
hypothesized that non-carbohydrate compounds such as -(Ϫ)-
quinic or (Ϫ)-shikimic acid might be considered as potential
mannose receptor ligands since they possess a conveniently
arranged vicinal diol (at the C-4 and C-5 position). Equally,
they might offer a greater stability than mannose in a biological
environment since the pyranose ring is replaced by a cyclo-
Results and discussion
Synthesis of the L-lysinyl cores
We decided to synthesise dendrimers based on -lysine trees, in
view of their biocompatibility and, in particular, their lack of
intrinsic immunogenicity.10 The use of such poly--lysine scaf-
folds has been extended more recently to the preparation of
glycodendrimers, mainly by Roy et al.11 However, the reported
syntheses have been modified for our purpose: the ε-amino
group of the first lysinyl residue was not incorporated into the
scaffold, in order to permit an ulterior linkage with fluores-
cent labels or with peptide antigens (Scheme 2). The syntheses
have been performed on solid support and monitored by
ninhydrin12 and 2,4,6-trinitrobenzenesulfonic acid (TNBS)13
tests. Boc-β-alanine (0.25 equiv.) has been anchored to a
4-methylbenzydrylamine resin (MBHA) 11 using N-[1H-
benzotriazol-1-yl(dimethylamino)methylene]-N-methylmethan-
aminium hexafluorophosphate N-oxide (HBTU)–HOBt as
acylating agents14 in order to adjust the loading at 0.1 mmol
gϪ1. The unchanged amino groups were then capped by acetyl-
ation. After acidic deprotection of the β-alanine amino group,
the peptidyl resin was coupled to a lysine, which was chain
protected with a permanent 2-chlorobenzyloxycarbonyl group.
A second lysine was introduced as its Nα,Nε-di-Boc derivative
and was deprotected by TFA treatment to furnish peptidyl resin
12. One third of 12 was allowed to react with chloroacetic
anhydride, obtained from chloroacetic acid and diisopropyl-
carbodiimide. N-Chloroacetylated wedges of the trees will
allow the ligation between the dendritic cores and -(Ϫ)-quinic
and (Ϫ)-shikimic acid derivatives. Compound 13 was obtained
following cleavage from the resin by HF–anisole treatment and
J. Chem. Soc., Perkin Trans. 1, 1999, 2967–2975
This journal is © The Royal Society of Chemistry 1999
2967

View Article Online
Scheme 1
subsequent purification by reversed-phase high-performance
atographic means and characterisation of the final dendrimers
liquid chromatography (RP-HPLC) in 52% overall yield.
Second- and third-generation cores 14 and 15 were obtained
by repeating coupling steps with Nα,Nε-di-Boc--lysine from
peptidyl resin 12 in 42% and 20% overall yield.
would not be compatible with the presence of partially isomer-
ised shikimic acid residues. Compound 2 could have been
used directly in sulfide ligation experiments but it appeared to
be slightly unstable during storage. We postulated that the
observed decomposition was occasioned by the free sulfhydryl
(thiol) groups. We sought to obtain 2 in an S-protected form to
alleviate this problem. A new synthesis based upon Fmoc chem-
istry17 was developed starting with Fmoc-glycine p-benzyloxy-
benzyl ester resin (Fmoc-Gly-O-Wang resin), 16b. -Cysteine
was introduced to the deprotected peptidyl resin as its S-tert-
butylthio derivative. This protecting group18 survives both the
synthesis and cleavage conditions while its removal can be easily
performed by reduction with trialkylphosphines. Such an
approach was preferred to the direct transformation of 2 into
the corresponding disulfide as one oxidation step is saved. Pep-
tidyl resin 7a was obtained after coupling with Nα,Nε-di-Fmoc-
-lysine and treatment with piperidine. Acylation of 7a with
unprotected shikimic acid followed by cleavage from the resin
(TFA–Me2S–H2O 95:2.5:2.5), and RP-HPLC purification,
furnished stable compound 1 in 14% overall yield. The same
strategy was applied for the preparation of compound 3, the
quinic acid analog of 1, from Fmoc-glycine o-methoxy-p-
(benzyloxy)benzyl ester resin (Fmoc-Gly-O-Sasrin),19 16c.
However, the coupling of unprotected -(Ϫ)-quinic acid with
peptidyl resin 7b proved to be abortive whatever the activation
used and led to the immediate formation of a known bicyclic γ-
lactone,20 coming from esterification of the C-1 carboxy group
with the C-5 hydroxy group. Preliminary protection of the
hydroxy groups was thus required to prevent quinic acid from
undergoing intramolecular cyclisation. -(Ϫ)-Quinic acid has
been peracetylated upon treatment with acetic anhydride in
Functionalisation of D-(؊)-quinic and shikimic acids
For their attachment to the lysine trees, commercially available
-(Ϫ)-quinic and shikimic acids should be functionalised by a
thiol group. For this purpose, they were coupled to a cysteinyl-
containing tripeptide on a solid support. We first devised a
route leading to compounds such as 2, ready for coupling with
the -lysinyl cores. Compound 2 was indeed obtained from
Boc-glycyl PAM resin 16a using the Boc/benzyl strategy15 and
HBTU–HOBt as acylating agents (Scheme 3). Peptidyl resin
16a was deprotected upon treatment with 50% TFA in DCM,
coupled with Boc--Cys(p-MeC6H4CH2)-OH and further
submitted to TFA deprotection. At this stage, an Nα,Nε-di-Boc-
-lysine residue was added to peptidyl resin to double the vector
valency and equally to increase more rapidly the dendrimer size.
After removal of the tert-butyloxycarbonyl protecting groups,
peptidyl resin 17 was actually coupled at both amino termini
with unprotected shikimic acid, preactivated for 30 seconds
with HBTU–HOBt–DIPEA (1:1:3 equivalents) in DMF.
Compound 2 was obtained after release from the resin by
HF–thiocresol–p-thiocresol treatment, which was followed by
RP-HPLC purification in 42% yield. The shikimic moieties
were perfectly stable under the strong acidic cleavage conditions
applied. In particular, no isomerisation of the double bond of
the natural isomer to the C-1/C-6 positions occurred.16 Such
stability is of singular importance since purification by chrom-
2968
J. Chem. Soc., Perkin Trans. 1, 1999, 2967–2975

View Article Online
Scheme 3 Reagents and conditions: i, TFA–DCM (50:50), 20 min;
then DIPEA–DCM (5:95), 3 × 1 min; ii, 4 equiv. Boc--Cys(p-
MeHC8H4CH2)-OH, HBTU–HOBt–DIPEA (4:4:12 equiv.), DMF,
45 min; iii,
4 equiv. Boc--Ly(Boc)-OH, HBTU–HOBt–DIPEA
(4:4:12 equiv.), DMF, 45 min; iv, 1 equiv./NH2 8, HBTU–HOBt–
DIPEA (1:1:3 equiv.)/NH2, DMF, 45 min; v, HF–p-cresol–p-thiocresol
(10:0.75:0.25) (v/w/w), 0 ЊC, 1.5 h; vi, piperidine–NMP (20:80), 20
min; vii, 2 equiv. Fmoc--Cys(SBut)-OH, HBTU–DIPEA (2:3 equiv.)
NMP, 40 min; viii, 2 equiv. Fmoc--Lys(Fmoc)-OH, HBTU–DIPEA
(2:3; equiv.) NMP, 40 min; ix, 2 equiv./NH2 8, HBTU–DIPEA (2:3
equiv./NH2, NMP, 1 h; x, TFA–H2O–Me2S (95:2.5:2.5), rt, 2 h; xi, 8
equiv. cyanuric fluoride, 1 equiv. pyridine, DCM, reflux, 2 h; xii, 2
equiv./NH2 18, 3 equiv./NH2 DIPEA, NMP, 1 h; xiii, TFA–iPr3SiH–
DCM (2.5:2.5:95), 4 × 5 min; xiv, MeONa, MeOH, rt, 3 h; xv, 4 ϩ 2
equiv. 10, 3.5 equiv. DIPEA, MeOH, reflux 48 h.
