Trivalent, Gal/GalNAc-containing ligands designed for the asialoglycoprotein receptor

Article abstract of DOI:10.1016/j.bmc.2008.03.017

A series of novel, fluorescent ligands designed to bind with high affinity and specificity to the asialoglycoprotein receptor (ASGP-R) has been synthesized and tested on human liver cells. The compounds bear three non-reducing, β-linked Gal or GalNAc moieties linked to flexible spacers for an optimal spatial interaction with the binding site of the ASGP-R. The final constructs were selectively endocytosed by HepG2 cells derived from parenchymal liver cells-the major human liver cell type-in a process that was visualized with the aid of fluorescence microscopy. Furthermore, the internalization was analyzed with flow cytometry, which showed the process to be receptor-mediated and selective. The compounds described in this work could serve as valuable tools for studying hepatic endocytosis, and are suited as carriers for site-specific drug delivery to the liver.

Full text of DOI:10.1016/j.bmc.2008.03.017

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry 16 (2008) 5216–5231
Trivalent, Gal/GalNAc-containing ligands designed
for the asialoglycoprotein receptor
Oleg Khorev, Daniela Stokmaier, Oliver Schwardt, Brian Cutting and Beat Ernst*
Institute of Molecular Pharmacy, Pharmacenter—University of Basel, Klingelbergstrasse 50, CH-4056 Basel, Switzerland
Received 6 July 2007; revised 26 February 2008; accepted 4 March 2008
Available online 7 March 2008
Abstract—A series of novel, fluorescent ligands designed to bind with high affinity and specificity to the asialoglycoprotein receptor
(ASGP-R) has been synthesized and tested on human liver cells. The compounds bear three non-reducing, b-linked Gal or GalNAc
moieties linked to flexible spacers for an optimal spatial interaction with the binding site of the ASGP-R. The final constructs were
selectively endocytosed by HepG2 cells derived from parenchymal liver cells—the major human liver cell type—in a process that was
visualized with the aid of fluorescence microscopy. Furthermore, the internalization was analyzed with flow cytometry, which
showed the process to be receptor-mediated and selective. The compounds described in this work could serve as valuable tools
for studying hepatic endocytosis, and are suited as carriers for site-specific drug delivery to the liver.
ꢀ 2008 Elsevier Ltd. All rights reserved.
1. Introduction
Kupffer cell receptor, the macrophage galactose lectin,
and the scavenger receptor C-type lectin (SRCL).4–7
Their binding properties were recently profiled by
Drickamer et al.8
The asialoglycoprotein receptor (ASGP-R) is located on
hepatocytes and is a Ca2+-dependent carbohydrate-
binding protein, or C-type lectin. It is expressed on
mammalian liver cells.1 Its main function is to maintain
serum glycoprotein homeostasis by the recognition,
binding, and endocytosis of asialoglycoproteins
(ASGPs), that is, desialylated glycoproteins with termi-
nal galactose or GalNAc residues. After internalization
via clathrin-coated pits and their fusion with endosomes,
the ASGPs are released in the acidic environment of the
endosome and transported to lysosomes for degrada-
tion, while the receptor is recycled back to the cell
surface.2,3
The affinity and specificity of the ASGP-R is a conse-
quence of oligovalent interactions with its physiological
ligands, a process termed cluster glycoside effect by Lee
et al.9 The receptor consists of two homologous sub-
units, designated H1 and H2 in the human system,
which form a non-covalent heterooligomeric complex
with an estimated ratio of 2–5:1, respectively. Both
subunits are single-spanning membrane proteins with a
calcium-dependent galactose/N-acetylgalactosamine rec-
ognition domain.10 Recently, the X-ray crystal structure
of the carbohydrate recognition domain (CRD) of the
major subunit H1 was elucidated.11
In addition to the ASGP-R, there are three additional
Gal/GalNAc-receptors in the C-type lectin family: the
Many studies have been performed with both natural
and synthetic carbohydrates to establish the structure–
affinity relationship for the ASGP-R. Baenzinger
et al.12,13 have shown that the human receptor exhibits
specificity for terminal Gal and GalNAc (with an
approximately 50-fold higher affinity for the latter) on
desialylated glycoproteins. Triantennary ligands dis-
played a higher affinity than their mono- and dianten-
nary counterparts. Furthermore, the studies led to the
conclusion that only the terminal residues are necessary
for specific recognition, and that the binding process
proceeds through a simultaneous interaction of 2–3 su-
Abbreviations: NIS, N-iodosuccinimide; TfOH, trifluoromethanesul-
fonic acid; NHS, N-hydroxysuccinimide; PyBOP, benzotriazol-1-yl-
oxytripyrrolidino-phosphonium hexafluorophosphate; Tris, 2-amino-
2-(hydroxymethyl)-1,3-propanediol; EEDQ, ethyl 1,2-dihydro-2-eth-
oxy-1-quinolinecarboxylate; DMTST, dimethyl(methylthio)sulfonium
trifluoromethanesulfonate; DIPEA, diisopropylethylamine; Hepes, 4-
(2-hydroxyethyl)-1-piperazine-1-ethanesulfonic acid.
Keywords: Asialoglycoprotein receptor (ASGP-R); Drug delivery;
Flow cytometry; Fluorescence microscopy; Fluorescent probes.
*
Corresponding author. Tel.: +41 61 267 15 51; fax: +41 61 267 15
52; e-mail: beat.ernst@unibas.ch
0968-0896/$ - see front matter ꢀ 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2008.03.017

O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
5217
gar residues with 2–3 binding sites of the heterooligo-
meric receptor. On the native receptor on the hepatocyte
cytosis, the ASGP-R has been validated as a potential
target for drug and gene delivery to the liver.7,18,19 As
an alternative to ex vivo gene transfer to the liver, which
requires invasive surgery,20 there is much interest in
in vivo protocols: (i) Wu et al.21 demonstrated successful
in vivo gene transfer to hepatocytes with poly-L-lysine
linked asialoorosomucoid, (ii) Hara et al.22–24 showed
that asialofetuin-labeled liposomes that encapsulate
plasmid DNA cause gene expression, and (iii) successful
gene transfer to hepatocytes using liposomal gene carri-
ers that possess synthetic galactose residues as a target-
able ligand for parenchymal liver cells has been
reported.25
˚
surface these binding sites are 25–30 A apart.
Studies on rabbit hepatocytes by Lee et al.,9,14 using syn-
thetic oligosaccharides, further reinforced the binding
hierarchy of polyvalent ligands: tetraantennary > trian-
tennary ꢀ diantennary ꢀ monoantennary. The IC50-
values for mono-, di-, tri-, and tetraantennary oligosac-
charides were found to be approximately 1 · 10ꢁ3
,
1 · 10ꢁ6, 5 · 10ꢁ9, and 10ꢁ9 M, respectively. In other
words, although the number of Gal residues/mol of li-
gand increased only 4-fold, the inhibitory potency in-
creased 1,000,000-fold. Because the fourth Gal moiety
present in the tetraantennary ligand does not markedly
enhance the affinity, it was assumed that the binding
requirements of the cell-surface receptor are largely sat-
isfied by the triantennary structure.15
In order to further exploit the ASGP-R for therapeutic
purposes, trivalent ligands with pendant Gal or GalNAc
residues connected by flexible spacers with appropriate
lengths to a common branching point were synthesized.
All these ligands incorporate 2-amino-2-hydroxymethyl-
1,3-propanediol (Tris) as the branching point (Fig. 2).
Kempen et al.26 synthesized the trivalent, Gal-termi-
nated ligand 1, where the carbohydrate moieties were di-
rectly linked to Tris. When 1 was labeled with
cholesterol and incorporated into liposomes, they were
mainly taken up by the Kupffer cells, via the Gal/Fuc-
recognizing receptor, and not by the parenchymal liver
cells via the ASGP-R.
The optimal distance of the Gal moieties in these
oligosaccharides was determined by binding assays with
synthetic carbohydrates representing partial structures
ofN-linked glycans,16 high-resolution NMR, and molec-
ular modeling studies.17 Based on these results, Lee
et al.9,16 presented a model for the optimal spatial
arrangement of the terminal sugar residues (Fig. 1).
Due to its specificity, predominant expression on hepa-
tocytes, and high capacity for receptor-mediated endo-
Therefore, a new generation of ligands with optimal
spacers was created. Biessen et al.27,28 extended the dis-
tance between the Tris branching point and the Gal res-
idues by using tetraethylene glycol spacers
˚
approximately 20 A in length. This indeed led to ligands
Branching point
Flexible spacer
with improved affinities (see 2, Ki = 0.2 lM, Fig. 2)
determined in a competition assay with 125I-labeled asi-
aloorosomucoid. In 1999, Sliedregt et al.29 designed a
second generation of cluster glycosides containing an
essential modification (see 3, Ki = 93 nM, Fig. 2). To en-
hance the chemical stability, the methylene acetal groups
in 2, which connect the spacers to Tris, were replaced by
acid stable ether bonds. Furthermore, the spacers were
no longer based on tetraethylene glycol to achieve the
appropriate spacing between the Gal residues, but rather
on a twelve atom fragment containing two amide bonds.
Finally, Rensen et al.30 combined the various features
from 2 and 3 to generate compound 4 (Ki = 2 nM,
Fig. 2), which exploited the expected 50-fold higher
affinity of GalNAc over Gal toward the ASGP-R.31
25 Å
Terminal
sugar residue
Gal
Gal
Gal
H1-/H2-CRDs
Stalk region
(coiled-coil)
To further improve the therapeutic profile of the previ-
ously reported ligands of the ASGP-R,26–31 we set out
to synthesize the optimal trivalent linker (19, Scheme
1) with reduced synthetic complexity, high in vivo stabil-
ity, and improved spacer flexibility. The resultant inter-
mediates 5a and 5b (Fig. 2), which possess terminal Gal
or GalNAc moieties, respectively, were then fluorescent-
labeled and tested for selective uptake by hepatocytes
using fluorescence microscopy and flow cytometry.
Moreover, since most of the previous research was done
on rat26–30 and mouse32 liver cells, and the final aim of
this research is liver-selective drug delivery in humans,
all our biological assays were performed using cell lines
of human origin.
Hepatocyte
cell membrane
Figure 1. Binding model for ASGP-R ligands in an optimal confor-
mation to the heterooligomeric receptor consisting of H1 and H2
subunits. Dashed line indicates the distance between the C-4 of each
Gal moieties; filled line represents approximate distance between
16
˚
branching point and C-6 of Gal (14–20 A). Adapted from Lee et al.

5218
O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
OH
HO
HO
O
O
H
N
O
O
N
H
OMe
OH
O
3
1 [26]
OH
HO
HO
O
O
O
H
N
O
O
O
O
N
H
O
O
OMe
OH
3
O
2 [27]
OH
HO
HO
O
O
H
N
H
N
NHCbz
O
O
N
H
OH
3
O
O
3 [29]
OH
HO
HO
O
O
H
N
O
NHAc
O
O
N
H
O
NH2
3
O
4 [30]
OH
HO
HO
O
O
R
O
O
O
O
NHCbz
N
H
3
5a, R = OH
5b, R = NHAc
Figure 2. Trivalent compounds 1–4 were specifically designed for, and tested on, the ASGP-R.26–30 Compounds 5a and 5b are the trivalent, Cbz-
protected intermediates introduced herein.
2. Results and discussion
2.1. Synthesis of fluorescent-labeled, trivalent ligands 6
and 7 and the negative control 8
The main structural features of the trivalent ASGP-R li-
gands 5a and 5b are as follows: (i) Tris is the central
branching point, (ii) the spacers are based on polypro-
pylene oxide, which combines flexibility with amphilici-
ty, (iii) the linkage between Tris and the spacers is a
hydrolytically stable ether bond, and (iv) the length of
the spacers can be easily varied.
Starting from 2-amino-2-hydroxymethyl-1,3-propane-
diol (Tris, 9), the polypropylene oxide spacers were
gradually extended by repetitive allylation/oxidative
hydroboration steps using 9-BBN in THF followed by
H2O2 and aqueous NaOH (Scheme 1). For the synthesis
of compounds 12 and 14, several allylation procedures
were examined using NaH, KOH, K2CO3 as bases in
various solvents (e.g., THF, DMF, dioxane), with and
without the addition of crown ethers, and quaternary
ammonium salts as phase transfer catalysts. All proce-
dures, including the literature procedure used to obtain
1034 in 68%, led to unacceptably low yields of approxi-
mately 40% for 12, along with a considerable amount
of a tetraallylated side product. The desired triallylated
compounds could finally be obtained in almost quanti-
tative yields with only traces of N-allylation, by employ-
ing liquid–liquid phase transfer catalysis.35 Thus, 12 and
14 were obtained in 95% and 90%, respectively, from the
corresponding triols using allyl bromide in refluxing
DCM/50% aqueous NaOH (1:1) with a catalytic amount
of 15-crown-5. Oxidative hydroboration and acetylation
gave 13 and 15 in excellent overall yields. The peracety-
lation step (!11, 13, and 15) was applied in order to
The glycine acylating the amino group of Tris in 4
(Fig. 2) has been replaced with Cbz-protected c-amino-
butyric acid, which upon deprotection furnishes a ver-
satile primary amino group for the attachment of
fluorescent labels and at a later stage therapeutic
agents. For our studies, the amino group was coupled
to Alexa Fluorꢁ 488 fluorescent label33 (!6 and 7,
Fig. 3), but in theory it could also be coupled to a ther-
apeutic agent. As a negative control for the fluores-
cence microscopy studies, and especially to
demonstrate the significance of the polypropylene oxide
spacers featured in our final compounds 6 and 7, we
also synthesized compound 8 (Fig. 3) The latter, in
contrast to 6 and 7, has only short spacers, and there-
fore does not fulfill the spatial requirements for triva-
lent binding to the ASGP-R.

