A sulfur tripod glycoconjugate that releases a high-affinity copper chelator in hepatocytes

Article abstract of DOI:10.1002/anie.201203255

Released in the cell: Three N-acetylgalactosamine units, which recognize the asialoglycoprotein receptor, were tethered through disulfide bonds to the three coordinating thiol functions of a sulfur tripod ligand that has a high affinity for Cu^I (see scheme). The resulting glycoconjugate can be considered as a prodrug, because after uptake by hepatic cells the intracellular reducing glutathione (GSH) releases the high-affinity intracellular Cu ^I chelator. Copyright

Full text of DOI:10.1002/anie.201203255

Angewandte
Chemie
DOI: 10.1002/anie.201203255
Bioinorganic Chemistry
A Sulfur Tripod Glycoconjugate that Releases a High-Affinity Copper
Chelator in Hepatocytes**
Anaꢀs M. Pujol, Martine Cuillel, Anne-Solꢁne Jullien, Colette Lebrun, Doris Cassio,
Elisabeth Mintz,* Christelle Gateau, and Pascale Delangle*
The copper ion is essential for life as a cofactor in a number of
vital processes. Its main role is to exchange electrons in
cuproenzymes needed for oxidative metabolism, neurotrans-
mitter and hormone biosynthesis, free-radical detoxification,
or iron absorption. However, high concentrations of copper
can be deleterious as they promote Fenton-like reactions,
thereby leading to oxidative damage of all the cell constitu-
ents, proteins, lipids, and nucleic acids. Therefore, the intra-
cellular copper concentration is strictly controlled so that
copper is provided to the enzymes that need it and does not
reach toxic levels.[1] Detoxification of the whole body occurs
in the liver, where copper is pumped towards the bile for
further elimination. The Wilsonꢀs disease patients whose
ATP7B gene is mutated lack the protein that is responsible
for pumping the copper ions out of the hepatocytes and suffer
from copper overload.[2] They need lifelong treatments by
chelators that help to lower dietary copper absorption and to
detoxify their body; unfortunately, these treatments have
harmful side effects and are not always efficient.[3] Since the
pool of intracellular copper is in the + I oxidation state, we
figured that a chelator that would enter the hepatic cells and
be specific for CuI could potentially represent an improving
alternative. Therefore, we are designing bifunctional mole-
cules that are able to both efficiently complex CuI and
specifically target hepatocytes.
a promising candidate for drug delivery into hepatocytes.[5]
ASGP-R recognizes terminal galactose (Gal) and N-acetyl-
galactosamine (GalNAc) with a better affinity for GalNAc
than for Gal.[6] Besides, sugar–protein interactions are greatly
enhanced by multivalent events, commonly known as the
“cluster glycoside effect”, which leads to a 102–103-fold
increase of the affinity for ASGP-R with each additional
carbohydrate from mono to triantennary structures.[7]
On the other hand, proteins involved in copper homeo-
stasis are outstanding sources of inspiration for the design of
efficient CuI chelators. In metallochaperones, CuI is mainly
bound by soft donors like cysteine thiolates in CXXC
sequences. This binding motif has led us to design peptides
including two cysteine residues;[8] these peptides show
a similar affinity for CuI as the metallochaperones (Kd ca.