Scheme 2 Reagents and conditions: i, 0.25 equiv./NH2 Boc-β-Ala-OH,
HBTU–HOBt–DIPEA (0.25:0.25:0.75 equiv.)/NH2, DMF, 1 h; ii,
Ac2O–DIPEA–DCM (5:10:85), 10 min; iii, TFA–DCM (50:50), 15
min; then DIPEA–DCM (5:95); iv, 4 equiv. Boc--Lys[(2-Cl)Z]-OH,
HBTU–HOBt–DIPEA (4:4:12 equiv.), DMF, 45 min; v, 4 equiv./NH2,
Boc--Lys(Boc)-OH, HBTU–HOBt–DIPEA (4:4:12 equiv.)/NH2,
DMF, 45 min; vi, 8 equiv./NH2 ClCH2CO2H, 4 equiv. DCC, DMF, 30
min; vii, HF–anisole (90:10), 0 ЊC, 1 h.
cific and non-specific interactions. However, a good candidate
should share a close structural similarity with the -(Ϫ)-
quinic and shikimic acid-derived dendrimers, i.e. all con-
structs must eventually be formed of an -lysinyl scaffold
branched through the same tripeptide to polyhydroxylated
molecules, which must be linked to the peptide via an amide
bond. In fact, the unrecognised dendrimer could be formed
from a polyhydroxylated compound whose stereochemistry is
related to -galactose. Since quinic and shikimic acid isomers
are not commercially available, we designed a synthesis from
-galactonolactone. Tripeptide 19 was obtained from peptidyl
resin 7a after acidic treatment (TFA–Me2S–H2O 95:2.5:2.5),
in 60% yield together with a dimer 20 (4%). Finally, inter-
mediate 19 was condensed with commercially available
-galactonolactone 1023 in refluxing MeOH to give 5 in 69%
yield.24
acetic acid in the presence of a catalytic amount of perchloric
acid to furnish 9 in 86% yield. This intermediate was trans-
formed into the corresponding acid fluoride 18 with cyanuric
fluoride in DCM at reflux.21 Subsequently, the fluoride was fur-
ther condensed with the peptidyl resin 7b in DCM in the pres-
ence of DIPEA. This activation is one of the most convenient
for the coupling of sterically hindered carboxylic acids.22
Release from the resin was carried out by treatment with 2.5%
TFA in DCM using triisopropylsilane as carbocation scavenger.
After lyophilisation, the crude residue was deprotected with
sodium methoxide in MeOH and purified by RP-HPLC to fur-
nish 3 in 59% overall yield. The coupling constants between
protons 3-H/4-H and 4-H/5-H of both quinic acid residues in 3,
determined by 1H NMR, were 2.9 Hz and 9.6 Hz, respectively.
These values are consistent with axial–equatorial and axial–
axial couplings and with a chair-like conformation of the
cyclohexanes having an equatorial amide group. Similarly, the
coupling constants between protons 3-H/4-H and 4-H/5-H of
both shikimic acid residues in 1, determined by 1H NMR,
were 4.6 Hz and 9.2 Hz, respectively. These values are com-
Synthesis of the hyper-branched L-lysinyl dendrimers
Having prepared the different building blocks, their final
assembly was undertaken via a straightforward two-step pro-
cedure. In fact, all glycodendrimer syntheses based upon the
nucleophilic substitution of chloroacetyl groups so far reported
have been achieved using per-acetylated glycosides in dry
organic solvents.11,25 Besides, only a few papers have related the
synthesis of glycodendrimers without using protecting
groups23,26 although one chemical and purification step, per-
formed on sophisticated molecules, is saved by adopting such a
strategy. Furthermore, the deprotection step is sometimes
patible with
a pseudo-equatorial orientation of hydroxy
groups at the C-4 and C-5 positions of the shikimic acids,
and support the structural analogy made between -(Ϫ)-
quinic and shikimic acids and -mannose. Having biological
assays in prospect, we need to prepare a priori structures
unrecognised by the mannose receptor. Such molecules will
indeed be helpful in allowing us to discriminate between spe-
J. Chem. Soc., Perkin Trans. 1, 1999, 2967–2975
2969

View Article Online
troublesome (e.g., partial deprotection and O→N migration
have been encountered during deacetylation).27 Protecting
groups might also alter the growth of dendrimers, especially
during attempts to obtain densely packed constructs.26b In pre-
liminary experiments, we noticed that the reduction and the
substitution could not be performed in one pot. In fact, every
nucleophile present in the reaction mixture was able to compete
with the free thiols for the substitution of the chloroacetyl
groups, including the liberated tert-butyl mercaptan, the
reducing agent [either tributylphosphine or the less nucleophile
tris(carboxyethyl)phosphine], and even the amino group of the
-lysinyl core when the experiment was conducted at insuffi-
ciently low pH. These results sharply contrasted with some
reported procedures in which substitution of N-chloro-
acetylated lysine trees by cysteine-containing peptides was per-
formed in the presence of a large excess of trialkylphosphines.28
These competitive reactions proved less efficient than the
desired substitution yet made purification procedures difficult
and resulted in a loss of overall yield. In our approach, the
disulfide bond of compounds 1, 3 and 5 was first reduced by
treatment with tri-n-butylphosphine, probably the best reducing
n
reagent,29 in a mixture of degassed PrOH–H2O 1:1 to give
intermediates 2, 4 and 6 (Scheme 1). The use of propan-1-ol as
co-solvent gave better results than other recommended solvents
such as trifluoroethanol.30 At the end of the reaction, the mix-
ture was carefully evaporated to dryness to remove any tert-
butyl mercaptan formed. Extraction of the crude residue was
avoided by preferential use of the minimum amount of phos-
phine though this resulted in a somewhat prolonged reaction
time (24 h). Intermediates 2, 4 and 6 (1.5 equiv. per chloro-
acetyl group to be substituted) were actually treated, without
further purification, with -lysinyl cores 13, 14 and 15 in a
degassed mixture of DMF–H2O (90:10) in the presence of
potassium carbonate. The use of DMF is known to increase the
thiol nucleophilicity and to limit disulfide bonding, forming
dimers.31 After 72–96 h, reactions were essentially complete as
revealed by the RP-HPLC profiles of the crude mixtures (see
Fig. 1a for one example). Dendrimers 21–27 were obtained after
RP-HPLC purification and further characterized by analytical
capillary zone electrophoresis (CZE) and electrospray ioniz-
ation mass spectra (ESI-MS) as illustrated for compound 26 in
Fig. 1b, 1c and Fig. 2. Loss of galactosyl chains in dendrimer
27, through ring closure to lactone, was observed during its
purification in an acidic buffer, which prompted us to switch to
a phosphate buffer system of a neutral pH.
Fig. 1 (a) RP-HPLC chromatogram of a crude mixture of compound
26 after lyophilisation. (b) RP-HPLC chromatogram of purified com-
pound 26. Chromatographic conditions: Beckman ultrapore C8 (300 Å,
5 µm, 4.6 × 25 mm). Flow rate 1 mL minϪ1, rt. Buffer A: 0.05% aq. TFA.
Buffer B: 0.05% TFA in CH3CN–H2O (80:20). Gradient 0–30% B
over 30 min (a); 0% B for 5 min then 0–25% B over 30 min (b). (c) CZE
profile of compound 26 (see Experimental section for conditions).