O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
5219
HO
HO
HO
[34]
b)
a)
NH2
O
10
a)
NHBoc
AcO
O
NHBoc
b)
3
3
9
11
AcO
O
O
NHBoc
O
O
O
NHBoc
O
3
3
12
13
a)
O
NHBoc
AcO
O
O
O
NHBoc
3
3
14
15
HO2C
NHCbz
17 [36]
c)
AcO
O
O
O
NH3 Cl
d)
3
16
O
e)
AcO
HO
O
O
O
NHCbz
N
H
3
18
19
O
O
O
O
NHCbz
N
H
3
Scheme 1. Reagents and conditions: (a) i—9-BBN, THF, rt, 24 h, then aq NaOH, H2O2, 0 ꢂC ! rt, 24 h; ii—Ac2O, pyridine, rt, 3 h, (11: 81%; 13:
87%; 15: 88%); (b) i—NaOMe, MeOH, rt, 24 h, quant.; ii—allyl bromide, 15-crown-5, DCM/50% (w/v) aqueous NaOH, reflux, 24 h, (12: 95%; 14:
90%); (c) 4M HCl in dioxane, rt, 30 min, quant.; (d) PyBOP, DIPEA, dioxane/DMF, rt, 24 h, 85%; (e) NaOMe, MeOH, rt, 4 h, 90%.
NH2
SO3
O
M
OH
HO
HO
O
O
H
N
SO3
NH2
O
O
O
O
N
H
OH
3
O
HO2C
6
NH2
SO3
O
M
OH
HO
HO
O
O
H
N
SO3
NH2
O
NHAc
O
O
O
N
H
3
O
HO2C
7
NH2
SO3
O
M
OH
HO
HO
O
H
SO3
NH2
O
N
O
N
H
OH
O
HO2C
3
8
Figure 3. Fluorescent-labeled, trivalent compounds 6, 7, and control 8; M+ are variable counterions.

5220
O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
facilitate purification and characterization of the inter-
mediate triols. The subsequent deacetylation of 11 and
group was cleaved yielding compound 26, which was la-
beled with Alexa Fluorꢁ 488 producing 7 in a 90% yield.
´
13 was achieved under standard Zemplen conditions.
For the elaboration at the N-terminus of 15, the Boc
protecting group was selectively removed using 4 M
HCl in dioxane leading quantitatively to compound
16. Subsequent condensation with the N-Cbz-protected
c-aminobutyric acid linker 1736 using PyBOP in DMF/
dioxane (1:3) and DIPEA as base yielded 18. In the final
As a negative control for cellular assays, compound 8
(Scheme 3) was synthesized via the acylation of Tris
(9) with N-Cbz-protected c-aminobutyric acid (17)36
using EEDQ in pyridine,38 yielding compound 27 in a
77% yield. The latter was then galactosylated with donor
2027 using NIS/TfOH as a promoter to give 28 in 51%.
After debenzoylation (!29), the Cbz group was cleaved
by hydrogenolysis to furnish compound 30, which was
subsequently coupled to the N-hydroxysuccinimidyl
(NHS)-activated Alexa Fluorꢁ 488 fluorescent label
yielding 8 in a 96% yield.
´
step, deacetylation under Zemplen conditions furnished
the trivalent glycosyl acceptor 19 in an overall yield of
27%, starting from Tris (9).
Galactosylation of 19 with ethyl 2,3,4,6-tetra-O-ben-
zoyl-1-thio-b-D-galactopyranoside (20)27 using DMTST
as a promoter furnished the trivalent intermediate 21 in
a 68% yield (Scheme 2). Debenzoylation (!5a) followed
by the cleavage of the Cbz protecting group gave 22,
which was coupled to the N-hydroxysuccinimidyl
(NHS)-activated Alexa Fluorꢁ 488 fluorescent label to
yield compound 6 in 81% yield. Alexa Fluorꢁ 488 was
found to be the optimal fluorescent label for our pur-
poses, combining high chemical and photostability with
high fluorescence intensity. An analogous sequence of
reactions was applied for the synthesis of 7. First, the
N-acetylgalactosamine trimer 24 (Scheme 2) was
obtained in 91% by glycosylating 19 with ethyl 3,4,6-tri-
O-acetyl-2-deoxy-1-thio-2-(2,2,2-trichloroethoxycarbon-
ylamino)-b-D-galactopyranoside (23).37 After the
cleavage of the Troc protecting group, the free amine
was directly acetylated to furnish 25. Upon deprotection
of the N-acetylgalactosamine moieties (!5b), the Cbz
3. Biological evaluation
The trivalent ligands 6–8 were examined for their selec-
tive binding to, and internalization by the ASGP-R
applying fluorescence microscopy and flow cytometry.
Two different cell lines of hepatic origin were used:
HepG2 cells derived from a human hepatocellular carci-
noma expressing the ASGP-R,39 and the human more
endothelial-like SK-Hep1 cells, which lack the
receptor.40
3.1. Fluorescence microscopy
The cells were incubated with the Alexa Fluorꢁ 488-la-
beled compounds 6, 7, or 8 for 1.5 h on ice to allow
binding of the compounds to the receptor while prevent-
OBz
BzO
O
BzO
SEt
OBz
OR1
R1O
R1O
20 [27]
O
O
d)
a)
O
O
O
O
NHR2
6
N
H
OR1
3
21 (R1: Bz; R2: Cbz)
5a (R1: H; R2: Cbz)
22 (R1: H; R2: H)
b)
c)
19
OR1
R1O
R1O
a)
O
O
d)
O
O
O
O
NHR3
7
OAc
AcO
AcO
N
H
NHR2
O
3
SEt
NHTroc
24 (R1: Ac; R2: Troc; R3: Cbz)
25 (R1: Ac; R2: Ac; R3: Cbz)
5b (R1: H; R2: Ac; R3: Cbz)
e)
23 [37]
b)
c)
26 (R1: H; R2: Ac; R3: H)
˚
Scheme 2. Reagents and conditions: (a) DMTST, 4 A MS, DCM, 0 ꢂC ! 10 ꢂC, 48–72 h, 68% for 21, 91% for 24; (b) NaOMe, MeOH/dioxane, rt,
4 h, 94% for 5a, 72% for 5b; (c) H2, Pd/C, EtOH/dioxane, rt, 24 h, 87% for 22, 95% for 26; (d) Alexa Fluorꢁ 488-NHS, DIPEA, 4 A MS, DMF/
˚
dioxane, rt, 4 days, 81% for 6, 90% for 7; (e) Zn dust, Ac2O, dioxane, rt, 24 h, 82%.

O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
5221
HO
O
a)
NH2
HO2C
NHCbz
HO
NHCbz
HO
HO
N
H
3
9
17 [36]
27
OBz
O
BzO
BzO
SEt
OR1
R1O
R1O
OBz
O
e)
20 [27]
NHR2
O
8
O
N
H
b)
OR1
3
28 (R1: Bz; R2: Cbz)
29 (R1: H; R2: Cbz)
30 (R1: H; R2: H)
c)
d)
˚
Scheme 3. Reagents and condition: (a) EEDQ, pyridine, 90 ꢂC, 24 h, 77%; (b) NIS, TfOH, 4 A MS, DCE/Et2O, 0 ꢂC, 1 h, 51%; (c) NaOMe, dioxane/
MeOH, rt, 6 h, 85%; (d) H2, Pd/C, MeOH, rt, 48 h, 87%; (e) Alexa Fluorꢁ 488-NHS, DIPEA, 4 A MS, DMF, rt, 4 days, 96%.
˚
ing unspecific uptake. In the washing step, unbound li-
gand was removed, and the cells were incubated for an
additional 40 min at 37 ꢂC to allow receptor-mediated
endocytosis of bound compounds to take place. The
specific uptake led to punctuate staining of the cells rep-
resenting endosomes containing the ligands, which were
visualized by fluorescence microscopy. HepG2 cells
showed specific uptake of 6 and 7, and only negligible
uptake of 8. The fluorescent content of the endosomes
can be distinctly seen (Fig. 4A and C) for compounds
6 and 7, respectively. Because the cells were grown and
incubated on glass coverslips, which were then mounted
upside down for visualization, enriched fluorescence can
only be observed in cytosolic areas, which are not
blocked by the nuclei. Panel E shows little or no such
fluorescent vesicles, since control compound 8 was not
internalized via the ASGP-R owing to insufficient spacer
length. As expected, no internalization into SK-Hep1
cells (which do not express the ASGP-R) could be ob-
served for compounds 6 and 7 (Fig. 4B and D). How-
ever, compound 8 showed a minor tendency to be
internalized by this cell line in an ASGP-R-independent
manner (Fig. 4F). Panels G and H show the autofluores-
cence of non-treated HepG2 and SK-Hep1 cells as
controls.
in which the uptake leads to a saturation hyperbola as
it is typical for a receptor-mediated process (Fig. 5).41
Uptake of compound 7 into HepG2 cells via the ASGP-
R at a concentration of 10 lM was competitively inhib-
ited by the presence of monosaccharide ligands: GalNAc
(IC50 = 4.55 0.32 mM) (Fig. 6A) and asialofetuin
(IC50 = 45.60 2.70 lM) (Fig. 6B), whereas the uptake
into SK-Hep1 was low and not affected by the presence
of asialofetuin.
In ASGP-R bearing HepG2 cells, uptake of control
compound 8 was low and proved to be unspecific as it
could not be inhibited by asialofetuin, a natural high-
affinity ligand of the receptor (Fig. 6C). ASGP-R-nega-
tive SK-Hep1 cells, on the other hand, evinced high up-
take of compound 8, unaffected by the presence of
asialofetuin (Fig. 6C) which could be explained by their
high endocytic activity that is usually associated with
endothelial cells.
4. Conclusion
Studies using fluorescent-labeled ligands for the ASGP-
R have been carried out before. Ishihara et al.42 pre-
pared fluorescein isothiocyanate-labeled, galactosylated
polystyrene ligands and analyzed their interaction with
the ASGP-R by flow cytometry. Wu et al.43 introduced
a new synthetic route, based on solid phase peptide syn-
thesis, toward fluorescent, synthetic, trivalent, N-acety-
lgalactosamine-terminated glycopeptides43 as ligands
for the ASGP-R.
3.2. Flow cytometry
The ASGP-R-mediated uptake of compounds 7 and 8
(negative control) was quantitatively evaluated by flow
cytometry (Figs. 5 and 6). Instead of performing the pre-
viously described steps (prebinding on ice, removal of
the excess, and internalization of bound compound),
the cells were continuously incubated with the test com-
pounds at 37 ꢂC and analyzed.
However, in this study we have introduced a set of no-
vel, fluorescent, trivalent, simplified oligosaccharide
mimics as ligands for the ASGP-R (6 and 7, Fig. 3).
These compounds not only comply with the afore-men-
tioned optimal ASGP-R ligand criteria, but also are syn-
thetically easily accessible and hydrolytically stable.
The median fluorescence intensity (MFI) of cells incu-
bated with compound 7 at concentrations ranging from
0.4 to 12.5 lM revealed low uptake of the compound
into SK-Hep1 cells compared to that of HepG2 cells,