[9]
10ꢀ16
)
and a high selectivity for CuI over ZnII, another
essential metal ion found in cells. Recently, one of these
peptides was functionalized with a cluster of carbohydrates to
target hepatocytes and shown to enter these cells and to
chelate intracellular copper.[10] Even more efficient CuI
chelators were obtained by favoring a CuS3 coordination
environment in mononuclear and cluster-type complexes, as
found in metallothioneins. Indeed, the tripodal ligand L2
(Scheme 1), which is based on a nitrilotriacetic acid (NTA)
moiety extended by three converging cysteines for metal
binding, shows a very high affinity for CuI, similar to that of
metallothioneins (Kd ca. 10ꢀ19),[11] and a large selectivity over
ZnII (108–109-fold).[12]
An efficient strategy to target hepatocytes is to use ligands
of the asialoglycoprotein receptor (ASGP-R), a hepatic lectin
that is chiefly expressed at the surface of these cells[4] and
Herein, we report on the glycoconjugate 1, which is
derived from ligand L2 and has the properties of L2 once it is
in the hepatocytes. Indeed, our synthetic strategy consists in
tethering to each chelating thiolate of L2 one carbohydrate
through a disulfide bond, which will be cleaved in the
reducing intracellular medium. Therefore, glycoconjugate
1 will acquire its specificity for CuI chelation by entering the
cells. Interestingly, among the structures proposed for optimal
ASGP-R recognition are tripods bearing on each arm a b-
linked GalNAc moiety that is kept 20 ꢁ away from the
branching point of the tripod by an ethylene glycol spacer.[13]
The tripodal architecture of ligand L2 is thus perfectly suitable
for the design of a triantennary glycoside cluster (1 in
Scheme 1). To ensure an optimal orientation of the three
terminal b-linked GalNAc moieties for recognition by ASGP-
R, an ethylene glycol spacer consisting of nine atoms has been
chosen to establish the 20 ꢁ connections between the
GalNAc and the branching point. The analogue 1-TAMRA,
bearing a carboxytetramethylrhodamine fluorescent unit, was
also synthesized to visualize the uptake into hepatocytes.
[*] Dr. A. M. Pujol, A.-S. Jullien, C. Lebrun, Dr. C. Gateau,
Dr. P. Delangle
INAC, Service de Chimie Inorganique et Biologique (UMR_E 3 CEA
UJF), Commissariat ꢀ l’Energie Atomique
17 rue des martyrs, 38054 Grenoble Cedex (France)
E-mail: pascale.delangle@cea.fr
Dr. M. Cuillel, Dr. E. Mintz
iRTSV, Laboratoire de Chimie et Biologie des Mꢁtaux (UMR 5249
CEA CNRS UJF), Commissariat ꢀ l’Energie Atomique
17 rue des martyrs, 38054 Grenoble Cedex (France)
E-mail: elisabeth.mintz@cea.fr
Dr. D. Cassio
INSERM UMR S757, Universitꢁ Paris-Sud
91405 Orsay (France)
[**] The financial support from the “Cluster de recherche Chimie de la
Rꢁgion Rhꢂne-Alpes” is duly acknowledged. We are grateful to
Ignacio Sandoval (CIBEREHD, Madrid, Spain) for his generous gift
of a polyclonal antibody anti-ATP7B and to Valꢁrie Nicolas
(Plateforme d’Imagerie Cellulaire, Chatenay Malabry, France) for
help with confocal analysis.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201203255.
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
a
10% decrease of the
amount of Cu complexed in
[CuI(bcs)2]3-. However, when
GSH was added together with
1, the amount of Cu bound to
the bcs ligand decreased dra-
matically to 8% of the total
copper, thereby evidencing the
release of the high-affinity che-
lator L2 after reduction of the
disulfide bridges of 1 into free
thiolates. For comparison, in the
presence of L2 only 2% of the
total copper is bound in the
[CuI(bcs)2]3- complex.[12b] These
experiments demonstrate that
the glycoconjugate
1 is not
a CuI chelator. However, the
reduction of 1 by GSH converts
it into an efficient CuI chelator,
thus confirming its potential
ability to bind CuI once entered
in the reducing medium of the
hepatic cells.
Scheme 1. Design and synthetic approach for the glycoconjugates 1, 1-Lys, and 1-TAMRA. Reagents and
yields: a) DMF; 36% (1) and 20% (1-Lys); b) TAMRA-NHS, diisopropylethylamine (DIEA), DMF; 16%
(1-TAMRA).
The synthetic strategy that leads to the glycoconjugates
(Scheme 1), consists in assembling the two building blocks L2
and 2 through three disulfide bonds in the last step. The
synthesis of the tripodal building block L2 was reported
recently[12b] and a similar procedure was used to obtain L2-Lys
from Na,Na-bis(carboxymethyl)-l-lysine, which affords a pri-
mary amine for the introduction of the fluorophore
(Scheme S1 in the Supporting Information). The GalNAc-
containing intermediate 2 was prepared from peracetylated
galactosamine, which was O-glycosylated through an oxazo-
line intermediate as the glycosyl donor, to favor the stereo-
selective formation of the b anomer (Scheme S2 in the
Supporting Information).[13c,14] The two tripods L2 and
L2-Lys were then coupled to three equivalents of intermediate
2 to afford glycoconjugates 1 and 1-Lys in 36% and 20%
yield, respectively (Scheme 1). The reaction of 1-Lys with the
N-succinimidyl ester of carboxytetramethylrhodamine
(TAMRA-NHS) gave the fluorescent analogue 1-TAMRA
in 16% yield after reverse-phase HPLC using a C18 column.