Conclusions
The substitution of N-chloroacetylated -lysinyl cores has been
extended to free polyhydroxylated molecules, resulting in the
synthesis of potential C-lectin ligand dendrimers having a val-
ency as high as 16. Their labelling or their linkage to antigens
Fig. 2 (a) Negative ESI-MS of compound 26; flow rate 3 µL minϪ1 at a concentration of 5 pmol mLϪ1 in 1% NH3 in CH3CN–H2O (50:50). (b) ESI-
MS true mass scale of compound 26.
2970
J. Chem. Soc., Perkin Trans. 1, 1999, 2967–2975

View Article Online
Table 1 Inhibition of binding of yeast mannan to concanavalin A by
p-nitrophenyl α--mannopyranoside and dendrimers 25, 26 and 27
Inhibitor
Con A IC50/µM
Man-α-OC6H4NO2-p
110
40
130
420
25
26
27
OH groups as with mannose receptors:32 higher branched
dendrimers 25 and 26 have, however, shown similar or higher
affinity than p-nitrophenyl α--mannopyranoside used as a
reference.
Surprisingly, weak affinity was also observed for dendrimer
27 (Table 1). To further investigate their activity, the dendrimers
will next be labelled with fluorescein isothiocyanate and evalu-
ated with dendritic cells’ mannose receptor.
Experimental
Materials and methods
Optical rotations were measured with a Perkin-Elmer 241
polarimeter; [α]D-values are given in units of 10Ϫ1 deg cm2
g
Ϫ1. Analytical and semi-preparative reversed-phase high-
performance liquid chromatography (RP-HPLC) separations
were performed on Shimadzu LC-6A and LC-4A systems on
a Beckman ultrapore C-8 (300 Å; 5 µm; 4.6 × 25 mm), or a
Hypersil hyperprep C-18 (300 Å; 8 µm; 15 × 500 mm), col-
umn at a flow rate of 1 or 3 mL minϪ1 with detection at 215
or 230 nm. Solvent system A: 0.05% TFA in water; solvent
system B: 0.05% TFA in 80% acetonitrile–20% water;
solvent system C: 0.05% TFA in 60% acetonitrile–40% water;
solvent system D: phosphate buffer 50 mmol, pH 6.95; solvent
system E: 50% phosphate buffer 50 mmol, pH 6.95–50%
acetonitrile. Time-of-flight–plasma desorption mass spec-
trometry (TOF-PDMS) spectra were recorded upon a Bio-Ion
20 Plasma Desorption Mass Spectrometer (Uppsala, Sweden),
and ESI-MS spectra on a Micromass Quatro II Electrospray
Mass Spectrometer. Compounds were verified for homogeneity
by analytical Capillary Zone Electrophoresis in a 75 µm × 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 at 30 ЊC using a 100
1
mM sodium borate migration buffer at pH 9.2. H and 13C
NMR spectra were recorded on Bruker DRX 300 or DRX 600
spectrometers. Chemical shifts are given in ppm and referenced
to internal TMS or sodium 3-(trimethylsilyl)-[2,2,3,3-d4]-
propionate (TMSP), when the spectra were recorded in H2O–
1
1
D2O (90:10). For the assignment of signals H, H–1H corre-
lation spectroscopy (COSY), total correlation spectroscopy
1
(TOCSY), nuclear Overhauser effect (NOE), 13C and H–13C
heteronuclear single-quantum correlation (HSQC) spectro-
scopy experiments were used. J-Values are given in Hz.
Preparation of the chloroacetyl poly-L-lysinyl cores
Each lysinyl core was synthesised at a 0.5 mmol scale starting
with 15 g of MBHA resin (initial substitution, 0.4 mmol gϪ1
)
(Senn chemicals), using the Boc/benzyl solid-phase peptide-
synthesis strategy. The coupling steps were performed using a 4-
fold excess of amino acid per amine to be treated via HBTU–
HOBt–DIPEA activation in DMF and monitored by the TNBS
and ninhydrin tests: typically, HBTU dissolved in DMF (4
equiv., 0.5 mmol mLϪ1), was added to a mixture of the amino
acid (4 equiv., 0.5 mmol mLϪ1), HOBt (4 equiv.) and DIPEA (8
equiv.) in DMF. After stirring for 1 min, the mixture was added
to the peptidyl resin (1 equiv.) swollen in DMF containing
DIPEA (4 equiv.), and mechanically shaken for 45 min. Follow-
ing filtration, the peptidyl resin was washed successively with
is made easier owing to the presence of a unique amino group
on these structures. In preliminary measurements, the binding
affinity of these constructs for concanavalin A has been deter-
mined by Enzyme Linked Lectin Assay as described by Pagé
et al.11c This lectin does not constitute a good model, since
formation of high-affinity complexes requires a greater inter-
action than simple binding with two trans di-equatorial vicinal
J. Chem. Soc., Perkin Trans. 1, 1999, 2967–2975
2971

View Article Online
DMF (3 × 2 min) and DCM (3 × 2 min). Cleavage of the Boc
protecting groups was achieved by treatment with TFA–DCM
(50:50; 1 × 2 min, 1 × 20 min), followed by washing with DCM
(2 × 2 min), neutralisation with DIPEA–DCM (5:95; 3 × 1
min), and washing with DCM (2 × 1 min). 0.25 Equivalents of
Boc-β-Ala-OH was first anchored to the resin followed by a
capping of the unchanged amino groups with Ac2O–DIPEA–
DCM (5:10:85) for 10 min followed by washing with DCM
(3 × 1 min), to diminish the loading. Every coupling was
followed by acetylation (10 min). Chain protection of the first
lysinyl residue introduced was secured by a 2-chlorobenzyl-
oxycarbonyl [(2-Cl)Z] group. The second lysinyl residue was
added as its Boc--Lys(Boc)-OH derivative. At this stage, one
third of the resin was deprotected and acylated using an 8-fold
excess of preformed chloroacetic anhydride, prepared via DCC
activation to provide the first level carrier core. The second- and
third-level cores were obtained by repeating the last two steps.
The cores were cleaved from the resin and deprotected by HF–
anisole (10:1; 11 mL per g of peptidyl resin), for 1 h at 0 ЊC,
precipitated in cold tert-butyl methyl ether, centrifuged, dis-
solved in water and lyophilised. Finally, the crude peptides were
purified by semi-preparative RP-HPLC [gradient: 100:0 to
50:50 (A:B) in 120 min], to furnish core 13 (160 mg, 52%); m/z
(TOF-PDMS) 498 (M ϩ H)ϩ; core 14 (215 mg, 42%); m/z
(TOF-PDMS) 928 (M ϩ Na)ϩ, 907 (M ϩ H)ϩ; core 15 (182 mg,
20%); m/z (TOF-PDMS) 1726 (M ϩ H)ϩ.
(1sn,3R,4sn,5R)-1,3,4,5-Tetraacetoxycyclohexanecarboxylic acid
933,34
To a suspension of -(Ϫ)-quinic acid (4 g, 21 mmol) in a 2:1
mixture of acetic acid–acetic anhydride (30 mL) was added one
drop of perchloric acid at room temperature. As the reaction
temperature increased to 50–60 ЊC, the mixture became clear.
The solution was stirred for a further 12 h, diluted with chloro-
form and then extracted succesively with saturated aq.