5222
O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
40
30
20
10
HepG2
SK-Hep1
0
0
5
10
15
Compound 7 (μM)
Figure 5. Titration of compound 7: adherent HepG2 and SK-Hep1
cells were incubated with compound 7 at concentrations ranging from
0.4 to 12.5 lM for 40 min at 37 ꢂC. MFI is the shift in median
fluorescence intensity from untreated to treated cells.
that the sugar residues have to be in an optimal spatial
arrangement in order to interact selectively and with
high affinity with the native ASGP-R. In final analysis,
we have demonstrated that compound 7 has a high po-
tential for use in site-specific delivery of therapeutic
agents (chemotherapeutics, DNA, etc.) to the liver.
The follow-up experiments are currently being
performed.
5. Experimental
5.1. General methods
NMR spectra were recorded on a Bruker Avance DMX-
C
500 (500 MHz) spectrometer. Assignment of 1H and 13
NMR spectra was achieved using 2D methods (COSY,
HSQC, TOCSY). Chemical shifts are expressed in parts
per million using residual CHCl3, CHD2OD, and HDO
as references. Optical rotations were measured using a
Perkin-Elmer Polarimeter Model 341. ESI-MS spectra
were measured on a Waters Micromass ZQ mass spec-
trometer. Reactions were monitored by TLC using glass
plates coated with silica gel 60 F254 (Merck) with the fol-
lowing mobile phases: (A) petrol ether/EtOAc (4:1); (B)
petrol ether/EtOAc (1:1); (C) petrol ether/EtOAc (3:7);
(D) EtOAc; (E) EtOAc/MeOH (9:1); and (F) DCM/
MeOH/H2O (10:4:0.8). Carbohydrate-containing com-
pounds were visualized by charring with a molybdate
solution (0.02 M solution of ammonium cerium sulfate
dihydrate and ammonium molybdate tetrahydrate in
aqueous 10% H2SO4). Compounds 6–8 were visualized
with UV light. All other compounds were visualized
with KMnO4 solution (2% KMnO4 and 4% NaHCO3
in water). Column chromatography was performed on
silica gel 60 (Fluka, 0.040–0.060 mm). Size exclusion
chromatography was performed on Sephadex LH-20
and Sephadex G-15 (Pharmacia). Methanol (MeOH)
was dried by refluxing with sodium methoxide and dis-
tilled immediately before use. Pyridine was freshly dis-
tilled under argon over CaH2. Dichloromethane
(DCM) and dichloroethane (DCE) were dried by filtra-
tion over Al2O3 (Fluka, type 5016 A basic). Tetrahydro-
Figure 4. Fluorescence microscopy images depicting the ASGP-R-
specific uptake of Alexa Fluorꢁ 488-labeled compounds. (A) Com-
pound 6 in HepG2 cells; (B) compound 6 with SK-Hep1 cells; (C)
compound 7 in HepG2 cells; (D) compound 7 with SK-Hep1 cells; (E)
compound 8 with HepG2 cells; (F) compound 8 with SK-Hep1 cells;
(G) control HepG2 cells; (H) control SK-Hep1 cells.
Both criteria are a prerequisite for a therapeutic applica-
tion at a later stage.
Moreover, using fluorescence microscopy and flow
cytometry, we have shown that compounds 6 and 7 ex-
hibit selective uptake by the ASGP-R on HepG2 cells
derived from human parenchymal liver cells—the major
liver cell type. The formation of distinct endocytic vesi-
cles could be clearly visualized. Furthermore, competi-
tion with asialofetuin, a naturally occurring serum
glycoprotein and known ligand of the ASGP-R, and
GalNAc confirmed the involvement of the ASGP-R in
the uptake of 7. Experiments using compound 8 have
further re-enforced the generally accepted assumption

O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
5223
noma) and SK-Hep1 (human liver adenocarcinoma) cell
lines were obtained from DSMZ (Deutsche Sammlung
fur Mikroorganismen und Zellkulturen). Both cell lines
¨
A
80
60
40
20
0
HepG2
SK-Hep1
were propagated in Dulbecco’s modified Eagle’s med-
ium (DMEM) high-glucose, without phenol red, supple-
mented with fetal bovine serum (FBS, 10%) 2 mM L-
glutamine, 100 U/ml penicillin, and 100 lg/ml strepto-
mycin (complete medium). During the incubation steps
of the cells outside the incubator, medium with a CO2-
independent buffer system was used (DMEM high-glu-
cose, without phenol red and FBS, containing 25 mM
Hepes).
0
0
0
0.3
0.3
0.3
1
3
10
→
30 100
GaINAc (μM)
5.2. General procedure A: the preparation of triacetylated
compounds 11, 13, and 15
B
C
80
60
40
20
0
HepG2
SK-Hep1
To
the
corresponding
triallylated
compound
(3.54 mmol) was added 9-BBN (0.5 M in THF, 38 ml)
dropwise. The solution was then stirred at rt under ar-
gon for 24 h. The mixture was cooled to 0 ꢂC, and aque-
ous NaOH (3 M, 39 ml) was added dropwise, followed
by the dropwise addition of H2O2 (30%, 8.9 ml). The
resultant mixture was stirred vigorously at rt for 24 h.
The mixture was saturated with K2CO3, and the organic
layer was separated. The aqueous layer was then ex-
tracted with THF (3· 80 ml), and the combined organic
layers were dried (Na2SO4) and concentrated under re-
duced pressure. The residue was dissolved in pyridine
(33.5 ml), acetic anhydride (33.5 ml) was added, and
the mixture was stirred at rt for 3 h. The mixture was
co-evaporated with toluene (200 ml), and the resultant
syrup was purified by silica gel chromatography to af-
ford compound 11, 13, or 15 as oil.
1
3
10 30 100
Asialofetuin (μM) →
80
60
40
20
0
HepG2
SK-Hep1
5.2.1. Tris(5-acetoxy-2-oxapentyl)-N-(tert-butyloxycar-
bonyl)-methylamine (11). According to general proce-
dure A, compound 1034 (1.21 g, 3.54 mmol) was
reacted with 9-BBN (0.5 M in THF, 38 ml), and then
treated with aqueous NaOH (3 M, 39 ml) and H2O2
(30%, 8.9 ml). After peracetylation, work-up, and chro-
matography on silica gel (petrol ether/EtOAc 8:2 ! 7:3;
1
3
10
30 100
Asialofetuin (μM) →
Figure 6. Competitive uptake of compound 7 at a concentration of
10 lM in the presence of either GalNAc (0.3–100 mM) (A) or
asialofetuin (0.3–100 lM) (B). The graphs represent the mean of
median fluorescence intensity (MFI) SD of three independent
experiments. (C) The uptake of control compound 8 at a concentration
of 10 lM in the presence of asialofetuin (0.3–100 lM) into HepG2 and
ASGP-R-negative SK-Hep1 cells.
1
Rf 0.52 B), 1.49 g (81%) of 11 was obtained. H NMR
(500 MHz, CDCl3): d = 1.42 (s, 9H, CMe3), 1.88 (m,
6H, 3· OCH2CH2CH2OAc), 2.04 (s, 9H, 3· OAc),
3.50 (t, J = 6.1 Hz, 6H, 3· OCH2CH2CH2OAc), 3.63
[s, 6H, C(CH2O)3], 4.13 (t, J = 6.5 Hz, 6H, 3·
OCH2CH2CH2OAc), 4.90 (s, 1H, NH); 13C NMR
(125 MHz, CDCl3): d = 21.1 (3C, 3· CH3C@O), 28.4
(3C, CMe3), 28.9 (3C, 3· OCH2CH2CH2OAc), 58.5
[C(CH2O)3], 61.7 (3C, 3· OCH2CH2CH2OAc), 67.8
(3C, 3· OCH2CH2CH2OAc), 69.5 [3C, C(CH2O)3],
79.1 (CMe3), 154.8 [N–(C@O)O], 171.1 (3C, 3·
CH3C@O). Anal. Calcd for C24H43NO11: C, 55.26; H,
8.31; N, 2.69. Found: C, 55.20; H, 8.24; N, 2.70.
furan (THF), dioxane, diethyl ether (Et2O), and toluene
were dried by refluxing with sodium and benzophenone.
Dry DMF was purchased from Fluka (absolute,
P99.8%) and was further dried over powdered 4 A
molecular sieves. Molecular sieves (4 A) were activated
in vacuo at 500 ꢂC for 2 h immediately before use. Alexa
Fluorꢁ 488 carboxylic acid succinimidyl ester (A20000,
mixture of isomers) was purchased from Molecular
Probes, Eugene, Oregon, USA. Zinc dust was activated
according to the standard procedures.44
˚
˚
5.2.2. Tris(9-acetoxy-2,6-dioxanonyl)-N-(tert-butyloxy-
carbonyl)-methylamine (13). According to general proce-
dure A, compound 12 (1.82 g, 3.54 mmol) was reacted
with 9-BBN (0.5 M in THF, 38 ml), and then treated with
aqueous NaOH (3 M, 39 ml) and H2O2 (30%, 8.9 ml).
After peracetylation, work-up, and chromatography on
silica gel (petrol ether/EtOAc 3:1 ! 3:2; Rf 0.32 B),
All cell culture media, supplements, and phosphate-buf-
fered saline (PBS) were purchased from Invitrogen, ex-
cept collagen type S from rat’s tail was obtained from
Roche Applied Science. Paraformaldehyde, NaN3, and
N-propyl gallate were obtained from Fluka. Bovine ser-
um albumin (BSA) was from Sigma and Mowiol 4-88
from Hoechst. HepG2 (human hepatocellular carci-
1
2.14 g (87%) of 13 was obtained. H NMR (500 MHz,