The Cu-chelating properties of 1 were evaluated in the
test tube, by competition with sodium bathocuproine disul-
fonate (Na2(bcs)). The [CuI(bcs)2]3- complex is formed as
shown in Equation (1) and is easily detected by its visible
absorption band (483 nm, 13300mꢀ1 cmꢀ1).[9a,10,12b] Thus, when
bcs ligand is added to any CuIL complex, as shown in
Equation (1), the more intense is the band at 483 nm, the
lower is the stability of the CuIL complex.
The efficiency of 1 was tested in two hepatic cell lines.
In vivo, hepatocytes are polarized cells, displaying apical
poles in contact with the outer medium, basal poles in contact
with the inner medium, and lateral poles in contact with
adjacent cells. In the liver, the outer medium is the bile that is
collected by a network of canaliculi formed by the union of
apical poles of adjacent hepatocytes, sealed by tight junctions
(Figure 2e). ASGP-R is embedded in the basal membrane in
Figure 1. Percentage of Cu in the complex [Cu(bcs)2]3- for samples
containing [Cu(CH3CN)4]PF6 (45 mm), bcs (100 mm) in phosphate
buffer (20 mm), pH 7.4 at 298 K, and a) 50 mm 1; b)50 mm 1 and 1 mm
GSH; c) 50 mm L2; d) 1 mm GSH.
contact with the blood. We checked that 1-TAMRA enters
HepG2 cells, the most commonly used human hepatic cell
line, and WIF-B9 cells, which are able to reconstitute
remarkably stable and polarized epithelia with functional
bile canaliculi within seven to ten days.[15] After two hours, the
dye was in all the cells and concentrated in a cell compartment
facing the canalicular spaces reconstituted by WIF-B9 cells.
Importantly, we also checked whether 1-TAMRA could enter
HeLa cells, a cell line deprived from ASGP-R.[16] This
3ꢀ
Cuþ þ 2 bcs2ꢀ ! ½CuIðbcsÞ2ꢁ
ð1Þ
As expected, 1 did not bind CuI as it contains three
disulfide bridges and no free thiolate functions: 100% Cu is
bound in a complex with bcs (Figure 1). Glutathione (GSH,
1 mm) was used to mimic the reducing intracellular medium:
in agreement with the low affinity of GSH for CuI, it induced
2
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Figure 2. Effect of Cu on ATP7B localization in WIF-B9 cells and efficiency of the glycoconjugate 1. a–d) ATP7B and HA4 (a canalicular membrane
marker) were detected by indirect immunofluorescence and imaged by confocal microscopy. In each panel are presented from left to right, the
Nomarski image, the green and red images of the sum of two confocal sections taken in the middle of the cell layer, and the merge image. On
the Nomarski images, bile canaliculi are indicated by blue asterisks. a) Cells kept in the basal culture medium for 5 h. b) Cells kept in the basal
culture medium plus CuCl2 (1 mm) for 5 h. c) Cells kept with CuCl2 (1 mm) for 2 h and further treated with 1 (10 mm) for 3 h. d) Cells kept with
CuCl2 (1 mm) for 2 h and further treated with C1 (10 mm) for 3 h. e) Scheme of polarized hepatic cells showing the localization of ATP7B (green) in
presence of high and low Cu concentrations. f) Dose-dependence effect of 1, C1, C2; the % of bile canaliculi surrounded by ATP7B is presented
as a function of the compound concentrations; all the experiments used in this curve were performed as in c) or d).
experiment showed that 1-TAMRA does not enter HeLa cells
under the conditions where it enters HepG2 and WIF-B9
cells, thereby illustrating the role of the hepatic receptor in
the cellular uptake of the glycoconjugate (Figure S2 in the
Supporting Information).
monitored by immunofluorescence.