NaHCO3 and water. The organic layer was dried over Na2SO4,
filtered and concentrated under reduced pressure. The residue
was finally precipitated by addition of pentane to give 9 (6.45 g,
86%); δH(300 MHz; CDCl3) 1.87 (1 H, dd, J5,6 10.2 and J6,6Ј
2972
J. Chem. Soc., Perkin Trans. 1, 1999, 2967–2975

View Article Online
13.6, 6-H), 1.96, 1.98, 2.03 and 2.08 (4 × 3 H, 4 s, 4 × CH3),
1.98–2.03 (1 H, m, 6-HЈ), 2.36 (1 H, dd, J2,2Ј 15.9 and J2,3 3.3,
2-H), 2.50 (1 H, dd, J2,2Ј 15.9 and J2Ј,3 3.2, 2-HЈ), 4.97 (1 H, dd,
J3,4 3.4 and J4,5 9.4, 4-H), 5.31 (1 H, ddd, J4,5 9.4, J5,6 10.2 and
J5,6Ј 4.2, 5-H), 5.50 (1 H, m, 3-H); δC(81.3 MHz; CDCl3) 21.0,
21.1, 21.2 and 21.4 (4 × CH3), 32.0 and 36.3 (2- and 6-C), 66.7,
67.7 and 71.3 (3-, 4- and 5-C), 78.8 (1-C), 170.1–170.5 (4 ×
CO2Me), 173.1 (CO2H); m/z (TOF-PDMS) 1102 (M ϩ Na)ϩ,
1080 (M ϩ H)ϩ, 1022 (M ϩ H Ϫ tBu)ϩ.
8.35 (1 H, d, Jα,NH 7.6, cysteinyl NH); δC(81.3 MHz; H2O–D2O
90:10) 22.7 (lysyl γ-C), 28.3 (lysyl δ-C), 29.4 [C(CH3)3], 30.9
(lysyl β-C), 37.5 (2 × 2-C), 39.4 (lysyl ε-C), 40.7 (cysteinyl β-C
and 2 × 6-C), 41.8 (glycyl α-C), 48.6 [C(CH3)3], 53.1 (cysteinyl
α-C), 54.4 (lysyl α-C), 66.7 and 66.8 (2 × 3-C), 70.8 (2 × 5-C),
75.4 (2 × 4-C), 77.2 (2 × 1-C), 172.5, 173.4, 174.4, 177.1 and
177.4 (5 × C᎐O); m/z (ESI-MS) 741 (M Ϫ H)Ϫ.
᎐
General procedure for solid-phase synthesis of compound 1 and
tripeptide 19
(1sn,3R,4sn,5R)-1,3,4,5-Tetraacetoxycyclohexanecarbonyl
fluoride 18
An Fmoc solid-phase peptide-synthetic strategy was adopted
for the preparation of these compounds. The syntheses were
performed on a 2 mmol scale starting with Fmoc-Gly-O-Wang
resin ester (substitution, 0.8 mmol gϪ1) (Novabiochem, Switzer-
land), with 4 or 2 equiv. of amino acids for the preparation of 1
and 19, respectively and HBTU–DIPEA as activating system in
NMP, and by monitoring the acylation and the deprotection
reactions during chain assembly using TNBS and ninhydrin
tests. Single-coupling protocol was performed. Shikimic acid (2
equiv. per amino group) was preactivated with HBTU–DIPEA
(1:1 equiv. per amino group) for 30 s in NMP and then added
to peptidyl resin 7a in NMP containing DIPEA (1 equiv. per
amino group). After 40 min, the solvent was removed by filtra-
tion, and the resin was washed successively with NMP (3 × 2
min), and DCM (3 × 2 min). After drying over P2O5 the com-
pounds were released from the resin upon treatment with TFA–
H2O–Me2S (95:2.5:2.5; 25 mL) at room temperature for 2 h.
The solutions were concentrated under reduced pressure, and
the residue were solubilised in water and lyophilised to furnish
the crude compounds.
To
a solution of -(Ϫ)-1,3,4,5-tetraacetoxycyclohexane-1-
carboxylic acid (500 mg, 1.39 mmol) in CH2Cl2 (6 mL) contain-
ing pyridine (112 µL, 1.39 mmol) was added dropwise cyanuric
fluoride (937 µL, 11.10 mmol). The reaction mixture was held at
reflux under N2 for 2 h. The mixture, from which a white pre-
cipitate had settled, was filtered through Celite and extracted
with water. Removal of the solvent from the organic layer after
drying with Na2SO4 provided an oil, which was further dried
under vacuum for several hours. The crude compound 18 was
characterised without purification: δH(300 MHz; CDCl3) 1.97–
2.05 (1 H, m, 6-H), 2.03, 2.06, 2.08 and 2.16 (4 × 3 H, 4 s,
4 × CH3), 2.45 (1 H, dd, J2,2Ј 15.4 and J2,3 3.7, 2-H), 2.50–2.66
(2 H, m, 2-HЈ and 6-HЈ), 5.08 (1 H, dd, J3,4 3.5 and J4,5 9.1,
4-H), 5.38 (1 H, ddd, J4,5 9.1, J5,6Ј 9.1 and J5,6 5.1, 5-H), 5.56
(1 H, m, 3-H); δc(81.3 MHz; CDCl3) 20.4, 20.6, 20.8 and 20.9
(4 × CH3), 31.9 and 35.4 (2- and 6-C), 66.0, 68.2 and 70.4 (3-,
4- and 5-C), 76.7 (JC,F 53.2, 1-C), 160.3 (JC,F 373.4, COF),
169.6, 169.7, 169.8 and 169.8 (4 × CO2Me).
N ꢀ,N ꢁ-Bis-[(3R,4S,5R)-3,4,5-trihydroxycyclohexenecarb-
onyl]-L-lysyl-[S-(tert-butylthio)]-L-cysteinyl-glycine 1. 1.64 g of
a pale yellow powder was obtained after lyophilisation. A sam-
ple of the residue (107 mg) was purified by semi-preparative
RP-HPLC [gradient: 100:0 to 50:50 (A/B), 110 min], yielding
15 mg (14%), of pure compound 1; [α]D25 Ϫ117 (c 0.19, H2O);
δH(300 MHz; H2O–D2O 90:10) 1.16 [3 × 3 H, 3 s, C(CH3)3],
1.14–1.35 (2 H, m, 2 × lysyl γ-H), 1.37–1.44 (2 H, m, 2 × lysyl
δ-H), 1.62–1.75 (2 H, m, 2 × lysyl β-H), 2.00–2.11 (2 H, m,
2 × 6-H), 2.59 (1 H, dd, J 11.3 and J 17.4, 6-H), 2.60 (1 H, dd,
J 11.1 and J 17.0, 6-H), 2.90 (1 H, dd, Jα,β 7.0 and Jβ,βЈ 14.0,
cysteinyl β-H), 3.04–3.16 (2 H, m, 2 × lysyl ε-H), 3.09 (1 H, dd,
Jα,β 4.9 and Jβ,βЈ 14, cysteinyl β-H), 3.58 and 3.59 (2 H, 2 dd, J3,4
4.6 and J4,5 9.2, 2 × 4-H), 3.73–3.76 (2 H, m, 2 × glycyl α-H),
3.83–3.88 (2 H, m, 2 × 5-H), 4.17 (1 H, m, lysyl α-H), 4.25–4.29
(2 H, m, 2 × 3-H), 4.52 (1 H, m, cysteinyl α-H), 6.22 and 6.28
(2 H, 2 br d, J2,3 4.0, 2 × 2-H), 7.86 (1 H, t, Jε,NH 5.4, lysyl εNH),
8.02 (1 H, t, Jα,NH 5.6, glycyl NH), 8.06 (1 H, d, Jα,NH 6.6, lysyl
αNH), 8.32 (1 H, d, Jα,NH 7.7, cysteinyl NH); δC(81.3 MHz;
H2O–D2O 90:10) 22.9 (lysyl γ-C), 28.3 (lysyl δ-C), 29.4
[C(CH3)3], 30.7 (lysyl β-C), 31.7 and 31.5 (2 × 6-C), 39.8 (lysyl
ε-C), 40.6 (cysteinyl β-C), 41.8 (glycyl α-C), 48.7 [C(CH3)3],
53.3 (cysteinyl α-C), 54.9 (lysyl α-C), 66.3 and 66.8 (2 × 4-C and
2 × 5-C), 71.9 and 72.0 (2 × 3-C), 131.1 and 132.1 (2 × 1-C),
133.1 and 133.7 (2 × 2-C), 170.5, 171.0, 172.7, 173.4 and 174.6
N ꢀ,N ꢁ-Bis-[(1sn,3R,4sn,5R)-1,3,4,5-tetrahydroxycyclohexane-
carbonyl]-L-lysyl-[S-(tert-butylthio)]-L-cysteinyl-glycine 3
An Fmoc solid-phase peptide-synthetic strategy was adopted
for the preparation of this compound. The synthesis was per-
formed on a 2 mmol scale starting with Fmoc-Gly-Sasrin ester
resin (substitution, 0.8 mmol gϪ1) (Bachem, Switzerland), 2
equiv. of amino acids, HBTU–DIPEA 2:3 equiv. in N-methyl-
2-pyrrolidone (NMP) and by monitoring the acylation and the
deprotection reactions during chain assembly using TNBS and
ninhydrin tests. Peptidyl resin 7b was acylated with quinic acid
derivative 18 (2 equiv. per amino group), with DIPEA (3 equiv.