5224
O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
´
CDCl3): d = 1.42 (s, 9H, CMe3), 1.78–1.91 [m, 12H, 3·
(OCH2CH2CH2)2OAc], 2.05 (s, 9H, 3· OAc), 3.45–3.50
(m, 18H, 3· OCH2CH2CH2OCH2CH2CH2OAc), 3.62
[s, 6H, C(CH2O)3], 4.15 (t, J = 6.5 Hz, 6H, 3· OCH2CH2-
CH2OAc), 4.93 (s, 1H, NH); 13C NMR (125 MHz,
CDCl3): d = 21.0 (3C, 3· CH3C@O), 28.4 (3C, CMe3),
29.0, 29.9 [6C, 3· (OCH2CH2CH2)2OAc], 58.4
[C(CH2O)3], 61.8 (3C, 3· OCH2CH2CH2OAc), 67.2,
67.9, 68.3, (9C, 3· OCH2CH2CH2OCH2CH2CH2OAc),
69.5 [3C, C(CH2O)3], 79.1 (CMe3), 154.8 [N–(C@O)O],
171.1 (3C, 3· CH3C@O). Anal. Calcd for C33H61NO14:
C, 56.96; H, 8.84; N, 2.01. Found: C, 57.67; H, 8.80; N,
2.56.
deacetylated under Zemplen conditions, and after
work-up, reacted with allyl bromide (7.5 ml, 87.9
mmol) under phase transfer catalysis conditions. After
work-up and chromatography on silica gel (petrol ether/
EtOAc 19:1 ! 9:1 ! 4:1; Rf 0.25 A), 2.25 g (95%) of 12
was obtained.1H NMR (500 MHz, CDCl3): d = 1.38 (s,
9H, CMe3), 1.79 (quintet, J = 6.3 Hz, 6H, 3·
OCH2CH2CH2O), 3.43–3.48 (m, 12H, 3· OCH2CH2-
CH2O), 3.59 [s, 6H, C(CH2O)3], 3.91 (m, 6H, 3·
CH2CH@CH2), 4.90 (s, 1H, NH), 5.17 (m, 6H, 3·
CH2CH@CH2), 5.86 (m, 3H, 3· CH2CH@CH2); 13C
NMR (125 MHz, CDCl3): d = 28.4 (3C, CMe3), 30.0
(3C, 3· OCH2CH2CH2O), 58.5 [C(CH2O)3], 67.4, 68.2
(6C, 3· OCH2CH2CH2O), 69.6 [3C, C(CH2O)3], 71.9
(3C, 3· CH2CH@CH2), 78.8 (CMe3), 116.7 (3C, 3·
CH2CH@CH2), 135.0 (3C, 3· CH2CH@CH2), 154.8
[N–(C@O)O]. Anal. Calcd for C27H49NO8: C, 62.89;
H, 9.58; N, 2.72; O, 24.82. Found: C, 62.66; H, 9.61;
N, 2.68; O, 24.87.
5.2.3.
Tris(13-acetoxy-2,6,10-trioxatridecyl)-N-(tert-
butyloxycarbonyl)-methylamine (15). According to gen-
eral procedure A, compound 14 (2.44 g, 3.54 mmol)
was reacted with 9-BBN (0.5 M in THF, 38 ml), and
then treated with aqueous NaOH (3 M, 39 ml) and
H2O2 (30%, 8.9 ml). After peracetylation, work-up,
and chromatography on silica gel (petrol ether/EtOAc
1:1 ! 3:7; Rf 0.1 B) 2.71 g (88%) of 15 was obtained.
1H NMR (500 MHz, CDCl3): d = 1.39 (s, 9H, CMe3),
1.75–1.88 [m, 18H, 3· (OCH2CH2CH2)3OAc], 2.01 (s,
9H, 3· OAc), 3.41–3.47 (m, 30H, 3· [(OCH2CH2-
CH2)2OCH2CH2CH2OAc]), 3.59 [s, 6H, C(CH2O)3],
4.12 (t, J = 6.5 Hz, 6H, 3· OCH2CH2CH2OAc), 4.90
(s, 1H, NH); 13C NMR (125 MHz, CDCl3): d = 20.9
(3C, 3· CH3C@O), 28.4 (3C, CMe3), 28.9, 29.6, 29.9
[9C, 3· (OCH2CH2CH2)3OAc], 58.4 [C(CH2O)3], 61.7
(3C, 3· OCH2CH2CH2OAc), 67.2, 67.7, 67.8, 67.9,
68.3, 68.4 [15C, 3· (OCH2CH2CH2)2OCH2CH2-
CH2OAc], 69.4 [3C, C(CH2O)3], 78.8 (CMe3), 154.7
[N–(C@O)O], 171.0 (3C, 3· CH3C@O). Anal. Calcd
for C42H79NO17: C, 57.98; H, 9.15; N, 1.61; O, 31.26.
Found: C, 58.08; H, 9.17; N, 1.70; O, 31.11.
5.3.2. Tris(9-allyloxy-2,6-dioxanonyl)-N-(tert-butyloxy-
carbonyl)-methylamine (14). According to general proce-
dure B, compound 13 (3.2 g, 4.6 mmol) was deacetylated
´
under Zemplen conditions, and after work-up, reacted
with allyl bromide (7.5 ml, 87.9 mmol) under phase
transfer catalysis conditions. After work-up and chro-
matography on silica gel (petrol ether/EtOAc
4:1 ! 1:1; Rf 0.64 B), 2.86 g (90%) of 14 was obtained.
1H NMR (500 MHz, CDCl3): d = 1.42 (s, 9H, CMe3),
1.78–1.87 (m, 12H, 6· OCH2CH2CH2O), 3.45–3.52
(m, 24H, 6· OCH2CH2CH2O), 3.62 [s, 6H,
C(CH2O)3], 3.96 (m, 6H, 3· CH2CH@CH2), 4.93 (s,
1H, NH), 5.22 (m, 6H, 3· CH2CH@CH2), 5.91 (m,
3H, 3·CH2CH@CH2); 13C NMR (125 MHz, CDCl3):
d = 28.4 (3C, CMe3), 29.9, 30.1 (6C, 6· OCH2CH2-
CH2O), 58.4 [C(CH2O)3], 67.3, 67.8, 68.4 (12C, 6·
OCH2CH2CH2O), 69.5 [3C, C(CH2O)3], 71.8 (3C, 3·
CH2CH@CH2), 79.0 (CMe3), 116.7 (3C, 3·
CH2CH@CH2), 134.9 (3C, 3· CH2CH@CH2), 155.0
[N–(C@O)O]. Anal. Calcd for C36H67NO11: C, 62.67;
H, 9.79; N, 2.03; O, 25.51. Found: C, 62.67; H, 9.72;
N, 2.10; O, 25.37.
5.3. General procedure B: preparation of triallylated
compounds 12 and 14
The corresponding triacetylated compound (4.6 mmol)
was dissolved in a solution of sodium methoxide in
MeOH (0.1 M, 40 ml), and the resultant solution was
stirred at rt for 4 h under argon. The solution was neu-
tralized with Dowex 50X8 (H+-form), and the solvent
was removed under reduced pressure to afford the de-
sired product in a quantitative yield as a colorless oil,
which was used without further purification.
5.4. Tris(13-acetoxy-2,6,10-trioxatridecyl)-methylamine
hydrochloride (16)
Compound 15 (680 mg, 0.781 mmol) was dissolved in
4 M HCl in dioxane (10 ml), and the resultant mixture
was stirred at rt under argon for 30 min. The solvent
was removed in vacuo to yield 16 (630 mg, quantitative)
The corresponding triol (1 mmol) was then dissolved in
DCM (5 ml) and the solution was added to a mixture of
50% aqueous NaOH (16 ml, w/v) and 15-crown-5
(19.8 ll, 0.1 mmol). Allyl bromide (1.64 ml, 19.1 mmol)
was then added, and the resultant mixture was refluxed
with vigorous stirring for 24 h. The mixture was cooled,
and the DCM (top) layer was separated, dried with
Na2SO4, and the solvent evaporated in vacuo. The resul-
tant syrup was purified by silica gel chromatography to
yield compound 12 or 14 as a yellow oil.
1
as an oil. H NMR (500 MHz, CDCl3): d = 1.68–1.77
[m, 18H, 3· (OCH2CH2CH2)3OAc], 1.92 (s, 9H, 3·
OAc), 3.35–3.44 [m, 36H, 3· (OCH2CH2CH2)2-
OCH2CH2CH2OAc, C(CH2O)3], 4.01 (t, J = 6.5 Hz,
6H, 3· OCH2CH2CH2OAc); 13C NMR (125 MHz,
CDCl3): d = 20.5 (3C, 3· CH3C@O), 28.6, 29.3, 29.6 [9C,
3· (OCH2CH2CH2)3OAc], 59.1 [C(CH2O)3], 61.6 (3C,
3· OCH2CH2CH2OAc), 66.7, 67.0, 67.4, 67.6, 68.3,
68.7 [18C, 3· (OCH2CH2CH2)2OCH2CH2CH2OAc,
C(CH2O)3],171.5 (3C, 3· CH3C@O); ESI-MS:
Calcd for C37H72NO15 (M+H)+: 770.49. Found: m/z
770.54.
5.3.1. Tris(5-allyloxy-2-oxapentyl)-N-(tert-butyloxycar-
bonyl)-methylamine (12). According to general proce-
dure B, compound 11 (2.39 g, 4.6 mmol) was