A
complementary
experiment with the detection of a Golgi protein,[19] instead
of the HA4 protein is shown in Figure S3 in the Supporting
Information. In agreement with previous work,[17] under basal
conditions where the copper concentration is low, ATP7B was
in the Golgi network between the nucleus and the apical
membrane (Figure 2a), whereas with copper ions (1 mm) in
the culture medium, ATP7B was mainly localized in a con-
tinuous crown around the canaliculus to excrete the excess of
copper into the bile (Figure 2b). The latter localization of
ATP7B bears all the hallmarks of the physiological needs for
copper detoxification and can therefore be used as an
By following the work of Hubbard and co-workers,[17] we
have recently used the localization of ATP7B, the protein
responsible for copper detoxification, as an intracellular
copper sensor in WIF-B9 cells.[10] WIF-B9 cells were incu-
bated under basal conditions or with added copper ions
(1 mm) and the localizations of ATP7B protein (in green) and
HA4 protein (a canalicular membrane protein[18] in red) were
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
evidence for the ability of our compounds to chelate excess of
intracellular Cu. Hence, ATP7B and HA4 localizations were
monitored in the presence of copper ions (1 mm) and either 1,
or C1, C2, L2ox as control compounds (Scheme 2 and the
Supporting Information). C1 is a molecule similar to 1 with no
sulfur atoms at all, unable to release the sulfur tripod. C2
mimics one arm of 1 and is able to release in the intracellular
medium low-affinity monothiol compounds that resemble
glutathione. L2ox is the oxidation product of L2, a dimer
containing three disulfide bridges but lacking the sugar
targeting units; L2ox could become a chelator if it entered
the cells.
be considered as a vector delivering L2, the active compound,
at the right place to chelate excess CuI. For the Wilsonꢀs
disease patients, 1 could represent an interesting concept of
a prodrug to be further developed.
Received: April 27, 2012
Published online: && &&, &&&&
Keywords: bioinorganic chemistry · cell recognition · copper ·
.
drug delivery · glycoconjugates
[1] S. Lutsenko, Curr. Opin. Chem. Biol. 2010, 14, 211 – 217.
[2] a) D. Huster, Best Pract. Res. Clin. Gastroenterol. 2010, 24, 531 –
539; b) P. Delangle, E. Mintz, Dalton Trans. 2012, 41, 6359 – 6370.
[3] a) B. Sarkar, Chem. Rev. 1999, 99, 2535 – 2544; b) O. Andersen,
Chem. Rev. 1999, 99, 2683 – 2710.
[4] a) M. Meier, M. D. Bider, V. N. Malashkevich, M. Spiess, P.
Burkhard, J. Mol. Biol. 2000, 300, 857 – 865; b) M. Spiess,
Biochemistry 1990, 29, 10009 – 10018.
[5] J. Wu, M. H. Nantz, M. A. Zern, Front. Biosci. 2002, 7, d717 –
725.
[6] J. U. Baenziger, Y. Maynard, J. Biol. Chem. 1980, 255, 4607 –
4613.
[7] a) Y. C. Lee, R. R. Townsend, M. R. Hardy, J. Lonngren, J.
Arnarp, M. Haraldsson, H. Lonn, J. Biol. Chem. 1983, 258, 199 –
202; b) Y. C. Lee, R. T. Lee, Acc. Chem. Res. 1995, 28, 321 – 327.
[8] a) P. Rousselot-Pailley, O. Seneque, C. Lebrun, S. Crouzy, D.
Boturyn, P. Dumy, M. Ferrand, P. Delangle, Inorg. Chem. 2006,
45, 5510 – 5520; b) O. Sꢂnꢃque, S. Crouzy, D. Boturyn, P. Dumy,
M. Ferrand, P. Delangle, Chem. Commun. 2004, 770 – 771.
[9] a) Z. Xiao, F. Loughlin, G. N. George, G. J. Howlett, A. G.
Wedd, J. Am. Chem. Soc. 2004, 126, 3081 – 3090; b) R. Miras, I.
Morin, O. Jacquin, M. Cuillel, F. Guillain, E. Mintz, J. Biol.
Inorg. Chem. 2008, 13, 195 – 205.
Scheme 2. Structures of control compounds.