per amino group), in NMP at room temperature for 1 h and
then washed successively with NMP (3 × 2 min), and then
DCM (3 × 2 min). This reaction was repeated using 1 equiv. per
amino group of both reagents. The compound was cleaved
from the resin upon treatment with CH2Cl2–TFA–triisoprop-
ylsilane (95:2:3; 4 × 20 mL) for 4 × 5 min. The solution was
concentrated under reduced pressure, the residue dissolved in
water and lyophilised to furnish a crude compound (1.40 g),
which was then dissolved in MeOH (25 mL). Methanolic 1 M
sodium methoxide was added dropwise to raise the apparent
pH ≈ 9. The reaction mixture was allowed to stand at room
temperature and was monitored by RP-HPLC. After 3 h, the
mixture was concentrated under reduced pressure. A sample of
the residue (100 mg) was purified by semi-preparative RP-
HPLC [gradient: 100:0 to 90:10 in 10 min, then 90:10 to 50:50
(A:B) in 70 min] to yield compound 3 (41 mg, 59%) following
lyophilisation; [α]D25 Ϫ73 (c 0.49, H2O); δH(300 MHz; H2O–D2O
90:10) 1.12 (3 × 3 H, s, 3 × CH3), 1.15–1.22 (2 H, m, 2 × lysyl
γ-H), 1.34 (2 H, quintet, Jγ,δ 7.2 and Jδ,ε 7.2, 2 × lysyl δ-H),
1.57–1.90 (10 H, m, 2 × lysyl β-H, 4 × 2-H and 4 × 6-H), 2.83
(1 H, dd, Jα,β 8.6 and Jβ,βЈ 14, cysteinyl β-H), 3.02 (3 H, m,
cysteinyl β-HЈ and 2 × lysyl ε-H), 3.33 (2 H, dd, J3,4 2.9 and J4,5
9.6, 2 × 4-H), 3.79 (2 H, d, Jα,NH 5.8, 2 × glycyl α-H), 3.80–3.90
(2 H, m, 2 × 5-H), 4.01 (2 H, br s, 2 × 3-H), 4.11 (1 H, dt, Jα,NH
7.0 and Jα,β 7.2, lysyl α-H), 4.54 (1 H, ddd, Jα,βЈ 6.0, Jα,NH 7.6 and
(5 × C᎐O); m/z (TOF-PDMS) 729 (M ϩ Na)ϩ.
᎐
L-Lysyl-[S-(tert-butylthio)]-L-cysteinyl-glycine bis(trifluoro-
acetate) salt 19. 1.20 g of a pale yellow powder was obtained
after lyophilisation. A sample of the residue was purified by
semi-preparative RP-HPLC [gradient: 100:0 to 85:15 (A/B) for
15 min, then 85:15 to 45:55 (A/B) for 100 min] to furnish
compound 19 (120 mg, 60%), and a dimeric compound 20 (15
mg, 4%); δH(300 MHz; H2O–D2O 90:10) 1.23 [3 × 3 H, 3 s,
C(CH3)3], 1.25–1.32 (2 H, m, 2 × lysyl γ-H), 1.59 (2 H, quintet,
Jγ,δ 7.6 and Jδ,ε 7.6, 2 × lysyl δ-H), 1.79–1.87 (2 H, m, 2 × lysyl
β-H), 2.89 (2 H, br s, 2 × lysyl ε-H), 2.96 (1 H, dd, Jα,β 7.2 and
Jβ,βЈ 14.0, cysteinyl β-H), 3.09 (1 H, dd, Jα,β 5.6 and Jβ,βЈ 14.0,
cysteinyl β-H), 3.88 (2 H, br s, 2 × glycyl α-H), 3.96 (1 H, t, Jα,β
ε
Jα,β 8.6, cysteinyl α-H), 8.02 (1 H, t, Jε,NH 5.9, lysyl NH), 8.11
(1 H, d, Jα,NH 7.0, lysyl αNH), 8.21 (1 H, t, Jα,NH 5.8, glycyl NH),
J. Chem. Soc., Perkin Trans. 1, 1999, 2967–2975
2973

View Article Online
6.4, lysyl α-H), 4.60–4.64 (1 H, m, 2 × cysteinyl α-H), 8.45 (1 H,
t, Jα,NH 5.6, glycyl NH), 8.80 (1 H, d, Jα,NH 7.2, cysteinyl NH);
δC(81.3 MHz; H2O–D2O 90:10) 21.4 (lysyl γ-C), 26.8 (lysyl
δ-C), 29.4 [C(CH3)3], 30.8 (lysyl β-C), 39.5 (lysyl ε-C), 40.7
(cysteinyl β-C), 42.1 (glycyl α-C), 48.8 [C(CH3)3], 53.3 and 53.5
Owing to the repetitive structure of dendrimers only selected
NMR data are reported.