O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
5225
5.5. N-{Tris[13-acetoxy-2,6,10-trioxatridecyl]methyl}-4-
(benzyloxycarbonyl-amino)-butyramide (18)
CH2(OCH2CH2CH2)2OH], 128.0, 128.1, 128.5, 136.7
(6C, C6H5), 156.7 [N–(C@O)O], 173.5 (C@O, amide);
ESI-MS: Calcd for C43H78N2O15Na (M+Na)+: 885.53.
Found: m/z 885.68.
Compound 16 (239 mg, 0.297 mmol), 1736 (70.4 mg,
0.297 mmol), and PyBOP (186 mg, 0.357 mmol) were dis-
solved in dioxane/DMF (4 ml, 3:1 v/v), and DIPEA
(229 ll, 1.34 mmol) was added. The mixture was stirred
at rt under argon for 24 h. The resultant solution was par-
titioned between DCM (15 ml) and H2O (15 ml). The
DCM layer was separated, and the aqueous phase was ex-
tracted with DCM (25 ml). The DCM fractions were com-
bined, dried (Na2SO4), and the solvent was removed
under reduced pressure. The resultant syrup was purified
by silica gel chromatography (petrol ether/EtOAc
1:1 ! 3:7 ! 0:1) to afford 18 (250 mg, 85%, Rf 0.22 D)
5.7. N-{Tris[13-(2,3,4,6-tetra-O-benzoyl-b-D-galactopyr-
anosyloxy)-2,6,10-trioxatridecyl]methyl}-(4-benzyloxy-
carbonylamino)-butyramide (21)
Compound 19 (151 mg, 0.173 mmol) and ethyl 2,3,4,6-
tetra-O-benzoyl-1-thio-b-D-galactopyranoside
(20)27
(670 mg, 1.04 mmol) were dissolved in dry DCM
˚
(10 ml), and the mixture was stirred with 4 A molecular
sieves (500 mg) at rt under argon for 2 h. The mixture
was cooled to 0 ꢂC, and DMTST (538 mg, 2.08 mmol)
was added. The reaction was stirred at 0 ꢂC for 24 h,
and then at 10 ꢂC for another 24 h under argon. The mix-
ture was then filtered and extracted with aqueous
NaHCO3 solution (10 ml, 1 M) and brine (10 ml). The or-
ganic phase was dried (Na2SO4), and the solvent was re-
moved under reduced pressure. The resultant syrup was
purified by silica gel chromatography (EtOAc/petrol
ether 1:1 ! 7:3 ! 1:0) to afford the desired product 21
(301 mg, 68%, Rf 0.15C) as a colorless solid. [a]D +72.9
1
as a yellow oil. H NMR (500 MHz, CDCl3): d = 1.76–
1.90 (m, 20H, 3· [(OCH2CH2CH2)3OAc, NCH2CH2-
CH2C@O]), 2.03 (s, 9H, 3· OAc), 2.07–2.19 (m, 3H,
NCH2CH2CH2C@O), 3.20–3.24 (m, 3H, NCH2-
CH2CH2C@O), 3.41–3.50 [m, 30H, 3· (OCH2CH2-
CH2)2OCH2CH2CH2OAc], 3.66 [s, 6H, C(CH2O)3],
4.13 (t, J = 6.5 Hz, 6H, 3· OCH2CH2CH2OAc), 5.07 (s,
2H, CH2Ph), 5.28 (br s, 1H, NH, Cbz), 5.86 (s, 1H, NH,
Tris), 7.28–7.34 (m, 5H, C6H5); 13C NMR (125 MHz,
CDCl3): d = 20.9 (3C, 3· CH3C@O), 25.8 (1C, NCH2-
CH2CH2C@O), 29.0, 29.8, 30.1 [9C, 3· (OCH2CH2-
CH2)3OAc], 34.4 (1C, NCH2CH2CH2C@O), 40.3 (1C,
NCH2CH2CH2C@O), 59.8 [C(CH2O)3], 61.8 (3C, 3·
OCH2CH2CH2OAc), 66.5 (1C, CH2Ph), 67.3, 67.7,
67.9, 68.4 [15C, 3· (OCH2CH2CH2)2OCH2CH2-
CH2OAc], 69.1 [3C, C(CH2O)3], 128.0, 128.1, 128.5,
136.7 (6C, C6H5), 156.6 [N-(C@O)O], 171.1 (3C, 3·
CH3C@O), 172.4 (C@O, amide). Anal. Calcd for
C49H84N2O18: C, 59.50; H, 8.56; N, 2.83; O, 29.11. Found:
C, 59.12; H, 8.36; N, 2.98; O, 29.56.
1
(c 1, CHCl3); H NMR (500 MHz, CDCl3): d = 1.64–
1.69, 1.76–1.85 [m, 20H, 3· (OCH2CH2CH2)3OGal,
NCH2CH2CH2C@O], 2.16 (m, 2H, NCH2CH2
CH2C@O), 3.18 (m, 2H, NCH2CH2CH2C@O), 3.22–
3.47 [m, 30H, 3· (OCH2CH2CH2)2OCH2CH2CH2OGal],
3.66, [s, 6H, C(CH2O)3], 3.68, 4.01–4.06 (m, 6H, 3·
OCH2CH2CH2OGal), 4.32 (m, 3H, 3· H5-Gal), 4.40
0
(dd, J5,6 = 6.7, J6;6 ¼ 11:3 Hz, 3H, 3· H6-Gal), 4.68
(dd, J5;6 ¼ 6:4, J6;6 ¼ 11:2 Hz, 3H, 3· H60-Gal), 4.81
(d, J1,2 = 7.9 Hz, 3H, 3· H1-Gal), 5.07 (s, 2H, CH2Ph),
5.27 (br s, 1H, NH, Cbz), 5.61 (m, 3H, 3· H3-Gal), 5.78
(m, 3H, 3· H2-Gal), 5.86 (s, 1H, NH, Tris), 5.99 (m,
3H, 3· H4-Gal), 7.22-7.26, 7.28–7.35, 7.37–8.09 (m,
65H, 13· C6H5); 13C NMR (125 MHz, CDCl3): d = 26.6
(NCH2CH2CH2C@O), 29.7, 29.8, 29.9 [9C, 3·
(OCH2CH2CH2)3OGal], 35.1 (NCH2CH2CH2C@O),
42.2 (NCH2CH2CH2C@O), 59.8 [C(CH2O)3], 61.9 (3C,
3· C6-Gal), 67.0 (CH2Ph), 67.4, 67.8, 67.9 [12C, 3·
(OCH2CH2CH2)2OCH2CH2CH2OGal], 68.1 (3C, 3·
C4-Gal), 68.4 (3C, 3· OCH2CH2CH2OGal), 69.1 [3C,
C(CH2O)3], 69.8 (3C, 3· C2-Gal), 71.2 (3C, 3· C3-Gal),
71.7 (3C, 3· C5-Gal), 101.8 (3C, 3· C1-Gal), 128.0–
136.7 (78C, C6H5), 165.2–166.0 (13C, 13 C@O). Anal.
Calcd for C145H156N2O42: C, 67.01; H, 6.05; N, 1.08.
Found: C, 66.51; H, 6.12; N, 1.13.
0
0
5.6. 4-(Benzyloxycarbonylamino)-N-{tris[13-hydroxy-
2,6,10-trioxatridecyl]methyl}-butyramide (19)
Compound 18 (231 mg, 0.233 mmol) was dissolved in a
solution of sodium methoxide in dry methanol (0.05 M,
20 ml), and the resultant solution was stirred at rt under
argon for 4 h. The reaction mixture was neutralized with
Dowex 50X8 (H+-form), and the solvent was removed in
vacuo. The resultant oil was purified by silica gel chroma-
tography (EtOAc/MeOH 95:5 ! 9:1) to afford 19
1
(181 mg, 90%, Rf 0.2 E) as an oil. H NMR (500 MHz,
CDCl3): d = 1.76–1.83 [m, 20H, 3· (OCH2CH2CH2)3OH,
NCH2CH2CH2C@O], 2.18 (t, J = 6.8 Hz, 2H, NCH2-
CH2CH2C@O), 2.62 (bs, 3H, 3· OH), 3.21 (m, 2H,
NCH2CH2CH2C@O), 3.43–3.51 (m, 24H, 3· OCH2CH2-
CH2OCH2CH2CH2OCH2CH2CH2OH), 3.58 [t, J =
5.8 Hz, 6H, 3· OCH2CH2CH2(OCH2CH2CH2)2OH],
3.67 [s, 6H, C(CH2O)3], 3.73 (t, J = 5.7 Hz, 6H, 3·
OCH2CH2CH2OH), 5.07 (s, 2H, CH2Ph), 5.45 (s, 1H,
NH, Cbz), 6.00 (s, 1H, NH, Tris), 7.28–7.34 (m,
5H, C6H5); 13C NMR (125 MHz, CDCl3): d = 25.7
(NCH2CH2CH2C@O), 29.8, 30.0, 32.0 [9C, 3·
(OCH2CH2CH2)3OH], 34.3 (NCH2CH2CH2C@O), 40.3
(NCH2CH2CH2C@O), 59.9 [C(CH2O)3], 61.7 (3C, 3·
OCH2CH2CH2OH), 66.5 (CH2Ph), 67.7, 68.1, 68.3
(12C, 3· OCH2CH2CH2OCH2CH2CH2OCH2CH2CH2-
OH), 69.2 [3C, C(CH2O)3], 69.7 [3C, 3· OCH2CH2-
5.8. 4-(Benzyloxycarbonylamino)-N-{tris[13-(b-D-galac-
topyranosyloxy)-2,6,10-trioxatridecyl]methyl}-butyra-
mide (5a)
Compound 21 (30 mg, 0.015 mmol) was dissolved in dry
dioxane (1 ml), and a solution of sodium methoxide in
methanol (0.1 M, 1 ml) was added. The resultant mix-
ture was stirred at rt under argon for 4 h, after which
it was neutralized with Dowex 50X8 (H+-form), filtered,
and the solvent was removed under reduced pressure.
The residue was purified by silica gel chromatography
(DCM/MeOH/H2O 10:3:0 ! 10:4:0 ! 10:4:0.1 !
10:4:0.2 ! 10:4:0.4) to afford 5a (14.6 mg, 94%, Rf 0.2

5226
O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
1
F) as a colorless solid. [a]D ꢁ5.73 (c 0.96, MeOH); H
NMR (500 MHz, MeOD): d = 1.73–1.88 [m, 20H, 3·
(OCH2CH2CH2)3OGal, NCH2CH2CH2C@O], 2.19 (t,
J = 7.3 Hz, 2H, NCH2CH2CH2C@O), 3.15 (t, J =
6.8 Hz, 2H, NCH2CH2CH2C@O), 3.44–3.75 [m, 54H,
3· H2-Gal, 3· H3-Gal, 3· H5-Gal, 3· H6-Gal, 3·
(OCH2CH2CH2)2OCH2CH2CH2OGal, C(CH2O)3, 3·
OCH2CH2CH2OGal-Ha], 3.82 (m, 3H, 3· H4-Gal),
3.95 (m, 3H, 3· OCH2CH2CH2OGal-Hb), 4.20 (d,
J1,2 = 7.4 Hz, 3H, 3· H1-Gal), 5.07 (s, 2H, CH2Ph),
7.29–7.35 (m, 5H, C6H5); 13C NMR (125 MHz, MeOD):
d = 27.5 (NCH2CH2CH2C@O), 31.0, 31.1 [9C, 3·
(OCH2CH2CH2)3OGal], 35.0 (NCH2CH2CH2C@O),
41.1 (NCH2CH2CH2C@O), 61.6 [C(CH2O)3], 62.4 (3C,
3· C6-Gal), 67.4 (CH2Ph), 67.9 (3C, 3· OCH2CH2-
CH2OGal), 68.8 (3C, 3· C[CH2O]3), 69.4 (3C, 3· C4-
Gal), 68.8, 70.2, 72.6, 75.0, 76.6 [24C, 3· C2-Gal, 3·
C3-Gal, 3· C5-Gal, 3· (OCH2CH2CH2)2OCH2CH2-
CH2OGal], 105.1 (3C, 3· C1-Gal), 128.9–129.5, 138.4
(6C, C6H5), 158.9 [N–(C@O)O], 175.6 (C@O); ESI-
MS: Calcd for C61H109N2O30 (M+H)+: 1349.71. Found:
m/z 1349.87.
cuo. The residue was purified by gel filtration on a Sepha-
dex LH-20 column (2.5 · 35 cm) using MeOH as an
eluant, then on an RP-18 column (H2O/MeOH stepwise
gradient 1:0–1:1) to yield 6 (2.2 mg, 81%) as a red solid
after final lyophilization ꢁfrom water. ESI-MS: Calcd for
C74H112N4O38S22ꢁðM=2Þ : 864.32. Found: m/z 864.85.
5.11. N-(Tris{13-[3,4,6-tri-O-acetyl-2-(2,2,2-trichloroeth-
oxycarbonylamino)-b-D- galactopyranosyloxy]-2,6,10-tri-
oxatridecyl}methyl)-(4-benzyloxycarbonyl-amino)-
butyramide (24)
Compound 19 (15 mg, 17.3 lmol) and ethyl 3,4,6-tri-
O-acetyl-2-deoxy-1-thio-2-(2,2,2-trichloroethoxycarbon-
ylamino)-b-D-galactopyranoside
(23)37
(54.6 mg,
104 lmol) were dissolved in dry DCM (1 ml) and stirred
˚
with 4 A molecular sieves (50 mg) at rt under argon for
2 h. The suspension was cooled to 0 ꢂC, and DMTST
(53.8 mg, 208 lmol) was added. The mixture was stirred
at 0 ꢂC for 24 h, then at 10 ꢂC for 48 h. The reaction was
quenched with triethylamine (50 ll), diluted with DCM
(5 ml), washed with brine (5 ml), dried (Na2SO4), and
the solvent was removed in vacuo. The residue was puri-
fied by silica gel chromatography (petrol ether/EtOAc
1:4 ! 1:9) to afford 24 (35.1 mg, 91%, Rf 0.32 D) as a
colorless solid. [a]D ꢁ3.6 (c 1.76, CHCl3); 1H NMR
(500 MHz, CDCl3): d = 1.76–1.84 [m, 20H, 3·
(OCH2CH2CH2)3OGalN, NCH2CH2CH2C@O], 1.96
(s, 9H, 3· OAc), 2.03 (s, 9H, 3· OAc), 2.14 (s, 9H, 3·
OAc), 2.20 (m, 2H, NCH2CH2CH2C@O), 3.22 (m,
2H, NCH2CH2CH2C@O), 3.43–3.56 [m, 30H, 3·
(OCH2CH2CH2)2OCH2CH2CH2OGalN], 3.60 (m, 3H,
3· CH2OGalN-Ha), 3,66, [s, 6H, C(CH2O)3], 3.85–3.90
(m, 3· H2-GalN, 6H, 3· H5-GalN), 3.97 (m, 3H, 3·
CH2OGalN-Hb), 4.05–4.19 (m, 6H, 3· H6-GalN, 3·
H60-GalN), 4.56 (d, J1,2 = 7.7 Hz, 3H, 3· H1-GalN),
4.66, 4.75 (A, B of AB, J = 11.7 Hz, 6H, 3· CH2, Troc),
5.08 (s, 2H, CH2Ph), 5.14 (m, 3H, 3· H3-GalN), 5.35
(m, 3H, 3· H4-GalN), 5.78 (d, J = 7.6 Hz, 3H, N–H,
GalN), 5.95 (br s, 1H, NH, Tris), 7.32 (m, 5H, C6H5);
13C NMR (125 MHz, CDCl3): d = 20.6, 20.7, 20.8 (9C,
9· CH3, AcO), 25.8 (NCH2CH2CH2C@O), 29.6, 29.7,
5.9. 4-Amino-N-{tris[13-(b-D-galactopyranosyloxy)-
2,6,10-trioxatridecyl]methyl}-butyramide (22)
Compound 5a (25 mg, 18.5 lmol) was dissolved in etha-
nol/dioxane (2 ml, 1:1 v/v), and Pd/C (10% Pd, 20 mg)
was added. The mixture was vigorously stirred under a
H2 atmosphere (1 atm) at rt for 24 h. The mixture was
then diluted with ethanol, filtered, and concentrated in va-
cuo to yield 22 as a colorless solid (19.5 mg, 87%). [a]D
ꢁ5.8 (c 1, MeOH); 1H NMR (500 MHz, MeOD):
d = 1.79–1.89 [m, 20H, 3· (OCH2CH2CH2)3OGal,
NCH2CH2CH2C@O], 2.31 (t, J = 7.1 Hz, 2H, NCH2CH2
CH2C@O), 2.87 (t, J = 7.3 Hz, 2H, NCH2CH2CH2C@O),
3.44–3.56 [m, 39H, 3· H2-Gal, 3· H3-Gal, 3· H5-Gal, 3·
(OCH2CH2CH2)2OCH2CH2CH2OGal], 3.60–3.75 (m,
15H, 3· H6-Gal, C[CH2O]3, 3· OCH2CH2CH2OGal-
Ha), 3.82 (m, 3H, 3· H4-Gal), 3.96, (m, 3H, 3·
OCH2CH2CH2OGal-Hb), 4.2 (d, J1,2 = 7.2 Hz, 3H, 3·
H1-Gal); 13C NMR (125 MHz, MeOD): d = 26.9
(NCH2CH2CH2C@O), 30.7, 31.0, 31.1 [9C, 3·
(OCH2CH2CH2)3Gal], 34.6 (NCH2CH2CH2C@O), 48.3
(NCH2CH2CH2C@O), 62.5 (3C, 3· C6-Gal), 67.9, 68.8,
68.9, 69.4, 69.7 [21C, 3· (OCH2CH2CH2)2OCH2CH2-
CH2OGal, 3· OCH2CH2CH2OGal, C(CH2O)3], 70.2
(3C, 3· C4-Gal), 72.6 (3C, 3· C2-Gal), 75.0 (3C, 3· C3-
Gal), 76.6 (3C, 3· C5-Gal), 105.1 (3C, 3· C1-Gal); ESI-
MS: Calcd for C53H103N2O28 (M+H)+: 1215.67. Found:
m/z 1215.91.
29.8
[9C,
3·
(OCH2CH2CH2)3OGalN],
34.8
(NCH2CH2CH2C@O), 40.7 (NCH2CH2CH2C@O),
59.8 [C(CH2O)3], 61.4 (3C, 3· C6-GalN), 66.6 (CH2Ph),
66.9 (3C, 3· C4-GalN), 67.4, 67.5, 67.6 [15C, 3·
(OCH2CH2CH2)2OCH2CH2CH2OGalN], 68.4 (3C, 3·
OCH2CH2CH2OGalN), 69.1 [3C, C(CH2O)3], 70.1
(3C, 3· C3-GalN), 70.5 (3C, 3· C5-GalN), 95.6 (3C,
3· CCl3, Troc), 101.6 (3C, 3· C1-GalN), 128.0–128.5
(6C, C6H5), 154.3, 154.5 [4C, 4· N–(C@O)O], 170.3–
171.2 (10C, 10· C@O). Anal. Calcd for
C88H132Cl9N5O42: C, 46.95; H, 5.91; N, 3.11. Found:
C, 47.22; H, 6.00; N, 3.03.
5.10. Fluorescent-labeled, Gal-terminated compound (6)
A stock solution containing compound 22 (10 mg,
˚
8.23 lmol), DIPEA (20 ll, 156 lmol) and 4 A molecular
5.12. N-{Tris[13-(2-acetamido-3,4,6-tri-O-acetyl-b-D-
galactopyranosyloxy)-2,6,10-trioxatridecyl]methyl}-(4-
benzyloxycarbonylamino)-butyramide (25)
sieves (25 mg) in dry DMF/dioxane (1 ml, 1:1) was stirred
at rt under argon for 2 h. The solution (500 ll) was trans-
ferred to a small vial containing Alexa Fluorꢁ 488-NHS
˚
(1 mg, 1.55 lmol), 4 A molecular sieves (25 mg), and a
Compound 24 (20 mg, 8.88 lmol) was dissolved in dry
dioxane (1 ml), and activated Zn dust (55 mg, 84.1 mmol)
was added, followed by acetic anhydride (272 ll,
2.66 mmol), and the reaction mixture was stirred at rt un-
stirring bar. The resultant mixture was stirred in the dark
at rt under argon for 4 days. The mixture was then diluted
with MeOH, filtered, and the solvents were removed in va-