After incubation for three hours with 1 (10 mm), the
pericanalicular distribution of ATP7B changed dramatically
as the green fluorescence localized at a large compartment
reminiscent of the Golgi network (Figure 2c), close to the
nucleus; this distribution resembles that observed in basal
conditions (Figure 2a). In similar experiments using C1 (10–
50 mm) or C2 (30–90 mm) or L2ox (50 mm), ATP7B distribution
did not change, thus indicating that the cells still needed
detoxification (Figure 2d and Figure S4 in the Supporting
Information). The proportion of bile canaliculi surrounded by
ATP7B increases with excess Cu concentration and is there-
fore related to the ability of the compounds to chelate toxic
excess copper and to cancel the detoxification process by
ATP7B. This proportion clearly decreased as the concentra-
tion of the sulfur tripod glycoconjugate 1 increased, whereas
the pericanalicular distribution of ATP7B is not sensitive to
the other compounds (Figure 2 f). These experiments dem-
onstrate that the two functions present in 1, the glycoside
cluster for ASPG-R targeting and the sulfur tripod for
efficient CuI chelation, are essential to promote the release
of a high-affinity intracellular CuI chelator once it is in hepatic
cells.
[10] A. M. Pujol, M. Cuillel, O. Renaudet, C. Lebrun, P. Charbonnier,
D. Cassio, C. Gateau, P. Dumy, E. Mintz, P. Delangle, J. Am.
Chem. Soc. 2011, 133, 286 – 296.
[11] P. Faller, FEBS J. 2010, 277, 2921 – 2930.
[12] a) A. M. Pujol, C. Gateau, C. Lebrun, P. Delangle, J. Am. Chem.
Soc. 2009, 131, 6928 – 6929; b) A. M. Pujol, C. Gateau, C.
Lebrun, P. Delangle, Chem. Eur. J. 2011, 17, 4418 – 4428.
[13] a) P. C. N. Rensen, S. H. van Leeuwen, L. A. J. M. Sliedregt,
T. J. C. van Berkel, E. A. L. Biessen, J. Med. Chem. 2004, 47,
5798 – 5808; b) O. Khorev, D. Stokmaier, O. Schwardt, B.
Cutting, B. Ernst, Bioorg. Med. Chem. 2008, 16, 5216 – 5231;
c) U. Westerlind, J. Westman, E. Tornquist, C. I. E. Smith, S.
Oscarson, M. Lahmann, T. Norberg, Glycoconjugate J. 2004, 21,
227 – 241.
[14] S. Nakabayashi, C. D. Warren, R. W. Jeanloz, Carbohydr. Res.
1986, 150, C7 – C10.
[15] C. Decaens, M. Durand, B. Grosse, D. Cassio, Biol. Cell 2008,
100, 387 – 398.
[16] J. H. Park, E. W. Cho, S. Y. Shin, Y. J. Lee, K. L. Kim, Biochem.
Biophys. Res. Commun. 1998, 244, 304 – 311.
[17] Y. Guo, L. Nyasae, L. T. Braiterman, A. L. Hubbard, Am. J.
Physiol. Gastrointest. Liver Physiol. 2005, 289, G904 – G916.
[18] A. L. Hubbard, J. R. Bartles, L. T. Braiterman, J. Cell Biol. 1985,
100, 1115 – 1125.
In conclusion, we demonstrated that the tripodal CuI
chelator L2 derived from NTA by an extension with three
cysteines, which exhibits an affinity for CuI as high as that of
metallothioneins, could be derived to enter hepatic cells and
complex intracellular copper. The glycoconjugate 1 can thus
[19] J. Saraste, G. E. Palade, M. G. Farquhar, J. Cell Biol. 1987, 105,
2021 – 2029.
4
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Bioinorganic Chemistry
Released in the cell: Three N-acetylga-
lactosamine units, which recognize the
asialoglycoprotein receptor, were teth-
ered through disulfide bonds to the three
coordinating thiol functions of a sulfur
tripod ligand that has a high affinity for
CuI (see scheme). The resulting glyco-
conjugate can be considered as a pro-
drug, because after uptake by hepatic
cells the intracellular reducing gluta-
thione (GSH) releases the high-affinity
intracellular CuI chelator.
A. M. Pujol, M. Cuillel, A.-S. Jullien,
C. Lebrun, D. Cassio, E. Mintz,*
C. Gateau, P. Delangle*
&&&& — &&&&
A Sulfur Tripod Glycoconjugate that
Releases a High-Affinity Copper Chelator
in Hepatocytes
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!