Tetramer 21 (3.39 mg, 61%) was obtained after semi-
preparative RP-HPLC [gradient: 100:0 to 70:30 (A/C), 70
min], as a white powder; δH(600 MHz; H2O–D2O 90:10)
1.31–1.46 (8 H, m, 8 × lysyl γ-H), 1.43–1.60 (6 H, m, 6 × lysyl
δ-H), 1.64 (2 H, m, 2 × lysyl δ-H), 1.68–1.89 (8 H, m,
8 × lysyl β-H), 2.17–2.23 (4 H, m, 4 × 6-H), 2.46 and 2.51
(2 × 1 H, 2 dt, Jα,αЈ 15.4 and Jα,β 6.2, 2 × alanyl α-H), 2.70–
2.80 (4 H, m, 4 × 6-HЈ), 2.89–2.95 (2 H, m, 2 × cysteinyl
β-H), 2.96–3.01 (2 H, m, 2 × lysyl ε-H), 3.06 (2 H, dd, Jα,β
5.3 and Jβ,βЈ 14.9, 2 × cysteinyl β-HЈ), 3.21 (2 H, m, 2 × lysyl
ε-H),), 3.25 (2 H, m, 2 × lysyl ε-H), 3.27 (1 H, d, J 16.5, CHH),
3.30 (1 H, d, J 16.5, CHH), 3.33 (2 H, s, CH2), 3.41–3.50 (2 × 1
H, 2 ddt, Jα,β 6.2, Jβ,βЈ 12.6 and Jβ,NH 6.2, 2 × β-alanyl β-H),
3.71–3.76 (4 H, m, 4 × 4-H), 3.95 (4 H, d, Jα,NH 6.3, 4 × glycyl
α-H), 4.00–4.05 (4 H, m, 4 × 5-H), 4.25 (2 H, dt, Jα,β 5.6 and
Jα,NH 6.3, 2 × lysyl α-H), 4.32–4.38 (4 H, m, 4 × lysyl α-H),
4.42–4.44 (4 H, m, 4 × 3-H), 4.58–4.62 (2 H, m, 2 × cysteinyl
α-H), 6.38 and 6.44 (2 × 2 H, 2 br d, J2,3 4.0, 2 × 2 × 2-H), 6.90
(1 H, s, NHH), 7.54 (br s, εNH2), 7.60 (1 H, s, NHH), 8.07 (2 H,
t, Jε,NH 6.8, 2 × lysyl εNH), 8.15 (1 H, t, Jα,NH 6.2, β-alanyl NH),
(cysteinyl α-C and lysyl α-C), 170.0, 172.7 and 173.8 (3 × C᎐O);
᎐
m/z (TOF-PDMS) 416 (M ϩ Na)ϩ, 395 (M ϩ H)ϩ.
N ꢁ-{L-Lysyl-[S-(tert-butylthio)]-L-cysteinyl-glycinyl}-L-
lysyl-[S-(tert-butylthio)]-L-cysteinyl-glycine tris(trifluoroacetate)
salt 20 (dimeric structure). δH(600 MHz; H2O–D2O 90:10) 1.34
[3 × 3 H, 3 s, C(CH3)3], 1.36–1.46 (4 H, m, 4 × lysyl γ-H), 1.55
(2 H, quintet, Jγ,δ 7.5 and Jδ,ε 7.5, 2 × lysyl δ-H), 1.70 (2 H,
quintet, Jγ,δ 7.5 and Jδ,ε 7.9, 2 × lysyl δ-H), 1.88–1.98 (4 H, m,
4 × lysyl β-H), 3.00 (2 H, br s, 2 × lysyl ε-H), 3.05–3.13 (2 H, m,
2 × cysteinyl β-H), 3.19–3.22 (4 H, m, 2 × cysteinyl β-H and
2 × lysyl ε-H), 3.91 (2 H, d, Jα,NH 5.8, 2 × glycyl α-H), 3.98 (2 H,
d, Jα,NH 5.8, 2 × glycyl α-H), 4.03–4.07 (2 H, m, 2 × lysyl α-H),
4.60–4.64 (2 H, m, 2 × cysteinyl α-H), 7.55 (br s, 4 × αNH2 and
2 × εNH2), 7.81 (1H, t, Jα,NH 5.9, lysyl εNH), 8.46 (1 H, t, Jα,NH
5.8, glycyl NH), 8.61 (1 H, t, Jα,NH 5.7, glycyl NH), 8.84 (1 H, d,
J
α,NH 7.4, cysteinyl NH), 9.02 (1 H, d, Jα,NH 7.6, cysteinyl NH);
m/z (TOF-PDMS) 772 (M ϩ H)ϩ.
ε
8.24 (1 H, t, Jε,NH 6.3, lysyl NH), 8.25–8.27 (2 H, m, 2 × lysyl
αNH), 8.31 and 8.32 (2 × 1 H, 2 t, Jα,NH 6.3, 2 × glycyl NH),
8.42 (1 H, d, Jα,NH 6.8, lysyl αNH), 8.47 (1 H, d, Jα,NH 6.3, lysyl
αNH), 8.61 (2 H, d, Jα,NH 7.6, 2 × cysteinyl NH); m/z (ESI-MS)
N ꢀ,N ꢁ-Bis-[(2R,3S,4S,5R)-2,3,4,5,6-Pentahydroxycyclohexane-
carbonyl]-L-lysyl-[S-(tert-butylthio)]-L-cysteinyl-glycine 5
1660.4 (M Ϫ H)Ϫ, 829.9 (M Ϫ 2H)2Ϫ
.
To a solution of 19 (140 mg, 0.22 mmol) in MeOH (5 mL) were
added DIPEA (137 µL, 0.79 mmol) and -galactonolactone
(160 mg, 0.90 mmol). Following 24 h at reflux, a further amount
of -galactonolactone (80 mg, 0.45 mmol), was added to the
reaction mixture and the reflux was maintained for 24 h. After
completion of the reaction, monitored by RP-HPLC, the sol-
vent was evaporated off under reduced pressure and the residue
was purified by RP-HPLC [gradient: 100:0 to 80:20 (A:B), 40
min], to furnish 5 (116 mg, 69%), following lyophilisation; [α]D25
Ϫ5.7 (c 0.77, H2O); δH(300 MHz; H2O–D2O 90:10) 1.17 [3 × 3
H, 3 s, C(CH3)3], 1.24–1.28 (2 H, m, 2 × lysyl γ-H), 1.30–1.48
(2 H, m, 2 × lysyl δ-H), 1.63–1.79 (2 H, m, 2 × lysyl β-H), 2.85
(1 H, dd, Jα,β 9.0 and Jβ,βЈ 14.0, cysteinyl β-H), 3.07 (1 H, Jα,βЈ 4.8
and Jβ,βЈ 14, cysteinyl β-HЈ), 3.11–3.16 (2 H, m, 2 × lysyl ε-H),
3.50–3.57 (6 H, m, 2 × 2-H and 4 × 6-H), 3.75–3.85 (6 H, m,
2 × glycyl α-H, 2 × 3-H and 2 × 5-H), 4.21 (1 H, dt, Jα,β 5.7
and Jα,NH 6.9, lysyl α-H), 4.27 and 4.35 (2 H, 2 br d, J3,4 1.1,
2 × 4-H), 4.58 (1 H, ddt, Jα,NH 7.8, Jα,β 9.0 and Jα,βЈ 4.8,
Tetramer 22 (5.98 mg, 66%) was obtained after semi-
preparative RP-HPLC [gradient: 100:0 to 75:25 (A/C), 25 min;
then isocratic] as a white powder; δH(300 MHz; H2O–D2O
90:10) 1.22–1.28 (8 H, m, 8 × lysyl γ-H), 1.34–1.43 (6 H, m,
6 × lysyl δ-H), 1.52 (2 H, m, 2 × lysyl δ-H), 1.57–1.64 (8 H, m,
8 × lysyl β-H), 1.65–1.99 (16 H, m, 8 × 2-H and 8 × 6-H), 2.33
(2 H, t, Jα,β 6.3, 2 × β-alanyl α-H), 2.69–2.84 (4 H, m, 2 × cystei-
nyl β-H and 2 × lysyl ε-H), 2.92 (2 H, dd, Jα,β 5.3 and Jβ,βЈ 14, 1
2 × cysteinyl β-H), 2.98–3.11 (6 H, m, 6 × lysyl ε-H), 3.15 and
3.19 (2 × 2 H, 2 s, 2 × CH2), 3.23 and 3.34 (2 × 1 H, 2 ddt, Jα,β
6.3, Jβ,βЈ 13.0 and Jβ,NH 5.7, 2 × β-alanyl β-H), 3.38 (4 H, dd, J3,4
3.1 and J4,5 9.7, 4 × 4-H), 3.76 and 3.77 (4 H, 2 t, Jα,NH 6.3,
2 × glycyl α-H), 3.87–3.96 (4 H, m, 4 × 5-H), 4.06 (4 H, br s,
4 × 3-H), 4.10–4.16 (4 H, m, 4 × lysyl α-H), 4.60–4.64 (2 H, m,
2 × cysteinyl α-H), 6.73 and 7.42 (2 × 1 H, 2 s, NHH and
NHH), 7.96 (1 H, t, Jα,NH 5.7, β-alanyl NH), 8.07 (1 H, t, Jα,NH
6.3, lysyl εNH), 8.10 (1 H, t, Jα,NH 6.9, lysyl εNH), 8.11 (2 H, t,
ε
cysteinyl α-H), 7.96 (1 H, t, Jε,NH 6.0, lysyl NH), 8.12 (1 H, t,
J
α,NH 6.3, 2 × glycyl NH), 8.12 (1 H, t, Jε,NH 6.8, lysyl εNH), 8.13
α
α
Jα,NH 6.1, glycyl NH), 8.15 (1 H, d, Jα,NH 6.9, lysyl NH), 8.16
(2 H, d, Jα,NH 6.8, 2 × lysyl NH), 8.25 (1 H, d, Jα,NH 6.7, lysyl
αNH), 8.29 (1 H, d, Jα,NH 6.3, lysyl αNH), 8.39 (2 H, d, Jα,NH 7.6,
2 × cysteinyl NH; m/z (ESI-MS) 1732.5 (M Ϫ H)Ϫ, 865.6
(1 H, Jα,NH 7.8, cysteinyl NH); δC(81.3 MHz; H2O–D2O
90:10) 22.7 (lysyl γ-C), 28.4 (lysyl δ-C), 29.4 [C(CH3)3],
30.9 (lysyl β-C), 40.5 (lysyl ε-C), 40.5 (cysteinyl β-C), 41.9
(glycyl α-C), 48.7 [C(CH3)3], 53.1 (cysteinyl α-C), 54.5 (lysyl
α-C), 63.7 (2 × 6-C), 69.7 and 69.8 (2 × 2-C), 70.4, 71.2, 71.3,
71.5 and 71.8 (2 × 3-C, 2 × 4-C and 2 × 5-C), 172.7, 173.7,
(M Ϫ 2H)2Ϫ
.