O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
5227
der argon for 16 h. The mixture was filtered and the sol-
vents were removed in vacuo. The residue was purified
by silica gel chromatography (DCM/MeOH stepwise gra-
dient 99:1 ! 93:7) to afford 25 (13.5 mg, 82%, Rf 0.13 E)
as a colorless solid. [a]D ꢁ12.2 (c 0.5, CHCl3); 1H NMR
(500 MHz, CDCl3): d = 1.76–1.85 [m, 20 H, 3·
(OCH2CH2CH2)3OGalN, NCH2CH2CH2C@O], 1.95 (s,
9H, 3· NAc), 1.99, 2.04, 2.14 (s, 27H, 9· OAc), 2.19 (m,
2H, NCH2CH2CH2C@O), 3.22 (m, 2H, NCH2CH2
CH2C@O), 3.44–3.48 [m, 30H, 3· (OCH2CH2-
CH2)2OCH2CH2CH2OGalN], 3.58 (m, 3H, 3· CH2O-
GalN-Ha), 3,67 [s, 6H, C(CH2O)3], 3.89–3.96 (m, 6H, 3·
H5-GalN, 3·CH2OGalN-Hb), 4.00 (m, 3H, 3· H2-
GalN), 4.09–4.18 (m, 6H, 3· H6-GalN, 3· H0-GalN),
4.64 (d, J1,2 = 8.3 Hz, 3H, 3· H1-GalN), 5.08 (s, 2H,
CH2Ph), 5.23 (m, 3H, 3· H3-GalN), 5.34 (s, 3H, 3· H4-
GalN), 5.41 (m, 1H, NH, Cbz), 6.05 (d, J = 8.6 Hz, 3H,
NHAc), 7.30–7.34 (m, 5H, C6H5); 13C NMR (125 MHz,
CDCl3): d = 20.6 (9C, 9· CH3, AcO), 23.3 (3C, 3· CH3,
AcHN), 25.9 (1C, NCH2CH2CH2C@O), 29.9, 29.8, 29.6
[9C, 3· (OCH2CH2CH2)3OGalN], 34.5 (NCH2CH2
CH2C@O), 40.3 (NCH2CH2CH2C@O), 51.3 (3C, 3·
C2-GalN), 59.8 [C(CH2O)3], 61.4 (3C, 3· C6-GalN),
66.5 (CH2Ph), 66.7 (3C, 3· C4-GalN), 67.0 (3C, 3·
OCH2CH2CH2OGalN), 67.7 [15C, 3· (OCH2CH2-
CH2)2OCH2CH2CH2OGalN], 69.0 [3C, C(CH2O)3],
70.0 (3C, 3· C3-GalN), 70.4 (3C, 3· C5-GalN), 101.2
(3C, 3· C1-GalN), 127.9, 128.0, 128.5,136.6 (6C, C6H5),
156.7, [1C,N–(C@O)O] 170.2–172.5, (13C, 13· C@O).
Anal. Calcd for C85H135N5O39: C, 55.15; H, 7.35; N,
3.78. Found: C, 54.84; H, 7.34; N, 3.61.
C(CH2O)3], 73.3 (3C, 3· C3-GalN), 76.6 (3C, 3· C5-
GalN), 103.1 (3C, 3· C1-GalN), 128.8, 128.9, 129.5,
138.4 (6C, C6H5), 158.9 [N–(C@O)O]; ESI-MS: Calcd for
C67H117N5O30Na(M+Na)+: 1494.77. Found:m/z 1495.41.
5.14. N-{Tris[13-(2-acetamido-b-D-galactopyranosyloxy)-
2,6,10-trioxatridecyl]methyl}-4-amino-butyramide (26)
Compound 5b (15 mg, 10.2 lmol) was dissolved in
methanol/dioxane (2 ml, 1:1 v/v), and Pd/C (10% Pd,
15 mg) was added. The resultant suspension was vigor-
ously stirred under a H2 atmosphere (1 atm) at rt for
3 h. The mixture was then diluted with methanol, fil-
tered, and concentrated in vacuo to yield 26 as a color-
1
less solid (13 mg, 95%). [a]D ꢁ1.49 (c 0.67, MeOH); H
NMR (500 MHz, MeOD): d = 1.78–1.82 [m, 20H, 3·
(OCH2CH2CH2O)3GalN, NCH2CH2CH2C@O], 1.99
(s, 9H, 3· NAc), 2.29 (t, J = 7.2 Hz, 2H, NCH2CH2-
CH2C@O), 2.82 (m, 2H, NCH2CH2CH2C@O), 3.48–
3.51 [m, 33H, 3· (OCH2CH2CH2)2OCH2CH2CH2O-
GalN, 3·H5-GalN], 3.55–3.63 (m, 6H, 3· CH2OGalN-
Ha, 3· H3-GalN), 3.67 [s, 6H, C(CH2O)3], 3.70–3.79
(m, 6H, 3·H6-GalN), 3.84 (m, 3H, 3· H4-GalN),
3.89–3.95 (m, 6H, 3· CH2OGalN-Hb, 3· H2-GalN),
4.35 (d, J1,2 = 8.4 Hz, 3H, 3· H1-GalN); 13C NMR
(125 MHz, MeOD): d = 23.1 (3C, 3· CH3, AcHN),
27.5 (1C, NCH2CH2CH2C@O), 31.0, 30.1 [9C, 3·
(OCH2CH2CH2)3OGalN], 34.7 (NCH2CH2CH2C@O),
41.2 (NCH2CH2CH2C@O), 54.3 (3C, 3· C2-GalN),
61.6 [C(CH2O)3], 62.5 (3C, 3· C6-GalN), 67.4 (3C, 3·
OCH2CH2CH2OGalN), 68.6, 68.8, 68.9 [15C, 3·
(OCH2CH2CH22OCH2CH2CH2OGalN), 69.4 (3C, 3·
C4-GalN]), 69.7 [3C, C(CH2O)3], 73.3 (3C, 3· C3-
GalN), 76.6 (3C, 3· C5-GalN), 103.1 (3C, 3· C1-GalN),
174.0, 175.2 (4· C@O); ESI-MS: Calcd for
C59H112N5O28 (M+H)+: 1338.75. Found: m/z 1339.22.
5.13. N-{Tris[13-(2-acetamido-b-D-galactopyranosyloxy)-
2,6,10-trioxatridecyl]methyl}-4-(benzyloxycarbonyl-
amino)-butyramide (5b)
Compound 25 (35 mg, 18.9 lmol) was dissolved in dry
dioxane (1 ml), and a solution of sodium methoxide in
methanol (0.1 M, 1 ml) was added. The mixture was stir-
red at rt under argon for 4 h, after which it was neutralized
with Dowex 50X8 (H+-form), filtered, and the solvent was
removed underreduced pressure. The residue was purified
on an RP-18 column (H2O/MeOH, stepwise gradient,
1:0 ! 2:3) to afford 5b (20.1 mg, 72%, Rf 0.1 F) as a color-
5.15. Fluorescent-labeled, GalNAc-terminated compound
(7)
A stock solution containing compound 26 (10 mg,
˚
7.47 lmol), DIPEA (50 ll, 292 lmol), and 4 A molecular
sieves (25 mg) in dry DMF/dioxane (1 ml, 1:1) was stirred
at rt under argon for 2 h. The resultant solution (500 ll)
was transferred to a small vial containing Alexa Fluorꢁ
1
less solid. [a]D ꢁ1.7 (c 1, MeOH); H NMR (500 MHz,
˚
MeOD): d = 1.75–1.82 [m, 20H, 3· (OCH2CH2CH2)3O-
GalN, NCH2CH2CH2C@O], 1.99 (s, 9H, 3· NAc), 2.20
(t, J = 7.3 Hz, 2H, NCH2CH2CH2C@O), 3.16 (t,
J = 6.8 Hz, 2H, NCH2CH2CH2C@O), 3.47–3.49 [m,
33H, 3· (OCH2CH2CH2)2OCH2CH2CH2OGalN, 3·
H5-GalN], 3.54–3.61 (m, 6H, 3· CH2OGalN-Ha, 3·
H3-GalN), 3.67 [s, 6H, C(CH2O)3], 3.72–3.79 (m, 6H,
3· H6-GalN), 3.83 (m, 3H, 3· H4-GalN), 3.89–3.95 (m,
6H, 3· CH2OGalN-Hb, 3· H2-GalN), 4.36 (d,
J1,2 = 8.4 Hz, 3H, 3· H1-GalN), 5.07 (s, 2H, CH2Ph),
7.24–7.35 (m, 5H, C6H5); 13C NMR (125 MHz, MeOD):
d = 23.1 (3C, 3· CH3, AcHN), 27.6 (NCH2CH2-
CH2C@O), 31.0, 30.1 [9C, 3· (OCH2CH2CH2)3OGalN],
35.0 (NCH2CH2CH2C@O), 41.1 (NCH2CH2CH2C@O),
54.3 (3C, 3· C2-GalN), 61.6 [C(CH2O)3], 62.5 (3C, 3·
C6-GalN), 67.4 (4C, 3· OCH2CH2CH2OGalN, CH2Ph),
68.9, 68.8, 68.6 [15C, 3· (OCH2CH2CH2)2OCH2CH2-
CH2OGalN], 66.4 (3C, 3· C4-GalN), 69.6 [3C,
488-NHS (1 mg, 1.55 lmol), 4 A molecular sieves
(25 mg), and a stirring bar. The mixture was stirred in
the dark at rt under argon for 4 days, then diluted with
MeOH, filtered, and the solvents were removed in vacuo.
The residue was purified by gel filtration on a Sephadex
LH-20 column (2.5 · 35 cm) using MeOH as an eluant,
then on an RP-18 column (H2O/MeOH stepwise gradient
1:0!1:1) to yield 7 (2.6ꢁmg, 90%). ESI-MS: Calcd for
C80H121N7O38S22ꢁðM=2Þ : 925.86. Found: m/z 926.05.
5.16. 4-(Benzyloxycarbonylamino)-N-{[tris(hydroxy-
methyl)]methyl}-butyramide (27)
Compound 1736 (3.01 g, 12.6 mmol) and EEDQ (3.26 g,
13.2 mmol) were stirred in dry pyridine (100 ml) at rt un-
der argon for 1 h. Tris (9) (1.33 g, 10.9 mmol) was then
added, and the resultant suspension was stirred at 90 ꢂC
for 24 h. The solvent was removed in vacuo, and the res-

5228
O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
idue was triturated with EtOAc and a few drops of
MeOH. The product was filtered off, washed with
EtOAc, then with cold water, and dried under high vac-
uum to yield 27 (2.88 g, 77%) as a white solid. 1H NMR
(500 MHz, MeOD): d = 1.79 (quintet, J = 7.0 Hz, 2H,
NCH2CH2CH2C@O), 2.28 (t, J = 7.4 Hz, 2H,
NCH2CH2CH2C@O), 3.16 (t, J = 6.8 Hz, 2H,
NCH2CH2CH2C@O), 3.73 (s, 6H, C[CH2O]3), 5.09 (m,
2H, CH2Ph), 7.28–7.35 (m, 5H, C6H5); 13C NMR
(125 MHz, CDCl3): d = 27.1 (NCH2CH2CH2C@O),
34.5 (NCH2CH2CH2C@O), 41.0 (NCH2CH2CH2C@O),
62.5 [3C, C(CH2O)3], 63.6 [C(CH2O)3], 67.4 (CH2Ph),
128.8, 129.0, 129.4, 130.6, 138.3 (6C, C6H5), 159.0 [N–
(C@O)O], 176.6 (C@O); ESI-MS: Calcd for
C16H24N2O6Na (M+Na)+: 363.15. Found: m/z 363.13.
2.83 mmol) was added. The resultant mixture was stirred
at rt under argon for 6 h, after which it was neutralized
with Dowex 50X8 (H+-form), filtered, and the solvent
was removed under reduced pressure. The residue was
purified on a Sephadex G-25 column, and, after the re-
moval of water in vacuo, washed with Et2O/DCM (ca.
2:1) to afford the desired product 29 (81.5 mg, 85%, Rf
1
0.11 F) as a white powder. [a]D ꢁ0.2 (c 1, MeOH); H
NMR (500 MHz, MeOD): d = 1.78 (m, NCH2CH2
CH2C@O), 2.22 (t, J = 7.1 Hz, 2H, NCH2CH2CH2
C@O), 3.16 (t, J = 6.9 Hz, 2H, NCH2CH2CH2C@O),
3.45–3.54 (m, 9H, 3· H2-Gal, 3· H3-Gal, 3· H5-Gal),
3.69 (dd, J5;6 ¼ 5:2, J6;6 ¼ 11:3 Hz, 3H, 3· H60-Gal),
0
0
0
3.76 (dd, J5,6 = 6.9, J6;6 ¼ 11:4 Hz, 3H, 3· H6-Gal),
3.81 (m, 3H, 3· H4-Gal), 3.93 (d, A of AB, J = 10.2 Hz,
3H, 3· CH2OGal-Ha), 4.27 (d, J1,2 = 7.6 Hz, 3H, 3·
H1-Gal), 4.32 (d, B of AB, J = 10.2 Hz, 3H, 3· CH2O-
Gal-Hb), 5.07 (s, 2H, CH2Ph), 7.29–7.35 (m, 5H, C6H5);
13C NMR (125 MHz, MeOD): d = 27.1 (NCH2CH2
CH2C@O); 35.0 (NCH2CH2CH2C@O), 41.2 (NCH2CH2
CH2C@O); 62.5 (3C, 3· C6-Gal), 67.4 (CH2Ph); 69.3 [3C,
C(CH2O)3], 70.4 (3C, 3· C4-Gal), 72.6 (3C, 3· C2-Gal),
74.9 (3C, 3· C3-Gal), 76.7 (3C, 3· C5-Gal), 105.5 (3C,
3· C1-Gal), 127.9–129.5 (6C, C6H5), 159.0 [N-(C@O)O],
175.9 (C@O); ESI-MS: Calcd for C34H54N2O21Na
(M+Na)+: 849.31. Found: m/z 849.44.
5.17. N-{[Tris(2,3,4,6-tetra-O-benzoyl-b-D-galactopyr-
anosyloxy)methyl]methyl}-(4-benzyloxycarbonylamino)-
butyramide (28)
Compound 27 (100 mg, 0.294 mmol) and ethyl 2,3,4,6-
(20)27
tetra-O-benzoyl-1-thio-b-D-galactopyranoside
˚
(753 mg, 1.17 mmol) were stirred with 4 A molecular
sieves (500 mg) in DCE/Et2O (20 ml, 1:1) at rt under
argon for 2 h. The mixture was cooled to 0 ꢂC, and
a solution of NIS (263 mg, 1.17 mmol) and TfOH
(0.117 mmol, 10.2 ll) in DCE/Et2O (10 ml, 1:1) was
added. The resultant mixture was stirred at 0 ꢂC under
argon for 1 h, upon which it turned deep brown. The
mixture was diluted with DCM (50 ml), filtered
through Celite, washed with Na2S2O3 (1 M, 25 ml),
followed by NaHCO3 (0.1 M, 25 ml), and dried with
Na2SO4. The solvent was removed under reduced
pressure, and the residue was purified by silica gel
chromatography (petrol ether/EtOAc 3:1 ! 1:1) to
yield 28 (311 mg, 51%, Rf 0.23 B) as a colorless solid.
5.19. 4-Amino-N-{[tris(b-D-galactopyranosyl-
oxy)methyl]methyl}-butyramide (30)
Compound 29 (27.2 mg, 32.9 lmol) was dissolved in
methanol (2.5 ml), and Pd/C (10% Pd, 25 mg) was added.
The resultant mixture was vigorously stirred under a H2
atmosphere (1 atm) at rt for 48 h. The mixture was then
filtered and concentrated in vacuo toyield 30 as a colorless
1
solid (19.9 mg, 87%). [a]D +3.1 (c 1, MeOH); H NMR
1
[a]D +39 (c 1, CHCl3); H NMR (500 MHz, CDCl3):
(500 MHz, MeOD): d = 1.89 (m, 2H, NCH2CH2CH2
C@O), 2.32 (m, 2H, NCH2CH2CH2C@O), 3.00 (m, 2H,
NCH2CH2CH2C@O), 3.43–3.50 (m, 9H, 3· H2-Gal, 3·
0
d = 1.65 (m, 2H, NCH2CH2CH2C@O), 1.86 (m, 2H,
NCH2CH2CH2C@O), 3.09 (m, 2H, NCH2CH2
CH2C@O), 3.52 (d, A of AB, J = 10.2 Hz, 3H, 3· CH2O-
Gal-Ha), 3.72 (m, 3H, 3· H5-Gal), 4.10 (d, J1,2 = 8.0 Hz,
H3-Gal,
3·
H5-Gal),
3.68
(dd,
J5;6 ¼ 5:1,
J6;6 ¼ 11:2 Hz, 3H, 3· H60-Gal), 3.72 (dd, J5,6 = 6.9,
0
0
0
0
3H, 3· H1-Gal), 4.32 (dd, J5;6 ¼ 7:4, J6;6 ¼ 11:2 Hz, 3H,
J6;6 ¼ 11:3 Hz, 3H, 3· H6-Gal), 3.77 (m, 3H, 3· H4-
3· H60-Gal), 4.35 (d, B of AB, J = 10.1 Hz, 3H, 3· CH2O-
Gal), 3.87 (d, A of AB, J = 10.1 Hz, 3H, 3· CH2OGal-
Ha), 4.25 (d, J1,2 = 6.8 Hz, 3H, 3· H1-Gal), 4.31 (d, B of
AB, J = 10.1 Hz, 3H, 3· CH2OGal-Hb); 13C NMR
(125 MHz, MeOD): d = 24.5 (NCH2CH2CH2C@O),
34.7 (NCH2CH2CH2C@O), 40.3 (NCH2CH2CH2C@O),
61.3 [C(CH2O)3], 62.6 (3C, 3· C6-Gal), 69.1 [3C,
C(CH2O)3], 70.8 (3C, 3· C4-Gal), 73.0 (3C, 3· C2-Gal),
75.3 (3C, 3· C3-Gal), 77.1 (3C, 3· C5-Gal), 105.8 (3C,
3· C1-Gal), 174.9 (C@O); ESI-MS: Calcd for
C26H49N2O19 (M+H)+: 693.29. Found: m/z 693.39.
0
Gal-Hb), 4.51 (dd, J5,6 = 6.2, J6;6 ¼ 11:2 Hz, 3H, 3· H6-
Gal), 5.02 (m, 2H, CH2Ph), 5.32 (s, 1H, NH, Cbz), 5.38
(dd, J4,3 = 3.3, J2,3 = 10.3 Hz, 3· H3-Gal), 5.62 (dd,
J1,2 = 8.1, J2,3 = 10.2 Hz, 3· H2-Gal), 5.75 (s, 1H, NH,
Tris), 5.84 (d, J3.4 = 3.2 Hz, 3H, 3· H4-Gal), 7.21–8.10
(m, 65H, 1 3· C6H5); 13C NMR (125 MHz, CDCl3):
d = 25.2 (NCH2CH2CH2C@O), 33.8 (NCH2CH2CH2C@
O), 40.1 (NCH2CH2CH2C@O), 59.4 [C(CH2O)3], 61.4 (3C,
3· C6-Gal), 66.4 (CH2Ph), 67.8, 67.9 [6C, C(CH2O)3, 3·
C4-Gal], 69.9 (3C, 3· C2-Gal), 71.0, 71.1 (6C, 3· C3-
Gal, 3· C5-Gal), 101.8 (3C, 3· C1-Gal), 127.9–136.7
(78C, 13· C6H5), 156.5 [N–(C@O)O], 164.9–172.6 (13C,
13· C@O). Anal. Calcd for C118H102N2O33: C, 68.27;
H, 4.95; N, 1.35. Found: C, 67.81; H, 5.08; N, 1.36.
5.20. Fluorescent-labeled, Gal-terminated control com-
pound (8)
A stock solution containing compound 30 (12 mg,
˚
17.3 lmol), DIPEA (50 ll, 291 lmol), and 4 A molecu-
5.18. 4-(Benzyloxycarbonylamino)-N-{[tris(b-D-galacto-
pyranosyloxy)methyl]methyl}-butyramide (29)
lar sieves (50 mg) in dry DMF (1 ml) was stirred at rt
under argon for 2 h. The resultant mixture (500 ll)
was transferred to a small vial containing Alexa Fluorꢁ
488-NHS (1 mg, 1.55 lmol), 4 A molecular sieves
(25 mg), and a stirring bar. The mixture was then stirred
˚
Compound 28 (241 mg, 0.115 mmol) was dissolved in dry
MeOH/dioxane (35 ml, 2.5:1), and sodium metal (65 mg,