Octamer 23 (2.25 mg, 42%), was obtained after semi-
preparative RP-HPLC [gradient: 100:0 to 80:20 (A/C), 35 min;
then isocratic] as a white powder; ESI-MS: Found: 3235.0.
Calc. 3235.6, m/z 1616.4 (M Ϫ 2H)2Ϫ, 1077.3 (M Ϫ 3H)3Ϫ
,
174.6, 175.8 and 176.7 (5 × C᎐O); m/z (TOF-PDMS) 772
᎐
(M ϩ Na)ϩ.
807.7 (M Ϫ 4H)4Ϫ
.
Octamer 24 (5.75 mg, 56%), was obtained after semi-
preparative RP-HPLC [gradient: 100:0 to 75:25 (A/C), 25 min;
then isocratic] as a white powder; ESI-MS: Found: 3379.0.
General procedure for the formation of sulfides
To a solution of compound 1, 3 or 5 (1.5 equiv. per chloracetyl
group to be substituted in the second step), in a mixture of
Calc.: M, 3379.7; m/z 1688.3 (M Ϫ 2H)2Ϫ, 1125.3 (M Ϫ 3H)3Ϫ
,
n
n
degassed PrOH–H2O 50:50 (1 mL), was introduced Bu3P (1
equiv.). Each mixture was stirred at room temperature under N2
for 24 h. The solvent was evaporated off under reduced pressure
and the residues further dried over P2O5 under vacuum for 15
min. To the crude, reduced compounds dissolved in a mixture
of degassed DMF–H2O (90:10; 500 µL) were added the
-lysine core 13, 14 or 15 (1–5 µmol, 1 equiv.), and the pH
adjusted to a value of 8–8.5 (paper) by adding solid K2CO3.
Each mixture was again stirred at room temperature for 72–96 h
and monitored by RP-HPLC. On completion, each mixture
was diluted in H2O, lyophilised and purified by RP-HPLC to
furnish compounds 21–27.
843.8 (M Ϫ 4H)4Ϫ
.
16-mer 25 (3.15 mg, 43%), was obtained after semi-
preparative RP-HPLC [gradient: 100:0 to 75:25 (A/C), 40 min;
then isocratic], as a white powder; ESI-MS: Found: 6382.0.
Calc.: M, 6383.0; m/z 1594.5 (M Ϫ 4H)4Ϫ, 1275.5 (M Ϫ 5H)5Ϫ
,
1062.8 (M Ϫ 6H)6Ϫ, 910.8 (M Ϫ 7H)7Ϫ
.
16-mer 26 (7.03 mg, 61%) was obtained after semi-
preparative RP-HPLC [gradient: 100:0 to 0:20 (A/C), 35 min;
then isocratic] as a white powder; ESI-MS: Found: 6671.0.
Calc. for M, 6671.2; m/z 2221.9 (M Ϫ 3H)3Ϫ
,
1666.7
(M Ϫ 4H)4Ϫ, 1333.1 (M Ϫ 5H)5Ϫ, 1110.8 (M Ϫ 6H)6Ϫ, 952.0
(M Ϫ 7H)7Ϫ
.
2974
J. Chem. Soc., Perkin Trans. 1, 1999, 2967–2975

View Article Online
8 K. Drickamer, Nature, 1992, 360, 183.
9 W. I. Weis, K. Drickamer and W. A. Hendrickson, Nature, 1992,
360, 127.
16-mer 27 (3.90 mg, 53%) was obtained after semi-
preparative RP-HPLC [gradient: 100:0 to 80:20 (D:E), 40
min; then isocratic], followed by desalting as a white powder;
δH(600 MHz; H2O–D2O 90:10) 1.26–1.43 (32 H, m, 32 × lysyl
γ-H), 1.44–1.62 (30 H, m, 30 × lysyl δ-H), 1.65–1.94 (34 H, m,
2 × lysyl δ-H and 32 × lysyl β-H), 2.48 (2 H, t, Jα,β 6.3, 2 ×
alanyl α-H), 2.89–3.01 (10 H, m, 8 × cysteinyl β-H and 2 × lysyl
ε-H), 3.11 (8 H, Jα,βЈ 4.0 and Jβ,βЈ 13.5, 8 × cysteinyl β-HЈ), 3.17–
3.22 and 3.24–3.32 (30 H, m, 30 × lysyl ε-H), 3.30 (8 H, s,
4 CH2), 3.33 (4 H, d, J 15.2, 4 × CHH), 3.37 (4 H, d, J 15.2,
4 × CHH), 3.40–3.43 and 3.49–3.57 (2 × 1 H, m, 2 β-alanyl
β-H), 3.65–3.73 (56 H, m, 8 glycyl α-H, 16 × 2-H and 32 ×
6-H), 3.83 (4 H, br d, Jα,NH 6, × 4 glycyl α-H), 3.86 (4 H, d, Jα,NH
5.9, 4 × glycyl α-H), 3.96–3.99 (16 H, m, 16 × 5-H), 4.01 (16 H,
br dd, J2,3 9.5 and J3,4 3.3, 16 3-H), 4.18–4.41 (16 H, m,
16 × lysyl α-H), 4.43 and 4.53 (2 × 8 H, 2 br s, 16 × 4-H), 4.58–
4.64 (8 H, m, 8 × cysteinyl α-H), 6.88 (1 H, s, NHH), 7.58
(1 H, s, NHH), 7.95–8.00 (9 H, m, 8 × glycyl NH and lysyl
εNH), 8.03 (1 H, t, Jα,NH 6.2, β-alanyl NH), 8.10–8.15 (9 H, m,
lysyl αNH and 8 × lysyl εNH), 8.18 (1 H, t, Jα,NH 6.2, lysyl εNH),
8.19–8.23 (4 H, m, 4 × lysyl εNH), 8.25–8.29 (10 H, m, 2 × lysyl
αNH and 8 × cysteinyl NH), 8.30 (1 H, d, Jα,NH 6.4, lysyl αNH),
8.32 (8 H, br d, Jα,NH 6.0, 8 × lysyl αNH), 8.40 (2 H, br d, Jα,NH
10 J. P. Tam, Proc. Natl. Acad. Sci. USA, 1988, 85, 540.
11 (a) R. Roy, D. Zanini, S. J. Meunier and A. Romanowska, J. Chem.