O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
5229
in the dark at rt under argon for 4 days. The mixture
was then diluted with MeOH, filtered, and the solvents
were removed in vacuo. The residue was purified on
an RP-18 column (H2O/MeOH gradient 1:0!19:1) to
yield 8 (1.8 mg, 96%) as a red solid after final lyophiliza-
tion from water. ESI-MS: Calcd for C47H59N4Na2O29S2
(M+2Na+H)+: 1253.25. Found: m/z 1253.59.
in PBS for 30 min at 4 ꢂC. After fixation, the coverslips
were washed abundantly with PBS and mounted upside
down, in a Mowiol 4-88 mounting buffer containing N-
propyl gallate, onto glass slides.
Selective cellular uptake of the Alexa Fluorꢁ 488-labeled
compounds was visualized using a Zeiss Axiovert 135
microscope with a 63· planapo objective (numerical
aperture = 1.4, oil) with the appropriate filter set (450/
490, FT 510, LP 520) equipped with a Zeiss AxioCam
MRm CCD camera run by AxioVision 3.1 imaging
software.
5.21. Ligand binding and internalization
5.21.1. Fluorescence microscopy. One day before the
experiments, the cells were seeded at a density of
2 · 105 cells/well into 12-well plates containing collagen
coated glass coverslips. The cells were washed once with
PBS, and then serum-starved for 30 min on ice in 1 ml of
DMEM containing 25 mM Hepes. They were then incu-
bated with 500 ll/well of the Alexa Fluorꢁ 488-labeled
compounds 6–8 (100 lM) in the same medium on ice
for 1.5 h in the dark. After the binding step, the cells
were washed carefully four times with cold PBS. Then
fresh, prewarmed, complete DMEM medium (1 ml/well)
was administered, and the cells were incubated for
40 min in an incubator at 37 ꢂC in a humidified CO2
atmosphere (5%, v/v), leading to the internalization of
the receptor-bound compounds into the cells. After the
internalization step, the cells were washed twice with
PBS and then fixed with 3% paraformaldehyde (PFA)
5.21.2. Flow cytometry. Cells were grown for 24 h in col-
lagen coated (80 lg/ml) 24-well plates at a density of
3 · 105 cells/well or in 96-well plates at 1.5 · 105 cells/
well.
The titration experiments were performed in 24-well
plates, the cell layers were first washed twice with cold
PBS before incubation with compound 7 at concentra-
tions ranging from 0.4 to 12.5 lM (1:2 serial dilutions)
in 200 ll of DMEM without FBS for 40 min at 37 ꢂC.
Then, the cells were washed twice with cold PBS, de-
tached and stripped from surface-bound compound by
incubating them in a mixture containing 0.025% trypsin
and 5 mM EDTA in PBS for 10 min on ice. The addi-
Figure 7. Example of flow cytometric analysis for the uptake of compound 7 into HepG2 cells. All experiments were analyzed in the same way. SK-
Hep1 cells, which are less susceptible to clumping, show less debris in dot plot A and sharper peaks in histogram C (data not shown).

5230
O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
8. Coombs, P. J.; Taylor, M. E.; Drickamer, K. Glycobiology
2006, 16, 1C.
9. Lee, Y. C.; Townsend, R. R.; Hardy, M. R.; Lo¨nngren, J.;
Arnarp, J.; Haraldsson, M.; Lo¨nn, H. J. Biol. Chem. 1983,
258, 199.
10. Bider, M. D.; Wahlberg, J. M.; Kammerer, R. A.; Spiess,
M. J. Biol. Chem. 1996, 271, 31996.
11. Meier, M.; Bider, M. D.; Malashkevich, V. N.; Spiess, M.;
Burkhard, P. J. Mol. Biol. 2000, 300, 857.
12. Baenziger, J. U.; Maynard, Y. J. Biol. Chem. 1980, 255,
4607.
tion of complete medium quenched this process. The de-
tached cells were collected and centrifuged at a speed of
1500 rpm for 3 min. Finally, the cells were fixed in 2%
PFA in PBS for 15 min on ice followed by an alde-
hyde-quenching step with 100 mM lysine in PBS for
10 min. The fixed cells were then washed once with
FACS buffer (PBS containing 0.5% BSA and 0.1%
NaN3) and resuspended in 200 ll of the same buffer
for measuring.
13. Baenziger, J. U.; Fiete, D. Cell 1980, 22, 611.
14. Connolly, D. T.; Townsend, R. R.; Kawaguchi, K.; Bell,
W. R.; Lee, Y. C. J. Biol. Chem. 1982, 257, 939.
15. Mammen, M.; Choi, S.-K.; Whitesides, G. M. Angew.
Chem. 1998, 110, 2908.
The competitive uptake experiments were performed
with cells grown in 96-well plates. Twenty microliters
of asialofetuin dilutions at concentrations ranging from
0.6 to 200 lM or GalNAc at 0.6–200 mM was added di-
rectly to the cells, immediately followed by the addition
of 20 ll of compound 7 or 8 diluted to 20 lM in DMEM
without FBS and then incubated for 60 min at 37 ꢂC in
the incubator. The cells were washed, detached, and
fixed as described above.
16. Lee, Y. C.; Lee, R. T. In Carbohydrates in Chemistry and
Biology; Ernst, B., Hart, G. W., Sinay, P., Eds.; Wiley-
¨
WCH: Weinheim, 2000; Vol. 4, p 549, and references cited
therein.
17. Bock, K.; Arnarp, J.; Lo¨nngren, J. Eur. J. Biochem. 1982,
129, 171.
18. Nishikawa, M. Biol. Pharm. Bull. 2005, 28, 195.
19. Wilson, J. M.; Grossman, M.; Wu, C. H.; Chowdhury, N.
R.; Wu, C. Y.; Chowdhury, J. R. J. Biol. Chem. 1992, 267,
963.
20. Grossman, M.; Raper, S. E.; Kozarsky, K.; Stein, E. A.;
Engelhardt, J. F.; Muller, D.; Lupien, P. J.; Wilson, J. M.
Nat. Genet. 1994, 6, 335.
21. Wu, G. Y.; Wu, C. H. J. Biol. Chem. 1991, 263, 14621.
22. Hara, T.; Aramaki, Y.; Takada, S.; Koike, K.; Tsuchiya,
S. Gene Ther. 1995, 2, 784.
23. Hara, T.; Aramaki, Y.; Takada, S.; Koike, K.; Tsuchiya,
S. Gene 1995, 159, 167.
Analyses were performed on a CyAn ADP flow cytom-
eter with Summit 4.1 software (Dako Cytomation). An
example for the analysis of HepG2 cells is given in Fig-
ure 7. The forward and side scatter gate R1 was set to
count 30,000 intact cells of each sample (Fig. 7, dot plot
A). In gate R2, the cells counted in gate R1 were dis-
criminated for doublets (Fig. 7, dot plot B). Histogram
C depicts an overlay of the log of fluorescence intensity
at 488 nm of untreated (gray) and treated HepG2 cells
(green) from gate R2. The uptake of a compound into
cells was evaluated by comparing the shift in median
intensity of fluorescence (MFI) between untreated cells
(background fluorescence) and treated cells.
24. Hara, T.; Kawasawa, H.; Aramaki, Y.; Takada, S.; Koike,
K.; Ishidate, K.; Kato, H.; Tsuchiya, S. Biochem. Biophys.
Acta 1996, 1278, 51.
25. Kawakami, S.; Yamashita, F.; Nishikawa, M.; Takakura,
Y.; Hashida, M. Biochem. Biophys. Res. Commun. 1998,
252, 78.
Further analysis and IC50 calculations were done with
GraphPad Prism 4 software.
26. Kempen, H. J. M.; Hoes, C.; van Boom, J. H.; Spanjer, H.
H.; de Lange, J.; Langendoen, A.; van Berkel, T. J. C. J.
Med. Chem. 1984, 17, 1306.
Acknowledgments
27. Biessen, E. A. L.; Beuting, D. M.; Roelen, H. C. P. F.; van
de Marel, G. A.; van Boom, J. H.; van Berkel, T. J. C. J.
Med. Chem. 1995, 38, 1538.
The authors thank the Swiss National Science Founda-
tion (SNF) for funding the research. We are grateful to
28. Biessen, E. A. L.; Broxtermann, H.; van Boom, J. H.; van
Berkel, T. J. C. J. Med. Chem. 1995, 38, 1846.
29. Sliedregt, L. A. J. M.; Rensen, P. C. N.; Rump, E. T.; van
Santbrink, P. J.; Bijsterbosch, M. K.; Valentijn, A. R. P.
M.; van de Marel, G. A.; van Boom, J. H.; van Berkel, T.
J. C.; Biessen, E. A. L. J. Med. Chem. 1999, 42, 609.
30. Rensen, P. C. N.; van Leeuwen, S. H.; Sliedregt, L. A. J.
M.; van Berkel, T. J. C.; Biessen, E. A. L. J. Med. Chem.
2004, 47, 5798.
R. Sutterlin, Drs. M. Durrenberger, and E. Casanova
¨
¨
for the technical support in fluorescence microscopy.
Furthermore, we thank M. Cavallari from the Experi-
mental Immunology Group of Prof. Dr. Gennaro De
Libero for his assistance with the flow cytometry exper-
iments. We would also like to thank W. Kirsch for high-
accuracy elemental analysis measurements.
31. Lee, R. T.; Lee, Y. C. Bioconjugate Chem. 1997, 8, 762.
32. Rensen,P.C.N.;Sliedregt,L.A.J. M.;Ferns,M.;Kieviet,E.;
vanRossenberg, S.M.W.;vanLeeuwen,S.H.;vanBerkel,T.
J. C.; Biessen, E. A. L. J. Biol. Chem. 2001, 276, 37577.
33. Haugland, R. P. Handbook of Fluorescent Probes and
Research Products, 9th ed.; Molecular Probes, Inc., 2002.
34. Segura, M.; Sansone, F.; Casnati, A.; Ungaro, R. Synthe-
sis 2001, 14, 2105.
References and notes
1. Ashwell, G.; Harford, J. Ann. Rev. Biochem. 1982, 51, 531.
2. Spiess, M. Biochemistry 1990, 29, 10009.
3. Geffen, I.; Spiess, M. Int. Rev. Cytol. 1992, 137B, 181.
4. Li, M.; Kurata, H.; Itoh, N.; Yamashina, I.; Kawasaki, T.
J. Biol. Chem. 1990, 265, 11295.
35. Dehmlow, E. V.; Dehmlow, S. S. Phase-Transfer Cataly-
sis, 2th ed.; Verlag-Chemie: Weinheim/Bergstr, West
Germany, 1993.
5. Nakamura, K.; Funakoshi, H.; Miyamoto, K.; Tokunaga,
F.; Nakamura, T. Biochem. Biophys. Res. Commun. 2001,
280, 1028.
36. Fosker, A. P.; Law, H. D. J. Chem. Soc. 1965, 5, 7305.
37. Ellervik, U.; Magnusson, G. Carbohydr. Res. 1996, 280,
251.
6. Hoyle, G. W.; Hill, R. L. J. Biol. Chem. 1988, 263, 7487.
7. Wu, J.; Nantz, M. H.; Zern, M. A. Front. Biosci. 2002, 7,
d717.

O. Khorev et al. / Bioorg. Med. Chem. 16 (2008) 5216–5231
5231
38. Villanueva, I.; Hernandez, B.; Chang, V.; Heagy, M. D.
Synthesis 2000, 10, 1435.
39. Schwartz, A. L.; Fridovich, S. E.; Knowles, B. B.; Lodish,
H. F. J. Biol. Chem. 1981, 256, 8878.
40. Heffelfinger, S. C.; Hawkins, H. H.; Barrish, J.; Taylor, L.;
Darlington, G. J. In Vitro Cell Dev. Biol. 1992, 28A, 136.
41. Lauffenburger, D. A.; Lindermann, J. J. Receptors:
Models for Binding, Trafficking, and Signalling, 1st ed.;
Oxford University Press: New York, 1993.
42. Yura, H.; Ishihara, M.; Kanatani, Y.; Takase, B.; Hattori,
H.; Suzuki, S.; Kawakami, M.; Matsui, T. J. Biochem.
2006, 139, 637.
43. Wu, Y. T.; Jiaang, W. T.; Lin, K. G.; Huang, C. M.;
Chang, C. H.; Sun, Y. L.; Fan, K. H.; Hsu, W. C.; Wang,
H. E.; Lin, S. B.; Chen, S. T. Curr. Drug. Deliv. 2004, 1,
119.
44. Vogel’s. Textbook of Practical Organic Chemistry, Pre-
ntice Hall, 5th ed., 1989, p 467.