Soc., Chem. Commun., 1993, 1869; (b) D. Zanini, W. K. C. Park and
R. Roy, Tetrahedron Lett., 1995, 36, 7383; (c) D. Pagé, D. Zanini and
R. Roy, Bioorg. Med. Chem., 1996, 4, 1949; (d) M. L. Palcic, H. Li,
D. Zanini, R. S. Bhella and R. Roy, Carbohydr. Res., 1998, 305, 433;
(e) R. Roy, W. K. C. Park and D. Zanini, Carbohydr. Lett., 1997, 2,
259; ( f ) A. R. P. M. Valentijn, G. A. van der Marel, L. A. J. M.
Sliedregt, T. J. C. van Berkel, E. A. L. Biessen and J. H. van Boom,
Tetrahedron, 1997, 53, 759.
12 E. Kaiser, R. L. Colescott and B. J. Williams, Anal. Biochem., 1970,
34, 595.
13 W. S. Hancock and J. E. Battersby, Anal. Biochem., 1976, 71, 261.
14 M. Schnölzer, P. Alewood, A. Jones, D. Alewood and S. B. H. Kent,
Int. J. Pept. Protein Res., 1992, 40, 180.
15 R. B. Merrifield (a) J. Am. Chem. Soc., 1963, 85, 2149; (b) Science,
1986, 232, 341.
16 Double-bond isomerisation from the C-1/C-6 to the C-1/C-2
positions has been realised in dry HF: R. Grewe and S. Kersten,
Chem. Ber., 1967, 100, 2546.
17 G. B. Fields and R. L. Noble, Int. J. Pept. Protein Res., 1990, 35,
161.
18 E. Wünsch and R. Spangerberg, in Peptides 1969, ed. E. Scoffore,
North-Holland, Amsterdam, 1971, p. 30.
α
α
6.3, 2 × lysyl NH), 8.43 (2 H, br s, 2 × lysyl NH); ESI-MS:
Found 6732.5. Calc. 6733.0; Found: 6555.0. Calc. for (M Ϫ
19 M. Mergler, R. Tanner, J. Gosteli and P. Grogg, Tetrahedron Lett.,
1988, 29, 4005.
20 H. O. L. Fischer, Ber. Dtsch. Chem. Ges., 1921, 54, 775.
21 L. A. Carpino, D. Sadat-Aalaee, H. Guang Chao and R. H.
DeSelms, J. Am. Chem. Soc., 1990, 112, 9651.
C6H10O6): 6554.9; m/z 2243.3 (M Ϫ 3H)3Ϫ, 1682.9 (M Ϫ 4H)4Ϫ
.
22 H. Wenschuh, M. Beyermann, E. Kraise, M. Brudel, R. Winter,
M. Schümann, L. A. Carpino and M. Bienert, J. Org. Chem., 1994,
59, 3275.
23 Synthesis of glycodendrimers based upon sugar-lactone opening
by polyamines has been described: K. Aoi, K. Itoh and M. Okada,
Macromolecules, 1995, 28, 5391.
Acknowledgements
This work was financially supported by the ANRS and by the
Fondation pour la Recherche Médicale (Sidaction) (to C. G.).
We are grateful to G. Ricart for recording ES-MS data, to J.-M.
Wieruszeski for the recording of NMR spectra and to S.
Brooks for proof-reading the manuscript.
24 M. Harnois-Ponton, M. Monsigny and R. Mayer, Anal. Biochem.,
1991, 193, 248.
25 (a) D. Zanini and R. Roy, J. Org. Chem., 1996, 61, 7348; (b) S. J.
Meunier and R. Roy, Tetrahedron Lett., 1996, 37, 5469.
26 (a) C. Kieburg and T. K. Lindhorst, Tetrahedron Lett., 1997, 38,
3885; (b) N. Jayaraman and J. F. Stoddart, Tetrahedron Lett., 1997,
38, 6767.
27 T. K. Lindhorst and C. Kieburg, Angew. Chem., Int. Ed. Engl., 1996,
35, 1953.
References
1 H. Lis and N. Sharon, Chem. Rev., 1998, 98, 637.
2 Y. C. Lee and R. T. Lee, Acc. Chem. Res., 1995, 28, 321.
3 (a) D. Zanini and R. Roy, in Carbohydrate Mimics. Concepts and
Methods, ed. Y. Chapleur, Wiley-VCH, Weinheim, 1998, p. 385;
(b) H.-F. Chow, T. K.-K. Mong, N. F. Nongrum and C.-W. Wan,
Tetrahedron, 1998, 54, 8543.
28 J. P. Defoort, B. Nardelli, W. Huang and J. P. Tam, Int. J. Pept.
Protein Res., 1992, 40, 214.
29 N. C. J. M. Beekman, W. M. M. Schaaper, G. I. Tesser,
K. Dalsgaard, S. Kamstrup, J. P. M. Langeveld, R. S. Boshuizen and
R. H. Meloen, J. Pept. Res., 1997, 50, 357.
30 L. Moroder, M. Gemeiner, W. Göhring, E. Jaeger, J. Musiol,
R. Scharf, H. Stocker, E. Wünsch, L. Pradayrol, N. Vaysse and
A. Ribet, Hoppe-Seyler’s Z. Physiol. Chem., 1981, 362, 697.
31 T. W. Muir, M. J. Williams, M. H. Ginsberg and S. B. H. Kent,
Biochemistry, 1994, 33, 7701.
4 (a) A. Avrameas, D. McIlroy, A. Hosmalin, B. Autran, P. Debré,
M. Monsigny, A. C. Roche and P. Midoux, Eur. J. Immunol., 1996,
26, 394; (b) P. D. Stahl, Curr. Opin. Immunol., 1992, 4, 49.
5 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
Heiden, D. Scheidegger, L. C. J. M. Oomen, T. H. M. Ottenhoff,
A. Tulp, J. J. Neefjes and F. Koning, Eur. J. Immunol., 1997, 27,
2426.
32 J. H. Naismith and R. A. Field, J. Biol. Chem., 1996, 271, 972.
33 R. Guenin and C. H. Schneider, Helv. Chim. Acta, 1983, 66, 1101.
34 Quinic acid derivatives have been named following Eliel’s
recommendations: E. L. Eliel and M. Bello Ramirez, Tetrahedron:
Asymmetry, 1997, 8, 3551.
6 The use of carbohydrate mimics tends to generalise. In our context,
the dendrimers could be seen as mimics of biomimetics. For a good
overview see: (a) Carbohydrate Mimics. Concepts and Methods,
ed. Y. Chapleur, Wiley-VCH, Weinheim, 1998; (b) E. I. Simanek,
G. J. McGarvey, J. A. Jablonowski and C.-H. Wong, Chem. Rev.,
1998, 98, 833.
Paper 9/04679H
7 W. I. Weis and K. Drickamer, Annu. Rev. Biochem., 1996, 65, 441.
J. Chem. Soc., Perkin Trans. 1, 1999, 2967–2975
2975